PDA

View Full Version : BUILD LOG: Quite an Unusual one



Boyan Silyavski
26-09-2013, 04:14 PM
Hi,

i am still in the middle of my 1st build for a friend and here comes the second already :cool:

This would be a brainstorm so , help and ideas will be highly appreciated. In fact i depend on it :semi-twins:

Basically with some hard earned money from my day job and a bit profit from the other build i want to start slowly to build this machine.

If i succeed it will be my dream machine:sentimental:

Before some months i asked the same question at the zone, but did not get any help there. So i have done some preliminary research alone.


So here are the details:

1. What i would like to achieve:

-Now comes the crazy part.
If its possible i would like to have a very sturdy and rugged router table with work dimensions 700x700x200mm. At the same time i would like to be able to mount a pneumatic impact hammer and form metal sheets on it. I would like to form a wok-like shape circle with aproximately 53-60cm diameter to shape it from flat to 15cm deep. The metal is soft, mild sheet 1.2mm thick. Now i shape them manualy.

Dont ask me why, its another 200 page explanation, but i want to be able to make this sheet into form using impacts, not any other method. Not spinning, not waterforming and so on. This is 100 percent shure. No doubt here. So big or small impact hammer head mounted on air rammer moved by CNC.
The impact are known frequency of 700 to 1000 impacts per minute and the impact head size ideally will be from 10 to 25, 30 mm in diameter and should not coincide, so there is a minimum speed that should be maintained.
For that task the main goal is not accuracy but repeatability! means the metal is springy, and so and so, what matters is the final shape to be same all over again. As i do it by hand now, and do it well, i am quite capable of programming the hits afters certain experimenting so that is not an issue.

Can this be done? Yes, i have seen a document where 30cm dish is achieved using unsupported round rails and 10mm hammer head.

What i want to achieve is good repeatability with a machine that will not break after a year doing this all day. And if at the same time i can use it for Wood and aluminum routing will be perfect. As i dont have money for 2/ even for one now


Here is a video how it is done by hand by a guy/not me, i do it for now with normal hammer/. The 1.2mm sheet is quite ductile in fact, not as people could imagine. The goal CNC-ing the process is repeating the same shape over and over again, in other words eliminating variables.
the thumper the guy uses on the video is bigger by one size that what i will use. The weight will be 4kg. He uses 35mm head , i would be happy even with smaller as i would CNC it, not be in a rush i mean.


https://www.youtube.com/watch?feature=player_embedded&v=VbR6fHoZrTQ





2. What i figured till now

I hope and i would like to be with moving gantry. Making a mill like, will be last resort.

Basically i have no problem with the general design. Big steel profile, sturdy table, double gantry as Johnathan build but with added 3rd rail somewhere higher as normal gantry for additional strength. Eventual Z box in the middle with 3 or even 4 No money will be saved for the sturdiness of the gantry and the table. Welded steel. I will design it so to resist a direct hit from a truck:topsy_turvy:

I said rails. Now here is where i need help.
After a lot of thought and checking specifications and so i came to 2 conclusions. There was a bit consideration for the price also.
I decided that my best bet will be HIWIN Roller guide square supported rails , medium to heavy preload, 30 size.

The other choice i meditated a lot , was a big SBR rails like 30-50 size with double Plain Frelon carriages which are not cheap. Same price as big Hiwin roller carriages.


Then the movement. Here i am lost. Ball screws possibly with 2 nuts each on every axis, also 2 ballscrews for x, 2 for y and 2 for Z. Driven by short belts and so.

I also contemplate the possibility to make a floating Z plate and dumpen it with adjustable bike suspension, to imitate the rebound of the human hand when doing the same process, that could further lower the stress on the rails.

I will start posting the current design and will need your help here :hysterical:

Also it would be interesting if there is a chance to make it a separate machine with some kind of cheap heavy duty stuff instead of expensive Hiwin roller motion.
moved by belts or so.

One thing to me is clear though. It should be moved by at least Closed Loop Steppers, which i thing will be best.

~andy
26-09-2013, 05:07 PM
I pity the guys neighbours.

Have you considered a fixed gantry design?

You can get circular rails if you want to rotate the circle.

Boyan Silyavski
26-09-2013, 07:05 PM
I pity the guys neighbours.

Have you considered a fixed gantry design?

You can get circular rails if you want to rotate the circle.

You should pity my neighbors too. Thats why i rented a house in the countryside. Cause its hours and hours.

I have considered. Unfortunately i live in a rented house. If i have to move a half tone machine just changing house, it would be quite difficult. Also the build could become quite expensive with moving tables and so.

Also here is how i think. If the rails can support such a job, then they will be ok for moving gantry if the gantry is extremely rigid.

I have considered round THK rails but that could lead to other problems, like reinforcing the rings where the sheet is mounted.
Now the rings are meant to be integrated in the table bed. I mean the sheet would be bolted around round hole in the table bed with 12x M10 screws.


The basic idea for the table bed made from 100x100 steel profile:

1024010241

JAZZCNC
26-09-2013, 07:22 PM
Other than the obvious fact it will need to be strongly built then I don't think it will be difficult make this work with but from what I see for it to work properly then you need the tool 90deg to the steel has it domes.? Affectively you need 5 Axis. . . . This would rule out Z axis like Jonathan's box design.

My approach for dual purpose machine would be to build a Sturdy gantry with something like 200x100x10. Then either have two separate Y axis one on each side each with own rails, screws, motors and Z axis or build a single double sided Y axis with Z axis on each side. My preference would probably be shared has it keeps cost and weight of gantry down but I would have rails on both sides.
Then the Rear Y axis would have heavy duty Z axis with short stroke and attached to that would be a stepper driven Rotary table which striker mounts onto creating the 4th axis. This would enable you to keep the striker 90deg to metal at any dome depth.

Next To hold the Metal and create the 5th Axis I would build a sturdy fixture on a Rotating table driven by stepper. Few ways this could done.? Either using standard rotary table driven by stepper and bolting Sturdy metal holding fixture onto it. . . .OR. . . . Make a Diy rotary table from an Old car or truck clutch Fly wheel and plates and bearings. Base plate would have bearing attached to it on it's perimeter and the fly wheel would sit on them and rotate on central spindle. It would be driven by stepper directly from the fly wheel ring gear. . . . . This table assembly would be removable from the main bed.

Now strictly speaking this wouldn't be 5 Axis has you'll only ever be using 3 Axis at same time. So you could make so it shares the same drives and just swap motor connections or fit Switching system to swap over. (would need powering down drives first)
It would go like this.!! . . Locate tool on centre of metal disc and park gantry and Y axis. These would now not move again so can unplug or switch to drive both rotary tables. Then swap or switch over front Z axis to rear Z axis.

Should be some Food for thought there.??

Ulsterman
26-09-2013, 08:06 PM
I would buy a rail package and Abb robot -----The 7th Axis is the table/carriage and carries the Robot to the Fixture -Numatic hammer on the head of the robot giving 6 axis control ?

Boyan Silyavski
26-09-2013, 08:10 PM
Other than the obvious fact it will need to be strongly built then I don't think it will be difficult make this work with but from what I see for it to work properly then you need the tool 90deg to the steel has it domes.? Affectively you need 5 Axis. . . . This would rule out Z axis like Jonathan's box design.
The overall angle of the shell/ the wok/ is 27 to 30 degrees from horizontal. meaning 53cm diameter and 14cm depth. That leads to 2 things:
1. A special hammer head with the desired angle to achieve the final result
2.As you say, more axis.
I calculated that for a 10-12mm diameter hammer head with specially designed angle it would not be a problem to form the shell. Thats how i saw they did it in on document. Though the impact frequency is unknown to me. Needless to say the shell will be greased or oiled so the head to be able to slip-stretch the sheet properly.



My approach for dual purpose machine would be to build a Sturdy gantry with something like 200x100x10.

Thats exactly my initial design intent



Then either have two separate Y axis one on each side each with own rails, screws, motors and Z axis or build a single double sided Y axis with Z axis on each side. My preference would probably be shared has it keeps cost and weight of gantry down but I would have rails on both sides.


Thats new idea to me, very grateful.



Then the Rear Y axis would have heavy duty Z axis with short stroke and attached to that would be a stepper driven Rotary table which striker mounts onto creating the 4th axis. This would enable you to keep the striker 90deg to metal at any dome depth.

I have thought of this but not sure if this is in my ability to design.Without mistakes, i mean.



Next To hold the Metal and create the 5th Axis I would build a sturdy fixture on a Rotating table driven by stepper. Few ways this could done.? Either using standard rotary table driven by stepper and bolting Sturdy metal holding fixture onto it. . . .OR. . . . Make a Diy rotary table from an Old car or truck clutch Fly wheel and plates and bearings. Base plate would have bearing attached to it on it's perimeter and the fly wheel would sit on them and rotate on central spindle. It would be driven by stepper directly from the fly wheel ring gear. . . . . This table assembly would be removable from the main bed.


I have to dig deeper about the clutch design then. I have contemplated something similar. Find a truck tire with the axle, fix it on a fixture 15degree inclined/half of the desired angle and made 2 axis only Y and Z. And rotate the tire by belt.
Believe it or not i could not find a tyre with axle here in Spain, for cheap from some scrapyard.

I have thought also of cutting 2 rings, similar to the rings i have for manual sinking, but at the outside cut with AT belt tooth/laser cut. Then flange them and lay them on a table where a lot of bearings are mounted in a circular pattern. The type of heavy load that are big balls only. I dont know the name in English.
But the thing will jump and i dont know how to fix it not to jump around.



Now strictly speaking this wouldn't be 5 Axis has you'll only ever be using 3 Axis at same time. So you could make so it shares the same drives and just swap motor connections or fit Switching system to swap over. (would need powering down drives first)
It would go like this.!! . . Locate tool on centre of metal disc and park gantry and Y axis. These would now not move again so can unplug or switch to drive both rotary tables. Then swap or switch over front Z axis to rear Z axis.

Thats ok for me.

But i don't understand in fact- what will i gain with the 4 rails on the gantry and 2 Z axis ? If i go with heavy duty preloaded roller guides i could use them at the same time for CNC and the impact hammer. if they f"""k i will just change them. or you mean to be able to do 2 things without swapping spindle with hammer? if just for avoiding swapping, then an over-sized Y could do the same job and park one of the Z out of the way.



Should be some Food for thought there.??



I will analyze and implement.

So Dean, my Big questions are:

Will Hiwin RGH 30 HA stand up to the task?

Will ballscrews with 2 nuts each stand up to the task and especially on Z, where additional vibrational creep has to be dealt with? Z with 2 ballscrews and 4 nuts should stand , yes? Ok, not forever, but at least for a time.

What size ballscrews on each axis? the normal speedy ones? 1605 and 1610 respectively?

~andy
26-09-2013, 08:20 PM
Out of interest, how are you planning to monitor the affect of the hammer?

Boyan Silyavski
26-09-2013, 09:03 PM
I would buy a rail package and Abb robot -----The 7th Axis is the table/carriage and carries the Robot to the Fixture -Numatic hammer on the head of the robot giving 6 axis control ?

I am saving for Bentley, so can not afford that one.


Out of interest, how are you planning to monitor the affect of the hammer?

To sink a shell is an art of itself. The idea is ideal shape with thicknesses distributed exactly where you want them, and its not linear. Repeating the result when hammering by hand is not possible, its difficult also to form a precise smooth shape .

Actually i am stretching the metal in a special way, not simply hammering it down. Every pass is different and the goal is to stretch and distribute the thickness where you want it without warping the sheet. As the sheet warps irreparably when not hammered properly. Ok, somewhat repairable, but with heat treating only. And it introduces other problems that have to be dealt with at further moment.

So implementing my hammering technique pass by pass will program it. Until i eliminate material variables, yes cold rolled steel is not so perfect as it sounds.

Butr if you mean literally, the answer is with 37db ear protection:toot:

JAZZCNC
26-09-2013, 11:43 PM
But i don't understand in fact- what will i gain with the 4 rails on the gantry and 2 Z axis ? If i go with heavy duty preloaded roller guides i could use them at the same time for CNC and the impact hammer. if they f"""k i will just change them. or you mean to be able to do 2 things without swapping spindle with hammer? if just for avoiding swapping, then an over-sized Y could do the same job and park one of the Z out of the way.

Yes was just to avoid swapping but if you have room for extra wide Y axis that would work OK





So Dean, my Big questions are:

Will Hiwin RGH 30 HA stand up to the task?

Yes easily. 30mm bearings are massive things and can handle large loads.

Will ballscrews with 2 nuts each stand up to the task and especially on Z, where additional vibrational creep has to be dealt with? Z with 2 ballscrews and 4 nuts should stand , yes? Ok, not forever, but at least for a time.

What size ballscrews on each axis? the normal speedy ones? 1605 and 1610 respectively?[/QUOTE]

Don't think 2 nuts would matter or help in this instance has reducing backlash isn't required.
Regards Z axis then Obviously two screws will take more punishment than one but single screw would just means changing sooner and less complicated.!

Yes standard screw pitch and either 16 or 20mm should be fine.

Wobblycogs
27-09-2013, 03:55 PM
Please take what I say with a pinch of salt as I'm new to this but I can't help feeling that if you are doing this on a regular basis you'd be better off building a dedicated machine for it. By starting with a regular CNC design you seem to need to provide five axes (6 if you have a rotating attachment on the bed).

The machine I'm thinking of is arranged like a lathe but with tail stock replaced with a three axis hammer. The tail stock would move in and out and the hammer up and down as well as swivel to ensure the hammering is always 90deg to the work piece.

John McNamara
28-09-2013, 01:24 AM
Hi Silyavski

Would a support under the hammer help?

Effectively an anvil that could be formed against. Ideally different anvils for forming groves or depressions of a certain shape.

This would also mean you could work rectangular or other shapes as well.

The anvil moves in unison with the top hammer so it should not be too difficult to incorporate it. (Although being able to rotate the anvil and the head in unison would be an advantage)

Have you seen this video?

Ford unveils new CNC prototyping process for metal parts | Manufacturers Monthly (http://www.manmonthly.com.au/news/ford-unveils-new-cnc-prototyping-process-for-metal)

Regards
John

Ulsterman
28-09-2013, 02:47 AM
Metal spinning and Flow forming lathe DENN model ZENN 80 CNC - YouTube (http://www.youtube.com/watch?v=3QOZWUsOXJA)
Have you considered flow forming ?------A lot of Automotive parts done this way now

The other option is the Robot because you don't need 7 axis with rail but only 6 -Part on the fixture and hammer on the Robot --2nd hand they are not so expensive now especially if single channel

John McNamara
28-09-2013, 09:18 AM
I guess the question is this machine a special purpose machine that has only one task making dished circular parts, or is it to be more general purpose?

Is the reason that only hammering will work, not flow forming as in Ulsterman's post or the Ford Link I posted, because of the sound qualities of the finished object when struck?

I guess the hammering introduces stresses that cause various resonance effects as in symbols.

Regards
John

Boyan Silyavski
28-09-2013, 03:40 PM
Hi guys,

Thanks for the interest and the help.

Let me answer some questions:

1. Please don't forget its a DIY machine and i don't have a machine shop, though can use some services where necessary. Also i don't have the budget for robots, CNC machining centers and so on. I am into making that musical instrument, but still learning, so as i am not a successful maker yet, the main purpose of the machine will be to sink the shells so i am not tired after 4 hours of precise manual sinking and so i have time to learn to tune it, which is much more difficult task.

2. 2 in 1?
I have a small belt driven CNC and i need bigger machine urgently. I am still thinking that if the machine is well designed, it can be a 2 purpose machine. Lets say 25% normal CNC and 75% special purpose.

3. What you call Flow forming is Supported ISF/incremental sheet forming/ . Unfortunately i aim for introducing particular stress in the material which can be achieved only with hammering. You guessed right that it affects dramatically the later capabilities of the musical instrument.

In fact support head bellow or a fixed form is not needed. They use this type of ISF for achieving precision and faster programming time. I dont care for that, as i know what i am doing and will take the material springiness in count, that's what they want to eliminate using the anvil.

That examples are another reason that i believe i could go without 5 or 6 axis. The air hammer will 1/2 ISF and 1/2 introduce stress meanwhile, carefully adjusting the pass depth and the pass pattern. With a properly designed tool head.


4. But lets look at the 5 or 6 axis design. I have no idea about these. What CAD and CAM? Now i use Aspire and mach3. I think they can not support this type of machining?

Also can somebody point me to a good DIY design of additional axis? I saw only 2000$ axis as the cheapest in the net

5.My worries about Z backslash are based on that the vibrating hammer will f///k the ball screw if let oscillate meanwhile back-slashing


6. So HIWIN 30 roller bearings. Do we rule out totally the Frelon SBR like supported plain bearings? Why they are not widely used on DIY CNC? They support a lot of load, are fast, self cleaning and support oscillating vibrations

7.And what about an rotary 4rth axis which rotates the mounted shell left and right only 15+15 degrees? Then always the head will be 90 degrees to the shell And the Z could be normal. I am talking something like my manual sinking stand? Driven by motor at the shafts

10257

Boyan Silyavski
26-10-2013, 05:22 PM
I am more and more inclined to make a special machine for the purpose. Anyways still contemplating it.

Nevertheless i have to start buying. So i decided to go backwards. The spindle first.

Need some help with the choice:

Now the big question is 3kw or 2.2kw spindle? I am more inclined to the 3kw. As a brand name inverter will be expensive this size, i am looking into the chinese vfds. Are they reliable enough, Dean mentioned somewhere that now they make them good. Should i know how to distinguish them or what? I see some have aluminum backs some not.

Wobblycogs
26-10-2013, 07:31 PM
I recently bought a 2.2kW spindle + VFD from Chai (ebay: linearmotionbearings2008) as part of a larger order. I've not had a chance to fire it up yet but it came well packaged and all appears to be in order. The price came out to be a few pounds more than the cheapest I could find but I found a number of very bad reviews for the cheapest seller (things not turning up, etc). From what I've read the spindles are normally ok, they all claim Japanese / German bearings but it's luck if they are. The standard VFD's are also normally ok but they seem to vary in quality more although I think the quality has improved a lot in recent years, it seems they either last for ever or go pop in the first few weeks.

Boyan Silyavski
01-11-2013, 12:00 PM
Guys, i have to start sooner this build as everyday i change my mind.:joker:

Now here is the deal. i have decided to make the machine with working area 1200 x 2500 and at the same time make it fast and rigid. Till now the problem was limited space and problems in moving the machine.
Now the funny part. I bought yesterday a towbar and am looking to buy a suitable platform trailer, so i will mount directly the machine on the trailer :hysterical:
I did not want to cut myself short just because i could note move it from home to home. So its decided.

I will definitely go with the Johnathan design of a gantry just make it suit my needs and abilities. maybe even add one more beam at one side like a standart gantry, as Dean suggested somewhere.

So here is where i need help:
When the machine was small it was clear that 1605 screw for the z and 1610 for X and Z.

Now some options for the gantry:
1. make it from 2 100x100x4 = whole gantry below 50kg

2. make it from 3 100x100x3 / 2soldered together/ - 1m=9kg, so whole gantry will be ~ 65 kg

3. make it from 3 100x100x7 = whole gantry 80-90kg

-So, the big question is how to move the gantry at 2600mm length?
-What screws, what motors?
-Servo maybe ? I know nothing about servo, can you point me to good money wise servo for the long axis?
-What drive relation 1:1 , 1:2, 1:3
-unfortunately i can not machine rotating ballnut. Could somebody help me here? any other options?


So now i am a bit confused between 2 options, make a sturdy beast 1250x1250 or make a sturdy beast 1250x2500 with your help. I mean an exemplary sturdy beast for the generations to come:beer:

Wobblycogs
01-11-2013, 12:08 PM
If you are going to put this in a trailer you'll need to make sure you add some way so you can jack it up on sturdy legs. If it's resting on the trailer wheels I think the machine will probably shake itself to bits.

Boyan Silyavski
01-11-2013, 12:12 PM
If you are going to put this in a trailer you'll need to make sure you add some way so you can jack it up on sturdy legs. If it's resting on the trailer wheels I think the machine will probably shake itself to bits.
I agree with you. i will take off the wheels and hard mount the trailer to the floor, and the machine will be hard mounted on the trailer. The trailer idea is to use it only when i change rented houses. Will not drive it around :pirate:

Boyan Silyavski
06-12-2013, 01:34 PM
Hi there. Almost finishing my design.

Basically i decided to go Big.

I designed quite an unusual gantry and Z with the main idea of eliminating as much overhang as possible, light as possible and easy to build as possible.

Another thing is that affected by Jonathan's build i could not help but contemplate an extreme machine with 3rd gantry beam. See picture 6. I believe Jazz sugested something similar, if i am not wrong.


The gantry would be driven by rotating nuts/2510 ball screw/ if Jonathan or somebody is kind enough to make them for me, i know free time is an issue there.

The gantry design as tested by me will have less than 0.01 flex and bend /twist/ and will weight 2 beams -40kg +Z weight or 55kg+Z weight if triple.

For the moment i am very inclined to invest in the Leadshine hybrid 220V AC drives, though maybe i will start with Geckos as i have them. Depends what would cost me for rails and ballscrews for 1250x2500. Still have not decided. But if i make it only 1250x1250, definetly the 3 beam design and the AC drives.


What worries me most is the bed, as i dont have money for 20mmm alu bed. I am contemplating how to make a strong cheap bed with strong channels.



Now please tell me what do you think? My main idea is to build something exceptional if possible, cheap as possible where possible :pirate:, and as simple as possible, so when i share the plans it could easily be repeated by other people.

PS. The Z design is just a sketch, is not finished, will be lightened and i will try to move the ball screw further to the center without messing with the spindle. Another point is that now the plan is for 100mm diameter spindle, as i assume most people would choose for such a machine.

109371093810939109401094110942

Wobblycogs
06-12-2013, 01:48 PM
I've also thinking about what to do for a bed with t-slots. My first thought was, like you, 20mm aluminium and machine channels into it. Then I saw the price of the piece of aluminium I'd need! It's probably the route I'll end up taking though as it's fairly simple.

The alternative option (which I've not priced up yet) is to get some aluminium profile, lay that on the bed so that it's very close to flat and then take a thin surfacing cut. You can then use the built in t-slots. Not sure how much cheaper that will be in the long run though. If you can get away with using regular or even thin walled profile I suspect it would be cheaper.

I've just about finished buying all the parts for my machine, cheap and simple are not available options :highly_amused:

Boyan Silyavski
06-12-2013, 02:03 PM
I've also thinking about what to do for a bed with t-slots. My first thought was, like you, 20mm aluminium and machine channels into it. Then I saw the price of the piece of aluminium I'd need! It's probably the route I'll end up taking though as it's fairly simple.

The alternative option (which I've not priced up yet) is to get some aluminium profile, lay that on the bed so that it's very close to flat and then take a thin surfacing cut. You can then use the built in t-slots. Not sure how much cheaper that will be in the long run though. If you can get away with using regular or even thin walled profile I suspect it would be cheaper.

I've just about finished buying all the parts for my machine, cheap and simple are not available options :highly_amused:

The DIYer biggest problem is that we never know what exactly the machine would be for, so we want it as multi functional as possible. At least that's me.

For now the best i thought of would be to do the bed in the following manner:

Most of it Plywood with nuts covered by expanded PVC sheet 10mm, covered by another sacrificial sheet of the same material. The area near me/ the first square meter/, made with the thick alu channel extrusion + sacrificial PVC, or all same but holes left for a toolplate that can be bolted over all directly to the supporting beams. Or the first square meter made from square bars, cut and bolted so that emulate tool plate. Even if not perfect, they will be strong to bolt to and the plastic over them will be resurfaced. I even think that that type of machine could scratch steel, so i could resurface them directly.

Boyan Silyavski
13-12-2013, 10:29 PM
After some thinking, calculating and so on, i believe i have the perfect compromise. Meaning- 1250x2500x200, very simple but robust Z, Z and gantry and bearings that will support the impact hammer. yes, i am still into the metal sheet forming idea. Atr least i would like to be able to try, without breaking the machine.

Basically with the design i achieved a 100mm/3kw/ spindle to have least possible overhang- the bit center from the gantry at ~ 80mm only. The spindle will be packed in aluminum rectangular of a kind. For the impact hammer a shell with the same dimension will be fashioned, so they could be easily interchanged.

I wanted roller blocks on the gantry and everywhere, but will be extremely expensive, and as i am not completely sure for the metal forming success, i decided on 2 roller rails +4 blocks on the Z and all else Hiwin 20 HG. For the sake of vibration and overbuilding i decided to put everywhere else 3 bearing blocks instead of 2. That will make everything rigid enough to withstand continuous vibration and at the end 4 blocks more from china will cost me only 150 euros more, which is irrelevant if the experiment succeeds or fails. having in mind that one roller blocks costs 100eu and 1 meter rail is double the HGR 20 rail price.

The roller blocks and rails for the gantry i found used for 150eu with shipping, so its a good deal.


Now the ball screw on the Z. Obviously with this extremely low overhang there is no place to fit the ball screw. So i fit it aside when i discarded the idea of third profile on the gantry, due to weight. So for the sake of vibrations of the impact hammer, i decided to fit 2 ballscrews. I believe even with one will work if not the hammer. As they are extremely cheap, i want just to be sure and overbuild here

What i want to say that if i had not made any big error, this design is quite sweet looking for a sturdy DIY machine.Take away the 3rd bearing blocks and the second ballscrews from the design/ due to the impact hammer/ and here you have quite simple but sturdy Z with only 80mm overhang to center of router bit from gantry. I am happy.

Any critique will be appreciated. Its not finished but i wanted to share.

PS. Dont worry for the holes placement and assembly, its under control :-) . Some minor changes will be made to the design so everything can fit in place without problems.


1098610987109881098910990

EddyCurrent
13-12-2013, 10:38 PM
Silyavski, I'm liking the Z design but I worry the whole thing will shake to pieces once the hammer gets going.

Boyan Silyavski
13-12-2013, 11:03 PM
Silyavski, I'm liking the Z design but I worry the whole thing will shake to pieces once the hammer gets going.

I really hope it would work. If you knew how much time, money, effort, sacrifice and Love went to that project, you will understand why i will take the risque.

If interested look here at this dissertation (http://www.dspace.cam.ac.uk/bitstream/1810/238665/1/Soren%20E%20Maloney%202010%20PhD%20thesis.pdf) . I know it by heart of course, but from page 53 onward is the interesting stuff. On page 55 you can see what machine they used and later the tool and the result. I mean if they did it like this, i should be able to do it even better with mine.

EddyCurrent
14-12-2013, 02:50 PM
I know you want a machine to do everything but I was thinking about this cheap 2 axis idea.
Hammer mounted in a steel frame, the hammer can pivot.
A linear actuator connected to the top of the hammer can pivot it to the required angle.
The base rotates under the hammer.

10991

This is not the one I would use but is just an example
http://www.ebay.co.uk/itm/18-STROKE-LINEAR-ACTUATOR-/170630116591?_trksid=p2054897.l4275

Boyan Silyavski
14-12-2013, 03:21 PM
I know you want a machine to do everything but I was thinking about this cheap 2 axis idea.
Hammer mounted in a steel frame, the hammer can pivot.
A linear actuator connected to the top of the hammer can pivot it to the required angle.
The base rotates under the hammer.

10991

This is not the one I would use but is just an example
18" STROKE LINEAR ACTUATOR | eBay (http://www.ebay.co.uk/itm/18-STROKE-LINEAR-ACTUATOR-/170630116591?_trksid=p2054897.l4275)

Eddy,
Thanks!
This is a very fresh idea. I would say the best idea till now.
It still will have to have Z and rotating axis. But this great idea eliminates the gantry and bearings blocks there. It will be best to use screw jack servo motor or similar instead the actuator.

I see a lot of +++:
Only 2 cheap big bearings will support the Z, where the hammer is, the enclosed structure will be more rigid and so on....

It seems doable. I will start working on this design and definitely will bring it to life when i have the big CNC machine ready.

By the way i have my eyes on big THK steel slew ring roller bearings in ebay. For the shell mount rotation. Will dig deeper these days. They ask like 500eur for 60cm diameter bearing assembly which is incredibly cheap. Though the one that suits me seems not to have mounting holes for the actual ring. Will investigate further.


However the Big question for me here is - How i would control that thing? What software? This is quite unclear to me.

EddyCurrent
14-12-2013, 03:31 PM
Just thinking very quickly about control, it may not be good enough for your requirement though. (trying to keep it cheap and simple)
Forget software.

The base would be a simple motor and gearbox
The screwjack would be simple motor operated.

The base would rotate at a constant speed.
At each full turn it will trigger a switch that will move the screwjack a small amount using a timer to run the motor.
2 limit switches will indicate when the hammer is at each end of it's travel.

OR

There will be one motor that drives everything.
The base and screwjack are connected by belts and pulleys to provide the correct speeds.

As the base rotates the hammer will slowly spiral into the centre.

OR

A VFD will control the screwjack
A simple motor will rotate the base

A tacho generator will be connected to the base so that it's output will be proportional to the rotation speed.
The output from the tacho will feed the reference (0 to 10v) of the VFD.
As the base rotates the hammer will move a small amount to spiral into the centre and if the base speed changes so will the screwjack speed.

Boyan Silyavski
14-12-2013, 03:34 PM
Just thinking very quickly about control, it may not be good enough for your requirement though. (trying to keep it cheap and simple)

The base would be a simple motor and gearbox
The screwjack would be simple motor operated.

The base would rotate at a constant speed.
At each full turn it will trigger a switch that will move the screwjack a small amount using a timer to run the motor.
2 limit switches will indicate when the hammer is at each end of it's travel.

It should be programmable. There are certain moments that the hammer needs to spiral towards the center, otherwise waves of material rize due to accumulated tension and at a certain moment everything will wrap if not controlled.

EddyCurrent
14-12-2013, 04:10 PM
Mach3 can do 2 axis so G code could control it and now I've seen the power of Mach3 Brains there's all kinds of things you could do.

EddyCurrent
14-12-2013, 11:01 PM
The hammer is now mounted on a gimbal and can be made to work in a circular motion by moving the X and Y axes.

No need for a rotating base so it can be made cheap and strong.
Mach3 can operate the X and Y axes.

10992

OR easier

A small X Y gantry mounted on top moves the end of the hammer so 50mm there would be maybe 300mm at the other end of the hammer.
Then Mach3 could be used in the normal way where you would use a 2D software to draw a reduced size pan with the required toolpaths.
A pneumatic proportional valve working on 0-10v (like a VFD) could be used to control air pressure to the hammer using M codes e,g, M3 to turn on the air
then for example S500 would tell the valve to deliver 5Bar, S200 = 2Bar , etc.
So the S codes could be put in the G code as required to change the hammer power.

10993

Boyan Silyavski
20-12-2013, 11:12 AM
Now back to reality. I need some help here please! With motor calculations. I read numerous threads here and at the zone, have done numerous calculations as far as i could but will shamelessly ask here again as i am stuck.Almost finished the drawings. The Z looks quite nice and now i am sure its possible, though some blocks will look like cheese due to the ridiculous hole placement:redface:

I though i had everything clear until yesterday i calculated the weights.

The Z weights 32kg and the gantry in total 120kg :black_eyed:

Part of the problem lies that i don't know what exactly i want:hysterical: as till now i have experience only with small weak machines.

1. I would like the gantry and the Z to be rigid as they are in the drawings
2. I would like the machine to be in the middle between resolution and speed. What i am not sure is what speeds i want/not rapids-cutting speed/. I will cut wood and aluminum on that machine. I chose 1610 to move the z left right and 2510 screw to move the gantry. I thought of 1:1, now i am thinking of 2:1, otherwise not only i am jumping in the so called "servo territory" but will have to jump in the "expensive big servo territory"

I have decided on the 3kw spindle. I am not sure but i believe i will be happy with sturdy heavy machine that could do 250IPM /6350mm per min/- cutting in wood at good if not full cutter depth. The weight should help with the depth and i assume with the aluminum finish. Aluminum i will be happy if i do it at flute diameter depth. It seems the machine will be sturdy enough for this, having in mind the very extremely low overhang with my design, enclosed frame spoindle and strong gantry +the quantity of bearings.

3. I thought of driving the Z with hybrid closed loop Leadshine Nema 23 3Nm, same the Z left right and 2x the same that move the gantry.

That was when i did not know about the actual weight.

Now the question is:
-do i make my machine smaller / which i would not like much, cause the money will be almost the same and missing eventual job on the long run could be very undesirable/

or

Keep it like this and figure how to drive it.


So definitely i need some help in this mess in order to evaluate options and make no mistake here.

The servos i am looking at are these: my friend Fred's BST store (http://www.aliexpress.com/store/314742/search?categoryId=100006834&SearchText=servo%2Bmotor)
He seems to have the lowest priced servos and drivers around and as i will buy the rails and screws from him, seems quite logical choice

Questions:
having in mind that i don't want to compromise one axis due to small savings of $ and ruin the integrity of the machine. Obviously on the long axis i will go with servos. And here BST store (http://www.aliexpress.com/store/314742/search?categoryId=100006834&SearchText=servo%2Bmotor) llloking at the prices of servos i am almost ready to fit servos on all axis, as they are almost the price of the Leadshine closed loops

1. The Z weight will stay the same. 10kg spindle + some aluminum, there is nothing more to do except lower the weight about 5-6kg which i believe is not worth it . That is clear. RM1605 screw . Should i stay with 3Nm leadshine Nema 23 closed loop motor 1:1, change the relation 2:1, or upgrade the motor to servo?
If servo, what reduction and how big the servo motor?

2. Z weight again same. RM1610 screw. Should i stay with 3Nm leadshine Nema 23 closed loop motor 1:1, change the relation 2:1, or upgrade the motor to servo? If servo, what reduction and how big the servo motor?

3. Gantry 120kg :black_eyed:, 2510, rotating ball nut both sides.
options i see:

- drive it with 2 servo motors, one each side. I am lost what power i need here. According to what i calculated if i am not mistaken could go with 2x 6N.M 1.8KW 3000RPM 110ST AC Servo Motor 110ST-M06030 + Matched Servo Driver / 310euro each . i am not sure though if the reduction here should be 2:1? Or if with 2:1 reduction i could go with smaller servos like the 4N.M 1.2KW 3000RPM 110ST AC Servo Motor 110ST-M04030 + Matched Servo Driver, where the price is 304euro, so as you see there is almost no price difference .

-drive it with 1 big motor and 25mm belt which i am almost sure i can fit to pass through the gantry beam
for a motor i believe the 130ST-M10025 servo motor and servo driver system 2.6KW 10Nm 2500rpm 10A / 375 euro/ would be ok. Again not sure if the ratio should be 2:1 as i assume.

4. Or i stay with closed loop steppers, but they arent cheap, i almost decided on servos.


5. I can not find what size in Nema are these servo motors so i could continue with the drawings, in fact no info on these cheap servos

PS. Found the documents at last for the servos PDF specs (http://www.knd.com.cn/knd/upload/accessory/20122/2012220141540788030.pdf)
it seems they are Xinje Servo drives and motors here PDF for the drive (http://www.xinje.hu/letoeltes/szervo/46-ds2-manual/download.html)s

110531105411055


What are your thoughts guys?? Please, don't be shy :redface: , tell me where i am wrong


here are some updated cheese drawings:

110471104811049110501105111052

Jonathan
20-12-2013, 01:04 PM
Answers to questions:

1) Ballscrews give a very large mechanical advantage. You do not need to worry about the weight of the Z-axis so much. See here (http://www.mycncuk.com/forums/diy-cnc-machine-building/4705-stepper-ballscrew-loading.html#post31916). You'll be fine with a generic 3Nm motor and 70V stepper driver. The closed loop ones are a good concept, but cost a huge amount more for little gain.

2) Don't use RM1610 - it doesn't gain anything.

3) I wouldn't go with the long belt idea, especially when using servos as servos work best when they are rigidly coupled to the load - i.e. without a long stretchy belt.

You may well be fine with two stepper motors - have you tried putting the values in this spreadsheet (http://www.mycncuk.com/forums/faqs-problems-solutions/1524-what-size-stepper-motor-do-i-need.html)?

If using one servo, the formula you need is this:

Je=M*(L/(2*pi))^2

Je is the equivalent inertia, L is the pitch of the screw (so 0.005m with 1:2 ratio) and M is the mass of the gantry.

Next find the inertia of the servo motor rotor (Jr), e.g. it says in the datasheet for 110ST-M04030 is 7.61kg-cm^2. You then need to find what 'inertia ratio' the motors will tolerate - typically it might be Je/Jr<5. So just use the above formula to calculate Je, divide it by Jr (7.61kg-cm^2) and if the result is less than 5 the motor is probably suitable.

e.g, M=120kg, L=0.5cm (1:2 ratio):
Je=120*(0.5/(2*pi))^2 (http://lmgtfy.com/?q=120*(0.5%2F(2*pi))%5E2)=0.76kg-cm^2

Je/Jr=0.76/7.61=0.1.
0.1 is much less than 5... so that servo is plenty big enough. Another way to look at it is rearrange the formula to find the mass a particular servo will drive, e.g:

M=Je/(L/(2*pi))^2
Je=Jr*5=7.61*5=38.05
M=38.05/(0.5/(2*pi))^2=6000kg(!)
Or if the ratio is 1:1, that's
M=38.05/(1/(2*pi))^2=1500kg
(So note the quadratic relationship - if you double the ratio the drive-able mass goes down by a factor of 4)

The rated speed of that motor is 3000rpm, so the gantry would go at 3000*0.01=30m/min with 1:1. You need to check the Je/Jr ratio - Je=Jr gets the most efficient energy transfer, but the drives do tolerate a mismatch so you need to find how much, as going with them equal generally results in an oversize motor. You should also include the inertia of the pulleys in the system - just work out their inertia (assuming cylinders is near enough) and add them to the Je value, remembering to take into account the drive ratio.

Either way, you can see that that servo is far bigger than you need... so I suggest you put the formulas in a spreadsheet and find which motor is best. Also try the stepper motor spreadsheet I linked to earlier, as I expect you'll find that there are suitable stepper motors available.

4) If you use stepper motors, and select the correct size motors, then they wont stall unless you crash the machine. The only time the closed loop servos help therefore, is if the machine crashes. For the same amount of money as the closed loop stepper you could get a much higher torque standard stepper motor and be certain that it wont stall.

Jonathan
20-12-2013, 01:31 PM
I forgot to mention...I'm not keen on the placement of your Z-axis ballscrew. If a large force is applied parallel to Z, the axis will begin to skew (i.e. rotate in the Y-Z plane), due to the ballscrew placement being off center. The effect can be reduced by increasing the spacing of the Z-axis linear bearings, or go back to two ballscrews, or get the ballscrew closer to the center. That's why on the new machine (http://www.mycncuk.com/forums/router-build-logs/6484-sufficiently-strong-machine.html) I placed the ballscrew intersecting the Z-axis ram - it got it closer to central, but admittedly still not ideal.

EddyCurrent
20-12-2013, 11:39 PM
Jonathan, good reply.

Bart
26-12-2013, 04:10 PM
Hi,
my first helping post here but I was reading for quite some time...
silyavski (http://www.mycncuk.com/forums/members/silyavski.html)
here is part of my excel spreadsheet that might be useful for you- - just type in your variables in green boxes and you should get all the date you need.
11071
I do not guarantee correctness of the results but I'm fairly confident that they are correct.
It's slightly more complicated then most spreadsheets that I've seen here so if you have any questions ask here.

Boyan Silyavski
26-12-2013, 04:21 PM
Hi,
my first helping post here but I was reading for quite some time...
silyavski (http://www.mycncuk.com/forums/members/silyavski.html)
here is part of my excel spreadsheet that might be useful for you- - just type in your variables in green boxes and you should get all the date you need.
11071
I do not guarantee correctness of the results but I'm fairly confident that they are correct.
It's slightly more complicated then most spreadsheets that I've seen here so if you have any questions ask here.

Thanks!
P[W] stands for calculated power per motor for X or for both motors?

Boyan Silyavski
26-12-2013, 05:19 PM
Some updates:

Motors:
Thanks for the help! Now things are more clear there.

The main option i will go most possibly is 4 servos x 400w or 2x400w +2x750w.

Obviously from the calculations if 2:1 then 2x200w+2x400w will be sufficient, but as the price difference between 200w,400w and 750w is merely 5-10 euro, i dont see a reason why not going properly bigger.

The Hybrid stepper option i discarded due to price similar to servos.

DC Servos and motors from Leadshine are almost the price of AC servos and motors , so 220v seems a lot better choice

Steppers are still an option, except for the long axis that moves the gantry. However i would like the Z also to be closed loop, due to the vibrations. Plus the price of decent driver and motor and PSU is almost as the price of AC servo, hence the servo decision.

Another point there is i am starting to doubt about the cheap servos, so i am trying to get for the same price 4 x ACH750 Leadshine AC Servo motor drives + motors. Bargaining with Chinese people takes time :-)

The Z axis

The single screw design i am almost 99% sure will be without a problem. Hoewver i decided not to risque due to the impact hammer head and changed the design. Lost something of the beautiful simplicity and acquired annoying amount of pulleys and idlers, but i could not design it better for now. Best would be with 2 stepper motors on 1 drive, i know,not a very good idea, but works/my flimsy cnc is that way on the gantry without any problem, but is flimsy, not rigid like this design/
So, for now that's how it would be:

1107211073

Rotating ball nut
At the end i found somebody help me with making it. Jeff, from the forum, who happens to live very near, here in Spain. Wow, what a work shop and machines he has, especially compared to my garage.

Now as i decided on Servo, may be even will go with double ball nut/all over the machine also/. 1.450kg said Fred yesterday, weights the 2510 double nut. I wonder what the inertia will be, he he.

I thought of integrating it into the gantry, but at the end decided to fix it aside, due to the possibility to lower significantly the screw length+ avoiding an accidental hit of the gantry to the screw mount

11074


Breakout board
-ESS and PMDX126 seems to be the first choice. What i don't like of this combo is the fact that i have to buy additional spindle control board

So i investigated quite a bit even looking at some second hand Galil motion boards on ebay.

Another good option seems the KFLOP. In fact it seemed ideal/8 axis they say +usb/ until i found that i have to know C+ and that half of its IO are 3v so i have to spend another 240$ for accompanying IO board.

What i need is very simple: cheap ethernet digital 8axis board :joker:

A question ?????
There is something i don't understand about boards and servos. Why all speak of the encoders to go in the BOB? Dont they go to the servo drive and the servo drive controls the closed loop? I read at Galil and KFLOP that the BOB closes the loop? Or am i mistaken?

EddyCurrent
26-12-2013, 07:27 PM
A question ?????
There is something i don't understand about boards and servos. Why all speak of the encoders to go in the BOB? Dont they go to the servo drive and the servo drive controls the closed loop? I read at Galil and KFLOP that the BOB closes the loop? Or am i mistaken?

I would always expect encoder feedback to go directly to the driver, you don't way any delay introduced, it should go directly to the summing junction that should also exist in the driver.
If feedback does go via the bob on any system, I would steer clear of it.

Bart
26-12-2013, 07:36 PM
P [W]
Stands for maximum (instantaneous) mechanical power of a motor when going at full speed and utilizing full torque.
This wont happen in real life as with stepper motors you'll have a safety margin of 1/2 so you will get only half of that power from a stepper motor. It's also important to bear in mind that you will utilize half of this power only in two instances. First instance, when you're accelerating to full speed at full allowable acceleration during fraction of a sec just before reaching max speed. Second, when cutting at full speed with full allowable force. Otherwise you're using only fractions of that power.

Cutting a steel is done between 1 and 2 m/min so if your Vmax is 10 m/min than you will use only 10-20% of Pmax and only if the gantry had to be pushed with full allowable force which is unlikely to happen.

In a cell where you input a torque of motor it was required to input a combined torque of motors for given axis. Now you just input a torque of motor and number of motors.

I also see that you are interested in power of motors, so I've attached one more spreadsheet inside to calculate various mechanical powers of motor, as well as some information on power and current drawn from power supply - to help choose PS.

11075

I don't know much about application of servo motors but from what i briefly read, the most important feature of a Servo Motors is a possibility to overload them up to 10 times which is not possible with stepper motors due to the nature of a driver.

If this overloading feature is possible than consider following.

For a 400 W servo motor that you've selected (assuming an overload factor of 5) you can get as much as 2000W for a fraction of a sec.
Now, if you open my spreadsheet (the new one) all the data there are for my design.
Equivalent Gantry Weight 290kg
Motors: 2x 1,5Nm -> maximum power of 150W per motor but steppers have to be used with some safety factor so as not to loose steps.

I've taken 1/2 so I'm Using only 75W of each motor. With that torque/power I can accelerate 290kg gantry to 10m/min (393 inches/min) in 55 mili sec. That's quite fast but with your servo and 2000W you will be able to accelerate it to the same speed in as little as 4,5 ms over a distance of 0,38 mm, ceteris paribus.

So what I want to say is: don't over-invest in motors.

Best regards

Clive S
26-12-2013, 07:38 PM
I would always expect encoder feedback to go directly to the driver, you don't way any delay introduced, it should go directly to the summing junction that should also exist in the driver.
If feedback does go via the bob on any system, I would steer clear of it.

Or perhaps to the motion control board (driver) .. Clive

EddyCurrent
26-12-2013, 08:17 PM
Or perhaps to the motion control board (driver) .. Clive

Clive, sorry I'm not sure what you mean ?
To clarify, by driver I meant motion control board. Feedback should go to that and not via the bob.

Clive S
26-12-2013, 08:32 PM
Clive, sorry I'm not sure what you mean ?
To clarify, by driver I meant motion control board. Feedback should go to that and not via the bob.

You are 100% correct of course, I was just trying to differentiate between a simple BOB and a full blown motion controller (driver) type.

Sorry for any confusion. ..Clive

EddyCurrent
26-12-2013, 08:51 PM
You are 100% correct of course, I was just trying to differentiate between a simple BOB and a full blown motion controller (driver) type.

Sorry for any confusion. ..Clive

I thought I'd had too many 'old speckled hens' :beer:

Jonathan
27-12-2013, 12:41 PM
Clive, sorry I'm not sure what you mean ?
To clarify, by driver I meant motion control board. Feedback should go to that and not via the bob.

There's actually a lot of debate on this topic when using LinuxCNC. Since LinuxCNC is real-time, you can implement the PID controller within the computer (instead of the motor drive), and have the computer output speed instead of position (step/dir) signals. That can work well, especially if you use a PCI card. Some people consider that better as then the computer is always tracking the following error... but just saying that is really oversimplifying the situation so if interested, Google it.

Boyan Silyavski
29-12-2013, 12:10 AM
After reading these days a lot of motion control manuals, i think i have it more clear. The majority of them have their own brand daughter boards and that's the main reason it seems that they can close the loop directly on the board, for servos and steppers. Also most of the expensive ones have possibility to input 2 encoder signals on 1 axis, so they can control backslash this way. Also it seems connected to some external servo drives, even if the servo drive closes the loop, it can transmit it to the board at the same time, for the previous reason or for some extra functions.

Now i just cant believe how little choice there is for 5 and up axis BOBs and combinations, especially to work with mach3. I almost discarded the Galil. Read in forums there is no support at all if you don't buy it new. Which is crazy for a product that is worth thousands. Also 99% at ebay sell them on As Is basis, but not for As Is money. I dont see my self paying hard earned cash for a board that is not guaranteed at least a week from the sale. Also read that many of them are possibly tweaked for a certain customer, so even less support. And later to have problems with the Mach3 plugin...I was tempted. And still i am.


Its unbelievable how poor the choice of BOB is today. Couldn't be that more clear what is needed:

-USB or Ethernet board,small sized, mach3 plugin, on board spindle control, on board 220AC/transformer, opto isolated inputs and outputs, 6-8 axis,4 relays,...

Combining all of these is almost impossible and starts to get expensive. Same price as second hand motion controller :soap:

EddyCurrent
29-12-2013, 11:29 AM
Silyavski, I know this is not what you are after but may be of use.

Got one of these the other day, great little unit and great price.

5V 1/2/4/8 12V 16 Channel Relay Module W/ LED Indicator Light For Arduino | eBay (http://www.ebay.co.uk/itm/5V-1-2-4-8-12V-16-Channel-Relay-Module-W-LED-Indicator-Light-Arduino-/360813650488?pt=UK_BOI_Electrical_Components_Suppl ies_ET&var=&hash=item92bad01216)

also these spindle boards work good and not too high a price
System4 (http://www.diycnc.co.uk/html/system4.html)

JAZZCNC
29-12-2013, 03:15 PM
Its unbelievable how poor the choice of BOB is today. Couldn't be that more clear what is needed:

-USB or Ethernet board,small sized, mach3 plugin, on board spindle control, on board 220AC/transformer, opto isolated inputs and outputs, 6-8 axis,4 relays,...

Combining all of these is almost impossible and starts to get expensive. Same price as second hand motion controller :soap:

If you don't need slaved motors and using steppers then have a look at the Csmio-IP-M controller or the Csmio-IP-S if you need slaved motors or using Servos with high count encoders and want the full package. Here CS-Lab - CSMIO - CNC Ethernet Controllers- inexpensive solutions - CS-Lab CNC controllers CNC Machining CNC software (http://www.cs-lab.eu/en/).

They are very high quality Ethernet Motion control units with industry standard 24V differential digital I/O so very good at handling electrical noise also come with very High quality Analogue spindle output built in. (Sorry to say this Eddy but those Spindle 4 boards are very erratic with speeds jumping around all over the place, used a couple now and not impressed.:thumbdown:)

No BOB or external Spindle board is needed all mounts on industry Standard Din rail with separate quality Din rail connectors. Simple to fit and install plus very very neat.

I've used just about every Motion control card now in the Diy range so Sub £300 and I can tell you 100% these knock the spots off any other.

I've fitted 2 now but only on new build machines so couldn't compare same machine performance against other Cards but unfortunately and very very frustratingly for ME that's just changed.!!
The Russian PLCM E3P I've been using on my own machine has just gone faulty, doing strange things and tripping with noise issues from VFD, either using USB or Ethernet connections.

Couldn't have happened at worse time has I've been flat out chasing my arse between Family health issues and getting stuff done plus now every where is closed for Xmas. Just lent my spare ESS smooth stepper to a friend who isn't answering his bloody phone, just sold the spare spare USB SS has never used it. . Arghhhh . .(PLCM E3P is now smashed into 1,0000000000000 bits.!!! and NO it gave me no satisfaction other than to vent.!)

No way was I going back to Parallel port, plus laptop doesn't have PP. So I've got Csmio-IP-M which was half fitted into a control box destined for machine currently !!.Trying.!! to finish but needs must so out it came temporarily until I can get another Card.

WOW . . . 20mins later it's wired up, software installed and working with full spindle control.
Worked Straight off Copy of my Old Mach3 profile with just a quick Plug-in swap and pin assignments with very little tweaking of motor tuning required. . . In fact this is what really impressed me has now can tune the motors much higher even thou the PLCM-E3P and CSMIO_IP_M are both supposed to be 100Khz.

I haven't had or got much time to play and don't need the speed anyway but the first dabble took my My Y axis from Max 11mtr/min & 950s/s Accel to stable 14Mtr/min at 1200s/s and that's with 5mm pitch screw direct drive on 3.5Nm Nema 34 motor at 75v. That's 2800Rpm from Nema 34 motor.!!!. . I was quite impressed with 2200rpm the PLCM allowed.!!

The Motors so sound very sweet even at these dizzy heights and the pulses are clearly very smooth has the movement is excellent even at very low feeds.
Wouldn't ever run at that speed has I don't need it but it just shows not all Cards are Equal and I'm positive this speed increase is down to the quality of the product has it's the only thing that has changed on my machine.! . . .Very impressive and I won't be coming off any time soon so I better get another ordered ASAP.

. . . . . HAPPY NEW YEAR EVERYONE. . . .Keep smiling.:yahoo:

Boyan Silyavski
11-01-2014, 11:05 PM
Lately spend too much time on choosing the BOB, motors and drives, which delayed my order for the ball screws and the rails. After all that talk, last night saw 2 interesting Galil listings on fleabay and something made me push the button and send best offer, thinking that if one accepts, will cancel the offer to the other.
Well, after i woke up this morning i found my self the happy future owner of 2 galil BOBs. i thought of excusing myself to one of them, then decided and payed them both. With the idea if one not working to use the other, or at least to see which one i like best. the other goes later for next machine or whatever.

So i bought the GALIL DMC-1880 PCI 8 axis Optima motion controller and the Galil DMC-2183 Ethernet/RS232 8-Axis Motion Controllers & 2 ICM-20105 Opto I/O .
Both seem to work with the Mach3 plugin / yes i checked before sending the offers/ .

I have some feeling that i made the right decision with these. What do you think?

Now i have to decide on the drives and motors. First of all now i can run steppers with encoder and servo motors directly closing the loop at the motion controller. Which in fact can lower the price of the build, implementing servos only for moving the gantry and steppers on the Z if i decide so.

No i am looking at a Galil AMP-19540 at ebay which coul drive directly 4 dc servos up to 500w each, costs only 100euro and connects directly to one of the BOBs i bought - the DMC-1880 . It seems this purchace can save me a lot of money on other drives and cheapen considerably the project, sourcing only the motors.

However i understand nothing of these type of servo and drives for me the description is like Chinese, i only see 500W :-) , here is from the description:

Galil's AMP-19540 is a 4-axis amplifier for driving brush or brushless motors up to 500 Watts. By interfacing directly to Galil's DMC-18x2, DMC-18x0, or DMC-18x6 controllers, it provides a cost-effective controller/drive solution for multi-axis applications. The AMP-19540 contains four transconductance, PWM amplifiers for driving brush or brushless motors. Each amplifier operates at 18V to 80V dc, up to 7 Amps continuous, 10 Amps peak. The AMP-19540 gain setting is easily configured with jumpers. The PWM switching frequency is 60 kHz.


The AMP-19520/40 is a brush/brushless trans-conductance PWM amplifier. The amplifier operates in torque mode and will output a motor current proportional to the command signal input.

Input Voltage: 18-80 VDC
Continuous Current: 7 amps
Peak Current 10 amps
Amplifier Gain 0.4, 0.7, or 1.0 A/V (jumper adjustable)
Switching Frequency 60 kHz (up to 175 kHz available--contact Galil)
Minimum Load Inductance 0.5 mH (low inductance option available)
Communtation Angle 120o (60o option available)


So, can some one help here please:

Can i find easily BLDC motors suitable for the drives, with hall sensors as far as i understood, new from ali express or from ebay? Are these drives any good, mean in 2014? And for my build?

Boyan Silyavski
14-01-2014, 12:41 PM
Silyavski, I know this is not what you are after but may be of use.

Got one of these the other day, great little unit and great price.

5V 1/2/4/8 12V 16 Channel Relay Module W/ LED Indicator Light For Arduino | eBay (http://www.ebay.co.uk/itm/5V-1-2-4-8-12V-16-Channel-Relay-Module-W-LED-Indicator-Light-Arduino-/360813650488?pt=UK_BOI_Electrical_Components_Suppl ies_ET&var=&hash=item92bad01216)

also these spindle boards work good and not too high a price
System4 (http://www.diycnc.co.uk/html/system4.html)

Thanks Eddie,

I will grab one of these, just have to figure how many relays i need.

Now that i have bought the Galil 1880 and have only 1 x 100 pin cable going out of it, i have to fashion something like a terminal breakout board where to connect all the I/O signals.
Any idea what cheapo components will fit best together? I am a bit lost here and dont have the money to buy the galil I/0 board. The connecting cable it seems is SCSI 100-Pin High-Density connector, though not completely sure.



Here below are some specs from the manual (http://www.galilmc.com/support/manuals/man1800.pdf)of the motion control/ Chapter 3 Connecting Hardware/ :

The DMC-1700/1800 provides optoisolated digital inputs for forward limit, reverse limit, home, and abort signals. The controller also has 8 optoisolated, uncommitted inputs (for general use) as well as 8 TTL outputs and 8 analog inputs configured for voltages between +/- 10 volts. 1X80 Controllers with 5 or more axes have 16 optoisolated uncommitted inputs, 8 TTL inputs, and 16 TTL outputs.

Wiring the Optoisolated Inputs Bi-Directional Capability. All inputs can be used as active high or low - If you are using an isolated power supply you can connect +5V to INCOM or supply the isolated ground to INCOM. Connecting +5V to INCOM configures the inputs for active low. Connecting ground to INCOM configures the inputs for active high. INCOM can be located on the DMC-1700/1800 directly or on the ICM-1900 or AMP-19X0. The jumper is labeled INCOM. ....


From what i understand everything important is opto isolated. The servos i bought seem to accept various signals, still have to fight with the manual to translate things like:

Position command can correspond to the following four types of output form.
Line Driver Output
+5V Open Collector Output
+12V Open Collector Output
+24V Open Collector Output
- Line drive output may correspond up to 450 kpps, and open collector output may
correspond up to 200 kpps.

It seems that the drives accept 5v signal, but recommend external resistor and 24v signal, for interference resistance.

Still have to figure what is line drive and what is position command, it seems to me line drive could be step and direction ans position command should be differential signal.

irving2008
14-01-2014, 05:36 PM
Post up a link to the servo manual and we may be able to give more guidance

Boyan Silyavski
14-01-2014, 06:11 PM
Thanks!
I discarded the Chinese servos due to some worry that will be difficult to tune. Decided on Mitsubishi servos with very good specs, from fleabay. But could not find a good looking ones, or were expensive/ yeah, sellers know/, or were incompatible with mach3 /again sellers know and try to fry you/ .

Yesterday i payed for 4x SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2 and CSMT-04BB1ANT3 motor 400W. Absolute encoders 2500, auto tuning, max 450khz pps, in and out encoder signals, and so on.


Here is the manual:http://www.cnczone.ru/forums/index.php?act=attach&type=post&id=8233 section6.1 and on

my head hurts from reading 10 different servo manuals :-)

Boyan Silyavski
25-07-2014, 01:51 PM
Hi, long time i put this off, but now that successfully finished my first build (http://www.mycncuk.com/threads/6457-Sturdy-and-Fast-all-Steel-CNC-my-first-build/page12) i am back on the track with my machine build.

Now i have all the electronics, motors and major part/linear rails and ball screws/ sleeping at my garage. This time i will build it much faster.

Having learned a lot from the first build, i reconsidered many things and am trying to finalize the drawing stage, so i can buy the steel profiles next week.
Many changes were done, simplifying, strengthening and all this having in mind the common materials here in Spain.

1. Decided that the machine will be on no long legs. It will rest on the floor.
the benefits:
-I work alone so lifting heavy materials will be a problem in the future. As it will be multi functional, there will be removable plasma torch and 1000x1000x100mm removable water bed also, for the not so rare cases i need some parts from steel done until i make my self proper plasma table.

This is how it will look:

12893

2. I wanted to avoid as much as possible laser cut pieces/$$$/ and use instead the profiles themselves as i would do the cutting at home which saves me quite a lot in fact. Only 3 types of steel will be used-100x100x4/the darker ones/, 100x100x3/ the lighter gray/ and 60x60x4/ the gantry verticals and reinforcement pieces/



The gantry support legs from 100x100x4 / so i can reach and screw the HIWIN blocks to the plate/:

12894 12895



The ball screw end supports from 2 pieces of UPN100 soldered together. I have the believe that this part is very important to be extremely sturdy in all directions as there wil be very heavy pull there, starting from the tightening and later when the gantry accelerates and so. Additional 2 100mm thick plates for reinforcement,one bellow and one for the face of the beam, so there will be no further doubt there :-).

12896


3. The rotating ball screw nut
Attached to 10m steel plate via screws , the plate is soldered directly to the gantry sides. I couldn't help and add 2 ribs to the plate, just in any case.

12897

basically the 30T HTD 5 pulleys are 90mm center to center, so 330m belt will fit according to this calculator http://www.bbman.com/catalog/belt-length-calculator.html
The problem here is that the servomotors have much larger central cylindrical part, so there is no more that +-3mm up and down for adjustment. that means the calc has to be very exact with servo motor, as oposed to steppers where i achieved +-6mm adjustment in both directions.


12904

The other thing is that i can not find Alu pulleys HTD5 30t. Only steel, i am not sure but it could impact the inertia as its 3 times more heavy.

4. Back to the gantry. In a way i am proud of it :joker:
Now the gantry is where i took liberties with the design, hopefully for good. As i said before there is extremely small overhang. Unluckily there was no 120x100 profile available, though when i started the drawing they said that there is . So i redesigned it again. 100x100x4 and 60x60x4. As i started the gantry from scratch and put a lot of thought in there, i believe it will not only be very successful but in the future will serve as an inspiration to all that want the benefit of closed aluminum gantry but don't have the means or money and go with steel. I mean - take away half of the reinforcements and vertical reinforcements /remember-i will mount impact hammer/ , and it still will be heavy duty normal gantry, couple this with the gantry sides like i did them, and you will have cheap very nice gantry that you need a way to cut the profile and solder +2 plates for the HIwins.
In short I believe that's the ideal High Heavy Duty DIY gantry. I hope you will like the design idea. Hidden and short screw, motor inside, all cut from profiles including the sides. In a normal HD variant, there will be less vertical profiles, less support triangles and one less profile each side of the gantry sides.


12898 12899

FYI : On this picture bellow you can see that the distance i left from front to the vertical supports is 100x100 -60x60/inside verticall supports/ = 40mm. But if the ball screw supports have to be mounted on the vertical bars, no additional plates as i made them, then the distance must be 43mm. Even so the vertical supports will protrude at the back only 3mm which can be masked by soldering, so 100x100 and 60x60 is ideal combination for HD Gantry

12900


This is how i envision a normal HD gantry, that will be comparatively very light for what it is and no need to go in the servo territory. I would like to hear critics or opinions on this :-):

12901 12902 12903


5. The Z
This is on what i am checking and working today. have to check again if the cheese holes are right on spot.

As i said above, with servos i can not design a plate with more than 3mm of movement in both directions to tighten the belt. So i had to do something i dont like and split the Z motor plate to be able to additionally adjust the belt. Still have to calculate again the exact length. And so the days go by...

Will update the Z pictures, when i finish counting cheese holes :-)


12905


6. The machine bed

Some changes here. Instead of vertically soldering 3 x 2 /100x100x3/ profiles i decided to lay them all in one plane, so the distances between them are very small-200mm. So the bed is really strong, even using cheap bed material.

12906




Quite a long post, but i am excited :butterfly:

John McNamara
25-07-2014, 02:34 PM
Hi Silyavaski

You have been busy with CAD, nice drawings, what cad package are you using?.
And great to see you have assembled the parts and drives....
How do you intend to join all the steel sections and the vertical bracing? By welding?
-
From memory you are going to use this machine with an impact hammer to make musical instruments.
-
Will it be an air operated hammer similar to one of these? https://www.google.com.au/search?q=air+hammer&source=lnms&tbm=isch&sa=X&ei=elfSU4GALcTy8QWeoIKACQ&sqi=2&ved=0CAYQ_AUoAQ&biw=1215&bih=817
-
I was using a small one hand one last week to break up some concrete. they cam deliver quite a blow, You can turn them down by using both a pressure and flow regulator. the settings will depend on what sort of material you are acting upon. even at a low setting the energy released is significant.
-
Tubular steel is very strong no doubt, however it can vibrate, particularly if the hammer is set to oscillate near the resonant frequency of the frame. Without very sophisticated FEA software it is hard to conceptualise just how a structure will behave. however you may find it advantages to consider the issues involved in advance and if vibration turns out to be a problem have a means of increasing the mass of the structure and what sort of damping measures you take to reduce the amplitude to an acceptable level.
.
Here is a link to an excellent paper on concrete filled steel structures used in machine design. http://www.mech.utah.edu/~bamberg/research/PrinciplesOfRapidMachineDesign/Principles%20of%20Rapid%20Machine%20Design.pdf
-
Regards
John

Boyan Silyavski
01-08-2014, 08:16 AM
Hi,
Now the question of the day is having in mind all the servos are 400w, do I drive all 1:1 or 2:1 ??? The x y encoders are 2048 and the z is 2500 ppr ,all the screws are xx10 and 1605 for the z
The servos will accept 100000 pps at step mode

Boyan Silyavski
13-08-2014, 06:50 PM
Last revision. Can not find a weak spot that is not bombproof. Starts to look more and more like a mill :frog:.

1. Decided to drivel all 20:30t, only the Z is 20:20t. This after exausting 3 day calculations, belt lengths,inertia, tooth engagement, speeds, etc. and at the end -common sense.
This is a separate thread of its own. I believe this the wisest decision between speed, acceleration, precision.


2.I decided to redesign the Z as it bothered me, especially with this long 1400mm belt. At the end common sense also helped here. At the end raised a bit the overhang without compromising anything so much. I design like this-until i see perfection at every detail.

So the Z. Tried to design it like a Russian tank from WW2. All plates are 20mm Aluminum.

- it would be covered against dust from all sides

13068 13069

-The cheese :hysterical: . The additional holes at the front are for mounting extra things, like: 3D plastic head, plasma torch, engraver, light and so on. The front plate on the moving box is separated in 3 parts. The lowest one is for the case when a so high object is machined that the spindle has to be all in, hence from there access to the nut of the spindle. the second part allows exchange of the spindle. The upper one stays there forever.

13070

-The linear movement and where is placed. I even decided against using it for the hammer. the idea of making a precision machine and the trying to brake it seems less and less attractive. So this one will make the other which will be purpose build 5-6 axis, but anyways as i have 2 ball screws i decided to fit them.

13071 13072 13073

-Reinforced the motor plate, this will hold the drag chain too. The belt now is very short as i like it, no extra pulleys and 6-7 teeth engagement each ball screw pulley. So seems ok to me.


13074



The rotating ball nut assembly. 20:30t hence 1.5:1 driven. Compromise due to common sense. At the end precision and power at the same time is the aim.

2048 encoder on the motor, so more or less 20m/min. I see on the motor specs that it can go up to 4500rpm, just the curve is straight till 3000rpm. The resolution will be ~0.003mm on all axis , 0.001mm on the Z/ 1605 screw and 2500ppr encoder/. That without any artificial gearing or micro stepping which is not needed it seems. Thats good, cause the Galil board is not micro stepping board.

13075 13076

Boyan Silyavski
13-08-2014, 07:07 PM
Hi Silyavaski

You have been busy with CAD, nice drawings, what cad package are you using?.
And great to see you have assembled the parts and drives....
How do you intend to join all the steel sections and the vertical bracing? By welding?
-
From memory you are going to use this machine with an impact hammer to make musical instruments.
-
Will it be an air operated hammer similar to one of these? https://www.google.com.au/search?q=air+hammer&source=lnms&tbm=isch&sa=X&ei=elfSU4GALcTy8QWeoIKACQ&sqi=2&ved=0CAYQ_AUoAQ&biw=1215&bih=817
-
I was using a small one hand one last week to break up some concrete. they cam deliver quite a blow, You can turn them down by using both a pressure and flow regulator. the settings will depend on what sort of material you are acting upon. even at a low setting the energy released is significant.
-
Tubular steel is very strong no doubt, however it can vibrate, particularly if the hammer is set to oscillate near the resonant frequency of the frame. Without very sophisticated FEA software it is hard to conceptualise just how a structure will behave. however you may find it advantages to consider the issues involved in advance and if vibration turns out to be a problem have a means of increasing the mass of the structure and what sort of damping measures you take to reduce the amplitude to an acceptable level.
.
Here is a link to an excellent paper on concrete filled steel structures used in machine design. http://www.mech.utah.edu/~bamberg/research/PrinciplesOfRapidMachineDesign/Principles%20of%20Rapid%20Machine%20Design.pdf
-
Regards
John


Thanks for the link.

Here is the comparison picture with the 3.2kw spindle. It weights next to nothing, compared to the spindle, which is a monster.

13083

here i started to put the electronics together. No power cable in front, all go back shielded by the alu plate. Later all in metal box, due to the occasional plasma jobs.

13082




The program i use is Sketchup, cause i am very fast at it. trying to learn also a proper Cad also at the same time. /NX9/.

TonyD
13-08-2014, 11:47 PM
Can't wait to see this come together, keep up the good work. :)

JAZZCNC
14-08-2014, 08:10 AM
I see no point in having 2 ballscrews to drive the Z axis.? Whats the thinking behind this.
Your just making the motors work harder and increasing costs for no practicle gain.? You'd be much better spending the money on a second ballnut and creating Zero backlash.!!

Boyan Silyavski
14-08-2014, 08:38 AM
I see no point in having 2 ballscrews to drive the Z axis.? Whats the thinking behind this.
Your just making the motors work harder and increasing costs for no practicle gain.? You'd be much better spending the money on a second ballnut and creating Zero backlash.!!

Hi Dean,

Yeah, what i was thinking, you got me there . I started the design half an year ago, and learned a lot during this time.The thing is that i wanted not only to have 2 ball nuts but also 2 extra ball screw supports, " 4 is better than 2 " to resist the vibration. Thats why the double ball screw. Also the double ball screw saves on the total Z length.

Now i am not so sure. Still its not too late. Now i see that i could have made the ball screw machined so that it accepts double support at the ends. I have an extra ball screw at home, longer, i mean. Then the only thing that needs changed will be to buy 2 longer Hiwin rails to compensate. Have to rethink that. There is still time.

Now going this way of thinking i start doubting if i should not make the whole z axis 100mm longer and have travel of 300mm instead of the 200mm as its now. Could just raise the vertical bars on the sides of the base with 100mm without compromising any integrity. They are ~80mm high right now. the thing is would i really need that extra 100mm and for what?

John McNamara
14-08-2014, 08:41 AM
For The Z or any other axis you cannot rely on a ballscrew to take any lateral load (at all) only a pure as possible long axis load is permissible.

You will need a pair of rails to carry the load. One ballscrew is fine sized according to the long axis load carrying capacity you will find on the manufacturers specification. Bear in mind the impact hammer will apply a large dynamic load to the ballscrew, You may need to use a higher rated screw to cope with this. I would contact the screw manufacturer and ask for advice.

When setting the rails and ballscrew up they must be perfectly collinear or you will be applying a forbidden lateral load as the paths interfere with each other.

Regards
John

JAZZCNC
14-08-2014, 08:49 AM
How does 2 screws same length give extra travel.? Edit: Ok re -read under stand now 2 ballnuts in parallel is less length than 2 nuts in series.! But still bad idea as it will be quite hard to ensure preload is correct to give zero backlash.

Unless your spinning the ballscrew at very very high speed then at this length vibration shouldn't be problem if supported at both end. If your really bothered then use fixed bearings at both ends and tension the screw.
Personally I would just use fixed at both ends with tension and double ballnuts, this will be more than enough.

John I don't think Boyan is trying to use ballnuts to increase strength just lessen backlash and vibration/whip on screws.!

John McNamara
14-08-2014, 09:01 AM
Reflecting on the dynamic load applied to the ballscrew by the Impact tool.
For many reasons I would prefer to use timing belts instead of ball screws. They do not suffer brinell pitting damage
This search came up with some useful links.
https://www.google.com.au/#q=brinell+pitting

I built this router with timing belts (The Z axis only in this machine is a ball screw but that could be changed)

I know the budget is finite but would it be possible to make a belt driven machine Or two machines One for impact work and one for Routing? Maybe sharing the electronics?

This Is the Router I built with a mate using timing belts for a non profit Men's Shed. http://www.machsupport.com/forum/index.php?topic=23730.0 There is a small video. It works really well and is used by the group so much there is a waiting list.

Regards
John

Boyan Silyavski
14-08-2014, 09:14 AM
For The Z or any other axis you cannot rely on a ballscrew to take any lateral load (at all) only a pure as possible long axis load is permissible.

You will need a pair of rails to carry the load. One ballscrew is fine sized according to the long axis load carrying capacity you will find on the manufacturers specification. Bear in mind the impact hammer will apply a large dynamic load to the ballscrew, You may need to use a higher rated screw to cope with this. I would contact the screw manufacturer and ask for advice.

When setting the rails and ballscrew up they must be perfectly collinear or you will be applying a forbidden lateral load as the paths interfere with each other.

Regards
John

Thanks John,

-Yes, we are talking all the time of long axis load. There will be no side forces or if there will be, they will be smaller than say cutting aluminum.

- the dynamic load of the hammer is not big at all, as the head of the hammer will be small, anyway, the hammer can not produce such a force as one can suppose.

-The collinearity will be way under specifications of ball screw or rails manufactures. I can easily make them collinear to something like 0.00mm



How does 2 screws same length give extra travel.? Edit: Ok re -read under stand now 2 ballnuts in parallel is less length than 2 nuts in series.! But still bad idea as it will be quite hard to ensure preload is correct to give zero backlash.

Unless your spinning the ballscrew at very very high speed then at this length vibration shouldn't be problem if supported at both end. If your really bothered then use fixed bearings at both ends and tension the screw.
Personally I would just use fixed at both ends with tension and double ballnuts, this will be more than enough.

John I don't think Boyan is trying to use ballnuts to increase strength just lessen backlash and vibration/whip on screws.!

Ok, thanks for the input.


The design as its now and i have the elements already at home, implements that the both screws have fixed bearing supports at both end.

So you say there will be no sense to make the design as you suggest with one ball screw, 2 nuts and 2 and 2 fixed bearing supports at both ends? I mean only one fixed end will be enough? The bearings doesn't seem very big to me.


Reflecting on the dynamic load applied to the ballscrew by the Impact tool.
For many reasons I would prefer to use timing belts instead of ball screws. They do not suffer brinell pitting damage
This search came up with some useful links.
https://www.google.com.au/#q=brinell+pitting

I built this router with timing belts (The Z axis only in this machine is a ball screw but that could be changed)

I know the budget is finite but would it be possible to make a belt driven machine Or two machines One for impact work and one for Routing? Maybe sharing the electronics?

This Is the Router I built with a mate using timing belts for a non profit Men's Shed. http://www.machsupport.com/forum/index.php?topic=23730.0 There is a small video. It works really well and is used by the group so much there is a waiting list.

Regards
John


I know about the ball deformation. It doesn't bother me though. The hammer head will be half 10mm diameter circle . The air in the hammer dampens the hit, the material is very stretchy too/1-1.2 mm normalized sheet/ , the whole sheet stretches also, not only at the point of hit. More or less the hit will equal to 1 bearing ball hitting the material. This would be spread to all the bearing balls in 2 ball screw nuts. I am not sure how many are they. means that its not like a hit on hard rigid surface


I was contemplating this belt driven. What i want right now is the following. To finish the machine as faster as possible. To try the hammer thing at least once. to draw conclusions and not break the precision of the machine in that process. To use the machine right away to make a new one, simpler, with homemade ridgid elements, belt driven as you say and most importantly 5-6 axis. I almost have all elements for a new machine, apart of this one.

JAZZCNC
14-08-2014, 09:27 AM
So you say there will be no sense to make the design as you suggest with one ball screw, 2 nuts and 2 and 2 fixed bearing supports at both ends? I mean only one fixed end will be enough? The bearings doesn't seem very big to me.

You miss under stand me I think.? . . . . . I would use 1 x screw with 2 fixed ends and tension the screw and if backlash was important then I'd use 2 ballnuts.
Correctly sized screw will have no issues with this setup if the forces are small like you predict.

Boyan Silyavski
14-08-2014, 09:58 AM
You miss under stand me I think.? . . . . . I would use 1 x screw with 2 fixed ends and tension the screw and if backlash was important then I'd use 2 ballnuts.
Correctly sized screw will have no issues with this setup if the forces are small like you predict.

Dean thanks for your time!

I started already redesigning it to go with one screw and 1 or 2 nuts as you suggested. That will save me for a beer or two. I understood you correctly, but messed with my English.

What i meant is if there would be additional benefit if i fix that screw both ends with double fixed end supports, meaning 2 supports each side chained together and then tightened by the nut at the end.

John McNamara
14-08-2014, 01:28 PM
I have seen Pre tensioned ball screws in large CNC machines, Also many where one end is fixed in preloaded angular contact bearings While the other is allowed to float in to allow for changes in temperature. The floating setup may be a less difficult setup to use, particularly if the frame is lightweight. If you pre tension you are imposing a substantial load on the frame, It would have to be stiff or it may be deflected out of alignment by the preloading stress, and the deflection would be likely to vary with temperature.

mitchejc
03-09-2014, 11:14 PM
Silyafski, I really don't know the first thing about making those drum surfaces but the other day I saw this video and thought if this kind of forming process might not work for you as an alternative to hammer forming the sheet metal.
https://www.youtube.com/watch?v=eh0ddqsx0KI&list=UUEXQkHxUA8P7aRhF_WqV0Ug

Boyan Silyavski
03-09-2014, 11:18 PM
Silyafski, I really don't know the first thing about making those drum surfaces but the other day I saw this video and thought if this kind of forming process might not work for you as an alternative to hammer forming the sheet metal.
https://www.youtube.com/watch?v=eh0ddqsx0KI&list=UUEXQkHxUA8P7aRhF_WqV0Ug

Exactly. Thats one of the reasons i made such strong Z and the cheese holes on the Z. I will not push the spindle bearings, instead will fit it somehow directly to the boxed Z.;

Boyan Silyavski
06-09-2014, 11:16 PM
Some updates. 2 very hot weeks, especially if you are working alone in 70% humidity, 40C temperature and soldering all day long +weekends and moving heavy stuff.


Had to clean all stuff from the garage. Draw white lines for the footprint. This machine will take a lot of space. Of course it rained right after i took everything out, did i tell you it hadn't rained like 11 months.

13315




Started with the gantry. decided to keep sides square, so later could flip and support easier the gantry. Do you see the blotch further on the floor. Guess what, just before starting all my boxes with paint fell on the floor and some spilled. A little diversion :-)


13316


I have mixed feelings about the Evolution saw. While the motor and blade is great/made in japan/ the saw is very very flimsy. Next project will be to throw away the base and do a new one. I wish i have bought the Makita. It was 100eur more expensive and don't come with blade- another 100, so thats why. Anyway, an essential machine for the project.

13317

Ventilator to cool myself and change of plans. The 60x60 profiles cut to size that were meant to support the rails were discarded. There was no way i can solder them from all sides. Instead i cut all of the 60x10 plate/6m in total/ and decided to separate them by the 60mm using at least the profiles i have already cut. Some carefull soldering from here and there. Very rigid result.

13318 13319 13320 13321 13322


As you see from the pictures above everything is soldered till the end. Now there was another momentary decision. No legs for now untill i decide something clever and the steel from the legs went just inside the sides of the machine. So stair like sides with no holes, there lie the rests of the profiles meant for legs.

13323


I cant explain how strong that machine frame feels. Including that bestial gantry. Vertical bend or twist- i did not even bother to measure it with the indicator. Its solid as a rock. starts to look like a Frankenstein thing from movie.

So today i finished working at 20.30 and couldn't help but move all elements with last available effort, as i wanted to see how it would look. I am a strong man, no lift or whatever help, so i am becoming stronger moving all day these pieces around while soldering. Did i tell you that all the soldered pieces weight around 500kg .


So here is my baby. Its huge. I am a bit surprised how big that is. Still has no name but i think it would be something military:

13324 13325 13326


So after ~15kg of wire, 2 CO2 bottles and some crazy days i start to see how will look. Still have no color in mind.

GEOFFREY
06-09-2014, 11:37 PM
looking like a really solid lump!!! How about drab (a sort of olive green) and "big bertha"? Well done. G.

EddyCurrent
07-09-2014, 10:11 AM
How about drab (a sort of olive green)

I remember it was Mid Brunswick Green that was the preferred colour year ago.

Clive S
07-09-2014, 10:52 PM
So here is my baby. Its huge. I am a bit surprised how big that is.I have found you a ball screw to fit it:-

13332

You might need bigger servo's though. ..Clive

Boyan Silyavski
09-09-2014, 12:15 AM
:hysterical: where did you find that ball screw picture?

Olive green is a good idea. Some pieces black or and dark grey. I tried with the other green Mid Brunswick Green, but did not like it
Unfortunately someone took the name big berta (http://bigberthacnc.blogspot.com.es/) . But lets make it alive, then the name.

Today started soldering the bed frame and bought the epoxy. As i stopped working at all except on the machine, money flows like water through my fingers. So i bought at the end the Spanish epoxy with very similar properties like the WS. from choice of three i choose the most similar. There were seemingly better though some data was not clear, from the one that could concern, i mean. 125eur with shipping for 5kg pack +1.6kg hardener.

Basically 90 min gel time, max isotherm 40C, compression strength 580kg/cm2, yellowish, contraction 0.1%, density 1.08 gr/cm3 which roughly translates to 6.5l so ~20eur /lit mix, I cross my fingers that it works ok.

13341


It seems that the machine will lay on the floor with no legs. have to figure where to fix some hooks for the future and most importantly some plates to lift it using car jack. To level it for the epoxy pouring. Cause now it is unliftable already.

Clive S
09-09-2014, 07:46 AM
:hysterical: where did you find that ball screw picture? I made it on my router with the 4th axis.:highly_amused: ..Clive

EddyCurrent
09-09-2014, 10:42 AM
el martillo de Thoryavski

Boyan Silyavski
15-09-2014, 11:32 AM
Today started soldering the bed frame and bought the epoxy. As i stopped working at all except on the machine, money flows like water through my fingers. So i bought at the end the Spanish epoxy with very similar properties like the WS. from choice of three i choose the most similar. There were seemingly better though some data was not clear, from the one that could concern, i mean. 125eur with shipping for 5kg pack +1.6kg hardener.

Basically 90 min gel time, max isotherm 40C, compression strength 580kg/cm2, yellowish, contraction 0.1%, density 1.08 gr/cm3 which roughly translates to 6.5l so ~20eur /lit mix, I cross my fingers that it works ok.

13341


It seems that the machine will lay on the floor with no legs. have to figure where to fix some hooks for the future and most importantly some plates to lift it using car jack. To level it for the epoxy pouring. Cause now it is unliftable already.

Now going cheap was really bad idea. Ok, it would have saved me 100euro on this size of structure...

But here is what happened:
The epoxy poured extremely well. It leveled well. And for 4 hours it was perfect. As gel time was 90min, i thought well, i cheated destiny. But no. After the 4rth hour it shrank in volume and there was not 20mm wide flat area in the center to munt the rail. I did the test 50mm wide channel, instead of the 40mm normal with WS 105+209. Read a lot why and it seems the reason is that the epoxy is low quality and contains non reacting solvents that vaporize when the hardening starts. On the long axis it was straight, so its not a mixing problem, anyways i mix it at least 5 minutes and change the can before pouring, so not my mistake. Repeated twice the test-same result.

So i took a risque and lost. My reasons were not that i am greedy, but the price difference multiplied by the future machines is quite some money. However i just returned the 5_1.6kg epoxy so the loss is about the cost of shipping, the time and 1 lit of epoxy.

Conclusions:
1.Seems the biggest unaccounted problem for that purpose is the shrinkage, in other words the purity of the epoxy. Hence the price of good epoxies :-). Otherwise flatness and surface quality is more a result of good mixing and following exact mixture to the gram and temperature when pouring.
2. Judging epoxies from their technical sheets means nothing. It can seem on paper right but poor in reality
3. Epoxies with very high viscosity >700 cps of mixture maybe contain something that will evaporate later

Left on the picture is WS 105+209 at the time of pouring and cured. Also the spanish epoxy untill 4h. Right is the shrinkage when 8h passed

13381

At the moment of pouring till 4h

13382

The problem after 8h of curing, a bit difficult to be seen but the drawing above shows it better. On the second picture the shrinkage lines can be seen at the reflection distortion . Not that it could not have worked as we are talking about 0.3mm here and on the long axis was straight, but not worth the additional fiddling.

13383 13384

Neale
15-09-2014, 10:37 PM
I don't think that you can blame the quality of the epoxy. It did what it was designed to do. I think that epoxies containing solvents are typically used for coating, so that they spread well - which is what your epoxy did! They are definitely not recommended for use as adhesives or fillers, mainly because of the shrinkage as the solvent evaporates. However, I didn't know this either until a couple of days ago when I read an article on the use of epoxies for adhesives and coating! Something to watch out for if anyone is looking for an alternative to the usual named brands.

Boyan Silyavski
16-09-2014, 10:43 AM
I don't think that you can blame the quality of the epoxy. It did what it was designed to do. I think that epoxies containing solvents are typically used for coating, so that they spread well - which is what your epoxy did! They are definitely not recommended for use as adhesives or fillers, mainly because of the shrinkage as the solvent evaporates. However, I didn't know this either until a couple of days ago when I read an article on the use of epoxies for adhesives and coating! Something to watch out for if anyone is looking for an alternative to the usual named brands.

Hi Neale,
This is specifically a casting epoxy, its not a coating epoxy. It specifically says 0.1% shrinkage in its technical sheet, and by no means its a cheap epoxy , its simply cheaper than imported WS epoxies.
I also read a lot of documents about epoxy and all say the same- 3 types of shrinkage, and exactly this type is due to non reacting solvents. This is not shrinkage due to temperature cause it happens after the initial hardening. From the company started to argue with me, saying this is not possible, it never happened, until the moment they understood that i know a lot of sh^t about epoxy and they agreed that this was because of the cheap solvents. That was the moment they agreed to break their policy of no return, once its opened.

Blackrat
16-09-2014, 06:39 PM
What about adding a filler like ATH ?

ba99297
16-09-2014, 08:22 PM
Hi Neale,
This is specifically a casting epoxy, its not a coating epoxy. It specifically says 0.1% shrinkage in its technical sheet, and by no means its a cheap epoxy , its simply cheaper than imported WS epoxies.
I also read a lot of documents about epoxy and all say the same- 3 types of shrinkage, and exactly this type is due to non reacting solvents. This is not shrinkage due to temperature cause it happens after the initial hardening. From the company started to argue with me, saying this is not possible, it never happened, until the moment they understood that i know a lot of sh^t about epoxy and they agreed that this was because of the cheap solvents. That was the moment they agreed to break their policy of no return, once its opened.
Syliavski i thought that a solvent free epoxy is a must at this procedure. Why you choose to buy an epoxy that contains solvent ( or i missaunderstood something?)

Vagelis

Boyan Silyavski
16-09-2014, 09:02 PM
What about adding a filler like ATH ?

I thought to do it , only had graphite at home. But truth is i want it perfect, so bought the West system epoxy at the end.


Syliavski i thought that a solvent free epoxy is a must at this procedure. Why you choose to buy an epoxy that contains solvent ( or i missaunderstood something?)

Vagelis

I chose to buy epoxy sin solvents but they sold me one with solvents :-)

ba99297
16-09-2014, 09:28 PM
You mean the solvent was in the epoxy before you make the mix?
Good for you that they admit their mistake and you finally get your money back.

Boyan Silyavski
14-10-2014, 03:26 PM
Hi,

The build is going on but i am stuck with epoxy problems. Mean BIG problems :beaten:.

Second time casting epoxy on the long sides and same problem. The area near the connecting bridges is sunk checked against straight edge. about 250mm wide visible inclination, not measured cause still not perfectly hardened but say 0.1-0.3mm . Each pour cost me 100euro :topsy_turvy: 2.5kg of WS epoxy mix.

Now for a woodworking machine this will not matter, but spending so much money on it i would like to be as Perfect as possible. Hell, at least 0.05 precision or better, real life, not exaggerated. I was striving to something like the first machine i made 0.01mm, knowing that on such a big scale it would be difficult but worth to try.

So to achieve this i depended too much on the epoxy.

Whats the f---g problem then? Everything is perfectly prepared. 21C in the garage, 45mm wide dams, 10min careful epoxy mix, all by the letter. First time i thought the dam was flimsy and contracted at certain points. Not so the second time, everything is reinforced with steel/the bridges/ and hot glue / the 10x10mm dams.

I gave it a lot of thoughts and the conclusion is : the 1600mm long bridges are the culprit. The epoxy contracts , so the bridge epoxy contracts and takes away epoxy from the rails at the exact moment when its solidifying. cause i looked 100 times after 1 hour of pour, 2, 3 hours of pour -it was ok. It happens on hour 4 to 6. Or maybe it was my mistake that the connecting bridges had to be both ends, so when all contracts , contracts equally, like a ring i mean. No money though for further experiments . I believe though that even if ring is formed the right angle between them is the culprit.

So yes, there is a problem with 3000mm rails and 1600mm distance between them. Inherent problem i mean. Not reported by any body till now but hey, did anybody who made such a big machine take care to measure it against DIN 874 straight edge? Possibly no.

I blamed the air trapped, blamed the supports, blamed this and that, but on a smaller scale i did not have any problem before, +fixed all possible doubts. So yes its inherent, until somebody prove me wrong.

I see only 2 possibilities from here on.

1. Will wait to dry for 3-4 days and mount the rails using the straight edge/ 1m/ and shim. Then will have to buy the 2000mm straight edge to continue with the build and check the rails against each other. But if there is a twist , 2000mm straight edge will not be enough or at least would be very difficult to straighten things up. To the precision i would like, i mean.

2. I knew it when pouring second time, but was not sure.:tan: . So may be third pour from scratch or half pour over this one. And on minute 45 / 21C/ cut the bridges simultaneously by some previously prepared method. Say a sliding plate or similar. The way in actual water channels the water flow is cut. Cause at this moment everything is perfectly leveled .


http://www.mycncuk.com/attachment.php?attachmentid=13608&stc=1

Clive S
14-10-2014, 05:31 PM
Would you mind stating which WS system you used. ie the hardener ..Clive

EddyCurrent
14-10-2014, 05:49 PM
I said here that my cross channels were 10mm wide but they should have been wider, I think that's part of the problem, maybe the channels should all be the same width.
http://www.mycncuk.com/threads/6565-Ready-Steady-Eddy?p=54474#post54474

Boyan Silyavski
14-10-2014, 07:41 PM
Would you mind stating which WS system you used. ie the hardener ..Clive

WS 105+209 of course.


I said here that my cross channels were 10mm wide but they should have been wider, I think that's part of the problem, maybe the channels should all be the same width.
http://www.mycncuk.com/threads/6565-Ready-Steady-Eddy?p=54474#post54474

Mine are 30mm wide. On the small machine they were 10mm and it worked. Maybe at such lengths should have been 45mm also, same as the rails channels.

If i had 3000mm straight edge i would have told you exactly what happened. Checking with the 1000mm one says the area near the cross channels is down.

Just measured it at the deepest spot 0.15mm difference.

I am starting to doubt if i should have semi filled the epoxy with something. It would have helped. And yes, WS 105+209 shrinks like crazy on 3000mm length. Visibly. You know that last time i fixed 5mm magnets to check levels. So it covered the magnets and when dried, one magnet was half out, so on 5mm thick/high all went down 0.5mm at least which is more than 10% if you know what i mean. The good thing is that it shrank equally wide wise. Not like the cheaper epoxy. But still 5kg pack
200euro...

PS. and no, it did not went down in between cause it was done so that it couldn't do it, so that's not the problem

jimbo_cnc
15-10-2014, 02:06 AM
Never done anything like that, but I'd have thought you'd want to:

1. As you say, cut the bridges. You'd do this as soon as any levelling flow has stopped.

2. How about cutting the 3000mm length into sections as well?

jimbo_cnc
15-10-2014, 02:11 AM
Someone should design a machine to level rails. Maybe screw a strip of aluminium to the rail top, and then have a travelling surface cutter that takes it's height from a laser (a rotating disco laser so both rails are surfaced to the same laser defined plane)

Tinkerer
15-10-2014, 06:21 AM
Silvaski, sorry to hear you've hit this snag, particularly after your success on the first machine. In an ideal world Jimbo is right, playing with epoxy is a necessary evil, which is better suited to machining. I'm beginning to wander whether I should weld the top half of my frame with rail beams and investigate whether there are machining services around to do the levelling. Knowing my luck I will have wasted several export batches which could go towards machining cost! I'm sure, the cost and availability of this alternative would likely dissuade me from this. The problem is you need a machine to level it, and the irony is that once finished, the machine you have built may be able to perform the task at hand!

Certainly your post and experience make me Dolby cautious when I get to this stage.

EddyCurrent
15-10-2014, 09:35 AM
Never done anything like that, but I'd have thought you'd want to:

1. As you say, cut the bridges. You'd do this as soon as any levelling flow has stopped.

2. How about cutting the 3000mm length into sections as well?

Daming the bridges after a time might be a good idea but you can't section the 3m sections otherwise it won't be level along the length.

mekanik
15-10-2014, 09:41 AM
Eddy's right has to be done in one pour, how about forget the bridge mix equal quantities of epoxy and pour both channels, when set find the highest rail and shim the lowest to match ?
Just a thought
Mike

HipoPapi
15-10-2014, 09:45 AM
Don't go to hard an your self Boyan. I find your dam idea to be a great one. The only drawback is that you lose the connection between the two sides hence they become two separate world's. I will definitely use this dam technic and will try to use something as thin as a raiser blade. Maybe cut open some beer cans and make some stripes of it. I think a thicker dam material could introduce a small unwanted wave. And as you seem to be nuts with precision you won't want that. :)
Sorry for your second fail but thanks for sharing as it is informative. Now I know it requires 21 degrees Celsius and in my case I see a problem there. In fact it even gives me some confidence as the result you have achieved is better than what I aim for in my build do to the nature of application.
Good luck with your next pour tho, if you go for one.

EddyCurrent
15-10-2014, 09:48 AM
As jazz has said before they don't necessarily have to be the same height but they do have to be level and in the same plane. Provided this criteria is met, when you skim the bed, this puts things right, and obviously the spindle must also be at right angles to the bed.

Neale
15-10-2014, 09:50 AM
Is there anything to be learnt from casting techniques? Arrange a large sacrificial reservoir close to the rails by widening the bridge locally but which can be cut away afterwards? However, I suspect that the shrinkage happens as the epoxy starts to gel, by which time it is not going to flow anyway. And what's happened to the "negligible shrinkage" property that we're told epoxy has?

I have a vested interest in this as I shall be putting epoxy on my own rails shortly...

EddyCurrent
15-10-2014, 10:11 AM
Remember these videos cncJim posted ? they talk about priming the job first with a mini pour.
http://www.mycncuk.com/threads/7190-Epoxy-pouring-advice?highlight=epoxy+video

Boyan Silyavski
15-10-2014, 11:14 AM
Eddy's right has to be done in one pour, how about forget the bridge mix equal quantities of epoxy and pour both channels, when set find the highest rail and shim the lowest to match ?
Just a thought
Mike

It will work, but at a later point the gantry have to be squared. It will make things extremely difficult then.





I am starting to think of something like that:

http://www.mycncuk.com/attachment.php?attachmentid=13611&stc=1 or http://www.mycncuk.com/attachment.php?attachmentid=13612&stc=1



One problem of cutting the bridge is that if i shut them off, that has to be complete, to avoid pulling. And at the same time any spill will kill the purpose.


There must be a way to do it properly, unluckily i dont know well the material/epoxy/. Sb who works all day with epoxy will know how. cause the key may be simply to make the bridge same, wider or even more thinner to avoid the pull. But which of them. When casting aluminum i know that the hole has to be as wide as possible, cause when all shrinks, the casting holes is like a reservoir , so it pull aluminum from its core and balances the thing.


Or may be even like this, so the rails will be thinner than the bridge and situated so that they will pull from it and play its role:

http://www.mycncuk.com/attachment.php?attachmentid=13613&stc=1




Now the question is if i decide to shim the 0.15mm, what to use, so that when the 150kg gantry goes over that it does not change. Mix epoxy when mounting and paint the bottom, then wait to dry a bit and tighten a bit when its elastic, so it stays put in that shape.

jimbo_cnc
15-10-2014, 11:34 AM
Trying to balance the flow with the shrinkage while it hardens sounds like a recipe for disaster.

Pour, flow, cut off, harden!

Cutting the rail into equal size/shape sections should give consistent shrinkage. If you do repeated bi-sections to cut the rail (eg into 4 or 8) then (with skill/slow cut) even the slight increase in pool height should be even in each pool.

edit:
More likely to work and easier: One big ring of equal cross section is the best answer. Ideally you would not have a 90deg corner near what will be the final end of the epoxy rail, a radius would be nice.

http://www.mycncuk.com/attachment.php?attachmentid=13618&stc=1

EddyCurrent
15-10-2014, 11:44 AM
Would this type of arrangement work, placed at the ends of the bridges where they meet the rail channels ?
http://ocw.mit.edu/ans7870/2/2.25/assignments/sec5/5-3/index.html
I'm thinking that when the epoxy is thin it will find the same height each side but as it hardens it will be more difficult for it to flow back into the bridge.

cncJim
15-10-2014, 11:46 AM
Hi silyavski,

Sorry to hear about your problems.

What about adding some nails/pins to the bridge before the pore? Could this maybe help the situation?

13615

Jim

JAZZCNC
15-10-2014, 12:01 PM
Not enough bridges IMO. I've only ever done one machine near that size with epoxy and I had 5 bridges with 2 people pouring epoxy.

To late know for you but for others like tinkerman this is why I don't like epoxy for large machines and use the adjustable top rail beam then shim as needed.

jimbo_cnc
15-10-2014, 12:04 PM
Another thought:

Are the bridge deeper than the side pools? If they are the same depth then the shrinkage should not cause flow, but if the bridge pools are deeper, then they will get inflow like you experienced because they will slump more.

Narrow deep bridges will suck less from the rails than wide deep bridges. So narrower would seem a good choice.

And all this assumes vertical walls.

HipoPapi
15-10-2014, 12:25 PM
I don't think painting the bottoms would do the job if you try to fix it. But how about leaving the channel sides in place or put some new ones on and after you mount the rails you pour a bit more to fill the gaps. And probably paint them as well just to make sure it gets in there.

jimbo_cnc
15-10-2014, 12:33 PM
How thin can you pour an epoxy rail? I've not used the stuff or seen a video. I 'm wondering if due to the above argument they should be done in 2 stages.

mekanik
15-10-2014, 01:20 PM
Hi Boyan
Don't understand this
It will work, but at a later point the gantry have to be squared. It will make things extremely difficult then.
I would have thought that the X rails would be level in both planes and secured to main frame before you started Gantry installation alignment ?
Mike

Boyan Silyavski
15-10-2014, 03:54 PM
Trying to balance the flow with the shrinkage while it hardens sounds like a recipe for disaster.

Pour, flow, cut off, harden!

Cutting the rail into equal size/shape sections should give consistent shrinkage. If you do repeated bi-sections to cut the rail (eg into 4 or 8) then (with skill/slow cut) even the slight increase in pool height should be even in each pool.

edit:
More likely to work and easier: One big ring of equal cross section is the best answer. Ideally you would not have a 90deg corner near what will be the final end of the epoxy rail, a radius would be nice.



That could be the proper way, though not understanding the process entirely, i prefer to make it like variant 3 of mine, so the rails finish in a wider bridges. It seems the thinner the bridge the faster it cures even if logic says otherwise, cause all say more epoxy at one place cures faster.




Would this type of arrangement work, placed at the ends of the bridges where they meet the rail channels ?
http://ocw.mit.edu/ans7870/2/2.25/assignments/sec5/5-3/index.html
I'm thinking that when the epoxy is thin it will find the same height each side but as it hardens it will be more difficult for it to flow back into the bridge.

I was having in mind something like this. As i said my worry here is that i must shut off totally the flow and remove the bridge at the moment, so any leak here will be a disaster. cause what if i shut the bridges off and leave them in place. If they harden earlier, they will pull the shut off plate toward them. So the story repeats or the plate have to be very thick. I will think again about that.


Hi silyavski,

Sorry to hear about your problems.

What about adding some nails/pins to the bridge before the pore? Could this maybe help the situation?



Jim


I think of adding gates and provisionally tack soldering steel profile for the bridge this time.


Not enough bridges IMO. I've only ever done one machine near that size with epoxy and I had 5 bridges with 2 people pouring epoxy.

To late know for you but for others like tinkerman this is why I don't like epoxy for large machines and use the adjustable top rail beam then shim as needed.

Before a week i was going to argue with you but now not :hysterical:. Until figure the proper way to do it.

5 bridges...that's a lot of epoxy. I would do it in a blink if i though the problem is there. And that the result would be as desired. The problem seems the length of the bridge. Say a good epoxy shrinks 0.1% on length. 0.1% from 1700mm =1.7mm so if you take 1.7mm from area of 200mm long this is exactly the missing 0.15mm epoxy

Do you remember depth and width of channels and bridges, especially length of the bridges and if all checked well against straight edge?


Another thought:

Are the bridge deeper than the side pools? If they are the same depth then the shrinkage should not cause flow, but if the bridge pools are deeper, then they will get inflow like you experienced because they will slump more.

Narrow deep bridges will suck less from the rails than wide deep bridges. So narrower would seem a good choice.

And all this assumes vertical walls.

Vertical walls. Same depth, i was very cautious after the first attempt. I agree that narrow bridges will suck less. I am almost sold on my variant 3 from the previous post, where the rails will suck from the bridges.


I don't think painting the bottoms would do the job if you try to fix it. But how about leaving the channel sides in place or put some new ones on and after you mount the rails you pour a bit more to fill the gaps. And probably paint them as well just to make sure it gets in there.

Considering all this, the very small gap and the time and effort, i just broke everything so will prepare for another pour


How thin can you pour an epoxy rail? I've not used the stuff or seen a video. I 'm wondering if due to the above argument they should be done in 2 stages.

You could pour according to me 3mm minimum. So yes, could be done 2 times, risking other stuff like not good enough spread and leveling due to specific gravity. further complicating things with the wait times. imagine if they each shrink in different way, due to temperature or humidity difference that day.


Hi Boyan
Don't understand this
It will work, but at a later point the gantry have to be squared. It will make things extremely difficult then.
I would have thought that the X rails would be level in both planes and secured to main frame before you started Gantry installation alignment ?
Mike


Both rails must be plain in the same plane. If they are equal in all other but one is higher than the other, no problem, that difference could be fixed with the gantry rails, but then you will have only one rail as a starting point for squareness . So yes, its much better all to be flat. So you have to deal with left right and squareness between the 2 rails when mounting. Cause there are 6 directions to move and fix, so avoiding 2 of them , makes things simpler.



And yes, i insist in this level to be as perfect as possible, cause in the future the plan is to surface other smaller machines inside, gantries and so. By fixing alu plates and then surfacing them. That will save each build 200 or more $$, so yes , its important. Not to speak for a long aluminum mold or so the $$$ i will make or miss, so i broke the epoxy and will do it again. Its pain in the a** but irrelevant to the money i have spent till now.

jimbo_cnc
15-10-2014, 05:25 PM
Vertical walls. Same depth, i was very cautious after the first attempt. I agree that narrow bridges will suck less. I am almost sold on my variant 3 from the previous post, where the rails will suck from the bridges.


damn! was sure I'd sussed it. Very confusing.

Why do they say large areas cure faster? does it generate it's own heat for curing?

If there was room for another 3mm pour on top of the old one, I would be tempted to do that.

Out of interest, what is the deepst and shallowest depth in the 2 rails on the previous pour?

Interesting you mention fixing aluminium plate to gantries and surfacing them. I had that thought today, someone should be selling ready made flat side gantries.

Boyan Silyavski
15-10-2014, 05:47 PM
damn! was sure I'd sussed it. Very confusing.

Why do they say large areas cure faster? does it generate it's own heat for curing?

If there was room for another 3mm pour on top of the old one, I would be tempted to do that.

Out of interest, what is the deepst and shallowest depth in the 2 rails on the previous pour?

Interesting you mention fixing aluminium plate to gantries and surfacing them. I had that thought today, someone should be selling ready made flat side gantries.

Theoretically more volume epoxy cures faster, cause the heat can not escape so fast, heating the mixture so then all cures faster.

WS 105-209 gives 1h gel time, though i found that 1h 1/2 is still ok, up to 2h when is okeyish almost but still not unusable. That at 22C .


As i carefully welded the frame and a couple of times leveled the structure, the depth is 4-6mm . If i do it next time all will be steel, dam 10x10mm welded , channel >45mm wide/cause will hit the bearing blocks other wise. perfect will be 8x8mm bar. Clean, fast, beautifull. No finishing needed and most of all on such a big table water can be applied earlier so level it perfectly. Before pouring i mean.

They sell mounting straight edges but 3m one willbe 3000euro. Of course then epoxy can be applied below and then the straight edge mold it pushing it down. Then remove and do it at the other side.

So i took out the epoxy. Will have to do it again it seems. I almost have it clear. Will update here, so you tell me what you think, meanwhile will search for somebody around who builds boats or something and knows about epoxies first hand, to see if help can be obtained.

What i came till now is very simple. Extend both rails both sides 30cm each, then bridge them. When dry cut the extra 30cm each side and hope for the best in the center :-). Have to contemplate it further but i feel that this will work. And apart from this i will know for sure if i can build one day 5x10 foot CNC :-) with epoxy. Or not.

FatFreddie
15-10-2014, 06:11 PM
Maybe a bit off the wall but have you thought of using plastic tubes between the two sides rather than bridges - these could be shut off with clamps once the epoxy has levelled. it would be an easy job to 3d print the necessary bits.

EddyCurrent
15-10-2014, 07:17 PM
Is there an accelerant for epoxy same as that for super glue ? maybe that could be used to make the bridge epoxy go off quicker ?

ba99297
15-10-2014, 08:33 PM
Syliavski I am sure you will find the solution. Your mind is sharp so I don’t worry. The only think I want to say is that some times we stack and we become maniac with some specs off the machines that we build. I would be the last one to talk here as I haven’t finished a machine yet. That doesn’t mean that I don’t have an opinion. So I think that the solution to your problem is more than obvious. It is #1 that you suggest. Go on with shimming. As I have read here in the forum many people use only shims. Why it would be such a big problem for you to use shims only in a small area ( near the bridges). Also have in mind that all these people who have already use the epoxy method, may never compare the straightness of the epoxy against a straight edge. The first introduction of that method that I saw here in the forum was Jonathans machine. But this machine ( as your first one ) was smaller. The scale is very important. Even a 3000mm straight edge may have (0,01%=0,3mm loss at accuracy ) I don’t want to introduce my self as the clever who knows everything. Most of the people here in the forum wouldn’t imagine the problem that you come up against. The only positive is that you give us a good lesson unfortunately by spending 200 euros. This is life. This is they way we learn, by making mistakes. No one could blame you. I think go on with shims is the best way to continue. For sure there is a solution about the epoxy method at this scale but if you want to find it you must pay. The most common characteristics that the DIYselfers have I think are stubbornness and persistence. We think that we can solve any technical issue and that make us some time loose time and money. It is useful some time to make a step back, see the things more clear and then go on again. I am no talking as an expert or as an experience constructor, I am talking as a friend ( even if I have never see you, you have helped me and that gives me the right to consider you friend ). Keep going

mekanik
15-10-2014, 08:36 PM
Just wondering if this might be worth experimenting with.
http://www.ebay.co.uk/itm/SP300-EPOSEAL-1-LITRE-PACK-UNIVERSAL-EPOXY-PRIMER-SYSTEM-F705-015-/291248298064?pt=UK_CarsParts_Vehicles_BoatEquipmen t_Accessories_SM&hash=item43cfc05850
It's almost like water so could contain a lot of solvent,but might be worth a try, i use it for sealing wood patterns and it takes about a day for any remnants in the mixing container to set solid, how does it compare price wise with the West Epoxy.

EddyCurrent
15-10-2014, 08:41 PM
@ba99297, very nice post :applause: I think we do know most of the answers it's just we don't have access to the right (hugely expensive and massive) machinery. For example if the frame could be put under a huge surface grinder the job would be done without epoxy and super accurate.
It also makes me wonder what happens when the frame is stood on an uneven floor, okay it's very rigid and the adjustable feet would be set as accurate as possible but would the frame move over time ? and by how much ? 0.1 mm ?

Tinkerer
15-10-2014, 09:51 PM
Eddy this is what I have been considering. Cost again could be an issue. A naive question, but I presume the starting point for accuracy is clamping the two top rail beams upside down to an accurate level reference point (precision benchtop) and welding top down (upside down)? I'd imagine working the other way will cause significant misalignment. Is this what everyone has done? I understand it will not get to micron accuracy, but it would aid initial accuracy before contemplating need for machining / epoxy?

JAZZCNC
16-10-2014, 12:07 AM
I've said it a 1000 times.!!! . . . Build in adjustabilty and between this, shimming and patience you will get the accuracy you require.!!
Most don't even know the accuracy they require and most are aiming far beyond there capabiltys to achive what they think the need.! When the reality is they don't need anywhere near the kind of accuracy they think they do.!! . . . . . For instance whats the point building a machine to 0.01mm accuracy for cutting woods that only need 0.1mm accuracy at best.?

Boyan if all you have is 0.15mm error over 3m length then I'd say your got it about as good as it gets with a single action. Now it's time to either shim or fill the low spot and scrape level. The important bit is that you know the 2 rails are on the same plane any surface discrepencys can be filled or shimmed out.
Look for epoxy putty and fill the low spot then scrape or sand it level checking with your straight edge.

IMO your expecting too much from epoxy over this length.!!

jimbo_cnc
16-10-2014, 12:41 AM
Some sensible comments here, bringing us back to reality. Repair the dip, fill, shim, whatever.

What matters is the rail flatness, not the epoxy flatness.

EddyCurrent
16-10-2014, 08:34 AM
That's what I did with mine, I used some 5 minute epoxy to correct a small dip and leveled back using a long straight edge. Once the rails are fitted it's going nowhere.
At the time it was referred to as, "a disaster", but that was wrong, it turned out fine.

GEOFFREY
16-10-2014, 08:38 AM
I think that Boyan has already removed the epoxy. G.

Blackrat
16-10-2014, 08:46 AM
what about heat ? on a hot day the frame is going to expand more than .1mm ?

jimbo_cnc
16-10-2014, 08:52 AM
what about heat ? on a hot day the frame is going to expand more than .1mm ?

It may do, but it will stay flat :)

Boyan Silyavski
16-10-2014, 09:19 AM
Hi, i agree about the reality thing and all comments. So i could have shimmed it easily. However it would not be woodworking machine only or mainly... If this was woodworking machine only i would have finished it already.

As i said before, one of the main purposes of this machine will be to surface under the rails smaller machines + surface gantries and Z. I don't have a mill, no money for a mill or place, so have to be clever and have good CNC capable of doing similar jobs in aluminum. So basically this machine will cut mainly aluminum , plastic and when nothing to do - wood. It will have mounted plasma torch and removable water bed 100mm deep, something like 1 square meter, so i could plasma cut some small parts. Yeah, one of the things that i will do in the future is make small machines by order or at least the important parts for them.

Hence the precision desired, required or dreamed of :-)


Will not change the WS epoxy, now that i am deep in it, to say. Will order another B pack. One more try and then will shim if unsuccessful. But i have to know. Can i or not make 5x10 machine using this method. Thats the real question.

In other words i am having fun. Now that i know that i can shim it. Plus i know that if i waste a ton of epoxy and make 5 bridges or additional long parallel bridge in the middle and interconnect all it would work, cause will simulate 4 small machines cast together. But where is the fun or the excellence in doing it like this. I want to find how with minimum effort will make the best result, its difficult to explain.


See below. Thats how i am thinking of to make the next try. bridges both sides only, no middle. Steel, soldered when dry rails mounted then rests carefully cut by hand . 200mm extension, so if anything bad happens it will be there. Rails will suck this time from the bridge, not like before. 140mm reservoirs extension, i cant explain, seems right to me. ~2.5 - 2.8kg mix/ 40mm wide -45mm wide channel/ .

Previous pours showed that rails length seems don't impede flow at this width and depth/ 5-6mm/ so only 2 bridges both ends. In fact i am almost 100% sure that could be done only with one end bridge, all seems so right, but i don't want to experiment right now with this.

What do you think?


http://www.mycncuk.com/attachment.php?attachmentid=13621&stc=1




PS. about the frame

Yeah i am conscious about the possible misalignment of the machine if it is moved on another premise. But hey, even my small lathe manual says that below should be perfectly flat and horizontal to achieve working accuracy. Its considered when designing the machine to be made so that this is irrelevant. Thats why the triple beams /100x100x4/ both sides-against bend in the middle , that's why the 3 double horizontal beams on bed -against twist. Also that's why the 4 legs are both ends and not say 20 percent inside, cause right now the frame bent what have to bend and combined with the sides and bed strength it will not bend more under 140kg gantry, in other words the gravity stress will work for me. Cause if i have put the legs 20% inside the frame length both sides so the frame would be strongest, when the gantry went one side it would lighten the middle and cause some vibrations.

So the answer is that when moved, legs must be leveled checking long rails against straight edge and even additional supports mildly fixed in the middle so they support, not push up. I told it before on the first build. Speaking about this and that but if you don't have the precision instrument to measure it... IMO a straight edge that could extend perpendicular to both long rails tops is a must for any precision build. I don't see how could be done without this.

EddyCurrent
16-10-2014, 09:25 AM
IMO a straight edge that could extend perpendicular to both long rails tops is a must for any precision build. I don't see how could be done without this.

You are absolutely correct with that.

jimbo_cnc
16-10-2014, 11:31 AM
I'm not sure I agree with your theory for putting the legs in the corners, but we won't worry about that now.

I'm interested in how to make it moveabe and still no twist. Obviously, making it very stiff is a good start. But what you really want is to know the reaction forces with the ground so you can install the machine elsewhere with same contact forces. The reactions don't need to be equal but equal would be sensible.

If you have screw adjusters on each foot, then the torque is an approximate measure of the contact force, so you would carefully go around the machine and even up the torque on each foot adjuster.

You want to do this before your epoxy pour.


Or maybe thick elastomeric pads, where you can measure the deflection as an indicator of reaction force. These might be good for vibration too. What is your plan for fixing to the floor? or does it just sit on the floor ?

Some numbers would be useful. Have you calculated the stiffness of the frame? eg if all the weight is on 2 opposite corners how much dip at the other 2 corners?

GEOFFREY
16-10-2014, 04:09 PM
If you made the centre of the bridges slightly higher than the sides would that not encourage the side epoxy to "draw" from the bridges? Obviously there must still be a sufficient depth to enable leveling. G.

ba99297
16-10-2014, 07:40 PM
Syliavski I dare to say that your last solution is not sure that will work as you imagine. I have never work on research but I think that a basic principle must be to make small steps every time. With your last solution you move the bridges from the middle to the edges. Who told you that the epoxy will be spreads the same way without the two middle bridges. I think 3000mm is too long distance for 5-7mm epoxy thickness. If I were you I will shim or for fill the gap with epoxy and then sand down as Dean suggest. If I had to recast I will use the same arrangement as your last casting try with the difference that I will make the bridges wider near the side tubes( where the rails will rest on). In other words I will put the reservoirs next to the spots that you mention the problem. As the epoxy has the tension to shrink at a specific spot, make your correction at that spot. See below to see what I mean.
http://www.mycncuk.com/attachment.php?attachmentid=13638&stc=1

Boyan Silyavski
16-11-2014, 11:46 PM
Hi there,
Long time no update. Long story short. Bought new epoxy and did it as I thought it would be right. It worked perfectly. As supposed the long rails sucked from the bridge. I went even further and made only 1 bridge. Yeah, the epoxy leveled without problems. I knew it.
This time I used Knauf channel 0.5mm 3m - 3€ from construction material site. Every attempt I am getting better for cheaper.
Also used 8 mm square bar around the epoxy. The channel was 4 mm wide so the bearings not to hit the bar

some photos. Will upload here also but now I am writing from my phone
https://plus.google.com/app/basic/photos/111571569713128538019/album/6082431536953046609?cbp=1b0j3yawglvr9&gl=es&sview=54&cid=5&soc-app=115&soc-platform=1&spath=%2Fapp%2Fbasic%2Fphotos&sparm=cbp%3Dpt4pyhrcs2kc%26gl%3Des%26sview%3D18%26 cid%3D5%26soc-app%3D115%26soc-platform%3D1%26pv%3Dhi%26pgpnum%3D1


Ps. Here are the photos:

http://www.mycncuk.com/attachment.php?attachmentid=13868&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13869&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13870&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13871&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13872&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13873&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13874&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13875&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13876&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13877&stc=1

mitchejc
18-11-2014, 05:33 PM
Wow, those epoxy sides came out very nicely. Also a great idea to use those 8mm square bars on the sides of the epoxy.

That base looks VERY solid and straight. I'm wondering how flat those X sides were before the epoxy, did you perhaps check that? The reason I'm asking is that I'm hoping to get away without having to use epoxy because I'm not getting something similar to the West Systems stuff here so I need a plan B.

Blackrat
18-11-2014, 06:25 PM
ive just done my gantry epoxy pour , i gotta say it works a bomb !! am pretty impressed how flat and smooth it has come out

i however used a ATH filler in my epoxy , makes the epoxy much harder

if you phone John at AMT composites in Capetown , and ask him about ampreg21 with the super slow hardner ...

mitchejc
18-11-2014, 06:48 PM
thx blackrat, much appreciated.

Boyan Silyavski
07-12-2014, 10:01 PM
I'm not sure I agree with your theory for putting the legs in the corners, but we won't worry about that now.

I'm interested in how to make it moveabe and still no twist. Obviously, making it very stiff is a good start. But what you really want is to know the reaction forces with the ground so you can install the machine elsewhere with same contact forces. The reactions don't need to be equal but equal would be sensible.

If you have screw adjusters on each foot, then the torque is an approximate measure of the contact force, so you would carefully go around the machine and even up the torque on each foot adjuster.

You want to do this before your epoxy pour.


Or maybe thick elastomeric pads, where you can measure the deflection as an indicator of reaction force. These might be good for vibration too. What is your plan for fixing to the floor? or does it just sit on the floor ?

Some numbers would be useful. Have you calculated the stiffness of the frame? eg if all the weight is on 2 opposite corners how much dip at the other 2 corners?

i am still not sure. When all is ready and un-leveled, we will see if i am right or wrong. If i had made a table and not all on the floor, the frame would be even stiffer, so that would not be an issue. But As you can see the garage is very small, lets say it will be the machine enclosure :-).

On paper if the table hangs in the air one side and somehow is supported only by one side/cantilever/ the bend will be around 0.03mm or less. In normal conditions with 500 kg gantry in the middle less than 0.03, so for the side twist.
So basically if i have not made any mistake all must be well under 0.05, which seems quite an achievement to me for 3000x1800frame, having in mind it has only 500kg steel in it.



Now my wife went a week before to Bulgaria on a business trip. It is our country of origin so i know where to make it cheaper than here, she managed to order the rest of the laser cut parts, the rotating nut parts, collect them , pack the and bring them to me. Meanwhile emptying on the way back all personal things of hers, cause this weight exactly 32 kg with the suitcase. I could only bow to her for the support.

http://www.mycncuk.com/attachment.php?attachmentid=14001&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14002&stc=1



I wonder how the 2 bearings go against each other? The thick inside ring side should be outside or facing each other? Somebody shed a light please!

I have to find my bloody caliper in the garage, cause the bearings don't enter without force at all. I called today and they said that they made all so that it must be press fit, something like 0.03mm more. I dont have an idea about that, will see how i fix all inside. First will have to bore the 30T pulley. Though actually i am still in a doubt to use the Chinese aluminum pulleys i bought from ali express or order new from belting online. cause some of the chinese pulleys the 20t ones were not straight, though i kept them for the sake of the price. 10x30t pulleys+10x20t pulleys=~86 euro. Having in mind i could not find cheaper than 13-15euro aluminum pulley 30t, one piece.
Unluckily, belting online had only steel ones and even they were getting expensive with the machining.


http://www.mycncuk.com/attachment.php?attachmentid=14003&stc=1

jimbo_cnc
07-12-2014, 10:25 PM
On axial type bearings, the fatter ring faces are where the thrust is applied. Because there is a taper or similar in there.

Jonathan
08-12-2014, 10:41 AM
Yes, you want the larger side of the inner rings facing outwards so that when you do up the nuts you apply a preload to the bearings. Don't put the bearings in the wrong way round, as they'll be pushed apart by the nut and probably be broken. Don't do the nuts up too tight else the bearings will fail prematurely and don't leave it too loose else the shaft can float in the housing.

From the machining marks in the picture it looks like the bearing bores were milled? If this is the case then it's a bit concerning - if their milling machine doesn't have zero backlash then the bore will be less round, which would distort the outer ring. It's also generally less accurate than boring, which is why bearing housings are usually bored.

When I make these I bore them on the lathe and make the bearings a light press fit, to get better rigidity and ensure the outer rings don't creep. Also, if you get more made I'd be inclined to increase the axial spacing of the angular contact bearings, as it's currently insufficient to get the best stiffness. You'll probably be OK, but the critical speed should be higher if the bearing spacing is increased a little.


[QUOTE=silyavski;64503]i am still not sure. When all is ready and un-leveled, we will see if i am right or wrong. If i had made a table and not all on the floor, the frame would be even stiffer, so that would not be an issue. But As you can see the garage is very small, lets say it will be the machine enclosure :-).

On paper if the table hangs in the air one side and somehow is supported only by one side/cantilever/ the bend will be around 0.03mm or less. In normal conditions with 500 kg gantry in the middle less than 0.03, so for the side twist.
So basically if i have not made any mistake all must be well under 0.05, which seems quite an achievement to me for 3000x1800frame, having in mind it has only 500kg steel in it.

Regarding the pulley selection, you know the dimensions of the rotating ballnut shaft and the rest can be estimated well enough, so work out the inertia with aluminium and steel pulleys and see if aluminium is worthwhile? Or post the following numbers here and I'll show you:

Mass of gantry (i.e. mass that both rotating nuts are moving).
Desired rapid feedrate.
Ballnut size (RM2510?).
Motor ratings.

I think you'll find aluminium pulleys aren't worth it, due to the relatively high inertia of the ballnut - but no point speculating...

Boyan Silyavski
08-12-2014, 03:45 PM
From the machining marks in the picture it looks like the bearing bores were milled? If this is the case then it's a bit concerning - if their milling machine doesn't have zero backlash then the bore will be less round, which would distort the outer ring. It's also generally less accurate than boring, which is why bearing housings are usually bored.
When I make these I bore them on the lathe and make the bearings a light press fit, to get better rigidity and ensure the outer rings don't creep. Also, if you get more made I'd be inclined to increase the axial spacing of the angular contact bearings, as it's currently insufficient to get the best stiffness. You'll probably be OK, but the critical speed should be higher if the bearing spacing is increased a little.



I was so tired yesterday, cause i had a flue, that i missed that. I guess i was overexcited .They were told to bore it on the lathe and i was promised that even it would be done in one pass without turning the block. I hope they did it right. Now i am worried a bit.




Regarding the pulley selection, you know the dimensions of the rotating ballnut shaft and the rest can be estimated well enough, so work out the inertia with aluminium and steel pulleys and see if aluminium is worthwhile? Or post the following numbers here and I'll show you:

Mass of gantry (i.e. mass that both rotating nuts are moving).
Desired rapid feedrate.
Ballnut size (RM2510?).
Motor ratings.

I think you'll find aluminium pulleys aren't worth it, due to the relatively high inertia of the ballnut - but no point speculating...


RM2510 nut
20t:30t
400w Samsung AC servo each side,
~100-120kg gantry


Decided to drivel all 20:30t, only the Z is 20:20t. This after exausting 3 day calculations, belt lengths,inertia, tooth engagement, speeds, etc. and at the end -common sense.
This is a separate thread of its own. I believe this the wisest decision between speed, acceleration, precision.

The rotating ball nut assembly. 20:30t hence 1.5:1 driven. Compromise due to common sense. At the end precision and power at the same time is the aim.

2048 encoder on the motor, so more or less 20m/min. I see on the motor specs that it can go up to 4500rpm, just the curve is straight till 3000rpm. The resolution will be ~0.003mm on all axis , 0.001mm on the Z/ 1605 screw and 2500ppr encoder/. That without any artificial gearing or micro stepping which is not needed it seems. Thats good, cause the Galil board is not micro stepping board.

Jonathan
08-12-2014, 06:10 PM
We can calculate a decent approximation to the inertia of all the parts in the system using this formula for the moment of inertia of a cylinder:
I=\frac{1}{2}\rho\pi r^4h

Just break down the parts into cylinders and add them all up - e.g. for the ballnut we can consider it as three cylinders - the flange, the smaller diameter bit and the hole.

To find the equivalent inertia of the 120kg gantry, you need to consider the effect of the ballscrew, so this formula is required (where L is the pitch of the screw):
I=m*\left( \frac{L}{2pi} \right)^2

The last thing to consider is the effect of the drive ratio - in your case 20/30. This causes the inertia of the parts on the driven shaft (the nut etc) to be scaled by this factor squared. The drive shaft inertia (i.e. motor + 20T pulley) is left as is and we work out the equivalent inertia as if everything was placed on the drive shaft.The total inertia of the system is therefore:

I=I_{drive}+I_{driven}*\frac{20}{30}

Putting all that together (see calculations attached) I get the following values for the total inertia of the system. I've had to guess the dimensions a bit and I don't know the inertia of the motor, so feel free to double check and make it more accurate if you wish.

motor pulley:nut pulley = inertia:
steel:steel = 4.54e-4 kg-m^2
steel:alu = 4.31e-4 kg-m^2
alu:steel = 4.42e-4 kg-m^2
alu:alu = 4.19e-4 kg-m^2

(For interest, equivalent inertia of the 120kg gantry is 1.36e-4 kg-m^2 - about the same as just the RM2510 ballnut!)

So as you can see, in your case the material you select for the pulleys doesn't make too much difference - even with both pulleys made from aluminium the gain is less than 10%. Lets check your not really close to the motor torque rating, else those few % might be valuable. The motor is rated for 400W at 3000rpm, so the torque rating is \frac{400}{2*pi*3000/60}=1.27Nm. Torque is the product of angular acceleration and inertia, so if we say you want 3m/s^2 acceleration (reasonably snappy for a machine this size) then you need 20/30*2*pi*a/L=1257 rad/s^2 acceleration so using the worst case (steel pulleys), 1257*4.54e-4=0.57Nm. Add to that the frictional losses using this (http://www.mycncuk.com/threads/7355-What-size-servo-motor-do-I-need?highlight=servo+motor+calculation) and you're still well below the continuous rating of the motor (let alone the peak), so I'd stick with steel pulleys as they will last longer. Just noticed I momentarily forgot you have two motors, so the required torque is actually even lower...

The other thing to check is the inertia ratio of the system (I.total/I.motor), to make sure it is within the ratings of the servo drive system otherwise the servos may not be able to control the machine stably. The motor is probably around 4e-5 kg-m^2, so your inertia ratio is around 10 at the moment. Do check what the servo system you have(? (http://www.ebay.co.uk/itm/SAMSUNG-400W-AC-SERVO-MOTOR-CSMT-04BB1ABT3-WORKING-/261275716517)) can tolerate as that's a bit high. If it's outside the recommended range you can change the pulley sizes to get a better match.

Boyan Silyavski
08-12-2014, 10:01 PM
Thanks Jonathan,
i remember i calculated all this at the time of choosing, again with your help there, but as i did not understand everything very well, it was like monkey playing with the calculator buttons.
Now i understand a little better :monkey:

It seems the actual ratio would be 12, same as the max allowed by manufacturer.

I still have no idea about the acceleration you are talking about, what will that mean in real life, but i would like the machine to be more or less snappy as you say :watermelon:

http://www.mycncuk.com/attachment.php?attachmentid=14015&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14016&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14018&stc=1

vargai
08-12-2014, 11:26 PM
I still have no idea about the acceleration you are talking about, what will that mean in real life, but i would like the machine to be more or less snappy as you say :watermelon:


This is good example for me being similar to mine situation.
I am wonder too what is the normal acceleration for a machine like this namely how fast would you like to reach the feed speed or the rapid speed.
My unchecked source warn to use over sized servo -
it says if you want to spin up to 3000 rpm a 400W motor during 10ms it require temporary 2500 W
It did not mention the inertia ratio that is as I see very important
factor.
At first sight this 10 ms is to fast for me -when I put together the mosaic now -you only have to think of a corner you mill .The machine has to stop and change the direction and run up again- and the surface quality depends on the acceleration here

Boyan Silyavski
09-12-2014, 12:56 AM
What Jonathan suggests "3m/s^2 acceleration (reasonably snappy for a machine this size)" should be quite snappy.

I could judge only from my lazy small machine and my first build which was for a friend, where i left the acceleration to 0.8m/s^2 cause after that it started to seem a bit dangerous.

Imagining that beast now to move the gantry from one end to other for 1 sec, thats would be quite a view. But as i understand it, the acceleration has more to do with corners, circles and so, than moving the gantry with tremendous speed from one side of the table to the other. Cause at the end i control the machine paths so generally i always strive to make them perfect, without wasted movement.

vargai
09-12-2014, 08:07 AM
To find the equivalent inertia of the 120kg gantry, you need to consider the effect of the ballscrew, so this formula is required (where L is the pitch of the screw):
http://www.mycncuk.com/vlatex/pics/34_ab15e68f900f95cf141f58d95e3a2ceb.png

Hi, I would like to apply this to my machine and wonder how it does come from ( see behind the formula)
That means this is not enough to count the reduced inertia on motor shaft with only the rotating parts?

120 kg*(0,01/2pi)e^2= 3,4 e-4 differs from your number.
Where I am wrong -Am I understanding everything well?
Thanks

vargai
09-12-2014, 08:50 AM
What Jonathan suggests "3m/s^2 acceleration (reasonably snappy for a machine this size)" should be quite snappy.
So at 600mm/min machining speed means 0,01m/s
t=v/a=0,0033s -it reaches the speed during 3ms- is quite snappy but no experience just feel
Actually I cannot see any point to apply high acceleration -rather high rapid speed if you want a fast maschine

BUT
My ignorance does not pull me back in making opinion :yahoo:

Jonathan
09-12-2014, 05:13 PM
Just spotted I forgot to attache the calculations to the previous post - I've attached them to this post instead.

14025

I'll make some general points which I think should answer the questions - do let me know if I've missed anything.



Acceleration, why have it?


A CNC machine rarely moves at a constant velocity, but does generally aim to move at a constant speed. The critical point here is the difference between velocity and speed - you can have the machine moving at a constant speed, whilst the velocity of each axis is varying greatly. Simple examples are cutting arcs or simply moving round a corner. This changing velocity means you have an acceleration.

The acceleration value you enter in the motor tuning is the limiting value - the controller should never command the drives to exceed it. You can therefore work out the minimum radius of a corner the machine can achieve for a given feedrate and feedrate. From the formulas for circular motion we can express the acceleration caused by the change in direction in terms of the feedrate and radius of the path:

a=\frac{v^2}{r}

For example, suppose you're cutting at 10m/min (e.g. cutting MDF) and you want to cut squares with 10mm radius corners, without the feedrate reducing at the corners, we can work out the required acceleration:

a=\frac{v^2}{r}=\frac{\left( 10/60 \right)^2}{0.01}=2.78m/s^2

You can play with the numbers to see when this matters - but as a general point we can say that the required acceleration is proportional to the velocity squared, so if you're mainly cutting at low speeds (e.g. cutting aluminium) then the acceleration is not so important. This should also explain the differences in toolpath you see between using constant velocity (G64) and exact mode (G61).


Inertia ratio


It seems their limit for the inertia ratio is set by the capacitance present in the motor driver dc-bus. When the motors are decelerated, some of the energy stored in the total inertia of the system is transferred into the capacitors, since the drivers are supplied from a simple rectifier which only lets power flow one way. The problem with this is it causes the voltage on the capacitors to rise and if the energy transferred is to high (due to high inertia ratio), the capacitors will be damaged. It's actually quite easy to work out roughly how much the voltage rises, just equate the energy stored in the inertia, E=\frac{1}{2}J\omega^2 to the capacitor energy, E=\frac{1}{2}CV^2 and re-arrange for voltage. The things we can do to alleviate this problem are add more capacitance, add a breaking resistor to dissipate the energy as heat or reduce the input energy by lowering the system inertia.
All of these seem reasonable options. Recall from my previous post that the equivalent inertia of the system depends on the square of the drive ratio, so by changing the drive ratio a small amount you can lower the inertia ratio. e.g. going to 18:30 instead of 20:30 would be sufficient. You should still be able to get 20m/min as the motor has sufficient torque at 3333rpm. A more interesting way round it could be to connect the DC-buses of all the servo drivers in parallel, as in general when one axis decelerates another is accelerating so will absorb the energy. You might still need a braking resistor to handle an e-stop event though.



Equivalent inertia of gantry


In my previous post I stated the following formula, without deriving or referencing it, which I agree was not good practice so I'll explain where it comes from:
I=m\left(\frac{L}{2\pi}\right)^2
When I use the word inertia, for a rotating system I mean the moment of inertia. For a mass moving linearly, such as the gantry, the inertial mass is equal to it's mass, m. The issue here is we need to incorporate this inertial mass into a rotating system, so we can sum the inertias seen by the motor shaft. The general formula to calculate the moment of inertia, for a mass (m) rotating about an axis at a distance r from that axis is:
I=mr^2
For every revolution of the ballscrew (or nut) the gantry moves a distance L (the screw pitch), so we can relate the angular speed of the ballscrew (or nut) to the linear speed of the axis as follows:
\omega=v\frac{L}{2\pi}
From the definition of angular speed we know that \omega=vr. We can combine these equations to find the equivalent inertia:
r=\frac{\omega}{v}
Substitute:
I=mr^2=m\left( \frac{\omega}{v} \right)^2
\frac{\omega}{v}=\frac{L}{2\pi}
I=m\left(\frac{L}{2\pi}\right)^2

We're not quite there though, as the ballscrew (or nut) isn't directly coupled to the shaft. This causes the angular velocity to be scaled by the drive ratio, lets call it \frac{N_1}{N_2}. If you include that in the above derivation you'll find the inertia is scaled by the square of this ratio, so:
I=m\left(\frac{L}{2\pi}\right)^2\left( \frac{N_1}{N_2} \right)^2

irving2008
10-12-2014, 04:12 AM
Deleted......

vargai
10-12-2014, 10:42 PM
I'll make some general points which I think should answer the questions - do let me know if I've missed anything.


Thank you Jonathan to clear this dark spot in my knowledge.

Boyan Silyavski
20-02-2015, 12:45 AM
Some updates. I a bit ashamed it goes so slow but in between other work, orders of stuff i do and working alone on the machine, it takes too long.


This is what you are up to, if you don't have straight edge. 3m rail. you move it left right from one end and from the center onward it stays in place. Not so bendy when short lengths but look at that...


https://plus.google.com/111571569713128538019/posts/P5Gu6PjqftB




A very lazy helper

http://www.mycncuk.com/attachment.php?attachmentid=14686&stc=1


This is what the empty epoxy bottles weight, in case you wonder or you loose power when measuring:

http://www.mycncuk.com/attachment.php?attachmentid=14687&stc=1 http://www.mycncuk.com/attachment.php?attachmentid=14688&stc=1


The key to epoxy leveling the gantry rails, squaring the gantry and other was to leave the plates that mount on the gantry bearings/the leg sides/ longer 10mm at the back side. Z plate was mounted, ball screw and then rails were squared, just to make sure all would fit.

http://www.mycncuk.com/attachment.php?attachmentid=14690&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14691&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14689&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14692&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14693&stc=1



I was impatient here, so i did all this alone. Moving the gantry up and down i mean. Starting to feel that i am training heavy lifting. Then gantry was mounted and squared again. gantry rails were checked for squareness against the long rails. Success.

http://www.mycncuk.com/attachment.php?attachmentid=14694&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14695&stc=1

Boyan Silyavski
20-02-2015, 12:47 AM
My "favorite" part - painting.
It took one day to clean the machine. One day to paint it 3 colors using what i had of paints meant directly over steel or oxide. A day to dry/it was cold. One day for second hand of these-white, silver and black. White looked best. One day to dry and then started the real paint.

Violet metallic, who would have guessed what color i would choose. Why that color? It needed to be different cause the machine deserves it /money, effort, result/. So i looked and i looked... And by the way i have got new Oakley Violet Iridium sun glasses a week before . So i decided i would look cool beside it :hysterical: and chose the color to be the same. Crazy, yeah?

http://www.mycncuk.com/attachment.php?attachmentid=14700&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14701&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14702&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=14703&stc=1

JoeHarris
20-02-2015, 01:17 AM
Ha ha! Like it.

mekanik
20-02-2015, 12:03 PM
Outstanding job Boyan, it won't be long now.
What model are your shades ? i was thinking of treating myself.
Regards
Mike

Boyan Silyavski
20-02-2015, 04:23 PM
Outstanding job Boyan, it won't be long now.
What model are your shades ? i was thinking of treating myself.
Regards
Mike

Thanks!

Oakley Style Switch, standard fit, clear frame,Violet Iridium/+spare black iridium/ OO9194-03. Possibly the only present i made to myself at least like from half an year ago. All else went in the machine or paying bills :anonymous:

njhussey
20-02-2015, 04:32 PM
Looking good Boyan, the machine that is :whistle:

I treated myself to some Oakleys a couple of years ago, Crossfire II's for everyday use and for flying my RC planes and Helis a pair of Half Jacket 2.0 with G30 IRIDIUM lenses...

GEOFFREY
21-02-2015, 01:22 AM
Whose a pretty boy then. I think you need the shades to shield from the glare off that machine!!! Looking great, as your work usually does. G.

JoeHarris
22-02-2015, 10:59 AM
How have you aligned the y axis rails? Apologies if this is explained elsewhere and I've missed it..?

Boyan Silyavski
22-02-2015, 07:24 PM
How have you aligned the y axis rails? Apologies if this is explained elsewhere and I've missed it..?


OK, the squaring. It was the main challenge at this size and the precision i would like to have. There was no skipping there. I had it clear before soldering the machine and the design was affected by it. I will try to explain it below. Every detail counts and i will try to explain it step by step.
The key to my way of thinking, apart from no skipping, was hard stop mounting, means i don't measure anything/cause i don't have the means to, instead i rely on straight edge, square and touching without gap. On the first machine i have made i managed to achieve 0.01mm more or less in all directions. This one was quite bigger challenge.

Another point is that i always strive for absolute alignment and straightness, not relative, meaning i don't skip relations or directions. Knowing of course that's not achievable in real life at home. But when you strive for the impossible the result is more than acceptable




Y axis is my long axis . Epoxy leveled, then mounted left rail permanently, against straight edge. Now it was to square the right rail. Then using precision square and straight edge aligned both ends. Then clamped there. Aligned middle, then clamped, aligned other ends then clamped. Marked the holes, un-clamped everything. Then again repeated the process against straight edge clamping here and there. Now i had small level problem even after spending so much time on leveling, this due to a small leak. Other thing on such size was that even using straight edge and precision square, at such length everything seems square but is not , at the end its up to you to feel the squareness scraping them against each other. i could say that was the most difficult job on the machine, and most time consuming.

X is my gantry rails where Z moves left right. First of all i left the plate on which the gantry steps upon the bearing blocks to protrude 10mm at the back. I mounted the plates but not the gantry. The plates were still not soldered. And i made sure using straight edge that they are perfectly on one line and square.Then i mounted the gantry on top of them. Then i soldered carefully on place, just fixing tacks. Then i dismounted the gantry and plates and soldered on the ground. Then mounted again up. Checked if all is OK. Then dismounted the gantry and put the gantry on its back over the Y rails. See pictures above. #143 (http://www.mycncuk.com/threads/6619-Quite-an-Unusual-one?p=67113#post67113)
The idea was the Y rails are now leveled with earth, cause i did not move the machine, so perfectly horizontal. So when i lay down the gantry on its back, using shims i can perfectly square the said plates to the Y rails, 90 degrees. So now i poured epoxy Under the X rails. The result was later when mounted, the X rails are perfectly 90 degree to the Y rails. Now the thing that was left was to mount the X rails properly and square them in the other direction to Y/parallel, meaning when all mounted left and right height of X rails to be same distance to bed/ .
So with the gantry still laying on its back and gantry legs at 90 degree to the Y rails, epoxy poured and dry, i measured same distance from the lower X rail to the gantry legs plates/which were on same plane from before, when i mounted them on machine and soldered gantry on its place./ . Used verier then tightened the bolt, so all was hard stop mounted. I could have drilled and mounted the lower rail now, but i just clamped it, cause i had max permission of +1mm up or down or the ball screw would not fit in gantry and Z plate. So basically i had to mount the ball screw on its place, let it have some future play for adjustment, then mount other rail and the Z plate. Some fiddling with all that and i knew where all should lay. So marked, dismantled all and using the vernier again mounted the lower X rail. Checked against straight edge, marked, drilled, tapped and fixed.
Now having one rail mounted permanently i had to mount again the ball screw, the Z plate and so i mounted the second rail.

So at the end of it i lifted the gantry on its place and using straight edge, precision square and the protruding gantry leg plates for reference i squared the gantry.

JoeHarris
25-02-2015, 08:51 PM
Nice work Boyan, figured you must have epoxied both axis. Time consuming buisness but must be satisfying to get it that precise!

Boyan Silyavski
01-04-2015, 03:24 PM
At the end i managed to move further the build. Thanks to my new friend and fellow forum member Jeff Potter who helped me a lot with machining all aluminum pieces for me.

Now what happened:

-i bought some sound insulating stuff so i can make the garage quiet, so nobody bothers me in the future when i am working. I mean the tourists that come now in the summer and rent the houses around me

http://www.mycncuk.com/attachment.php?attachmentid=15054&stc=1

-tested this and that and at the end finished the prototype of the fog-less mister and tested it while milling the aluminum pieces for the Z


https://www.youtube.com/watch?v=c07M_-QUVUE

-tapped and started putting the Z box together , 5mm tap stem fits perfectly in the 6mm drill guide


http://www.mycncuk.com/attachment.php?attachmentid=15055&stc=1


-The rails were meant not to obstruct the back plate. the idea was to use the back plate as a guide to mount them perfectly flat and square


http://www.mycncuk.com/attachment.php?attachmentid=15056&stc=1


-mistake in the design. How the hell i was going to screw the ball screw nut supports. So i decided to cut a slot and at the end they will be screwed with 3 screws only.

http://www.mycncuk.com/attachment.php?attachmentid=15057&stc=1

-as everything was machined very precisely and as i helped Jeff make his machine work in 3D mode, all fit perfectly together
http://www.mycncuk.com/attachment.php?attachmentid=15058&stc=1 http://www.mycncuk.com/attachment.php?attachmentid=15059&stc=1 http://www.mycncuk.com/attachment.php?attachmentid=15060&stc=1

-the idea of the shortest front plate is if i lift and hide the spindle to its highest possible position inside the box, so i could reach and change tools without problems

http://www.mycncuk.com/attachment.php?attachmentid=15061&stc=1

-the Z looks like a bad ass missile warhead

http://www.mycncuk.com/attachment.php?attachmentid=15062&stc=1 http://www.mycncuk.com/attachment.php?attachmentid=15063&stc=1

Boyan Silyavski
01-04-2015, 03:33 PM
Some more. I hope you like detailed photos of builds. /there is 10 image limit and i like taking photos :-)/

-The rotating ball nuts were press fitted and all worked fine. Though i have to change small things in the design, things that made me think a lot how to mount all of it together. And how it will be dismounted in near future if i decide to repair the belt...

By the way i decided to start making together with Jeff and selling the assembly in the near future, for forum members and other. Redesign it a bit for any specific case and machine it properly. It is a result of the fact that i spend too much time fiddling with that assembly and investigating. And trying to find who will do it for me, and worrying if they did it properly and repairing their mistakes...
So i would announce it at its proper place in the forum, hence will stop to brag here.

http://www.mycncuk.com/attachment.php?attachmentid=15065&stc=1

-tadaaaa, second mistake. The 30t pulley is too much in front so hits the Z back plate

http://www.mycncuk.com/attachment.php?attachmentid=15066&stc=1


-mistake repaired easily, thanks god

http://www.mycncuk.com/attachment.php?attachmentid=15067&stc=1

- dry test looking for mistakes

http://www.mycncuk.com/attachment.php?attachmentid=15068&stc=1

-ready to "surface" straight the side plates

http://www.mycncuk.com/attachment.php?attachmentid=15069&stc=1

-i knew the sound proofing of the garage will pay off. why this always happens at lunch time when all normal people rest. had to take the bigger "surface" hammer and "surface" the plates for about half an hour

http://www.mycncuk.com/attachment.php?attachmentid=15070&stc=1

-that's better

http://www.mycncuk.com/attachment.php?attachmentid=15071&stc=1

-another dry test and squaring together
http://www.mycncuk.com/attachment.php?attachmentid=15072&stc=1

Boyan Silyavski
01-04-2015, 03:48 PM
Wow, 3 posts, it seems i had a lot to show and say :hysterical:

-the Z again, 2 spot welds to hold the side plates in place

http://www.mycncuk.com/attachment.php?attachmentid=15073&stc=1

-thats why i did not bother before too much rectifying the back Z plate. No i pushed it flat to the whole straight assembly, so all fell in place. Then spot welded CAREFULLY

http://www.mycncuk.com/attachment.php?attachmentid=15074&stc=1

-I know Dean, OTT and too much but that cost me only 30euros more. I promise never to do that again.

http://www.mycncuk.com/attachment.php?attachmentid=15075&stc=1

-squaring the Z in both directions against the straight edge that lies flat on long rails. this took a couple of hours to do it properly until satisfied

http://www.mycncuk.com/attachment.php?attachmentid=15076&stc=1

-another consequence of the mistake i made here

http://www.mycncuk.com/attachment.php?attachmentid=15077&stc=1



- That must be the most bad ass DIY Z axis i have seen on single gantry machine :cheerful: , OTT but make me proud and smiling when i look at it, so i assume it was worth it :victorious:, did i say it looks like a missile ready to launch ?


See that big, no - huge mill vice? Thats a gift from Jeff for the build. have to rethink this and that for the bed. I never imagined i will have some day such a beauty. On other hand its not fair some people to have 3 of them :joker:

http://www.mycncuk.com/attachment.php?attachmentid=15079&stc=1

mitchejc
01-04-2015, 07:55 PM
Well done sylyavski! That Z looks VERY strong, I like it a lot.

toomast
01-04-2015, 08:15 PM
EPIC.
You sir are raising the bar for all of us :)

JAZZCNC
01-04-2015, 11:44 PM
Don't get the point of 2 ballscrews on Z axis Boyan but it sure looks Cool.!. . . Mega OTT but very Cool.:applouse:

Boyan Silyavski
02-04-2015, 06:16 AM
Thanks guys!



Don't get the point of 2 ballscrews on Z axis Boyan but it sure looks Cool.!. . . Mega OTT but very Cool.:applouse:

Thanks Dean,

I will explain. So people don't get over excited and start doing machine with double ball screw assemblies on the Z . As i said i will not do that again.

First i wanted to use a pneumatic hammer to shape the 1mm metal sheet. So i thought better 2 ball nut assemblies than just double ball screw nut, cause so the vibration will be spread on 4 fixed ball screw supports instead of 2.

Then as we spoke it became clear that 1 ball nut assembly is ok, especially if fit with double ball nut.It also became clear that maybe i will not use at all the hammer but make separate machine.

But i saw that because of the Z rail length /i had them already/ i will loose Z travel, don't remember exactly but instead of 200 it was to be 160. Or i had to change this and that and move the upper bearing blocks on the Z plate too much down, which would make pointless the so strong Z box


And as all was designed for 200mm travel i decided to live with the 2 ball screws

So yeah, it was a mistake , not a design wonder, just cost me 1 ball screw more. One year ago i had no idea how big can be 2505 screw, how angular contact bearings work, that i could have made custom end supports and so on. Slowly evolving...


There are also some other small mistakes on the gantry design that i would not repeat, most of them are OTT because of the hammer idea. Not that big mistakes but unnecessary waste of time and energy. I know now how to make all much simpler and faster. Same strong.

Jonathan
02-04-2015, 11:03 AM
Brings back memories...

http://www.mycncuk.com/attachment.php?attachmentid=15081&stc=1

Anyway. How have you ensured alignment of the Z-rails - both their spacing (parallel to Y) and parallelism? I can't see any features for alignment, so either you milled reference edges in after assembly, spent time with a DTI and straight edge, or (I doubt) didn't do it? It almost looks like the spindle is just clamped by the plates with the arc-section cutout? If so I'd be a bit concerned about maintaining the accuracy of the rail alignment.


Then as we spoke it became clear that 1 ball nut assembly is ok, especially if fit with double ball nut.It also became clear that maybe i will not use at all the hammer but make separate machine.


I've probably mentioned this before, but I'm not convinced that a double not would actually work better than 1 nut here. In a double nut system the nuts are preloaded against each other, so a force is applied the load on one nut decreases whilst the force on the other nut increases. For any force above the preload force, all the force will be transferred to one nut.

Either way, as you implied, in general it's more economical to use a larger ballscrew (or larger rails) than multiples *unless* the spacing of the linear bearings is such that racking is an issue. I mentioned it here (http://www.mycncuk.com/threads/6484-A-sufficiently-strong-machine?p=50915#post50915) regarding rails and stiffness, but similar reasoning applies to ballscrews.

Boyan Silyavski
02-04-2015, 01:16 PM
Brings back memories...

that creation of yours is exactly what inspired me for my Z



Anyway. How have you ensured alignment of the Z-rails - both their spacing (parallel to Y) and parallelism? I can't see any features for alignment, so either you milled reference edges in after assembly, spent time with a DTI and straight edge, or (I doubt) didn't do it? It almost looks like the spindle is just clamped by the plates with the arc-section cutout? If so I'd be a bit concerned about maintaining the accuracy of the rail alignment.



First of all all steel there was laser cut and holes laser drilled, which helped a lot /0.01-0.03mm accuracy of all cut pieces/. Then the aluminum pieces were machined <0.01mm

So knowing that the aluminum pieces were precise, i mounted first the rails on the separate side pieces, checking them to be tight corner fit against another piece of aluminum and straight edge. Later i mounted the 3 sides of the box using , straight edge, precise squares and digital caliper. So now i knew they were parallel to each other in all planes, /same distance/. Then i checked if the spindle enters and it was tight fit just easy enough to slide.

That told me till here everything is perfect/ cause the spindle was machined before 99.8mm OD. We were at the point of machining it and checked it against DTI and it seemed it was not necessary. That is the "good" original spindle with the 4 bearings.

Then knowing that the spindle box is mounted together under 0.01mm i just screwed to the side bearing blocks the side steel plates checking them flat and square against flat surface. Then i used some small clamps to not let them move back and forth.

So as all was mounted together maintaining that precision i laid it down carefully on top of the back Z plate. Now as you can guess the Z plate was not perfectly flat. So i spot welded the side plates/together with the aluminum assembly/ to the Z plate, but only at one side that so the previous assembly is square left right. See pictures above. I used also some clamps not to let it twist left right. Then i force clamped all down to the steel back Z plate , so now it was square and flat / cause the steel side plates wer cut by laser 0.01mm , remember? and i have mounted them perfectly flat against flat plate/ .


So now all was 0.01mm mounted together. I just carefully soldered spot by spot, careful not to heat the bearing blocks.


Later i took out all the aluminum assembly together with the bearing blocks and painted the steel assembly , masking the sides where the bearing blocks were meant to enter later So at certain places there is no paint, to ensure tight fit and not spoil the accuracy.

Same must be done for the back at the place where the ball screw nut housing on the gantry will be fixed to the back Z plate. I just left 1mm space in the drawing, to be shimmed, but due to 1mm less epoxy height, it was tight fit.

What i would say that a sincere judging helps here, one should not lie to himself and want to speed things, cause these are the moments where accuracy may be lost and accumulated.

So I fit the steel Z assembly to the gantry and squared it using the vertical Z bearing block surfaces to a straight edge across the long rails. Then i resquared the gantry ball screw, loosening it, going one end, tighten, going another end tighten and repeat. This to have the Exact same travel as the rails

So when i fit the aluminum assembly i had only to square it front back/ again using the long straight edge across long rails. See photo above. Then tightened the bearing blocks to the . That simple.

Something should be clear though. everything was designed to fit together using only 2 precision squares and 1m and 2m straight edges. No dial indicator was necessary at any moment. Everything was judged by the tight fit of the squares to surfaces and straight edges. I did so with the first machine i build and achieved 0.01mm precision.

So basically i will not use DTI until the moment i have the machine running. from experience it seems i can notice with naked eyes 0.05mm gaps without problem against the straight edge/ the epoxy story/ so i am happy that everything is tight fit.

When all ready i will check with dti and straight edge and report. But for now it seems i maintain the same precision as the first build and on that size of machine, well ... i am happy

JAZZCNC
02-04-2015, 02:10 PM
So basically i will not use DTI until the moment i have the machine running. from experience it seems i can notice with naked eyes 0.05mm gaps without problem against the straight edge/ the epoxy story/ so i am happy that everything is tight fit.

Now now Boyan.!!! . . You contradict your self about lieing to one's self here with this statement.!! . . . . Your fooling your self if you think you can see "Five one hundreth's Gaps" with naked eye over the length or width of this machine.

Not saying your machine isn't or won't be accurate because I know it will or you will make it so. But I must say think your going to be surprised when time comes to cutting just how far out of Tram that spindle will be if you don't put a DTI on it. I also think it will be Bitch to setup looking at the pics.!!

Boyan Silyavski
02-04-2015, 04:30 PM
What i meant was i can spot the 0.05-0,1 mm gap between a straight edge and something, say rail or other flat surface. Not that i can spot if the machine is straight or not over all its length or width. I spent a week just making sure the long rails are mounted straight and flat, not counting the epoxy.



Yeah, without DTI its difficult to say more.


What do you mean to tram the spindle? Its trammed already using 2m straight edge and squares. There will be no more tramming until the moment i cut some square aluminum shapes and first i make sure the gantry is squared by the limit switches and later do some stuff and see if there is something off judging by the final dimensions of the work.

Then the bed would be surfaced or head trammed first, but only if i am off by a great deal, depends on what results tell me. It is very easy in fact to tram it all as i always followed your advice of adjust ability making all through holes bigger.

Now about the DTI. I could clamp it now on one side rail and move the head assembly up down and see what DTI says but ... ok i went and clamped it. It says nothing, i mean i mounted all very well only using the squares and edge. Unbelievable but true.

I will power the whole thing then check and readjust if necessary, cause its very difficult to move all by hand. long distances i mean and i get bored.

At the end even if vertically is off with 0.1mm on 200mm travel/ which is not the case/, this is nothing. on 20mm aluminum sheet the inclination would be 200/20=10, so 0.01/10=0.01mm

JAZZCNC
02-04-2015, 07:32 PM
You will tell instantly if the spindle is trammed correctly or not when you surface material with a wide cutter. If off then you'll get steps and it doesn't take much error to get steps. The pictures make it look difficult to get at bolts if shimming is needed etc.

Boyan Silyavski
11-04-2015, 02:43 PM
Now comes one of the last parts. I frankly admit i was quite indecisive there, from the beginning of the design. So it was time to sit on my arse and make some calculations and think the things clearly.

The working area / table bed. May be it a theme for a whole new topic. i see many people are wondering how to do it differently than plywood, especially if they mean to cut aluminum and cool it.

Now lets be clear i am not talking about the machine bed structure nor the sacrificial board. I am talking about the working bed/fixing area and method which is in between the frame and the sacrificial board and which is water/coolant resistant and is strong enough for moderate aluminum milling

So lets see the options. prices are Spanish, but may be more or less the relations will keep in different countries. Some thickness prices are not known but can be extrapolated. 1kg of steel is ~1 euro, 1kg of aluminum 6082 is ~5euro,

-phenolic 10mm thick 65euro/m2 so as i see it needs to be minimum 20mm best 30mm thick ...
-marine ply wood 15mm thick ~36-50 euro per m2 , i see it needs to be at least 30mm thick.. so its good for a spoil board but expensive for top
-marine ply wood with phenolic top for truck flooring, now this is a good one, but its even more expensive and the phenolic will be very thin to surface

Now instead of writing here i have made some drawing with data for the better options:

http://www.mycncuk.com/attachment.php?attachmentid=15125&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=15126&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=15127&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=15128&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=15129&stc=1


I have spent all morning in calculation and last days rethinking the things. So for me its clear:

My choice will be UPN140. I have the bed clearance for it. It is cheap and very strong for the money. Like the last photo, 2 welded together, with T channel and DIY Tee nuts as per drawing. So i will drill a lot of holes and use them for fixture and big sheets will be fixed with clamps that will pass trough the channels. It will be screwed to the table with nuts that pass between the connection of 2 of the UPN. I will have to find a resistance welder from somewhere for clean job.



Note: Moment of inertia is x axis cause its in C position when measures are given, so a bit misleading. Anyway, its in the good for us direction.

last thing that i have not clear is if i will machine it both sides and machine the bed bellow, so all is flat when mounted and dismounted if needed. I am seriously starting to think to somehow mount a grinder on the Z or surface all using small bit, slow speed and coolant. i said before no steel work with that machine, but its quite tempting. And from what i have seen people have done it with much weaker designs.

One thing i know is that on top of that UPN140 i will use then thin phenolic sheet or even better scrap pieces, for sacrificial board. At leats on some part of the machine. On other i will use cheap expanded PVC sheet.

I hope that these calculations money/material/ strength will help people to evaluate properly what exactly bed they need.

Notes
-UPN profile can be cast filled with cement which will make things even better. Non crack cement for table tops and garage floors is enough for that.
-aluminum T channels are damn expensive so i did not even consider them. Even steel channels are very expensive, crazy.


PS2.
One question . For the the threaded holes to be closed to keep them clean , what is the name of the screw that is like a grub screw but is for flat screw driver?

toomast
12-04-2015, 08:00 AM
PS2.
One question . For the the threaded holes to be closed to keep them clean , what is the name of the screw that is like a grub screw but is for flat screw driver?
I believe it is called SLOTTED GRUB SCREW.

Boyan Silyavski
19-04-2015, 07:11 PM
OK. Some development.

Did i tell you that i plan to add plasma to the machine. Front half for machining, back half of table for plasma. Or whatever life commands, :greedy_dollars:

I know, old China HF plasma no good, combined machines no good... I can say yes, shoot 2500 euro and tomorrow a brand new Hypertherm Powermax 45 will be at my door.

First i have to know if that will pay off where i live, for what i do and so on. However instead i decided to start with what i have at hand, slowly make my way around. I did not skip the THC though as i believe it will be extremely important especially on a combined CNC table like mine.

I did my research and decided to go with the Neuron THC. I skipped the simple THCs where there were no advanced controls. Vietnamese and similar were out of question due to lack of support or support in some strange language.

The ones who are into plasma cutting know the choice is not so big money wise and feature wise. Razordance THC is only for Hypertherm, then there is the Proma, mini THC that looked promising, Cadcnc and Neuron. Or you pay big dollar. Cadcnc and Neuron seemed the most advanced. I as many times at the point of buying something from Cadcnc but the fact that he is in USA, his stuff works bundled together, some of his ridiculous policies mad em not do it. So the Neuron it was.

I will make a separate review when all is working, but i would say the documentation and the support from Andrey till now is outstanding. And to tell you the truth that the most important thing when you start doing something that you do for first time and you don't know nothing about.

So the Neuron THC lite came at home after customs delay /Russia is not EU you know /.

Thats the Plasma, tig, arc. German controlled Chinese machine with 5 year warranty.

http://www.mycncuk.com/attachment.php?attachmentid=15167&stc=1

First thing i changed the cover screws cause i decided i would open it quite frequently :-)

http://www.mycncuk.com/attachment.php?attachmentid=15168&stc=1

Then i twisted some pairs of wire for the voltage divider /1mm2 >600v wire, as per recommendation of Neuron manufacturer / :

http://www.mycncuk.com/attachment.php?attachmentid=15169&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=15170&stc=1

Then i blowed the dust of the insides, by the way on the box says to do that 2 times per month
http://www.mycncuk.com/attachment.php?attachmentid=15171&stc=1

Here is my DIY drier for the plasma, painting and for blowing dust. No body would want to blow some drops of water on electronics, right. 10" filter box and 1/2kg of desiccant. Black plastic tube in the middle with a lot of holes so air can pass. On the right is a last stage lubricator outlet for air tools. Thre dry air outlet is in the middle

http://www.mycncuk.com/attachment.php?attachmentid=15172&stc=1

The insides, with the Neuron mounted, arc ok mounted , RJ45 ethernet communication cable. meanwhile i bout and learned to use cable connector making tools. as it seems Ethernet is the future.

http://www.mycncuk.com/attachment.php?attachmentid=15173&stc=1

The shed in front of the garage, the welding table and the machine bed arrived. I alsmost finished all till today, though run out of gas, otherwise i would have been welding right now. Will take more pictures when ready. the rage saw cut the UPN 140 like a charm. I was for worried for a moment when all arrived cause i forgot to check if the saw could do that wide

http://www.mycncuk.com/attachment.php?attachmentid=15174&stc=1

i need a welding table to make my welding table, isnt that funny

http://www.mycncuk.com/attachment.php?attachmentid=15175&stc=1

D.C.
12-07-2015, 12:07 AM
OK. Some development.
Will take more pictures when ready. the rage saw cut the UPN 140 like a charm. I was for worried for a moment when all arrived cause i forgot to check if the saw could do that wide

Amazing build Boyan, those rage saws are quite the beast aren't they.

Do you have an update and more pics of how the build is progressing, this has got to be one of the best build logs I've seen, it would be shame for it to stop before we get some video of the finished thing in action!

Boyan Silyavski
12-07-2015, 08:40 AM
HI,

yes i have an update.

First one of the piles of steel at the previous photos now is welding and fabrication table (http://www.mycncuk.com/threads/8600-Welding-bench?highlight=welding+table), the other pile with the long pieces now is a shed in front of my garage where the machine is. At the link above you could get idea how to fix the UPN to the bed

https://lh6.googleusercontent.com/FvzYAZ5RfqaigavYmmeDrAGPORmdcgElEx17HR-RrAc=w1358-h727-no


https://lh6.googleusercontent.com/-a9Mnl0u_43M/VaIQvVw72MI/AAAAAAAAGW8/qsCvtUme-qQ/w1358-h764-no/20150712_085906.jpg




Now about the machine itself. Cables mounted and ready to run. I had some problems fitting the cables cause i made the error underestimate how big cable chain i need. I had some serious cable chain but it seems i needed even bigger.
Some of the encoder cables were shorter so i had to invent here also. Most cheap was to use HDMI cable, i hope it lasts some time.

Now the biggest problem. I can still not make run the machine. Reason being that although all is connected and programmed from Mach 3 side and from Galil side it would not move.

So i read 10 times the servo and the Galil manual / each 400 pages/. At the end frustrated i started trouble shooting. Checked 1 by one all cable connections cause the servos were tripping, cleaned contacts...this took a couple of days. At the end i discovered that when programming the servos somehow i missed to program correctly the motor name of 2 of them, hence the problems. So more or less all is ready here, still have 1 enable wire to 1 of the servos not working well so i have to rewire that.
Jogged the servos from the software and tuned them for a second.
That servos are incredibly fast, smooth and silent. The machine moves like crazy fast super silently. I directly jogged the Z assembly left right on the gantry 10-15m per min. Its scary.
Good that the frame is balanced horizontally. yeas i have 10 cm deviation on the floor so i used 30-35 square wood pieces leveled them and the machine is on top of them.

So the machine should be running any minute but 2 weeks i couldn't manage to make the Galil board work. Yes, it works communicates, switches signals and so, but still not able to make it speak / pulse/ to the servos. And on top of that i burned the COM port from changing cables. So i could not now change the servo type of signal expectancy. GRrrrrrrrrr!! Now have to wait for new COM board cause where i live...

I asked about help with the Galil at Mach forum, Galil forum, nobody... So after 2 weeks i discover that i have to issue MT command to change the type of pulse it sends. It may be the greatest board but with nonexistent support, no schematics of connection... maybe i did a mistake not contacting directly support, but from what i saw in their forum-please read the manual its explained there. yes its f***ng explained but its 400 pages manual and i am reading 3 of them at once and on the way learning new terms that i had no idea yesterday what that even means, so a bloody connection diagram for every application example and step by step instruction would be great for stupid people like me and the 20 engineers that asked the same questions...

So i am still stuck, last effort these days, then it goes off the machine cause i will receive 2 Pokeys boards, so one goes there. Will leave the Galil apart, mount it in another box , study it and use it in another machine. And then bloody sell it :-)

So any day soon it will be fully alive, then i will move the gantry around and mount and level the ready UPN140 pieces for the working table bed to be level, Thats why i need to move the Z around so i measure and shim where necessary.

https://lh6.googleusercontent.com/-DQB48AJUSIw/VaIQxqx8DYI/AAAAAAAAGWM/CO-FR9TMLtw/w1358-h764-no/20150712_085936.jpg


Here where is the PC, the VFD and so on, at the entrance. Switch on the machine,mount piece, Mach3, program run, then go out and close garage door. The garage=Big enclosure for the machine

http://www.mycncuk.com/attachment.php?attachmentid=15692&stc=1

toomast
28-07-2015, 08:21 AM
Looking good!
How did you end up mounting UPN140 for the bed? I mean how did you get it level? Or you just have them laying on the machine for the moment?

eurikain
17-08-2015, 08:46 PM
I honestly think anything you touch turns into gold ;)

Boyan Silyavski
21-08-2015, 11:08 AM
Looking good!
How did you end up mounting UPN140 for the bed? I mean how did you get it level? Or you just have them laying on the machine for the moment?


Drilled the UPN, soldered oversized nuts to form a bed, drilled machine and fixed with bolts. Same as what i did here http://www.mycncuk.com/threads/8600-Welding-bench



Thanks for the kind words :-)


Have not updated it long time. Its almost finished, now rectifying some design mistakes. Will speak about them later.

Here is a video i made yesterday of test runs at 10000mm/min acceleration 3000, gantry weights 180-160kg or sth like that. Its scary. Its snappy, so object achieved so far.

Waiting for the new cable drag chains, last were a bit small so now i bought bigger from Russia and should be any day here

Enjoy:


https://www.youtube.com/watch?v=ohU3zJLcR1E

toomast
21-08-2015, 11:14 AM
COOL!
Watch out for ESD.

CanonGuy
05-11-2015, 01:29 AM
love this build! Any videos of it cutting yet? :thumsup:

Lachlan
05-11-2015, 11:37 AM
Boyan Absolutely awesome build as per usual! Actually it qualifies as CNC porn I think! I am soo jealous well done!!!!

(Way off topic) One thing I worry about is the roof of your extension has no gutter! So all the water is going to run down the wall and into everything you have leaning or hanging on the wall. Am I missing something?

Oh and more videos or at least photos of chips flying when you can!!!:eagerness:

AndyGuid
05-11-2015, 12:11 PM
Maybe it's a case of the rain in Spain falling mainly on the plain and Boyan living in the Spanish mountains where it seldom rains?:hysterical:

Chaz
28-11-2015, 04:42 PM
Drilled the UPN, soldered oversized nuts to form a bed, drilled machine and fixed with bolts. Same as what i did here http://www.mycncuk.com/threads/8600-Welding-bench



Thanks for the kind words :-)


Have not updated it long time. Its almost finished, now rectifying some design mistakes. Will speak about them later.

Here is a video i made yesterday of test runs at 10000mm/min acceleration 3000, gantry weights 180-160kg or sth like that. Its scary. Its snappy, so object achieved so far.

Waiting for the new cable drag chains, last were a bit small so now i bought bigger from Russia and should be any day here

Enjoy:


https://www.youtube.com/watch?v=ohU3zJLcR1E

What motors / gear ratios?

Boyan Silyavski
02-12-2015, 05:45 PM
HI guys,
sorry have not been here a while. Thing is that have been busy with small orders last 3 months and i just couldn't find the time to finish the machine.


love this build! Any videos of it cutting yet? :thumsup:

Soon.



Boyan Absolutely awesome build as per usual! Actually it qualifies as CNC porn I think! I am soo jealous well done!!!!

(Way off topic) One thing I worry about is the roof of your extension has no gutter! So all the water is going to run down the wall and into everything you have leaning or hanging on the wall. Am I missing something?

Oh and more videos or at least photos of chips flying when you can!!!:eagerness:

Thanks! Rain goes down on the wall and then on ground. Distance between wall and shed is 2-3cm. reason being is that even if i made it closed using PU foam, water will still drip due to the wall type. So i did not bother


Maybe it's a case of the rain in Spain falling mainly on the plain and Boyan living in the Spanish mountains where it seldom rains?:hysterical:

In fact before i made the shed it was 11months drought here. I finished the shed- and it could not stop raining.

But main reason for shed is not rain only but sun, cause at summer is very hot here. Everyday.


What motors / gear ratios?

400w AC servo motor/ballscrew = 2:3 /20t:30t/ so 1:1.5 meaning top speed at 3000RPM will be 20m/min though that motors could easily spin 5000RPM but there the power drops.



Last Christmas orders, so i send my wife to her mother, switch off my phone for a week and finish the machine /hopefully/

Main problem after all was not aligning the ball screws well to the rotating nut assembly/ due to chinese nut out of spec size/ , so now thats what i have to do .then cables and box.

Chaz
02-12-2015, 08:02 PM
HI guys,
sorry have not been here a while. Thing is that have been busy with small orders last 3 months and i just couldn't find the time to finish the machine.



Soon.




Thanks! Rain goes down on the wall and then on ground. Distance between wall and shed is 2-3cm. reason being is that even if i made it closed using PU foam, water will still drip due to the wall type. So i did not bother



In fact before i made the shed it was 11months drought here. I finished the shed- and it could not stop raining.

But main reason for shed is not rain only but sun, cause at summer is very hot here. Everyday.



400w AC servo motor/ballscrew = 2:3 /20t:30t/ so 1:1.5 meaning top speed at 3000RPM will be 20m/min though that motors could easily spin 5000RPM but there the power drops.



Last Christmas orders, so i send my wife to her mother, switch off my phone for a week and finish the machine /hopefully/

Main problem after all was not aligning the ball screws well to the rotating nut assembly/ due to chinese nut out of spec size/ , so now thats what i have to do .then cables and box.

Thanks. What make of servo / drivers, not sure if this was in the thread somewhere ...

Boyan Silyavski
02-12-2015, 08:58 PM
Thanks. What make of servo / drivers, not sure if this was in the thread somewhere ...

Samsung , ebay , Korea

Chaz
02-12-2015, 10:27 PM
Samsung , ebay , Korea link please?

Boyan Silyavski
03-12-2015, 08:45 AM
The shop (http://stores.ebay.com/pkusedparts/)

SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMQ-04BA1ANM3 400W w/CABLE thats what i bought for Z

SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMT-04BB1ANT3 400W w/CABLE thats for the gantry

Now the first motor is shorter and fatter and the second is smaller and longer, more like a typical stepper. Both are ok. When buying servo motors and drives make sure they are 220V AC, motor is standard inertia and not some specialized motor for a special purpose, that drive works in STEP DIR mode, especially take care with Iskawa as not all will work ok for CNC control. Make sure you buy them with cables as 1 cable is >100 euro or more

I was between these and Panasonic/ a bit more expensive/ but decided on price. The best thing about these drives is Autotuning. You push the button and it starts jogging for 3 seconds and that's it-its tuned.

So my recommendation is the SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMT-04BB1ANT3 400W w/CABLE as it more standard size. and smaller shaft or servo motor plus automatic tuning.

Another good options are the Minawa /Panasonic series /


The other money wise option is Chinese servo drive drive MCAC808 and 400w servo: 60ASM400 which last i checked were ~180USD if you buy lot/4/ from Shenzhen Just Motion Control Electromechanics Co.,Ltd. Write to Molly/remove space/ molly @jmc-motion.com
Tell her Boyan from Spain send you there so my 30 email exchanges with her help you with the price.
I would use these motors in my prototype machines and hopefully later. These are probably the cheapest servos ever directly from a manufacturer. If you send me your mail i will save you needless conversations and send you some documents about these servos.

Hope that helps. There are other options but these are my conclusions as a balance of features and moneywise

Chaz
03-12-2015, 10:23 AM
The shop (http://stores.ebay.com/pkusedparts/)

SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMQ-04BA1ANM3 400W w/CABLE thats what i bought for Z

SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMT-04BB1ANT3 400W w/CABLE thats for the gantry

Now the first motor is shorter and fatter and the second is smaller and longer, more like a typical stepper. Both are ok. When buying servo motors and drives make sure they are 220V AC, motor is standard inertia and not some specialized motor for a special purpose, that drive works in STEP DIR mode, especially take care with Iskawa as not all will work ok for CNC control. Make sure you buy them with cables as 1 cable is >100 euro or more

I was between these and Panasonic/ a bit more expensive/ but decided on price. The best thing about these drives is Autotuning. You push the button and it starts jogging for 3 seconds and that's it-its tuned.

So my recommendation is the SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMT-04BB1ANT3 400W w/CABLE as it more standard size. and smaller shaft or servo motor plus automatic tuning.

Another good options are the Minawa /Panasonic series /


The other money wise option is Chinese servo drive drive MCAC808 and 400w servo: 60ASM400 which last i checked were ~180USD if you buy lot/4/ from Shenzhen Just Motion Control Electromechanics Co.,Ltd. Write to Molly/remove space/ molly @jmc-motion.com
Tell her Boyan from Spain send you there so my 30 email exchanges with her help you with the price.
I would use these motors in my prototype machines and hopefully later. These are probably the cheapest servos ever directly from a manufacturer. If you send me your mail i will save you needless conversations and send you some documents about these servos.

Hope that helps. There are other options but these are my conclusions as a balance of features and moneywise



Thanks.

Chaz
14-01-2016, 03:52 PM
Any progress with this machine?

Boyan Silyavski
14-01-2016, 06:23 PM
Any progress with this machine?

:hopelessness:
I need 5 days to finish it and could not find the time from a couple of months. Either i am on a vacation with wife, very busy with orders or tired. Hopefully very soon, as i start needing it.

left long ball screw is unmounted and the ball nut needs centering. Cause the bloody Chinese made it bigger than the drawing so it did not fit right in the rotating nut. Change one servo cable and make the enclosure box

Scott Damman
03-03-2016, 05:24 PM
:hopelessness:
I need 5 days to finish it and could not find the time from a couple of months. Either i am on a vacation with wife, very busy with orders or tired. Hopefully very soon, as i start needing it.

left long ball screw is unmounted and the ball nut needs centering. Cause the bloody Chinese made it bigger than the drawing so it did not fit right in the rotating nut. Change one servo cable and make the enclosure box


Just have to say killer work! This thread has greatly helped me in the start and development stages of my build. Im excited to see yours up and running.

Thought I would post to keep track of your progress.

Boyan Silyavski
19-04-2016, 11:15 PM
Some small progress.

At the end i found the cable problem that was making all time my Z axis servo trip. I measure all- fine, test the machine and the Z trips. measure again all connections fine. Come on.... Eliminated connectors and stuff, and at the end it was the cable. Some times giving error some times not.


Have found proper steel box for the electronics and painted it. Soon all will go there.


These days have made some test runs and had problems with the Pokeys57CNC board. Working a test program and after 15 min or so disconnects from mach 3. Had to reinstall all, update firmware and today at the end was working fine for the moment


Another problem with this board is the Spindle speed control. This could be my mistake though to lack of understanding. I read 10 times at least http://www.machsupport.com/forum/index.php/topic,32353.0.html and https://www.machsupport.com/forum/index.php?topic=22727.0. So i have in my pulley ratio min 1000 max 24000 and ratio 1 and at the PWM screen of the plugin 1000hz refresh and i figured 1/240=0.004667 multiplier. So the problem being is even with mach3 stopped there is 3v from board and could not be lowered using the micro, did i say "MICRO" adjusting pot. Obviously i have to do something in the VFD adjustments, other wise from 12000 to 24000 rpm works fine with that numbers. I hope i am not wrong


So i have a couple of jobs queued already and no time to play a lot. So basically i will do some jobs as it is, cables are short to use full table length but i dont need that for the mooment.


Now remember what i have always said that i will not use it to mill steel. contrary to that i decided the first ever cut to be on steel, just to check if i could surface the table :-)

Opened HSMadvisor and as i dont have here any more suitable bits i plugged a 6mm 2 flute carbide, adjusted manualy the spindle speed and gave it a go. At 1mm depth of cut though, i did not want to get overexcited on the first run:


18240


Well, it cut it dry without problems, but a spilled a bit of oil on the table , ha ha, The misting cooler is still not ready for that machine, though i have all the parts and the tubes are on place.

About the cut. No vibrations on the machine. Perfect surface, well it seems i have trammed well the head. Thanks god. As if you remember i did it using straight edge and precision squares. I did not have Dial Indicator at that moment. Just used it later to confirm.

The cut may not look great on picture, as its magnified, but believe me-it is mirror like and scratching with finger is like scratching glass.


18241


here is the video of the first cut. My mobile amplified the sound, its quiet in reality, but if you listen to it you will hear not any ringing or whatsoever from the machine.



https://www.youtube.com/watch?v=PruNXn3vUQ4


I decided to stop for today as i have no water pump connected to spindle. And came here to brag a bit :yahoo:

JoeHarris
21-04-2016, 07:49 PM
I take it you are not going to Surface the whole bed with that cutter?!

ImanGH
19-10-2016, 06:06 PM
Hi boyan
are you using epoxy for y axis? if yes how use?
thank's

feldah
08-12-2018, 04:21 AM
Hi there,
Long time no update. Long story short. Bought new epoxy and did it as I thought it would be right. It worked perfectly. As supposed the long rails sucked from the bridge. I went even further and made only 1 bridge. Yeah, the epoxy leveled without problems. I knew it.
This time I used Knauf channel 0.5mm 3m - 3€ from construction material site. Every attempt I am getting better for cheaper.
Also used 8 mm square bar around the epoxy. The channel was 4 mm wide so the bearings not to hit the bar

some photos. Will upload here also but now I am writing from my phone
https://plus.google.com/app/basic/photos/111571569713128538019/album/6082431536953046609?cbp=1b0j3yawglvr9&gl=es&sview=54&cid=5&soc-app=115&soc-platform=1&spath=%2Fapp%2Fbasic%2Fphotos&sparm=cbp%3Dpt4pyhrcs2kc%26gl%3Des%26sview%3D18%26 cid%3D5%26soc-app%3D115%26soc-platform%3D1%26pv%3Dhi%26pgpnum%3D1


Ps. Here are the photos:

http://www.mycncuk.com/attachment.php?attachmentid=13868&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13869&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13870&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13871&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13872&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13873&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13874&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13875&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13876&stc=1

http://www.mycncuk.com/attachment.php?attachmentid=13877&stc=1Sorry to pick up that very old thread. But as I am new can't write messages. Could you explain how you did the dam beside the rails. You welded it on. But how did you make sure it's properly sealed so no Epoxy went under it. For the bridge I understood you took a channel from Knauf.

I was considering to upgrade my CNC with epoxy where I drilled the holes already. To prevent the epoxy to float in the holes. I see two options do it as you did and fill all holes with silicone. Or weld a channel over all holes.
Thanks a lot
Felix.

Sent from my MI 6X using Tapatalk

Neale
08-12-2018, 10:55 AM
A good way to seal the leaks around a dam before you pour your epoxy is to make up a very small amount of the epoxy and using a small brush or similar just go round the inside of the dam brushing the epoxy into place. it will run into the holes or bridge the smaller holes but because there is not a large quantity behind it, will not run right through and leak out the bottom. You only need to leave it long enough to stop flowing easily before you pour the main epoxy batch - you do not need to leave it for days to fully harden. I made my dam from gaffer tape, stuck to the sides of the rails so that the edge of the tape was above the rails. Biggest problem was peeling it away from the epoxy when the epoxy had hardened. Next time I shall try pulling it away before the epoxy has fully hardened!

I had to strip my first load of epoxy off the rails and start again, but I had already drilled the rails. I put small pieces of Sellotape over the holes and poured the epoxy over the top. I could see where the holes were and then just drilled through the epoxy after it had set. Worked for me!

feldah
08-12-2018, 11:06 AM
Ok think that'll do. Maybe to be double sure sure I will also use silicone. I imagine that would be frustrating seeing your epoxy flushing in a not properly sealed hole. :-)

Sent from my MI 6X using Tapatalk

Boyan Silyavski
26-12-2018, 08:21 PM
First time i did it not entirely right. With time discovered new ways and next time i did it welded 6mm or 8mm bar stock, distance 40mm between. Have to check if it was 8mm as it may touch the Hiwin 20 block. Better read my second build log to see how i poured a couple of times epoxy on a 3m machine untill i figure the perfect way. iin short i discovered that the most important part is to make the damned dam longer like 15 cm at least than the actual machine to compensate for the shrink and that way only one bridge is enough, and the bridge must be then outside the machine. So when all dries and shrinks, the bridge will not eat from the actual ways but from the extra length. If you know what i mean

feldah
27-12-2018, 05:20 AM
Yes thanks. I read that post already. I will follow it. I have a vacuum chamber where I can stir epoxy under vacuum. I am excited to see if it will make a difference in accuracy. Good job btw.

Sent from my MI 6X using Tapatalk

Boyan Silyavski
02-01-2019, 08:43 PM
Yes thanks. I read that post already. I will follow it. I have a vacuum chamber where I can stir epoxy under vacuum. I am excited to see if it will make a difference in accuracy. Good job btw.

Sent from my MI 6X using Tapatalk

No difference if you stir it properly, without introducing too much air. Remember you torch it later also. But i agree that vacuum is best

NordicCnc
23-04-2020, 05:46 AM
The shop (http://stores.ebay.com/pkusedparts/)

SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMQ-04BA1ANM3 400W w/CABLE thats what i bought for Z

SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMT-04BB1ANT3 400W w/CABLE thats for the gantry

Now the first motor is shorter and fatter and the second is smaller and longer, more like a typical stepper. Both are ok. When buying servo motors and drives make sure they are 220V AC, motor is standard inertia and not some specialized motor for a special purpose, that drive works in STEP DIR mode, especially take care with Iskawa as not all will work ok for CNC control. Make sure you buy them with cables as 1 cable is >100 euro or more

I was between these and Panasonic/ a bit more expensive/ but decided on price. The best thing about these drives is Autotuning. You push the button and it starts jogging for 3 seconds and that's it-its tuned.

So my recommendation is the SAMSUNG SERVO MOTOR DRIVER CSDJ-04BX2, CSMT-04BB1ANT3 400W w/CABLE as it more standard size. and smaller shaft or servo motor plus automatic tuning.

Another good options are the Minawa /Panasonic series /


The other money wise option is Chinese servo drive drive MCAC808 and 400w servo: 60ASM400 which last i checked were ~180USD if you buy lot/4/ from Shenzhen Just Motion Control Electromechanics Co.,Ltd. Write to Molly/remove space/ molly @jmc-motion.com
Tell her Boyan from Spain send you there so my 30 email exchanges with her help you with the price.
I would use these motors in my prototype machines and hopefully later. These are probably the cheapest servos ever directly from a manufacturer. If you send me your mail i will save you needless conversations and send you some documents about these servos.

Hope that helps. There are other options but these are my conclusions as a balance of features and moneywise



Hello Boyan.

Nice machine I have to say. Looks very stiff! How is it running these days? Do you have a lot of work for it or mostly hobby stuff?

Anyway I wanted to ask how you went about auto-tuning the axis with double servo motors (Y-axis in my eyes, not the gantry). Auto-tuning is a very nice feature in any servo motor software, but I have not figured out how to tune an axis with double servo motors.

Also how did you go about calculating the servo motor size? Perhaps you have explained this in a post somewhere, but I couldn't find it! :)

Thanks!

Boyan Silyavski
23-04-2020, 09:21 AM
Hello Boyan.

Nice machine I have to say. Looks very stiff! How is it running these days? Do you have a lot of work for it or mostly hobby stuff?

Anyway I wanted to ask how you went about auto-tuning the axis with double servo motors (Y-axis in my eyes, not the gantry). Auto-tuning is a very nice feature in any servo motor software, but I have not figured out how to tune an axis with double servo motors.

Also how did you go about calculating the servo motor size? Perhaps you have explained this in a post somewhere, but I couldn't find it! :)

Thanks!

Thanks!

I have sufficient work for the machine. Not as much as it deserves though. I make some product on it, sometimes for other people and sometimes cut letters for signs.

I tuned the Z and used same parameters on the other axis, as more or less motors a properly calculated for the loads on each axis, working at optimum.

Then one day i figured that if i move left or right motor 5mm, its not a problem as the gantry is like 2 meters. So there is play enough. Then i entered in Samsung software and limited the movement to 5mm each side and tuned the axis like that. With a short movement. I say 5mm just to be on the safe side, it could have been 10-15mm, on such a long gantry this is like nothing, before the other side tries to engage. In Real life if i switch off 1 motor i can move the gantry front/back even with one motor only .Of course autotuning stresses the thing, so better in the short safe range. Check all that range adjustment on the Z firts, so you dont bend something.

The servo motor size Jonathan or Dean helped me, its on the first pages somewhere. But basically 400 or 750W ac servo at 230vac are the thing for a machine like that. I went with 400w as ballnut is geared 3:2 so at 3k rpm of motor i am at 20k mm / min and my motor power is multiplied by 1.5 at that ratio. My motors support up to 6k rpm so ...

As far as i understand the things, Panasonic, Samsung or other brand servos are way better than the chinese ones, so if i was going with China servos would have went with 750w ones, as price is irrelevant also on the chinese servos. No big difference. But remember. All time i am talking about AC servos driven by 230VAC drives.

NordicCnc
23-04-2020, 09:53 AM
Sounds good that you have work for the machine!

Thanks Boyan, for all the answers.

I like your idea and it makes sense to perform auto-tuning like that.

Myself I have build log (build in real life not started yet) if you are interrested to check it out.

Anyway I was thinking about those servo sizes as well. I will take your advice and get Delta 750W servos! Perhaps also 2020 (plan was 2010 with 400W servos) ballscrews to get a crazy speed of 30-40m/min, depending on what gear ratio I choose (3:2 or 2:1).

Not sure if I ever will need that speed but nice to have some room for experimenting.



Skickat från min SM-A530F via Tapatalk

Bongo
10-07-2020, 08:06 PM
Hi Boyan. Inspiring. After reading this thread I had to join the forum, just to say how cool your machine is and how impressed with your work I am. Thanks so much for sharing. I am partway through building a beast-sh 1.2 x 2.4m steel and epoxy machine, which I plan to run with similar Samsung servos to you. I am just debating whether to spin the 2 long ballscrews or their ball nuts to eliminate the potential problem of whipping.
Like you I am aiming at snappy ;)
Anyway, I wonder if you would be able to share some of the details of how you designed / constructed that area where the servos, belt drive the ballnuts and attach to the gantry ends?
Thanks again for your inspiring work!

Boyan Silyavski
13-07-2020, 11:14 AM
I am partway through building a beast-sh 1.2 x 2.4m steel and epoxy machine, which I plan to run with similar Samsung servos to you. I am just debating whether to spin the 2 long ballscrews or their ball nuts to eliminate the potential problem of whipping.
Like you I am aiming at snappy ;)
Anyway, I wonder if you would be able to share some of the details of how you designed / constructed that area where the servos, belt drive the ballnuts and attach to the gantry ends?
Thanks again for your inspiring work!

Hi,
will take photos

Definitely go rotating nuts. Normally i run it at 8000 mm/min for wood as faster wood will chip. But for foam or 3d jobs i run it 16.000 or faster. And being snappy means no slowing in corners. Especially important if doing deep trochoidal toolpaths. Meaning a very serious difference in time. Not that a 10.000 mm min max speed machine is not enough for a normal work, but if you have an order of say 100 doors or something the difference will be like 4 or 5 more days to do the job. And talking about profiling the panels only. Not 3d job

Bongo
13-07-2020, 02:30 PM
Hi,
will take photos

Definitely go rotating nuts. Normally i run it at 8000 mm/min for wood as faster wood will chip. But for foam or 3d jobs i run it 16.000 or faster. And being snappy means no slowing in corners. Especially important if doing deep trochoidal toolpaths. Meaning a very serious difference in time. Not that a 10.000 mm min max speed machine is not enough for a normal work, but if you have an order of say 100 doors or something the difference will be like 4 or 5 more days to do the job. And talking about profiling the panels only. Not 3d job

Hi Boyan, thank you for the good information. Yes I want to aim for some of those speeds :tyrannosaurus: without it being a disaster if I don't achieve - This is mainly ambitious hobby stuff, but would like to have the option to put it to work for money.

I guess you are not a fan of ballscrew damper type things like this - https://www.youtube.com/watch?v=NWB6FAJCPhA ?
Even if I ddn't have a problem with whipping, I would be a bit concerned about the rotating inertia of the ballscrews if I went the driven screw rout.
The two long ballscrews I have are 32mm dia. 5mm pitch and 2300mm stroke. They are only rolled c7 accuracy (claimed) ones from China, but will hopefully work for what I want. Maybe I would have gone 15 or even 20mm pitch if I was doing over, but I started this a long time ago... I bough end supports with them - only cheap ones though so not worried if I don't end up using them...

I look forward to seeing the photos of your machine's design in this area. Am almost certain I will copy it quite closely. What size belts did you go for? Any leads on best place to get belts and pulleys?
Thanks again for your guidance on this - priceless info from an experienced builder!

PS.
I just pulled the trigger and bought 4 of the 400W AC Samsung servos and drivers, similar to the ones you use, from dy-global. Fingers crossed they get here ok.

Boyan Silyavski
13-07-2020, 05:23 PM
Hi Boyan, thank you for the good information. Yes I want to aim for some of those speeds :tyrannosaurus: without it being a disaster if I don't achieve - This is mainly ambitious hobby stuff, but would like to have the option to put it to work for money.

I guess you are not a fan of ballscrew damper type things like this - https://www.youtube.com/watch?v=NWB6FAJCPhA ?
Even if I ddn't have a problem with whipping, I would be a bit concerned about the rotating inertia of the ballscrews if I went the driven screw rout.
The two long ballscrews I have are 32mm dia. 5mm pitch and 2300mm stroke. They are only rolled c7 accuracy (claimed) ones from China, but will hopefully work for what I want. Maybe I would have gone 15 or even 20mm pitch if I was doing over, but I started this a long time ago... I bough end supports with them - only cheap ones though so not worried if I don't end up using them...

I look forward to seeing the photos of your machine's design in this area. Am almost certain I will copy it quite closely. What size belts did you go for? Any leads on best place to get belts and pulleys?
Thanks again for your guidance on this - priceless info from an experienced builder!

PS.
I just pulled the trigger and bought 4 of the 400W AC Samsung servos and drivers, similar to the ones you use, from dy-global. Fingers crossed they get here ok.



There is a rotating ball nut thread in the forum, most info is there including free designs from me and other people


15mm HDT 5 belts

I use Belting online for belts and pulleys. Once bought 20 crappy chinese pulleys so -never again


i use this calculator (https://www.bbman.com/belt-length-calculator/) for the centre to centre distances and other stuff


I can send you the CAD files if needed for the rotating nut , ball screw is 25 size, screw should be manually greased well as it heats when things get faster

Bongo
13-07-2020, 05:28 PM
Hi Boyan, yes, thank you, I have occupied the last three hours digesting that other thread on the rotating nut design :) Looks good.
I would much appreciate the CAD files to tinker with - there was a link in the other thread, but I think it has expired.
Thanks for the guidance regarding the belts - that probably saved me a lot of pain.

Edit: Your ballscrew was 10mm pitch if I'm not mistaken? Would it seem reasonable to run a 1:1 pulley on mine to the 400W servos because pitch is 5mm? Thanks again!

Boyan Silyavski
13-07-2020, 05:40 PM
I guess you are not a fan of ballscrew damper type things like this - https://www.youtube.com/watch?v=NWB6FAJCPhA ?

If you see a 250kg gantry moving at 20k mm min at 3000mm/sec you would laugh at the speed shown on the video.

nhan863240
18-04-2023, 11:15 AM
Hi Silyavski,

I have to dig this old thread up as I have a tricky question (for me) regarding your frame.

I see that you welded your frame and everything including parts of the Z assembly. Did you do some stress relieve? If not, how the residual stress affect the accuracy, given that you have been using it for some years?

Thank you for sharing! Not only on this thread but others, I learnt a lot from you (particularly about belt and pulley) from another thread.

nhan863240
18-04-2023, 06:04 PM
Deleted