DIY CNC Design. criticism wanted.

[joel0407]

Ok guys, I've wanted to do this for a long time. I think I've been a member here since 2012 or something. My interest comes and goes. (It usually goes due to funds).

I've finally bitten the bullet and started a build. I've purchased and received these ball RM1605 screws in 350mm, 800mm and 2 x 1000mm also SBR16 rails in 300mm, 800mm and 1000mm pairs.
Ballscrew RM1605-1000mm Ballscrew &BF12/BK12 End support &6.35*10mm Coupling 007 | eBay
2 SBR16--300mm Liner rail RM1605-350mm Ballscrew &BF12/BK12 &6.35*10mm Coupling | eBay
SBR16--800/1000mm Liner rail RM1605-800/1000 Ballscrew &BF12/BK12 &Coupling

I've tested using M6 x 50mm cap head socket screws (machine thread) in 18mm marine ply and I'm more than happy with the strength of the joint so that's my intended construction.

I have spent an unknown number of hours on my design over about the last 4 weeks and I think I have it at the stage I can show you guys and have you thrown stones at it so to speak. Almost everything about the design is like it is for a reason. Things like obviously it's best if the linear bearings are spaced as far apart as possible to prevent racking but then we'd have no travel so we come to a happy medium where the bearings are spaced as much as possible to prevent racking but still allow a reasonable travel. I have gone with 300mm spacing on the 1000mm X axis rails, 200mm on the 800mm Y axis rails and 200mm on the 300mm Z axis rails.

The X Gantry ends slope back so the router bit end up positioned dead centre of the X gantry bearings.
I have 2 ball screws on the X axis to hopefully completely remove racking from this axis as it is also the widest.
The bed is lower in the centre so the ball screws for the X axis are at the sides of the bed about the average height that I intend to work at to further minimize racking.
I know 5mm lead ball screws aren't ideal but neither is the cost of 10mm lead ball screws. It's a trade off for the cheaper price.
I really like some of the T-Slot designs but I think T-Slot is over priced for what it is. My bed is supported on each end by aluminum angle bolted to 80x40mm aluminum RHS.
I think it's a little bit different but for the Z axis, the rails are attached to the Z Gantry and they slide in the bearings that are fixed to the Y Gantry. So the rails go up and down rather than the bearings. The reason being is I was trying to keep the bearings spaced at 200mm. This meant that the Z and Y axis bearing ended up back to back and I couldn't bolt them on.
The unusual arrangement of rails on the Y Axis with one facing up and one forward is a result of me being undecided to which way to face them. I know the linear rails will hold more force directly up and down rather than across the bearing. The problem I have getting my head around is that each axis although only moving in a single axis itself, it bears the load of the other 2 axis. For example if the Y axis moves left to right the force is between the 2 rails at the ball nut however the drag is at the face of the bit in the router. This would try and rotate the Y gantry. This would make it best to mount both rails either up or down. But then if the X axis was to move forward or back then the drag being at the router bit would try and pull the Y Gantry out at the bottom or in at the bottom. This would make it best that the rails be mounted forward or back. Ultimately I tried a few designs with the rails at 45 degrees with one either side of the router body. I think this was the best design but then I ran out of space for the Z axis.

I have made contact with another member on here local to myself here in Perth, Australia. He has said he can cut the front, back, X Gantry centre and the X Ball screw support. Those are all pieces that will be too big for my machine to cut. All the other small pieces, I will hand cut as best as I can from MDF then once I have the machine up and running. I will use it to re-cut the parts from 18mm Marine ply. Disassemble and reassemble.

I have purchased 5 axis break out board and I will have a dedicated computer most likely with a PCI parallel port card. I know this is not the ultimate these days however as I read from someone else, their are plenty still around that work with these breakout boards. At the cost it's not that much of a loss to up grade it later.

I am considering AM882 drivers and NEMA23 23HS33 motors. These motors have an uncanny resemblance to the leadshine 57HS22 motors. The specs are almost identical and I suspect the 23HS33 maybe a Chinese copy. I will purchase these on Thursday.
Nema 23 CNC Stepper Motor 2.83Nm(400 oz.in) 4.0A 23HS33-4008S - US$

Power supply wise, I haven't decided. I keep going up in voltage to gain speed for rapids but I think the calculations for these motors limits me to 48v. I like the look of the speed at 60v though. Maybe I should have coughed more cash for 10mm lead screws LOL.

I must admit that part of my initial thoughts were that it wouldn't matter having slow screws as it's just a hobby machine and I could wait. Now I have read a little and know a little about feeds and speed, I was starting to worry about being too slow. I think I'm in the ball park with the motors I've chosen though and later down the track I do have further design plans. I'd like to flip the motors around and contain them in the base and gantries rather than them hanging out the ends as the current design. I'll do this by running a toothed belt between the motors and the screws. This will then give me the flexibility to tune the setup a little with pulley ratios.

These are the drawings that I've drawn from scratch in Fusion 360. I hadn't used Fusion 360 previously but I have been CAD drawing for over 16 years. Mostly only as a hobby but I do have a university certificate in CAD and did it as my employment for about 12 months when I started.

I'm looking for you guys with more experience to have a look and tell me problems I might have. Tell me what's bad about it. I won't be offended. Throw stones so to speak.

Happy Days




Similar Threads:

• Alibre design projects - help wanted
• Newbie- first design, after constructive criticism
• New Machine Build- Just want some constructive criticism
• Design Work wanted
• Design Engineer wanted

[Biggs427]

I don't have much experience so I won't throw any rocks, but here's my take on your design.

1 - Gantry uprights are too thin imo.

2 - The gantry bridge might be a bit thicker and if you want less flex, I'd do a torsion box instead of the "webbing"

3 - Your base should also be a torsion box on which you attach your rails. Your whole machine sits on it, it needs to be stiff.

4 - If the stepper specs are right (1.8mH), you "should" be able to spin the stepper to 1000rpm which would give you around 200IPM. I spin mine at that speed with an AM882 and 40v PS. My machine rapids at 400IPM with 10mm pitch ballscrews.

Nice job for a first pitch. Took me 2 month designing my second machine.

[joel0407]

Thanks for your reply mate. Experienced or not, it's all good. Opens discussions anyway.

1 - Gantry uprights are too thin imo.

They are 18mm Marine Ply. I can stand in the centre of 1m x 300mm with minimal deflection so I think these should be fine as they are only 354mm tall.

2 - The gantry bridge might be a bit thicker and if you want less flex, I'd do a torsion box instead of the "webbing"

Again, Marine ply is pretty strong. Much stronger than MDF. Yes some parts will be cut from MDF for the initial build but then they will be changed out once I use the machine to cut the parts from Marine Ply. It's not fully boxed in but assembled channel. There is a 50mm support across the top and 40x40 aluminum angle across the bottom. It was meant to be the same 50mm or 18mm marine ply across the bottom but having the rail on the front means I cant bolt through. With the angle, I'll bolt through and sandwich the Gantry Bridge.

3 - Your base should also be a torsion box on which you attach your rails. Your whole machine sits on it, it needs to be stiff.

Yep. Got that. I couldn't draw wire cable. LOL. So I'll be running cables diagonal between the corners with turn buckles in the middle to tighten it up. I'm hoping I can use them to make slight adjustments should the base not be perfectly square.

4 - If the stepper specs are right (1.8mH), you "should" be able to spin the stepper to 1000rpm which would give you around 200IPM. I spin mine at that speed with an AM882 and 40v PS. My machine rapids at 400IPM with 10mm pitch ballscrews.

Thanks. I'm still not completely across steppers yet. Like I said, I'm hopping to be in the ball park and then adjust a little to get it right. Just as long as I don't end up replacing all the electronics, I'll be happy.

Nice job for a first pitch. Took me 2 month designing my second machine.

Thanks again. I'm just hoping nothing is too seriously wrong for it not to work at all then I can make improvements once it's built. I really like how I can construct things with Fusion 360 and make sure it's all going to fit together at least. I found this interference test pretty good.

[joel0407]

This is my test piece for the construction. I tried drilling and tapping but it took out too much wood and stripped easily. Tried not tapping but the ply split. M6 tap needs a 5mm drill. 5.5mm drill untapped worked best. I added PVA glue for good measure and the screws can still be removed easily. I'm just not sure this is going to work very well for the MDF pre-build.

[pirate323i]

I have read of people using superglue soaked into the "threads" on MDF to improve its properties...
(No personal experience)

Sent from my SM-G930F using Tapatalk

[joel0407]

I have read of people using superglue soaked into the "threads" on MDF to improve its properties...
(No personal experience)

Sent from my SM-G930F using Tapatalk

Yep. I've done that with balsa. At a mush smaller scale though. I think I'd need near a bottle of CA per hole. LOL. I'm going to cut some MDF and test this afternoon.

[pirate323i]

Yep. I've done that with balsa. At a mush smaller scale though. I think I'd need near a bottle of CA per hole. LOL. I'm going to cut some MDF and test this afternoon.

Haha, good luck!

Sent from my SM-G930F using Tapatalk

[louieatienza]

This is my test piece for the construction. I tried drilling and tapping but it took out too much wood and stripped easily. Tried not tapping but the ply split. M6 tap needs a 5mm drill. 5.5mm drill untapped worked best. I added PVA glue for good measure and the screws can still be removed easily. I'm just not sure this is going to work very well for the MDF pre-build.

Fasteners are usually weakest being secured to the edge of a board. One reason is that the core woods are not usually as hard as the skin woods. Another reason is that when not tapping the holes, this puts pressure on the plies, causing them to eventually split as the threads displace the wood instead of fitting in cut threads. This is one reason you may see MDF and plywood builds use barrel nuts. It may be better to pre-drill, countersink, and use a SPAX screw, which are self-tapping and take out less wood than the M5 bolts.

I don't think your machine is large enough that you'll find your screws to be slow. However I think you underestimate the power of the screw/stepper combo as they will twist and tweak your frame no matter how stiff you think that marine grade plywood is. Many similar builds here, and almost on all of them reinforcements were added to the gantry risers. You really could use a smaller stepper, somewhere in the 280-350oz-in range. And that drive is way overkill unless you get them dirt cheap.

Finally, I'd reconsider the mounting method for the steppers. They will get hot, and mounting them to a wood product allows no heat dissipation. It's one reason why manufacturers recommend the steppers be mounted to metal motor mounts.

[Biggs427]

To give an idea, here's the link to my second build: http://www.cnczone.com/forums/cnc-wo...rch-frame.html

Im running 200oz stepper direct on the 10mm pitch screws.

I had these motors on my old machibe with single 3/4 plywood uprights and it was twisting like crazy even if I reinforced it with baltic birch.

[joel0407]

To give an idea, here's the link to my second build: http://www.cnczone.com/forums/cnc-wo...rch-frame.html

Im running 200oz stepper direct on the 10mm pitch screws.

I had these motors on my old machibe with single 3/4 plywood uprights and it was twisting like crazy even if I reinforced it with baltic birch.

I suspect draw slides and and single screw on the X axis might have been part of the problem there.

I'm just looking over your second build.

[joel0407]

To give an idea, here's the link to my second build: http://www.cnczone.com/forums/cnc-wo...rch-frame.html

Im running 200oz stepper direct on the 10mm pitch screws.

I had these motors on my old machibe with single 3/4 plywood uprights and it was twisting like crazy even if I reinforced it with baltic birch.

A couple of questions about your second build. Any reason you put the rails higher than your work height? Does it cause you any problems your router being so far out in front of the bearings?

[joel0407]

This is the problem I see having the X screw above the work.

Putting the screw as close as possible to the work height minimizes the rotational force.

[ger21]

I agree that the screws should be OK on that size machine. With AM882's, you can run in the 50-60V range, which should get you good speed.
For comparison, I can get 200ipm with 1/2-8 2 start acme, and 250oz motors at 24V. The acme is only about 60% efficient, vs 90-95% efficiency with ballscrews. With higher voltage, ballscrews, and better drives than I'm using, you might be able to see 250-300ipm.

Regardless of how strong you think your plywood is, it will flex. You don't want ANY flex at all. I'd at least double the side plates, and make a torsion box gantry. At the very least, make the gantry 2-3 layers thick, with no cutouts.
Another option is to make more cutouts in your gantry piece, and add a skin on each side of 1/4" baltic birch. This would make it a thin torsion box, and would greatly increase stiffness. As it's a smaller machine, this will probably be adequate.

I'd add some triangular braces or square side panels on the Z axis plate. As drawn, it'll flex quite a bit the way the spindle is mounted.

[joel0407]

[QUOTE=ger21;2063250]I agree that the screws should be OK on that size machine. With AM882's, you can run in the 50-60V range, which should get you good speed.

Yep. I'm going the AM882 drivers mainly for the extra voltage handling. Sure there are plenty of cheaper options but not many that will handle the voltage.

Regardless of how strong you think your plywood is, it will flex. You don't want ANY flex at all. I'd at least double the side plates, and make a torsion box gantry. At the very least, make the gantry 2-3 layers thick, with no cutouts.
Another option is to make more cutouts in your gantry piece, and add a skin on each side of 1/4" baltic birch. This would make it a thin torsion box, and would greatly increase stiffness. As it's a smaller machine, this will probably be adequate.

I'm starting to get worried by the number of people that doubt the strength of the marine ply. I can stand on this stuff with the 50mm back to create a U channel and get no flex. I'm doubting the router would handle 85kg of force. I'm also using 16mm supported rail. I've seen other machines use unsupported rail (with flex mind you). I'm sure my marine play gantry will exceed some of the T-Slot gantries that I've seen

I'd add some triangular braces or square side panels on the Z axis plate. As drawn, it'll flex quite a bit the way the spindle is mounted.

Thanks. I'd missed that.

[Biggs427]

I suspect draw slides and and single screw on the X axis might have been part of the problem there.

I'm just looking over your second build.

The drawer slides were a problem, but to mitigate this I had the slides preloaded and I added a moving knot.

Even then I could measure deflection. When the force was applied along the Y axis, there were deflection. So yes, the slides were part of the problem, but not the only problem. Uprights were also part of it.

That's why I chose to build my second machine base on a Solsylva (spelling?) The shorter uprights gives good results.

Unless I missed it, what are your expectations for your new machine in terms of capabilities?

[joel0407]

Unless I missed it, what are your expectations for your new machine in terms of capabilities?

Honestly. Initially, I only wanted to cut 6mm Pin Ply wood. LOL. That was it. I'm sure I'll use it for what ever I can though. I'm liking some of the wooden gear machines I've seen. Most of the gears are cut from MDF but I might do something to look better from wood.

Happy Days

[NIC 77]

I'd add some stiffening to your gantry risers and your z plate.

Get rid of those lightening holes in your gantry and make the gantry a box structure with ribs on the inside and a back piece.

If this was made from 18mm aluminum, I'd say OK, but since it is wood, I say you need to think more along the lines of a torsion box for all of the different pieces. Actually, if it was aluminum I'd still say to get rid of those lightening cuts in the gantry.

Strength doesn't matter to a CNC machine. It's all about stiffness. A couple changes to make it more boxy with ribs IMO is better.

Thin countersunk wood screws with a small pilot hole and glue to put the box pieces together I should think. Yep, tapping wood for M6 screws, bad idea. Use wood screws and glue for wood!

You'll have to think about how you will attach the rails etc. I would think that nuts, washers, and lock washers on the back side would be best. Not sure if you will have access to the backs of all of your pieces. Perhaps Ger or Biggs can give some advice on this.

I think you put the ball screw that moves the Z assembly in the correct place. I have looked at pics of big VMC machines that use huge servos and cut steel, and the ballscrew is in between the gantry rails, but closer to the bottom rail, just like you have it. I'm doing the same thing on my build, which is steel frame.

I should think that you will be able to cut whatever thickness of wood you have clearance for, you just won't be able to do it in a single pass.

I wouldn't worry too much about the weight of your moving parts. You are making this out of wood. Even if your gantry assembly weighs 200 lbs, you should be just fine. If you are thinking that you need to make it as light as possible to get the best acceleration, I would say that it will not be stiff enough to handle the acceleration.

IMO, back to the drawing board. Happy Days!

[louieatienza]

Yep. I'm going the AM882 drivers mainly for the extra voltage handling. Sure there are plenty of cheaper options but not many that will handle the voltage.

Well yes according to the spec sheet the high inductance of the stepper in series winding would allow it to run at pretty high voltages, though again, since you have no way for the steppers to dissipate heat, I probably wouldn't go all out on the voltage.

I'm starting to get worried by the number of people that doubt the strength of the marine ply. I can stand on this stuff with the 50mm back to create a U channel and get no flex. I'm doubting the router would handle 85kg of force. I'm also using 16mm supported rail. I've seen other machines use unsupported rail (with flex mind you). I'm sure my marine play gantry will exceed some of the T-Slot gantries that I've seen

Just because you can stand on it doesn't mean it can't twist. And the steppers, in conjunction with the Z axis acting as a lever arm, will twist that gantry. I do think the rest of the machine is fine for its intended purpose, but from my own experience as well as the dozens upon dozens of builds here that started off as having a single-layer gantry upright, the uprights always end up reinforced.

[joel0407]

Thanks for the comments all. Keep them coming.

"Louieatienza" The motors will handle more voltage in series but then the inductance goes up considerably. I'll have to check the calculators but I suspect that I wouldn't gain any speed running it at 60v series over 48v parallel. Yes I'll pick up on torque but I don't think that I'm really short on that.

"TWIST"
Now I'm racking (pun intended) my brain over this. I'm going to guess that all that have problem with this have been running single screw on the X axis.

If I stand at the end of the table and the X gantry is moving away from me. The force in the router bit is toward me. Now the Y Gantry attaches to the X Gantry at 4 places. 2 on the top, on the same plane and 2 on the bottom, on the same plane. So with the X gantry moving away, force on the bit toward me, the force on the bottom rail is toward me and the force on the top rail is away. It's irrelevant if the Y Gantry is toward the left or right as there is an even push from each ball screws on each X Gantry side.

If we reverse the above and have the gantry moving toward me rather than away. Every thing is just opposite. The only force is forward and back on the top and bottom rails. Now I find the rails already very strong the short distance of 1m and than I'm backing that with 18mm ply sheet and then backing that with 50mm x 18mm support on the top and 40x40 aluminum angle at the bottom.

I just struggle to see where the flex will come from.

Should I be doing this with a single ball screw on the X axis. Firstly bellow the drag point of the bit then if the Y gantry is at it's extreme then you are relying on the bearing spacing. I could see this twisting the gantry no end.

Happy Days.

[TumbiUmbiTek]

May I ask what controller software you thinking of using? Mach3, UCCNC or LINUXCNC
Cheers