My 1st Hybrid Steel/Ally (1150 x 850 work area) (comments/advice please)

[lebies]

Hey all,

Thanks for the great input so far, I'm sure there'll be more as we go ... :).

mitchejc
Ah, good to see another South African on the forum...

Thanks man,
been lurking for a while now, and now I'm sorry I didn't get my act together and post a lot earlier.
Thanks for the links, already learned from them...

I'm really glad to hear the steppers are good enough, was a bit concerned there. I've used their smaller steppers on a project and also found them to be very good.
Yep, Greg's stuff is very good, and he's a very knowledgeable and approachable guy as well. I'll definitely look into Wendy's. I sourced the local (jo'burg) Hiwin agent and got what I thought to be very reasonable prices from him. My main concern with China was the shipping time, but your deal is great as far as that's concerned. Also, I've heard that a lot of the Hiwin stuff coming from China is not the same quality that comes from the Taiwan factory. Apparently they pick up the stuff (from Hiwin) that doesn't quite make Hiwin's standard, and sell that. i.e. it's not reject, just not what you'd get from the original manufacturer. I'd be interested in hearing how yours go for you.

As for the 4th Axis, you guys are all right, I'm probably overstepping the bounds. I'm going to reduce the 4th axis radius to around 75mm (1/2 it). Then again I hear what you're saying, and will look into the whole software thing in a lot more depth ... even if I have to look at writing something to accommodate (on the longer term).

EddyCurrent
There are other things but for a start these jump out at me...

Thanks for this.
1. I'll take your good advice and source the suggested spindle and collet.
2. I am going to reduce the working radius on my 4th axis (hence the longer travel) and go to around half of it. I'll also increase the distance between the blocks as suggested. What should I look at for a minimum please?
3. Hopefully jazzcnc does comment, I'd be very interested in hearing his suggestions.
4. I started out on 20mm rail originally, only to be told by the manufacturers that 15 would more than suffice in terms of loading, etc. As I downsize the Z axis, I'm going to also downsize on the other axis as well. Not a whole lot, but a fair amount.
5. With the reduced Z axis and added other suggestions, I don't think I want to go adjustable table (for now anyways).

routercnc
That is a large machine to also cut aluminium...

Agreed. The main materials I'll be working with will be hardwood and perspex. For the occasions I'll want to work ally, it probably doesn't justify it anyway.

You have a very large offset from the Z axis to the Y rails - ...

There's another way I can do the Y-axis, and that's to use 2 of those profiles atop of each other, i.e loose the center piece - but that's if I insist on going that route. If I do, although simpler, it's a hellova lot more expensive and again, probably not worth the expense. I'll look at going back to an earlier design which was an all steel machine (and stop being so damned lazy :) ). In that version, I got the the ratio down much better.

The end plates on the gantry look very thin.

My bad! I should have said the end plates are 5mm steel plate, as are the gantry cross plate (bottom plate) and 4th axis mount plate.

Will you add a bed to this?

Yep, 32mm MDF, and I'll add 2 additional cross members to stiffen it up as well as suggested.

X axis motors

Yes, I got THAT one wrong, didn't I :). Not quite sure what I was thinking of on this one, probably one of my earlier version designs!

Clive S
If you are going to slave your x axis you might have to re think Linuxcnc ...

Thanks for pointing this one out to me. I'll rehash it and also better research it.

Thanks all so far for the great feedback. With you all pointing me in the right directions I'm really looking forward to getting this thing done.

[EddyCurrent]

I'll also increase the distance between the blocks as suggested. [I]What should I look at for a minimum please?

Well mine have a gap of 95mm, that is inner face to inner face, I calculated that as the best for my machine.

With regard to the bed, I urge you to use birch plywood rather than MDF. I used two layers of 19mm and sealed it all with a 50/50 mix of polyurethane varnish and danish oil. That's after leveling the bed by running some G code to run the router over it with a large end mill cutter.

With regard to aluminium, I thought I might do some cutting of this myself but when you consider the requirements for lubrication/coolant, a wood cutting router just does not have the appropriate bed arrangement.
You might want to think about that in your design before starting the build i.e. an aluminium T slot bed might be a better alternative.

[Boyan Silyavski]

With regard to aluminium, I thought I might do some cutting of this myself but when you consider the requirements for lubrication/coolant, a wood cutting router just does not have the appropriate bed arrangement.
You might want to think about that in your design before starting the build i.e. an aluminium T slot bed might be a better alternative.

One of the best ways to cut aluminum is to use old fridge compressor that blows cheap airbrush that spays mist of alcohol at the cutter. The mist evaporates instantly and leaves everything dry and cool. So coolant is not a problem.

[lebies]

One of the best ways to cut aluminum is to use old fridge compressor that blows cheap airbrush that spays mist of alcohol at the cutter. The mist evaporates instantly and leaves everything dry and cool. So coolant is not a problem.

Cool (no pun intended) idea :) ...

After all your initial input (for which I'm very grateful thanks), here's take 2:

- Z-Axis travel now 100mm (all dims +-)
- Y & Z axis now welded plate (mainly 5mm with 8mm spindle mount plate.
- Used 100x50x4 to mount X and Y-axis rails onto
- Widened guide rail blocks all-round

A pic paints a 1000 words ... so here ya go (2ns is just the Y-bed detail construction):



As always, your feedback valued :)...

[lebies]

Well mine have a gap of 95mm, that is inner face to inner face, I calculated that as the best for my machine.

With regard to the bed, I urge you to use birch plywood rather than MDF. I used two layers of 19mm and sealed it all with a 50/50 mix of polyurethane varnish and danish oil. That's after leveling the bed by running some G code to run the router over it with a large end mill cutter.

Done ... thanks.

With regard to aluminium, I thought I might do some cutting of this myself but when you consider the requirements for lubrication/coolant, a wood cutting router just does not have the appropriate bed arrangement.
You might want to think about that in your design before starting the build i.e. an aluminium T slot bed might be a better alternative.

Have rehashed my thinking. I'll be able to do it on my updated version but it's no longer a high priority.

I wanted to use the al T-slot bed, but I'll buy it piece-meal for future. This stuff is extremely expensive around here - very, very nice, but damned costly for now :).

Thanks for your help, appreciate.

[Boyan Silyavski]

Hi,
just read all you said and need carefully.

Some thoughts on your design that come to my mind, just what i would do., Having in mind the results you desire. I will try to resume so you dont have to look million threads for pieces of information:

- Beef up everything to 80x80x3 and the gantry to 2x 80x80x4 soldered together, or keeping your aluminum gantry design take a look at other threads to see how people do it properly with say 20mm thick plates
Beefing the structure is the cheapest thing to do on a machine. So why not do it. I have done extensive calculations and assure you that this table design is the stiffest and cheapest and lightest possible and at the same time easier to make, simple. Also doing so you avoid the necessity of diagonals. I am talking about 80x80 as ideal, to 100x100 as overbuild. I do it 100x100. You can find more info here , Tonys CNC is very good example of what you can achieve with this type of structure, of course accustomed to your needs + alu gantry

-Make the Z travel 200mm, its good compromise between stiffness and length. Take a look at how i did it here post #109 and how looks finished from 12mm steel plate a bit up at the same page. Your design will not work for that length/ or any/. The alu profile is too flimsy. Make the plate exactly as i did from 12mm steel, or 20mm aluminum at least . The plate should have ribs that dont allow it to bend in the long axis directions. In my design the ribs are the formed on the back by the steel that supports the rails and the rails themselves. Either way in front or back it must be ribbed.
The Z in question reaps fully extended through aluminum like a butter with 1/4 router bit and surfaces alu glass like finish.

-normally for multi material machine people here on forum use 1610 on all axis and 1605 on Z

-15 size linear rails is in the ball park of amateur machine. most machines that want pro results start fro 20 size linear rail, that is assumed the typical in a well made diy build. The price is almost the same.

-Hiwin from China is original. And probably the most used rail that i know of. Dont be mistaken here. Contact Fred / BST automation/ from my name and he will beat any price of linear stuff you see on ebay or aliexpress+ offers perfect machining service. You can source almost all for a build from him, including spindle and VFD. Stuff that is not listed at his store i mean.
China ball screws are acceptable, cheap and at the end do the job perfectly. Usually the quality of a build suffers from many other stuff than to concentrate on the quality of ball screws.

-on the table bed the beams should be from 200 to max 300mm separated. At 300mm you must fit sth at least 6cm thick not to vibrate. So as the steel is the cheapest thing in a build i don't see why not do it properly. Proper alu bed is very expensive. So do your self a favor and make it right.

- Now that you want 4rth axis to be 300mm OD is not a problem with 200mm axis and 2 beams separated on bed say 350mm where you can fit the axis and when not in use cover it with sheet.

-If you were inspired by Deans/Jazzcnc/ or Johnathans build to design the structure, have in mind that these machines use toolplate as table bed that makes thing very stiff.
So its not just paste and copy thing.

-Speaking about bearing spacing of bearing blocks i see as absolute minimum 260mm end to end outside measured, having in mind The long, not short Hiwin 20 carriages which each is 90mm long. That on all axis and in all directions! . Compromise can be made on the back Z plate width as Eddy said above and on the front Z plate width. i dont see though how the Z plate could be wider less than 150mm without compromising something

-that drivers you point to will Not work. you need 550-650W 70VDC non regulated power supply, and drive the motors at that Voltage 70VDC not 50VDC, so you need drivers like Leadshine AM882 which typically is the most recommended drive here. 3Nm Low Induction motors is what you need and drive them at 4.2A with the drivers at 70 VDC.
With the before mentioned you can drive all with short HTD 5M 15mm short belts and 20T pulleys 1:1, To calculate the center distances and other measures http://www.bbman.com/catalog/belt-le...alculator.html

-There are many BOB and when you search to buy happens that all of them have some trade off. Money must be spend here if yo want smooth motion. Ethernet is the way. ESS+126 PMDX board. You can buy first the board only if money is an issue. I use other board due to other reasons but it also works with ESS.

Hope that summary helps you save time fiddling and shortens the time till when chips will be flying around.

Last note is to analyse and not simply copy, cause simple stuff like simple looking structure has a lot of thought gone into it and changing little things can compromise many things, as usually the designs are quite perfected by themselves.

[lebies]

Hi,
just read all you said and need carefully.

Some thoughts on your design that come to my mind, just what i would do., Having in mind the results you desire. I will try to resume so you dont have to look million threads for pieces of information:

Thanks for this detailed answer, very much appreciated.

I'm not a mechanical engineer, so the best way for me to learn is by burning my fingers. To do so I guess I must expose my weaknesses but have the balls to put it out there and trust folks with experience will come to the rescue - as has been the case here (thanks all). Collectively that makes of us a better, stronger community. I look forward to when I can start contributing and in so doing free up the folks with their vast knowledge to spend more (quality) time innovating and have drones like me help newbies get going .

On to the business at hand.

Not sure you saw my revisions (we posted at around the same time) but I've made some significant changes to the original. I went back to one of my 1st designs and have decided to go all steel.

- Beef up everything to 80x80x3 and the gantry to 2x 80x80x4 soldered together, or keeping your aluminum gantry design take a look at other threads to see how people do it properly with say 20mm thick plates
Beefing the structure is the cheapest thing to do on a machine. So why not do it. I have done extensive calculations and assure you that this table design is the stiffest and cheapest and lightest possible and at the same time easier to make, simple. Also doing so you avoid the necessity of diagonals. I am talking about 80x80 as ideal, to 100x100 as overbuild. I do it 100x100. You can find more info here , Tonys CNC is very good example of what you can achieve with this type of structure, of course accustomed to your needs + alu gantry

Nothing to loose and much to gain, will implement.

-Make the Z travel 200mm, its good compromise between stiffness and length. Take a look at how i did it here post #109 and how looks finished from 12mm steel plate a bit up at the same page. Your design will not work for that length/ or any/. The alu profile is too flimsy. Make the plate exactly as i did from 12mm steel, or 20mm aluminum at least . The plate should have ribs that dont allow it to bend in the long axis directions. In my design the ribs are the formed on the back by the steel that supports the rails and the rails themselves. Either way in front or back it must be ribbed.
The Z in question reaps fully extended through aluminum like a butter with 1/4 router bit and surfaces alu glass like finish.

Done. I cut it down to even less. In fact 150mm diameter is probably more than enough for my needs.

-normally for multi material machine people here on forum use 1610 on all axis and 1605 on Z

Thanks, will do.

-15 size linear rails is in the ball park of amateur machine. most machines that want pro results start fro 20 size linear rail, that is assumed the typical in a well made diy build. The price is almost the same.

I'm going Hiwin 20 on everything now.

-Hiwin from China is original. And probably the most used rail that i know of. Dont be mistaken here. Contact Fred / BST automation/ from my name and he will beat any price of linear stuff you see on ebay or aliexpress+ offers perfect machining service. You can source almost all for a build from him, including spindle and VFD. Stuff that is not listed at his store i mean.
China ball screws are acceptable, cheap and at the end do the job perfectly. Usually the quality of a build suffers from many other stuff than to concentrate on the quality of ball screws.

Fred and I are speaking and sounds like we'll come up with a solution. Ridiculous thing is, for a set of ballscrews at $90, DHL want to charge $92 for shipping!
If we're going to find a good shipping solution between us, my costs drop to the extent I can use Servos and an alu bed.

-on the table bed the beams should be from 200 to max 300mm separated. At 300mm you must fit sth at least 6cm thick not to vibrate. So as the steel is the cheapest thing in a build i don't see why not do it properly. Proper alu bed is very expensive. So do your self a favor and make it right.

I'm going to buy the alu bed piece-meal over the next few months, so I'll work to it in good time and give it high priority.

- Now that you want 4rth axis to be 300mm OD is not a problem with 200mm axis and 2 beams separated on bed say 350mm where you can fit the axis and when not in use cover it with sheet.

Will now go less radius as stated, but I can't quite picture what you actually mean here :).

-If you were inspired by Deans/Jazzcnc/ or Johnathans build to design the structure, have in mind that these machines use toolplate as table bed that makes thing very stiff.
So its not just paste and copy thing.

I was definitely inspired by their builds, hence the strategy on the alu bed in good time. Also, I've now got the message - go overboard on the base, it's worth it!

-Speaking about bearing spacing of bearing blocks i see as absolute minimum 260mm end to end outside measured, having in mind The long, not short Hiwin 20 carriages which each is 90mm long. That on all axis and in all directions! . Compromise can be made on the back Z plate width as Eddy said above and on the front Z plate width. i dont see though how the Z plate could be wider less than 150mm without compromising something

Thanks to Eddy (and locating his stiffness calculator) I was able to learn a whole hellova lot, and improve on my original design. Mostly though, many of the critical principles were highlighted for me to learn from and/or about.

-that drivers you point to will Not work. you need 550-650W 70VDC non regulated power supply, and drive the motors at that Voltage 70VDC not 50VDC, so you need drivers like Leadshine AM882 which typically is the most recommended drive here. 3Nm Low Induction motors is what you need and drive them at 4.2A with the drivers at 70 VDC.
With the before mentioned you can drive all with short HTD 5M 15mm short belts and 20T pulleys 1:1, To calculate the center distances and other measures http://www.bbman.com/catalog/belt-le...alculator.html

-There are many BOB and when you search to buy happens that all of them have some trade off. Money must be spend here if yo want smooth motion. Ethernet is the way. ESS+126 PMDX board. You can buy first the board only if money is an issue. I use other board due to other reasons but it also works with ESS.

Great! I'm going to use this combo myself. Just busy sourcing motors to match.

Hope that summary helps you save time fiddling and shortens the time till when chips will be flying around.

Last note is to analyse and not simply copy, cause simple stuff like simple looking structure has a lot of thought gone into it and changing little things can compromise many things, as usually the designs are quite perfected by themselves.

As stated above, once these things have been pointed out, it's easy enough to pick up on. Without Eddy's sheet (and/or knowledge thereof) I'd have been fumbling around in the dark for quite some time I can tell you. As a suggestion (very humbly), a place to put all such info in it's own right on the main board?

Thankfully there's a community of very helpful folks to nudge one in the right direction, and once you've found the proper info, you can actually do precisely what you recommend .

Thanks again for the detail, really appreciate and it'll definitely shorten my build time.

[Boyan Silyavski]

Thanks, will do.

Fred and I are speaking and sounds like we'll come up with a solution. Ridiculous thing is, for a set of ballscrews at $90, DHL want to charge $92 for shipping!
If we're going to find a good shipping solution between us, my costs drop to the extent I can use Servos and an alu bed.

Thats why the suggestion to buy everything from him, cause shipping costs are high, so they could be spread among all pieces. i have done the calc and seems cheaper to buy all, pay shipping, duty, than buy from Europe, where i live. 1/3rd cheaper. you will have to do your calc though. Dont forget to tell him to state lower value.

Will now go less radius as stated, but I can't quite picture what you actually mean here :).

i mean this, the green is the 300mm diameter workpiece

[lebies]

Thanks for this detailed answer, very much appreciated.

I'm not a mechanical engineer, so the best way for me to learn is by burning my fingers. To do so I guess I must expose my weaknesses but have the balls to put it out there and trust folks with experience will come to the rescue - as has been the case here (thanks all). Collectively that makes of us a better, stronger community. I look forward to when I can start contributing and in so doing free up the folks with their vast knowledge to spend more (quality) time innovating and have drones like me help newbies get going .

On to the business at hand.

Not sure you saw my revisions (we posted at around the same time) but I've made some significant changes to the original. I went back to one of my 1st designs and have decided to go all steel.



Nothing to loose and much to gain, will implement.




Done. I cut it down to even less. In fact 150mm diameter is probably more than enough for my needs.



Thanks, will do.



I'm going Hiwin 20 on everything now.



Fred and I are speaking and sounds like we'll come up with a solution. Ridiculous thing is, for a set of ballscrews at $90, DHL want to charge $92 for shipping!
If we're going to find a good shipping solution between us, my costs drop to the extent I can use Servos and an alu bed.



I'm going to buy the alu bed piece-meal over the next few months, so I'll work to it in good time and give it high priority.




Will now go less radius as stated, but I can't quite picture what you actually mean here :).




I was definitely inspired by their builds, hence the strategy on the alu bed in good time. Also, I've now got the message - go overboard on the base, it's worth it!



Thanks to Eddy (and locating his stiffness calculator) I was able to learn a whole hellova lot, and improve on my original design. Mostly though, many of the critical principles were highlighted for me to learn from and/or about.



Great! I'm going to use this combo myself. Just busy sourcing motors to match.



As stated above, once these things have been pointed out, it's easy enough to pick up on. Without Eddy's sheet (and/or knowledge thereof) I'd have been fumbling around in the dark for quite some time I can tell you. As a suggestion (very humbly), a place to put all such info in it's own right on the main board?

Thankfully there's a community of very helpful folks to nudge one in the right direction, and once you've found the proper info, you can actually do precisely what you recommend .

Thanks again for the detail, really appreciate and it'll definitely shorten my build time.

Thats why the suggestion to buy everything from him, cause shipping costs are high, so they could be spread among all pieces. i have done the calc and seems cheaper to buy all, pay shipping, duty, than buy from Europe, where i live. 1/3rd cheaper. you will have to do your calc though. Dont forget to tell him to state lower value.




i mean this, the green is the 300mm diameter workpiece

Gotcha, thanks.

Been buying all my electronics from China for a while now, so have learned a few tricks :).

I'm thinking of pairing the AM882 with 86HS35 steppers. Good match?

Not sure I should belt-drive both X-axis ballscrews from one motor, or drive each individually.
If I go belt-driven both from 1 motor, I'll go look out around here for some examples.

If I go individual drives, what size motor should I be thinking of?