PDA

View Full Version : BUILD LOG: All steel, no welding, 2000x1300x250mm



eurikain
18-08-2015, 10:52 PM
Hi !

This is a story about an all steel, 2000 x 1300 x 250 mm work area CNC router, where nuts and bolts replace weldments.

The idea is quite simple and the concept is not new and has been tested and verified to work on a small steel design (http://www.machsupport.com/forum/index.php?topic=23730.0). I decided to scale it up to 10mm steel sheets.

But my goal was not only to remove the need of welding, but also to integrate a linear shaft support so that round supported rails can be made cheaper. Almost all parts will be plasma cut from 10mm steel sheets. A few parts, like the Z axis moving plate and the ball nut attachment plates will be cut from 15mm steel sheets to reduce binding or avoid bracing. Linear rails are standard 20mm stainless steel shafts. and will be welded to the supports (I haven't found a way to fix them using nuts and bolts that would fit with my design, and I don't believe it would be a good way to do it).

Here are a few renders of the complete Z axis design (calculated weight is 9kg without spindle):

http://www.mycncuk.com/attachment.php?attachmentid=15907&stc=1
Complete Z axis.
http://www.mycncuk.com/attachment.php?attachmentid=15906&stc=1
Inside the frame.
http://www.mycncuk.com/attachment.php?attachmentid=15905&stc=1
This image shows the steel bolt and nut connection, plus the integrated linear shaft support. You can also see how the connections affect the part by cutting the shape of the bolt and nut, and adding joints on both sides of the bolt. I actually made smart components that achieve this. I have a few different versions and sizes for the connections:
- simple: cuts the shape of the bolt and nut
- single: cuts the shape of the bolt and nut + a "tooth" on one side (sizes are small, medium or large)
- double: cuts the shape of the bolt and nut + a "tooth" on both sides (sizes are small, medium or large)
All cuts have a play of .1mm. Bolts are M10 x 35mm with matching nut (standard, grade or pitch don't really matter as there aren't any tapped holes).
http://www.mycncuk.com/attachment.php?attachmentid=15908&stc=1
Back of the Z moving plate with balll nut support. All made from 15mm steel sheet.

Machine specification:

Workarea:
X: 2000mm
Y: 1300mm
Z: 250mm

Weight:
X: ?
Y: 28kg without gantry sides and without Z
Z: 9kg without spindle

Linear motion:
X: SBR20UU + SFU2010 2000mm
Y: SBR20UU + SFU1610 1300mm
Z: SBR20UU + SFU1605 500mm (because it came with supports and couplers, and was cheaper than shorter lengths)

Linear motion electronics:
X: 2 x Nema 23 425oz/in driven by 2 x Leadshine AM882
Y: 1 x Nema 23 450oz/in driven by Leadshine AM882
Z: 1 x Nema 23 425oz/in driven by Leadshine AM882

Spindle:
2.2kw water cooled + VFD

Breakout board:
Basic 5 axis board, 4 limit switches, with Mach3 support (https://buildyourcnc.com/item/electronicsAndMotors-3axis-425-elcombo), from BuildYourCNC.

Power supply:
DIY unregulated 70V 10A, following Boyan and Clive advices (http://www.mycncuk.com/threads/8966-Which-PSU-and-drivers-do-I-need).

Where I got the parts:
- Ballscrews + nuts + end supports + couplers are from AliExpress seller Kevin Ho at ROB BEARING AND MECHANICAL PARTS (http://www.aliexpress.com/store/334759). Contact him if you need custom lengths (that's what I did).
- Nema 23 425oz/in + breakout board are from buildyourcnc.com 3 axis Nema 23 electronics kit (I purchased them 4 years ago).
- Nema 23 450oz/in is from a random seller on AliExpress.
- spindle kit is originally from another AliExpress seller (http://www.aliexpress.com/item/Water-Cooled-Spindle-Set-2-2KW-Spindle-Motor-Matching-2-2KW-Inverter-80mm-clamp-water-pump/2041841053.html) but I found the shipping price too high, so I asked Kevin Ho (the guy from whom I bought the ballscrews) if he could get the kit for me with a better price and he gave me a $60 discount (this is calculated using kit price on August 18th of 2015, which is the day I am writing these lines).
- AM882 drives are from here (http://www.aliexpress.com/item/inStock-Free-Shipping-Leadshine-2-Phase-Digital-Stepper-Drive-AM882-With-SENSORLESS-Stall-Detection-Max-80/32258905084.html). Be careful to make orders with maximum THREE drives or you'll pay for shipping. I did an order of three and another of one.
- SBR blocks are from here (http://www.aliexpress.com/item/Promotions-4-pcs-SBR20UU-Linear-Bearing-20mm-Open-Linear-Bearing-Slide-block-free-shipping-20mm-CNC/1765771803.html). Again, be careful to make orders of maximum TWO packs, or you'll pay for shipping. I did an order of two and another of one.

Currently shipping:
- spindle kit
- 3 nema 23 couplers (made a mistake by ordering 8mm couplers instead of 6.35, so had to re-order).

Modelling on:
Solidworks 2014 SP2
Windows 7 Home Basic
Eight year old HP Pavilion DV6500, 2GB RAM, small NVIDIA graphics chipset

Your input is welcome and very valuable to me ! I will make the DXF files when done modeling the entire machine, and when most of you will agree with the general design. Plus, there's no hurry to get the parts, as the company I work with will have them cut in the next 24 hours after they receive the files.

Special thanks to Boyan Silyavski, Clive S, toomast, all members that have answered my questions, and all members that contribute to this amazing forum.

I will start the design of the gantry tomorrow. I am too tired of clicking the Solidworks "rebuild" button for today.

Cheers,
Eurikain.

Clive S
19-08-2015, 07:38 AM
It's certainly impressive I can see a lot of work has gone into this.
I do worry about the straightness of the steel to get an accurate machine. May I ask why you have not gone with square profile rails? I look forward to watching this develop. :welcoming: ..Clive

eurikain
19-08-2015, 11:04 AM
It's certainly impressive I can see a lot of work has gone into this.

Hi Clive !
Thanks ! Yes, you are right, probably more than 30 hours of modeling so far, because I went through several designs, and had never used SW before, so I had to learn everything from zero ! But now I've got most of it figured out :D Great program to work with, though sometimes a bit frustrating.


I do worry about the straightness of the steel to get an accurate machine.

I understand, I am not sure myself as it is an experiment. In the project I linked in my original post (http://www.machsupport.com/forum/index.php?topic=23730.0), they got a .1 play on joints, 0 play on holes (that's what I've got), and they cut the plates on a .001 accuracy machine, so plays were correct when cut. The machine that will cut my parts has more than .001mm so I'm not really worried about cut straightness. if all edges, joints, teeth and pockets are cut straight I wouldn't see how there could be any straightness error :/ Where would they come from ? Are you talking about the steel straightness itself ? May be I can ask them to surface the sheets from which they cut the parts on both sides ? Would that be a fix ?


May I ask why you have not gone with square profile rails?

Price :) ! The goal was to make things cheapER. 12 SBR20UU blocks cost me $60. Stainless steel shifts will be less than $50 for all axis. Shaft supports are integrated to parts and don't add any cost to the machine but a few seconds of cutting.


I look forward to watching this develop. :welcoming: ..Clive
Thanks ! I am very excited to get the model ready and working so I can send the parts to the cutter :) !

Wal
19-08-2015, 01:24 PM
Price :) ! The gold was to make things cheapER. 12 SBR20UU blocks cost me $60. Stainless steel shifts will be less than $50 for all axis. Shaft supports are integrated to parts and don't add any cost to the machine but a few seconds of cutting.

Hi Eurikain.

I've used TBR20's on my build and they work well etc. BUT, if I was building it again I'd use profile rail. Why? Rigidity (my rails aren't really long enough for this to be an issue for me) and ease of maintenance. Have a look at the attached pic:

http://www.mycncuk.com/attachment.php?attachmentid=15909&stc=1

The bearing housing in the block will work it's way loose from time to time despite the top and side grub screws that are meant to hold it in place (you can only nip 'em up so far before the unit binds). Once the bearing housing has loosened it will, over time, rotate and eventually the sharp corners of the cylinder will scrape along the rail supports. This isn't a catastrophe, nor should it lead to a catastrophic failure, but it's an annoyance that'll cost you time and an extra thing to watch out for. No biggie for a hobbyist like myself, but if downtime's something you want to avoid then do yourself a favour and buy profile rails...!

Wal.

eurikain
19-08-2015, 01:58 PM
Hi Wal, and thanks for the input :)

I appreciate the advice ! Though I think I won't worry much about this as this is my first build and as I don't intend to machine things that often (I'm still a student and the machine will only be accessible on weekends). I went with SBR rails because in my humble opinion replacing a $5 SBR block every so often is not such a problem :) The machine design is done so that you can take it easily apart whenever you want and put it back together without loosing starigthness, rigidity or anything, as steel won't move when mounting / unmounting.

May be in a later version of this machine I will switch to profile rails and HWIN blocks though :)

Wal
19-08-2015, 02:16 PM
That's fair enough. Just thought I'd give you a heads up. Best of luck with the build!

Wal.

eurikain
19-08-2015, 02:55 PM
That's fair enough. Just thought I'd give you a heads up. Best of luck with the build!

Wal.

Very nice of you, that's what this forum is for and I'm looking forward to reading all you guys' caution advices and design improvement comments :very_drunk: !

eurikain
20-08-2015, 10:15 AM
Hey mates ! (hahahahaha for those of you who will get it)

I spent yesterday modeling the Y axis. Today I will make the gantry sides to fix the Y axis, fix its motor, and connect Y and Z if I have the time.

Here is a render of the Y axis with a section view (calculated weight is 28kg):
http://www.mycncuk.com/attachment.php?attachmentid=15913&stc=1

Wal
20-08-2015, 01:12 PM
Looking at your renders - in reality, your rails will be supported along their entire length, won't they? Not at intervals, like you've got them in the images..?

Wal.

Clive S
20-08-2015, 01:30 PM
Looking at your renders - in reality, your rails will be supported along their entire length, won't they? Not at intervals, like you've got them in the images..?

Wal.Wal according to his first post he intends to weld them to the supports so not using conventional rails. ..Clive

eurikain
20-08-2015, 01:40 PM
Wal according to his first post he intends to weld them to the supports so not using conventional rails. ..Clive

That's right :) They will be supported at intervals. That's because they are integrated in the design. There is no efficient way to plasma cut full support profiles. Or you would need to cut the profile from lets say 20mm steel and stack all the supports next to each other, but that's just too much of a pain and would lead to alignment inaccuracy.

Supporting them at intervals won't cause binding of the 20mm plain stainless steel shaft, considering that the interval length stays reasonable. Current interval is 220mm on Y axis and 240mm on Z axis.

Wal
20-08-2015, 03:15 PM
>That's right :) They will be supported at intervals.

Ah, okay. You'll need to keep a close eye on those creeping bearing housings then, if one of those housing edges were to catch on a support it could spell trouble, although you'll probably hear them knocking/scraping before it does. Bear in mind that the machine could be working on a small part of a project, going back and forth many times over (with the bearing housing silently creeping out of place but not having to pass a support strut...) before moving on to a different part of a table... If it's crept enough to catch on a support strut... Disaster.

You might want to 'arrest' the housings so that they don't rotate past a certain point, ie. with a blob of epoxy - I don't mean for you to glue 'em in... Just a blob or two on the outside of the housings that stops them rotating past a certain point... [I should probably do the same with mine...]

Wal.

Wal
20-08-2015, 06:41 PM
Wait - weld...? Weld the supports to the rails...? Surely you'll be asking those rails to bend...?!

routercnc
20-08-2015, 07:00 PM
I've just jumped into this thread and I would think twice about welding the rails. They will almost certainly bend out of alignment with the heat and shrinkage, even I suspect with TIG which can be more delicate.

It may also remove the hardening from the local area although that will probably be the least of your problems.

Don't want to knock your ideas but I would seriously check whether this is a good strategy - so buy a very small length of rail, put a bearing on it with a cheap laser pointer aimed a a wall. Mark the spot, then weld the rail to a something to replicate the support post and see what happens to the spot. Then decide if this approach is going to work before spending money on all the laser cutting . . .

To be honest, supported round rail would be a much better option and it is easy to bolt down via the flanges.

Wal
20-08-2015, 07:08 PM
>To be honest, supported round rail would be a much better option and it is easy to bolt down via the flanges.

+1 - Even if you insisted on not using the supports that the rails comes with, you can unbolt them and use the pre-drilled/tapped holes in the rail to attach your own supports...

eurikain
20-08-2015, 09:17 PM
Thanks for the replies ! It's nice to see the support !

You are right, welding the rails will cause bending. What about brazing ?

Pass the next paragraph if you're not interested in my totally subjective opinion and prefer getting to the more relevant part of my post :D

Now, here is the problem I have with the fully supported rails: I have looked into that option for quite a while, but nobody at AliExpress will accept to ship 2000mm round supported rails to me... I asked already, they shipped the SFU2010 2000mm lengths without any problem, but won't take the risk of sending those rails over 1500mm. Now that's only one side of the problem... the other one is that shipping cost is just a waste of money... there is no real added value to those rails except the fact that they're round stainless steel rails mounted on stainless steel supports. Now this is just my opinion, and I know that using them would be the best option, but I just feel really bad spending $300 on shipping on a $100 rail, that's just not enough cost effective for me and not in the philosophy of this machine :(


Even if you insisted on not using the supports that the rails comes with, you can unbolt them and use the pre-drilled/tapped holes in the rail to attach your own supports...
Now that's a winner :) But, I can get those wholes drilled and tapped in a 20mm stainless steel shaft, here in Belgium, with $0 shipping cost. I could include that into my design as well ! Do you have a picture of a supported round rail where shaft and support are separated ?

toomast
20-08-2015, 09:41 PM
Something fresh and new!

Your calculated weight numbers seems to be very low! 10mm and 15mm steel is heavy stuff...

My experience with laser and plasma cutting tells me that positional accuracy is always good but edge is never 90 degrees. Plasma is worse than laser. It also gets worse with thicker materials. Most certainly you will have to use straightening tools since sheets are not flat. But good rubber hammer will get it fixed :D
You have to somehow make the openings for M10 nut pressfitted otherwise the nuts will fall out. That especially when you assemble the gantry...or you have another solution for it?

As for the rails...somehow I have a feeling that you will end up with HIWIN stuff.

Blackrat
20-08-2015, 10:05 PM
Please dont go with the round rails !

it will ruin what is looking like a nice design

thats all i have to say :S

Wal
20-08-2015, 10:09 PM
Do you have a picture of a supported round rail where shaft and support are separated ?

I don't, I'm afraid, but here's a link to some data etc. for my TBRs:

http://www.zappautomation.co.uk/mechanical-products/precision-supported-round-rail-386/tbr20-supported-roaund-rail.html

IIRC even though the holes were spaced equally, there was no control over how far the first/last bolts fell from the ends of the rail - it didn't matter as I didn't need to separate the shaft from its support and I had full control over where I was putting my own mounting holes in the supporting base. I imagine it matters a bit more for you, but you could always buy over-length and then cut down so that the holes fall where you need 'em to...

>As for the rails...somehow I have a feeling that you will end up with HIWIN stuff.

Yes yes! (...and I like my round rails, but profile is the way to go..!)

Wal.

routercnc
20-08-2015, 11:11 PM
On the rail lengths problem I know you can butt individual profile rails together to make longer rails but not seen it done on supported rails. I wonder if you cut off a piece of the supporting extrusion you get with supported rails (say 200mm) and use it to bolt 2 shorter rails together ie a joining piece. Because the support is continuous and has a shallow radius on the top which the round rails sits in it should help align the rails to each other. You could then order 2x 1m rails . Not tried it but something to think about.

Sven
21-08-2015, 08:38 AM
Okay, allow me to state it again:

Using unsupported shaft in this machine will be a waste of money and effort!
Building this thing in steel allows it to be a good performer, using unsupported shafts will allow flex, making the results no better then the same thing built of plywood.

Then two general remarks on the design:

1: It seems to me that you will not be able to assemble this thing, as the nuts will be falling out all the time.
2: If you add tabs+rectangular holes here and there to position parts, you can probably glue the thing together, if you choose the right kind of epoxy.
If not using epoxy you can still use an occasional spot weld to fix a tab in its hole.

Boyan Silyavski
21-08-2015, 03:11 PM
Dont forget to leave space for adjustments. Even if precisely laser cut, sheets bend, etc.

I would say oversize all holes by 2mm diameter

eurikain
23-08-2015, 11:11 AM
Something fresh and new!
Thanks ! I apreciate the compliment :D !


Your calculated weight numbers seems to be very low! 10mm and 15mm steel is heavy stuff...
I know, I have the same feeling, but I made sure that all parts were assigned non allied steel... so.. I don't really know what to think :/ Only the shafts are set to allied stainless steel.


But good rubber hammer will get it fixed :D
What do you mean ? I know what a rubber hammer is (at least I believe so ^^), but do you think that'll be enough to straighten it ?


You have to somehow make the openings for M10 nut pressfitted otherwise the nuts will fall out. That especially when you assemble the gantry...
Actually, the spacing between the nut and the part is .1mm on each side. The reason for that is that I want to compensate cutting accuracy, plus tolerance for bolts (I've checked a sample and they're not always the exact same diameter, weirdly).


or you have another solution for it?
I do ! Good old tape ! That will hold the nut while inserting the bolt :)


Please dont go with the round rails !
it will ruin what is looking like a nice design
thats all i have to say :S
I hear you :) Look at the end of this post for an answer ;)


I don't, I'm afraid, but here's a link to some data etc. for my TBRs:
http://www.zappautomation.co.uk/mechanical-products/precision-supported-round-rail-386/tbr20-supported-roaund-rail.html
Thanks a lot !


IIRC even though the holes were spaced equally, there was no control over how far the first/last bolts fell from the ends of the rail - it didn't matter as I didn't need to separate the shaft from its support and I had full control over where I was putting my own mounting holes in the supporting base. I imagine it matters a bit more for you, but you could always buy over-length and then cut down so that the holes fall where you need 'em to...
True, it does matter for me, but as you said, I can always buy over length and then re-cut to fit my needs.


On the rail lengths problem I know you can butt individual profile rails together to make longer rails but not seen it done on supported rails. I wonder if you cut off a piece of the supporting extrusion you get with supported rails (say 200mm) and use it to bolt 2 shorter rails together ie a joining piece. Because the support is continuous and has a shallow radius on the top which the round rails sits in it should help align the rails to each other. You could then order 2x 1m rails . Not tried it but something to think about.
That sounds like a good idea. I'll think about it, but I am worried about the moment my SBR blocks will roll on that junction... that might cause a cut "hick up", couldn't it ?


Okay, allow me to state it again:
Using unsupported shaft in this machine will be a waste of money and effort!
Building this thing in steel allows it to be a good performer, using unsupported shafts will allow flex, making the results no better then the same thing built of plywood.
Who said shafts were unsupported ? They're supported every 220mm. Go ahead and try to bind/bend a 220mm long 20mm thick plain stainless steel shaft. Yeah, probably with a 200kg Z axis, still... As currently designed, the Z will probably not exceed 15kg. Add cutting forces, let's say you go up to 40kgs (and I'm being generous ^^), considering the Z is supported in 4 points on 2 shafts, I honestly don't think binding will be an issue. But go to the end of this post for a proper answer on that :)


Then two general remarks on the design:
1: It seems to me that you will not be able to assemble this thing, as the nuts will be falling out all the time.
The solution is to assembly it smartly. Actually, two solutions exist: 1. use tape to hold the nuts in place while inserting the bolts. 2. assemble parts in an order that will allow nuts to fall on the ground instead of inside a closed space.


2: If you add tabs+rectangular holes here and there to position parts, you can probably glue the thing together, if you choose the right kind of epoxy.
If not using epoxy you can still use an occasional spot weld to fix a tab in its hole.
I like that advice very much :) I see this as a big Mechano (if you guys know that toy brand). What kind of epoxy are you thinking about ?


Dont forget to leave space for adjustments. Even if precisely laser cut, sheets bend, etc.
I would say oversize all holes by 2mm diameter

Good idea if laser/plasma cutting. Water jet cutting might be a solution.

General answer for the four main concerns:

1. Shafts aren't the way to go, they're bad, bad, extremely bad, up to the point that you'd be willing to whip me if I use them :D

People, I hear you ! But plear hear me too :) I know most of you like the HWIN rails / blocks, and I know for a fact that they are far better than SBR/TBR blocks with round rails. However, my original idea was to integrate the supports in the design, and I don't want to change it. BUT, I am going to compromise with you all: I will keep the integrated shaft supports, go with shafts and SBR blocks (I have already bought them a month ago, plus I'm interested to see how it turns out). HOWEVER, I will add to the model holes for HWIN rails and blocks. IF shafts turn out to be a bad decision or if I'm just not happy with the result, I will cut off the supports, buy the HWIN rails / blocks and just mount them on like nothing happened. I think that will make you all happy ! This is not a waste of money since I've already bought the SBR blocks, and shafts will be and additional cost of $50. This way, I git it a shot, and if it works, it's a $600 minimum savings for me. Otherwise, it's a $0 loss as I will happily sell those blocks to someone here on the forum, and if no one wants them I will keep them for whichever exotic project I decide to work on :)

Now, considering this, can anyone tell me what size HWIN rails/blocks I would need for this machine, so I can get a quote from my favorite AliExpress seller Kevin Ho ?

2. Nuts will fall when assembling.

Tape will hold them from falling.

3. Welding shafts on the supports will make them bind.

Sure, I trust your experience. I will ask the welder shop to make a test with two fake supports and one short shaft. We will then measure the binding. However, I think that you're worrying too much on this. Binding will happen if the shaft gets too hot, which can be avoided by doing very short runs (few seconds long). If all else fails, I will get the shafts and parts drilled and tapped by a profesionnal turner and let him mount them. And if that doesn't work, I'll go with option 2 proposed in point 1 (cut off supports and go with HWIN).

4. Plasma cutting and laser cutting will cause parts to move / bind / shrink, in other words change and cause misalignment.

Sure. I contacted the cutter on Friday and they said they're going to use a water jet cutter. I asked about accuracy and they said they can go up to .01mm accuracy. I also asked about cut straightness, because I've heard that water jet cutting looses accuracy on the bottom of the cut. For example, if you want to cut a 10x10mm square out of 10mm steel, the machine will cut 10.01mm x 10.01mm on top but 10.02 x 10.02 on the bottom. The cutter I called said that it is true but that effect can be overcome if you sonfigure the machine correctly. When cutting the parts, they (and I'll be there to check !) will do some test cuts until they have the right setup.

Alright, so here are my takes on what you guys have said in the past few days. Modeling is going great, gantry is almost done. I've had to stop modeling for two days because I traveled a bit :) Renders of the model will come soon :)

I am very thankful to all of you that are commenting the design and giving advice ! I surely will read you, and try to find the best (function and cost wise) solutions to the issues you raise.

Cheers from a warm sunny Belgium !

JoeHarris
23-08-2015, 03:36 PM
Why not just tap the holes?!!

JoeHarris
23-08-2015, 03:38 PM
If money is a concern on this - how are you affording waterjet/ laser cutting of parts?! I would trade this for better rails and bearings...

eurikain
23-08-2015, 05:28 PM
Why not just tap the holes?!!
Which one ?


If money is a concern on this - how are you affording waterjet/ laser cutting of parts?! I would trade this for better rails and bearings...
I know, sounds like a crazy idea ^^. Laser cutting won't be so expensive as I have industrial prices because I am working with a contact of mine that works for a company that make thousands of orders at the cutting shop. But I do agree with you.. I honestly don't know if the effort I put into modeling this thing would be worth the money gained on welding, but I think so.

When you say "trade this for rails and bearings", are you suggesting to use weldments instead of nuts and bolts ?

EDIT: for your information, I just asked for a HIWIN 20 quote from my AliExpress seller. We'll see what comes out.

JoeHarris
23-08-2015, 07:54 PM
All of the holes where you have the flat-pack IKEA type fixings - why not take the stronger and more conventional engineering approach of tapping? Personally I would invest some time in learning to weld as the cost of hundreds of fasteners will be excessive?

eurikain
23-08-2015, 08:15 PM
Hi !

I actually already know how to weld, and have access to a free welders shop, but I haven't learned enough I think... But I would still need to get the parts laser cut, though I would not need such a tight tolerance, right ?

PS: rail length needs to be 2400mm. I doubt anyone will ship those lengths to me :s When working with HIWIN rails, can you combine shorter rails to make one ?

JoeHarris
23-08-2015, 08:32 PM
Hi !

PS: rail length needs to be 2400mm. I doubt anyone will ship those lengths to me :s When working with HIWIN rails, can you combine shorter rails to make one ?

I believe so. Good luck with the build!

eurikain
23-08-2015, 09:29 PM
Thanks ! BTW, bolts and nuts won't be such a cost worry ! I have them at bulk price from a company I know that makes railway supports. They have millions of bolts of all sizes in their stock :D

eurikain
24-08-2015, 04:32 PM
Alright, moving forward with the assembled and complete gantry (Y axis, Z axis and gantry sides). I still need to mirror the right gantry side but that's a detail compared to the amount of hours I put into making the sides and assembling everything. Solidworks can be a pain sometimes, and the 2GB RAM of my good old HP laptop are starting to struggle a bit.

Gantry weight is calculated to 39 kg without the second gantry side (@toomast: this is strange, but I've checked again that parts were set to the correct material, may be I can PM you the model and you can check if I have done something wrong ?). The Y axis is securely fastened to the gantry sides and contributes quite a lot to its rigidity.

http://www.mycncuk.com/attachment.php?attachmentid=15933&stc=1
Assembled gantry with sides and Z.

http://www.mycncuk.com/attachment.php?attachmentid=15932&stc=1
Gantry side with motor mounted on left side plate. Notice the extruded cut in the Y axis front plate to make room for the motor.

http://www.mycncuk.com/attachment.php?attachmentid=15931&stc=1
Gantry side. Y axis is fixed at 14 points. Weirdly, I had rendered this with fasteners showing, but they did not appear on the render.

General remarks:


I got a quote from my seller. He is asking for $886 for the following (shipping not included):
- 2 x 2400mm HIWIN 20 rail
- 2 x 1600mm HIWIN 20 rail
- 2 x 700mm HIWIN 20 rail
- 4 x HGH20CA blocks
- 8 x HGW20CA blocks

Shipping is not included because he does not know how to ship the 2400mm rails. He said the maximum length is 2000mm, but he will be looking for other ways to ship. Does anyone know if it is possible to assemble two HIWIN rails to make one (kind of like @routercnc suggested for round rails) ? I would be worried about the moment where the blocks slide on the junction...

Let's say we find a solution for shipping. You're looking a $1100 at least. Trying the round rails first is just a must now. Unless one of you guys is willing to sponsor me :yahoo:, I'm ready to sacrifice machine quality to the price of 1000 Euros with which I will gladly go on a vacation with my girlfriend :redface:.
I haven't gotten a fastener count yet but once the X axis is added I will post one. I think we're looking at 150 bolts and nuts. X axis will most likely add another 100. But I'm being very generous on these numbers so don't take me at my word.
Great surprise... DHL taxes for the leadscrews have arrived today :beaten:... That's and additional cost of $70.
Spindle has shipped yesterday night and is on its way, first class on FedEx airlines :nevreness:. Taxes for that package will be cheaper of course, I'm guessing around $40.


Next steps:


Finish the model (X axis, bed, limit switches, spindle clamp holes). The X axis will be modeled like the Y and Z axis with integrated shaft supports and fasteners. The bed will be made of 100x100mm square steel profile rails, I think (because it's just cheaper and faster to make, assemble and weld, as there are no complicated parts / holes / cuts).
Make DXF files and send them to different cutters for quotes (i'll take the cheapest). Cutting will be done with a .04mm maximum tolerance (anything better is a plus but not a must).
Order AM882 (200€).
Make PSU (100€).
Put it all together !
Configure, test and report precision :)

After modeling for 20 hours in the past two days, and writing this post for one more hour, I think I've deserved a break, so I'll go make myself a mojito now.
Cheers :beer: !

Boyan Silyavski
24-08-2015, 05:20 PM
i don't want to discourage you but as i have some experience doing similar assemblies from laser cut parts:

-going from square to round rails is a very bad idea at that point. Why waste effort and money then to make all that rigid structure. If you go with round rails then make a simpler square profile structure, no need to make that laser cut monster

- BST automation at aliexpress (http://es.aliexpress.com/store/314742), Fred is a very good seller, he could do all that stuff, machining screws, send 3m well packed ball screws, lower value, etc. tell him Boyan sends you, he gives me best prices compared to ebay and aliexpress, furthermore you could order all from him, including VFD ans spindle, he has more than he lists, just email him, no problem with errors or warranties

-the way you fix your motor and ball screw, that moves the Z left right, take care to be able to adjust that

-as i made some similar stuff i am telling you you will need at least 12 heavy duty clamps to be able to straighten all and solder. plus a big plastic hammer at least, i would say you will need a press. see what i did here for the Z (http://www.mycncuk.com/#155) #155 (http://www.mycncuk.com/threads/6619-Quite-an-Unusual-one?p=68417#post68417) and imagine at your scale of things how difficult that will be in reality if plate is not straight after cutting

eurikain
24-08-2015, 06:26 PM
i don't want to discourage you but as i have some experience doing similar assemblies from laser cut parts:
No worries, you are not discouraging me at all ! I appreciate very much the effort you put in giving me advice, and I am willing 100% to listen to you !


-going from square to round rails is a very bad idea at that point. Why waste effort and money then to make all that rigid structure.
Yes, I've decided to go round rails, no doubts. My next build with higher budget will get those HIWIN stuff.


If you go with round rails then make a simpler square profile structure, no need to make that laser cut monster
What do you mean ? I don't picture it very well :/


- BST automation at aliexpress (http://es.aliexpress.com/store/314742), Fred is a very good seller, he could do all that stuff, machining screws, send 3m well packed ball screws, lower value, etc. tell him Boyan sends you, he gives me best prices compared to ebay and aliexpress, furthermore you could order all from him, including VFD ans spindle, he has more than he lists, just email him, no problem with errors or warranties
Right, I've seen you guys all buy from him. I'll note that and make sure I order from him next time. I won't forget to tell him it's from you so I can get best prices. It's nice of you to share !


-the way you fix your motor and ball screw, that moves the Z left right, take care to be able to adjust that
Okay, I'll make slots for the motor holes. I have more room on the Y axis. Z is a bit tighter so not sure I can have big slots there.


-as i made some similar stuff i am telling you you will need at least 12 heavy duty clamps to be able to straighten all and solder. plus a big plastic hammer at least, i would say you will need a press. see what i did here for the Z (http://www.mycncuk.com/#155) #155 (http://www.mycncuk.com/threads/6619-Quite-an-Unusual-one?p=68417#post68417) and imagine at your scale of things how difficult that will be in reality if plate is not straight after cutting
Sounds good. But, what kind of clamps are you talking about ? Can you give me an example ? Maybe I can ask the cutter to surface the steel ? I'll give them a call tomorrow and ask what they can do to minimize the distortion effect due to heat. Water jet cutting would probably solve this issue. Post-heating / pressing the plates as well (?). But as you said, a good rubber hammer and clamps will do it with enough patience.

Boyan Silyavski
24-08-2015, 08:20 PM
Sounds good. But, what kind of clamps are you talking about ? Can you give me an example ? Maybe I can ask the cutter to surface the steel ? I'll give them a call tomorrow and ask what they can do to minimize the distortion effect due to heat. Watergate cutting would probably solve this issue. Post-heating / pressing the plates as well (?). But as you said, a good rubber hammer and clamps will do it with enough patience.

No need for water jet cutting. cause the plate itself is not straight in the planar dimension.

talking about gantry


What i am saying is that wyou will need big woodworking clamps to clamp together the front plate to the cut of the up and down plates. As the cut will be straight you have to push well the fron and back plate so they are flat against that edges. Which could be a challenge even with clamps . Do you get what i mean? You may be will need a press and then clamp. But when you have them at hand you will see.

Cause if you dont clamp them well against that edge prior to soldering, you will need to epoxy level below the rails.

eurikain
24-08-2015, 11:54 PM
Alright I understand what you are saying now :) !

But a more general understanding of what you are saying is (correct me if I'm wrong):


If you go with round rails then make a simpler square profile structure, no need to make that laser cut monster
+

Cause if you dont clamp them well against that edge prior to soldering, you will need to epoxy level below the rails.
= Forget the IKEA-like nuts and bolts to fasten the plates. Just weld it together. May be make bigger joints so I have larger spots to weld.

?

eurikain
25-08-2015, 11:05 AM
@toomast: This is sort of embarrassing, but after applying steel 304 to all my parts (except standard parts like leadscrews and stuff) solid works calculates 280kg (where 57 kg for Z axis)......... oO ?? Is that even possible ?

eurikain
25-08-2015, 02:55 PM
Important updates:

Following Boyan's advice, I am switching to a new Z / Y / gantry design, adapting my old one, removing nuts and bolts. I am also switching to ALUMINIUM instead of steel, as this thing is way too heavy. The new design lightens it but going to aluminium will be the solution. I'll be using 20mm sheets, unless you guys think 15mm is better to reduce weight. I'll be adding pockets at some places to lighten the whole thing.

I called the cutting company. They said parts will be water jet cut (I have no choice on that, but no big deal, I've got good prices anyway) with .25mm accuracy.

I keep the round rods, no changes on that. Instead of being supported by profiles, they will be supported by M8 bolts every 70mm plus at both ends.

Here are the pictures of the Z axis, evaluated to 17kg, but that should decrease with pocketing.
http://www.mycncuk.com/attachment.php?attachmentid=15947&stc=1
Complete Z.

http://www.mycncuk.com/attachment.php?attachmentid=15946&stc=1
Z without front plate.

Cheers.

Boyan Silyavski
25-08-2015, 03:20 PM
Important updates:

Following Boyan's advice, I am switching to a new Z / Y / gantry design....


What are you doing right now is Not following my advice. Of course its your machine, i am just trying to help. Seems you a missing my biggest point. What i told you before:

Keep the design with square rails! Save from all else. Dont compromise on that. Dont exchange that for a vacation:joker:



Ok i understand - expensive, you want to change the design to round rails. Then read some similar build logs and design much lighter and simpler frame.


Here is my advice:

-make the Z strong. 20mm aluminum plate is Not strong enough if not braced. Ideal will be 12-15mm steel plate, braced and done like #8 here (http://www.mycncuk.com/threads/8778-Driving-the-gantry-on-Z-axis?p=70040#post70040)where i explain much better and simpler Z design with the same effort

-make gantry from profiles cut on a Rage saw and soldered. gantry sides also. Look my second project in signature for the gantry sides / 100x100 profile/ Design all machine from one profile, say 80x80x4mm or 100x100x3mm , no fancy plates and crap where you waste money. Order 6m steel box profiles, cut all on Rage saw and solder.

From here you will save the money for the square rails. Difference between square rails and other type of linear movement is like from Ferrari compared to Trabant


-My direct advice is follow my first build and change gantry sides like i did them on my second. Thats the cheapest and strongest machine at the same time possible.

Wal
25-08-2015, 08:32 PM
I keep the round rods, no changes on that. Instead of being supported by profiles, they will be supported by M8 bolts every 70mm plus at both ends.

Stop. Think about that. Then think about it again and again until this crazy idea is banished from your head..!

(Out of morbid curiosity, how were you ever going to make that work..?)

Wal.

toomast
27-08-2015, 07:44 AM
@toomast: This is sort of embarrassing, but after applying steel 304 to all my parts (except standard parts like leadscrews and stuff) solid works calculates 280kg (where 57 kg for Z axis)......... oO ?? Is that even possible ?

Yeah, thats more like it. Steel is heavy stuff.
I see you have a lot of feedback with good advice. It is not easy to decide the correct path ;)

eurikain
27-08-2015, 10:14 AM
Sounds good, I'll think about it and rethink everything :)

@Wal: It works in SW :) I just didn't think steel was going to be that heavy x) Obviously this is due to the size of the machine, not much about the thickness of the steel. 1300mm Y travel implies at least 1700mm gantry width. Using 100x100x3 profiles will definitely be the key to lightness :)

Clive S
27-08-2015, 11:09 AM
Using 100x100x3 profiles will definitely be the key to lightnessI would go for a thicker profile say 5mm so that you can tap it ok. I used 60x60x5 profile

Wal
27-08-2015, 11:36 AM
@Wal: It works in SW :) I just didn't think steel was going to be that heavy

Euri, I'm talking about supporting the rails with 8mm bolts... That just sounds like a really bad idea to me. How are you proposing to make this work without introducing kinks into the rails as you torque them up? Fair enough, if you could guarantee that each and every tapped hole in the plate was IDENTICAL and furthermore the tapped holes in the rails were IDENTICAL then (despite it still being a really bad idea) it probably could work, but why? Why why why..? This is about the silliest way of doing it..! Okay, mounting the rails on blue-tack would be sillier, but not by much. Just think about it for a moment - I'm not trying to disparage, but to me this method makes no sense at all..!!

Wal.

Lee Roberts
27-08-2015, 03:23 PM
Euri, I'm talking about supporting the rails with 8mm bolts... That just sounds like a really bad idea to me. How are you proposing to make this work without introducing kinks into the rails as you torque them up? Fair enough, if you could guarantee that each and every tapped hole in the plate was IDENTICAL and furthermore the tapped holes in the rails were IDENTICAL then (despite it still being a really bad idea) it probably could work, but why? Why why why..? This is about the silliest way of doing it..! Okay, mounting the rails on blue-tack would be sillier, but not by much. Just think about it for a moment - I'm not trying to disparage, but to me this method makes no sense at all..!!

Wal.
I have to agree with Wal on this, it really looks like your making life hard for yourself for no other reason than to compromise the integrity of the machine.

It's nice to design somthing totally bespoke, never seen before and so on but I too really would encourage you to look at this from a different perspective, setting aside what the machine will look like cosmetically, I say cosmetically because we really can't see there is advantage to your rail fixture method and it's just going to serve you a handycap.

.Me

Wal
27-08-2015, 04:00 PM
Yep, It's not just achieving the consistency in your method that'll give you a problem - the height of the rail will also depend on the torque of the bolts - as such it's going to make measuring for the ball-nut bracket fitment a real pain... You need to be designing around what you know - data that is readily available, not "Oh, I'll just put these rails in using the maddest method I can come up with and then see about my ballnut bracket later.."

Heh, the rails will probably resemble a roller-coaster by the time everything's torqued up anyway. Using this method the ballnut might just move smoothly along a couple of inches of track...

Honestly, don't pursue this inviable method..! It'll be a waste of your time and money. Enough said.

Wal.

D.C.
27-08-2015, 05:45 PM
Eurikain,

Listen to the sound of bearings moving in this video:

http://www.youtube.com/watch?v=bGbWBDtzydw

It sounds bad.
Very bad.
Very, very bad.

I'm guessing this video or something like it is where you got the idea for DIY supported rails.
It is a bad idea.
Very bad.
Very, very bad.

The reason why everyone is giving you negative feedback is because it is a bad idea, so ditch it and move on. If it helps any, hiwin rails from hobby machines have a good second hand market value. When you outgrow your machine you can recover a lot of the price by selling them on.

PS I'm the resident village idiot, annoying people with my hair-brained ideas, even I figured out that DIY supported rails was a non-starter.

eurikain
29-08-2015, 07:44 PM
Hi ! Sorry for the late reply, I've been a bit busy :)

I hear you all, and I will take your advice. I know, DIY supported rails are not the way to go. I hear you. Even if this is not going to work, I think that the profile method was going to work, just not with steel, or made differently with short pieces. I think DIY rail is possible, but you need to get the supports milled from ALU or steel and then tap them. What I am just explaining here is that ROUND fully supported rails don't seem efficient to me cost / strength wise. I've run a couple simulations in SW and that confirmed my idea. You could semi-support a round rail with at least 1/2 of the material used in fully supported chinese rails. Of course, getting those parts machined may be expensive, unless you've got a CNC within hand's reach, or know someone that can get them milled.

That said, supporting the rails with bolts is a silly idea, I agree.

eurikain
06-09-2015, 09:03 PM
@Boyan
I requested a quote from my seller Kevin and from Fred at BST automation for SBR rails.
Kevin asks for $305, Fred for $345 ;)

Lee Roberts
06-09-2015, 10:03 PM
Now ask fred if he can get any lower or at the least price match ?

.Me

eurikain
06-09-2015, 10:40 PM
That's an idea :)

JoeHarris
07-09-2015, 07:11 PM
@Boyan
I requested a quote from my seller Kevin and from Fred at BST automation for SBR rails.
Kevin asks for $305, Fred for $345 ;)

I don't know Kevin but Fred is a reliable seller.

Boyan Silyavski
08-09-2015, 08:16 AM
I don't know Kevin but Fred is a reliable seller.

Exactly. i wouldn't trust sb else to pack me a 3m ball screw from China.

JoeHarris
09-09-2015, 08:14 PM
You want to know that if something isn't right that they will sort it out. Having bought from China a few times now - Fred has been the most hassle free. One faulty item out of 18 and he sent a replacement a few days later with no quibbles.

eurikain
09-09-2015, 11:38 PM
The 2m ballscrew was packed very well. It was supported at the bottom with a plywood plank. I know fred puts them in a box. Next time I order I will definitely go with him. Plus, his English is excellent compared to what I've seen on AliExpress so far :D

Mark Needham
19-11-2015, 02:23 AM
Good Morning. How is the build coming on.
Just to let you know, I have used the Chinese SBR rail and guide. 1000mm long.
Been in use some 18 months now. Faultless operation, I have had no trouble at all.http://g02.a.alicdn.com/kf/HTB1nvRMIXXXXXbtXFXXq6xXFXXX3/SBR20-linear-guide-set-1pc-of-SBR20-rail-L500mm-20mm-linear-guide-cnc-router-part-2pcs.jpg_120x120.jpg
The rail is supported, so why in your drawings is it not?
I have attached a quick photo from an Ali site, just to confirm the style of rail that you are using.
I just happened on this site, and I am curious, as to how you got on.
Cheers,
Mark Needham

Lee Roberts
28-06-2017, 12:51 PM
What happened next?