PDA

View Full Version : BUILD LOG: Aluminium FEA optimalized CNC router



PotatoMill
25-04-2016, 10:37 AM
I few weeks ago, I decided to get a CNC router to cut foam, carbon fiber and maybe some aluminum. I had also decided not to build one myself, because I knew router building often end up as the new hobby instead of the hobby you originally built the router for. Then someone directed me to this forum and all my plans changed.

I am now planning to build an aluminum-framed router that will eventually have capability to cut different ferrous alloys. In addition, I hope I can do some high-speed tool-paths in aluminum. In the early stages of the design, I planned to use aluminum profiles for the build, without any need for welding, but profiles are expensive and did not have the stiffness I wanted. Then I went on to steel, but that got to heavy. So now I am going to use 60x60mm t=4mm aluminum square tubes welded together. I have to practice my aluminum welding some more before I start with the build. I have done FEA on most of the build, where I started with requirements of a frame stiffness of 10-20 N/μm (10'000 N/mm - 20'000 N/mm). Deflection is proportional with force so I have checked the stiffness with lower forces and then multiplied up to 10'000 N and 20'000 N. I have done many iterations on the frame and the FEA images under is only the most relevant for the current design.

Planned working area is 1000x500x250mm

Complete assembly
18301


The plan is to use 20mm profile rails and 16mm ballscrews on all the axis. With dual 1000mm drive one the X-axis. I'm going to use nema23 stepper motors but i'm still uncertain on the torque requirements I have to do some more calculations there. Though it seems like most of the steppers have a very high corner speed and to utilize all the torque and speed of the motors they should be geared down about 3:1. Any thoughts on this?


Base assembly (500N loads)
18300

18291
Stiffness; 223 N/um. I tried some different setups for vertical loads on the base. 5 vertical supports on the sides gave 4.8 times more stiftenes on the worst case compared to 3 vertical supports.
18292
Stiffness; 53 N/um. No need for cross bracing on load in the X axis it's more than stiff enough.
18294
The weak link where the forces in the Y axsis, this was fixed with some side bracing
18293
Stiffness; 51 N/um. The side bracing gave 4.7 times higher stiffness. Success! (I tried a lot of different potential setups before deciding on this one)



Gantry (100N loads)
18302
The whole weight of the gantry including the Z-axis came to 32kg, i'm going to see if I eventually can cut the weight down. I had hoped for something in the range of 20kg.
18296
Single aluminium tube thick gantry. Were within 10 N/μm but not 20.
18295
Stiffness; 32 N/um. Double thickness improved the stiffness 2.7 times, and the weight gain was small compared to the gain in stiffness.



Z - axis (100N loads)
183051830318304

The Z -axis ended up on 12kg. I have looked for ways to reduce the weight but often the stiffness suffers.

I'm going to use a 2.2kW water-cooled spindle. However if it is going to be use for ferrous alloys I have to gear it down to reasonable speeds. The plan is to use a belt reduction, but it's probably going to be a later update.

Under is some pictures of the results on the Z-axis. This is clearly the weakest link on the whole machine. It's no point making the the rest of the machine stiff if this part has 1/10 of the rigidity of the rest og the machine. The two vertical bracings on the front improved the stiffness but it's still not inside my requirements. The best solution to this that I have seen is Routercnc's MK4 design. However I dont have a CNC machine to build a CNC machine and I don't want to over-complicate my first build. It would be great if someone could come with input and suggestions on the design before I start ordering stuff, or it could become very expensive.
18297
Stiffness of the Z-axis in the X direction (worst load direction) 3.9 N/um.
18299
Stiffness; 0.85 N/um
18298

magicniner
26-04-2016, 03:53 AM
Have you done an analysis of resonant frequencies with the gantry and z in a variety of positions?

toomast
26-04-2016, 08:16 AM
Nice pictures. I would advice you to make bed analysis as well. Consider the surface you will mount the hole machine. On floor or on table. Maybe you need adjustable legs. This will probably change all calculations you made so far.

routercnc
26-04-2016, 08:52 AM
Hi Potatomill

Some good work there using analysis to explore the design. Some comments -

Although aluminium is 3 times lighter than steel it is also 3 times less stiff. Using steel in the base assembly, where the extra weight is of no consequence and is actually very beneficial, might negate the need to add the extra triangular bracing. Especially if you can go for high wall thickness.

Don't forget that the stiffness you are aiming for (10-20 N/um) is for the WHOLE machine, from tool tip to bedplate. If you design each sub-assembly to be 10-20 N/um you will get somewhat less stiffness in total because they will behave like a set of springs in series. In other words each sub-assembly needs to be significantly stiffer than 10-20 N/um so that when added together the total stiffness is in the 10-20 N/um range.

The Z axis is always a challenge with this style of machine. Try 20-30mm thick aluminium plate and see what that does. I suspect it will not give what you want especially at 450mm long. So another option is to use a piece of RHS (steel) with say width similar to the current plate width, depth 80-100mm, and length 450mm. Then house the spindle inside with cut outs or access holes to the spindle as required.

I think the limiting factor of the design, if you really are trying to get 10-20 N/um will be the torsional stiffness of the gantry as seen by the tool tip (i.e. Z axis and gantry modelled together with load applied at tool tip).

Good luck with it all . . .

PotatoMill
26-04-2016, 10:19 AM
Have you done an analysis of resonant frequencies with the gantry and z in a variety of positions?
Not yet but eventually I’ll get to it. I don’t know how accurate it will be because of all the simplifications i need to do on the analysis. Frames like this have a tendency to vibrate quite a lot.


I would advice you to make bed analysis as well. Consider the surface you will mount the hole machine. On floor or on table. Maybe you need adjustable legs. This will probably change all calculations you made so far.
I'm not completely satisfied with the base fixturing myself. I do not have the most ideal location for my machine so I have to move it twice a year. Through a door up some stairs. Hence, the plan is to put weight on it when it is in place. If I find a better location for the machine, I will change the design accordingly.


Although aluminium is 3 times lighter than steel it is also 3 times less stiff. Using steel in the base assembly, where the extra weight is of no consequence and is actually very beneficial, might negate the need to add the extra triangular bracing. Especially if you can go for high wall thickness.
I totally agree. Steel would be first choice if I find a good place for the machine. The Young's modulus for steel is certainly better than for aluminum.


Don't forget that the stiffness you are aiming for (10-20 N/um) is for the WHOLE machine, from tool tip to bedplate. If you design each sub-assembly to be 10-20 N/um you will get somewhat less stiffness in total because they will behave like a set of springs in series. In other words each sub-assembly needs to be significantly stiffer than 10-20 N/um so that when added together the total stiffness is in the 10-20 N/um range.
20 N/um is an optimistic goal. I had to do a lot of simplifications in the analysis. The worst case for the base is currently at 51 N/um and the gantry 33 N/um. As you pointed out, the twisting of the gantry, the deflection will increase with the length of the Z-axis. The base will not be as stiff either due to the fact that it's fixed to the ground in the analysis. I forgot to write it in the original post, but 500N loads where used for the base and 100N for the gantry. Adding it now. Deflection is about proportional with force, so I just multiply it up to find the stiffness in N/um.


The Z axis is always a challenge with this style of machine. Try 20-30mm thick aluminium plate and see what that does. I suspect it will not give what you want especially at 450mm long. So another option is to use a piece of RHS (steel) with say width similar to the current plate width, depth 80-100mm, and length 450mm. Then house the spindle inside with cut outs or access holes to the spindle as required.
I will explore this setup, and some other potential setups and post what I find out.

PotatoMill
02-05-2016, 09:19 PM
Update on the design. I have finally managed to get sufficient stiffness in the Z cart (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart). The solution was no surprise, maximizing second moment of area, and... steel. I also moved the Z blocks down as far as I could and placed all the Z blocks over the Y blocks. I am now planning to use a 100x200mm t=4mm square steel tube. The Z cart (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart) became twice as heavy, now 20 kg (44 lbs). However, about 20 times stiffer. The large cut on the backside is so I can get more space for the ball-screw and eventually remove the spacers for the Z blocks. I did test if the cut would compromise the design but it held up well to loads in all directions and torsional load.
18373

Now that the Z cart (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart) is good, torsional stiffness in the gantry is the problem. I am also doing the analysis on the whole gantry with Z and Y carts. The connections between the blocks and rails is not correct. They are just bonded. I do not know of any good ways in SolidWorks to make them in an efficient way, without messing around with spring connections. Therefore, it is a gross simplification, but it should give an approximate overview of the deformation.
18374
The image is with both gravity and a 100 N (22.4 lbs) load in the X direction, a potential worst-case load. It gives a deflection of 0.014 mm (0.55 thou). With some vibrations, joints, and the base, the deflection will be bigger. At this point, the structural frame is not the weak link any more, but ball-screws, rails, steppers, drivers, and so on. The stiffness is at 13N/um now. As I am a newbie on this, I wondered if this is a good point to settle?

The gantry, Z and Y carts weight about 40 kg (37.8 lbs) with a aluminium gantry. And 65 kg (143 lbs) with a steel gantry. So I guess steel gantry is not out of the question.

I also looked into changing the design on the gantry, by using two parallel beams and having the Y and Z cart (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart) mounted in the middle. Like Routercnc's MK4 router design. This had many advantages on the rigidity and weight distribution.
18375
However, this would require rails on both sides of the Z cart (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart) and I cannot find any good way to prep the surface and mount them accurately. With the classic gantry design, I can use epoxy, but with the parallel beam design, the Y and Z cart becomes a problem. The design would also be a lot more complex with parallel ball-screws on Y. In addition, four rails on Z or parallel ball-screws here too. And I have enough problems to worry about already, and on a first build it's best to keep it simple.

njhussey
02-05-2016, 09:51 PM
If you've got the Z blocks over the Y blocks how the heck are you going to do one set up??!

PotatoMill
02-05-2016, 10:00 PM
Magic I suppose. I know there is a wizard over at cnczone. The setup was mainly for testing out forces and deflection, to do many iterations and test concepts I did not go trough all the details. Actually I think i'm going to use normal blocks on one side and the wider carriage on the other side. I'm going to post a image of a better assembled design soon.

PotatoMill
05-05-2016, 01:38 AM
Little update on the Z and Y cart (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart) design, after a lot of iterations i'm now converging on a better solution for the design. However there are still changes to make and improve upon, and I hope on some feedback on my design.
http://www.cnczone.com/forums/attachment.php?attachmentid=319178&d=1462404085&thumb=1 (http://www.cnczone.com/forums/attachment.php?attachmentid=319178&d=1462404085)
A image of the whole thing. With the 100x200mm t=6 steel square tube in the front.
http://www.cnczone.com/forums/attachment.php?attachmentid=319180&d=1462404166&thumb=1 (http://www.cnczone.com/forums/attachment.php?attachmentid=319180&d=1462404166)
New setup on the stepper motor for the Z axis with a belt and pulley system. The Stepper is also mounted so it's easy to tighten the belt. It's also easy to pluck out of the machine.
The end of the ball screw it fixed with a custom double ball bearing fixture. Using 10mm thick SKF 6201 ball bearings.
http://www.cnczone.com/forums/attachment.php?attachmentid=319182&d=1462404357&thumb=1 (http://www.cnczone.com/forums/attachment.php?attachmentid=319182&d=1462404357)
Njhussey, after some back and fourth I managed to make space for the Mounting of the Z and Y blocks over each other.
http://www.cnczone.com/forums/attachment.php?attachmentid=319184&d=1462404679&thumb=1 (http://www.cnczone.com/forums/attachment.php?attachmentid=319184&d=1462404679)
Here is the square tube see-trough so it's possible to see the mounting of the spindle and the ball screw mechanism. I had to make a cut in the square tube from the top to make space for the ball screw. This did not however weaken the rigidity of the Z cart (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart) significantly.

PotatoMill
05-05-2016, 03:29 PM
Update: Pictures is not good for viewing a design so I have placed it on GrabCAD for everyone to view in 3D. https://grabcad.com/library/z-cart-cnc-1it's currently only the Z and Y c (http://www.ebay.com/sch/Home-Garden-/11700/i.html?_nkw=cart)art since i'm working on changing the gantry and the base.

routercnc
05-05-2016, 07:37 PM
It's an interesting design and if you really want to do the Z axis that way then I would add a reinforcement panel on the front outer face of that RHS box to spread the load of the spindle out to the edges (I assume the spindle is mounted just on the front face and not sandwiched between the front and rear faces?). Say a 8-10mm thick aluminium plate as wide as the RHS and enough height to pick up on the 4 spindle mounting bolts. Otherwise the loads will go into the weaker membrane panel and not be transferred to the stiff edges. If you simplified your loading condition on the analysis this will not show up.

Same could be applied to the rails, I would add a thin strip of say 6-8mm thick steel, by 30mm wide on the inside of the RHS directly where the rails mount and screw the rail bolts right through into it.

Also in post#9, first picture, the upright bracket holding the ballnut to the Z axis is too thin. It needs to be approaching the length of the ballnut and then bolted down in 4 places. It should be a 'boss' not a 'bracket'. Also the plate it sits on is on the small side if you are trying to maximise performance. This needs a better connection to the upright Y axis plate. As drawn it will twist when loaded. Maybe a similar plate parallel to the current plate but above the ballnut would work trapping the ballnut in-between 2 plates.

Keep exploring, it's good to see something different every now and again.

PotatoMill
05-05-2016, 08:18 PM
It's an interesting design and if you really want to do the Z axis that way then I would add a reinforcement panel on the front outer face of that RHS box to spread the load of the spindle out to the edges (I assume the spindle is mounted just on the front face and not sandwiched between the front and rear faces?). Say a 8-10mm thick aluminium plate as wide as the RHS and enough height to pick up on the 4 spindle mounting bolts. Otherwise the loads will go into the weaker membrane panel and not be transferred to the stiff edges. If you simplified your loading condition on the analysis this will not show up.

Same could be applied to the rails, I would add a thin strip of say 6-8mm thick steel, by 30mm wide on the inside of the RHS directly where the rails mount and screw the rail bolts right through into it.

Also in post#9, first picture, the upright bracket holding the ballnut to the Z axis is too thin. It needs to be approaching the length of the ballnut and then bolted down in 4 places. It should be a 'boss' not a 'bracket'. Also the plate it sits on is on the small side if you are trying to maximise performance. This needs a better connection to the upright Y axis plate. As drawn it will twist when loaded. Maybe a similar plate parallel to the current plate but above the ballnut would work trapping the ballnut in-between 2 plates.

Keep exploring, it's good to see something different every now and again.

The spindle mount is bolted from the front and the rear of the large square steel tube. However I'm not satisfied with it because the mounting of the plate can maybe drag the clamping around the spindle apart. I also need to make tramming possible.

The nice thing with the steel square tube is all the space for mounting parts. If I eventually want to add a 3:1 belt reduction on the spindle, for steel. (whisfull thinking)

Plates on the inside for the rails sounds like a good idea.

The upper ballnut connection is made of 3 layers of 10mm aluminium plates, where the lower one is bolted to the large steel tube. The grabCAD model should show it better. Or I'm misunderstanding something.

Thanks for the comments it have helped me removing a lot of weak points.

Edit: I see I forgot to add the mounting holes for the backside of the spindle attachment on the square tube.

PotatoMill
07-05-2016, 11:39 PM
Now the whole thing begins to come together again. I have done some more FEA on the whole frame. I will just add that they are approximations and gives a rough overview. Joints and mounting points will add to the deflections.

18408
Base:
I did some test with and without the side supports. The side supports make the base a lot more rigid. However, without the side support the deflection on the base is still insignificant compared to the gantry and the y and z carts. I also looked at how steel would increase the rigidity, but since the deflection on the base already was insignificant on the total frame, the increased stiffness from the steel was very small on the whole frame. The base is about 57kg with aluminum and 105kg with steel.
18407
Gantry:
The gantry have a lot to say on the stiffness of the frame especially torsion that add to deflection on loads in the x axis. The gantry is changed from a frame with a lot of 60x60mm square tubes to a single 300x200mm t=6 rectangular tube. The stiffness of the new setup is about the same, but it makes the assembly a lot easier. Which means there is less areas to make mistakes the rigidity of the machine will suffer from. I had to go for steel on the gantry to get it rigid enough, an aluminum rectangular tube was just not enough. Casting aluminum could be possible, but that will again make it more complex and prone to errors. The gantry is still movable on its own, but it is now at 50kg. :suspicion: The Y and Z cart is at 20kg so in total the gantry and Y and Z cart will be 70kg.


The design on GrabCad is now updated https://grabcad.com/library/z-and-y-cart-gantry-cnc-1. However, the ball screws for moving the gantry is not finished yet. And I see that mounting of the rails for the Y cart makes it impossible to attach the Y cart so I have to find a solution for that. Moreover, it would be great if someone have a good suggestion.

magicniner
08-05-2016, 02:58 AM
Given that the additional cost and effort is low I'd add the side supports anyway as there may be a significant reduction in resonance under circumstances for which you have not calculated.

PotatoMill
08-05-2016, 03:10 AM
Given that the additional cost and effort is low I'd add the side supports anyway as there may be a significant reduction in resonance under circumstances for which you have not calculated.
Indeed, that's why I will probably build them anyway, but it can also be easily added later. The only minus is the added with, if I can get it trough a door on wheels whitout removing the gantry and tipping it on the side it would be nice. Else than that it is only positive with the added rigidity for unforseen deflections. Hovewer the base is already more than strong enough, it's the torsion on the gantry that is the weak point.

PotatoMill
07-06-2016, 07:47 PM
I'm currently considering making a stationary gantry design. Since the weight of the gantry pluss the z and y cart comes inn at 80kg (mostly in steel).
The cons will be;
-Longer machine
-Longer rails
-Potensial for sag in the bed

Pros
-Less weight to move for the steppers
-Simpler design
-Less twist on the z axis, the largest factor for displacement

The reason i'm posting this is because it probably a lot more pros and cons than I can come up with, that I hope someone here will add.

PotatoMill
17-06-2016, 07:42 AM
So I have been working on redesigning the whole machine. I'm currently going for a stationary gantry design.
18718
Due to the fixed gantry design the heaviest moving part is now the base plate with 30kg. I'm going to use 20mm HIWIN profile rails on all axis.
There will be one lead screw on each axis, no need for two on the X any more due to the stationary gantry. Here i'm going to use 1605 ball-screws. To drive the axis i'm currently looking at using servos in the 200-400W range. There is not the same amount of information on servos av steppers so I hope someone here have some knowledge on servos. Seems like JMC produce the cheapest servos and drives, but I cant find a lot of data on them. If there servos from JMC are good enough i'm probably going to use them. BST automation also seems to have a ok selection on servos, but again there is little documentation to find.

Lee Roberts
17-06-2016, 07:21 PM
BST automation also seems to have a ok selection on servos, but again there is little documentation to find.

If you send them a msg Fred will be more than happy to help and send you what you need.

.Me

magicniner
17-06-2016, 07:56 PM
18718


It looks like you have the rails on the base and the carriages moving on the X-axis.
If you have the carriages fixed on the base and the rails moving under the X-axis you can have maximum support and rigidity where you need it most at all times.

remrendes
17-06-2016, 10:33 PM
I can give some info about that. I have one 60ASM200 at home. My computer is not enough fast for this resolution, so that is why it looks slow on a video. The electric gear ratio is 1:1. https://www.youtube.com/watch?v=qyrEdBIpqx8

magicniner
17-06-2016, 10:39 PM
I can give some info about that. I have one 60ASM200 at home. My computer is not enough fast for this resolution, so that is why it looks slow on a video. The electric gear ratio is 1:1. https://www.youtube.com/watch?v=qyrEdBIpqx8

Go on then, Giz a Clue?

remrendes
17-06-2016, 10:43 PM
A square tubes looks not enough strong for me for aluminium machining.
I'm working on this frame for alu, with a smaller motors than what you want.
18722

You should move the X ballscrow to lower. It is better between the 2 rails.
Everything looks welded. How can you do the end machining for a rails and others?

magicniner
17-06-2016, 10:52 PM
Info about what specifically?
I can't see who's post you are addressing and though it's obvious to you it's not obvious to me :-(

PotatoMill
17-06-2016, 10:57 PM
The design is kinda crude, it was a fast iteration. I'll probably get some time in the weekend to upgrade it so the details is more understandable. The i'm going to add end plates to the large gantry square tube. Most of the machine will be welded. I have planned to bolt the gantry and base together at the bottom of the gantry legs. I guess I need to remove the aluminium part in the title of the post sinse it's mostly steel now.


If you send them a msg Fred will be more than happy to help and send you what you need. Thanks for the tips, seems like Fred is a helpful guy.


It looks like you have the rails on the base and the carriages moving on the X-axis.
If you have the carriages fixed on the base and the rails moving under the X-axis you can have maximum support and rigidity where you need it most at all times.
I agree, I changed it back and forth several times. It became a compromise. Rails on the base would make sure the support is maksimum and equal at all times. However the moving base would be longer and the leverage from the longer base would put more stress on the machine. The weight of the moving part would also be higher.



I can give some info about that. I have one 60ASM200 at home. My computer is not enough fast for this resolution, so that is why it looks slow on a video. The electric gear ratio is 1:1.https://www.youtube.com/watch?v=qyrEdBIpqx8 Nice video, you got any info in addition?

remrendes
17-06-2016, 10:59 PM
Info about what specifically?
I can't see who's post you are addressing and though it's obvious to you it's not obvious to me :-(

Sorry, 60ASM200 is a JMC servo motor, what I mean.

remrendes
17-06-2016, 11:05 PM
Nice video, you got any info in addition?

I have eveything :) I can give a contact number as well, I have seen a factory in China. They are pretty good.
If I will not forget, I'm going to upload a training video what I have from them.

PotatoMill
18-06-2016, 01:28 PM
The cost of servos is going to be quite high, so i'll probably for closed loop steppers instead. However the ClearPath servos seems tempting. Is there any good reason to use servos instead of closed loop steppers?

Fenza
18-06-2016, 10:40 PM
The cost of servos is going to be quite high, so i'll probably for closed loop steppers instead. However the ClearPath servos seems tempting. Is there any good reason to use servos instead of closed loop steppers?

Closed loop steppers don't have the speed or torque capabilities of servos and are sold as an 'economical' alternative. The only good reason to use closed loop steppers is because they are cheaper.

Some cheap servos like the JMC can be similar price to most closed loop steppers, it'll be interesting to know if anyone has tested how their performances compare.

Cheers,

Fenza

remrendes
19-06-2016, 05:27 PM
The cost of servos is going to be quite high, so i'll probably for closed loop steppers instead. However the ClearPath servos seems tempting. Is there any good reason to use servos instead of closed loop steppers?

Both have a feedback to driver, so both are better than a normal steppers.
A cost of the servos is higher than what you think. E.g. you have to reduce the rpm, with 1:2 - 1:3 transmission. 1x zero backlash pulley is approx 30-40, one belt can be 15-20 and you need 3 set. With steppers, you just need a coupling.
Unfortunately I started with servos, but I could get a cheap set from China.

PotatoMill
19-06-2016, 06:22 PM
Thanks for the quick response Fenza and Remrendes

According to a quick estimate I did I dont think i'm going to need any more torque than what these closed loop steppers deliver; http://www.aliexpress.com/item/new-original-Leadshine-nema23-2NM-Hybrid-servo-kit-HBS507-573HBM20-1000-Closed-loop-stepping-motor-drive/32566531363.html?spm=2114.30010308.3.20.uQElOe&ws_ab_test=searchweb201556_10,searchweb201602_5_10 037_10017_301_507_10033_10032,searchweb201603_1&btsid=ae7e7cbb-232f-411d-acba-82b3fb4b4813

Estimates done with

-40kg weight of moving part
-130N force of cutter (excessive)
-20N friction
-85% efficiency
-10 m/s^2 acceleration (more than I think i'll need, but I haven built any CNC machines before so i'm not certain on this.)
-5mm lead

Using T = (F*l)/(2*PI*n) where F is total force, l is lead per revolution, n efficiency
I get 0.44N well within

However I do know that theory and experience dont always match up. And in real life there could be some factor that make the estimate wrong.

Fenza
19-06-2016, 06:49 PM
Thanks for the quick response Fenza and Remrendes

According to a quick estimate I did I dont think i'm going to need any more torque than what these closed loop steppers deliver; http://www.aliexpress.com/item/new-original-Leadshine-nema23-2NM-Hybrid-servo-kit-HBS507-573HBM20-1000-Closed-loop-stepping-motor-drive/32566531363.html?spm=2114.30010308.3.20.uQElOe&ws_ab_test=searchweb201556_10,searchweb201602_5_10 037_10017_301_507_10033_10032,searchweb201603_1&btsid=ae7e7cbb-232f-411d-acba-82b3fb4b4813

Estimates done with

-40kg weight of moving part
-130N force of cutter (excessive)
-20N friction
-85% efficiency
-10 m/s^2 acceleration (more than I think i'll need, but I haven built any CNC machines before so i'm not certain on this.)
-5mm lead

Using T = (F*l)/(2*PI*n) where F is total force, l is lead per revolution, n efficiency
I get 0.44N well within

However I do know that theory and experience dont always match up. And in real life there could be some factor that make the estimate wrong.

Hi PotatoMill your math certainly is correct but it doesn't account for the full picture. I use Kollmorgen MOTIONEERING to calculate more precisely what kind of motors are required, it's a little tricky to understand in the beginning but sticking with it pays off massively! http://www.kollmorgen.com/en-us/service-and-support/technical/motioneering-online/ here's the link to their site. I use their older application at the bottom of that page, it looks like they have a new online tool available now.

Kollmorgen have a huge range of motors available and the MOTIONEERING app helps point you to which is suited for your machine, I just use the data from that motor to base my search for a cheap servo / stepper system off of.

Also always remember that steppers holding torque is not the same as a servo rated/max torque.

A good rule of thumb that I use (though I'm not entirely sure it's correct) is that a 3Nm stepper is roughly equal in power to a 100W Servo.

Hope this helps!

Fenza

remrendes
19-06-2016, 07:00 PM
Estimates done with

-40kg weight of moving part
-130N force of cutter (excessive)
-20N friction
-85% efficiency
-10 m/s^2 acceleration (more than I think i'll need, but I haven built any CNC machines before so i'm not certain on this.)
-5mm lead


With 5mm lead is not the best ballscrew for you. You shoud use the motors up to 500rpm. Under this, you nearly can get a full holding torque.

PotatoMill
19-06-2016, 07:29 PM
It would probably be best to make a 3:1 gearing to utilize the holding torque best. Or is it possible to get cheap ballscrews with shorter than 5mm lead?

remrendes
19-06-2016, 08:41 PM
It would probably be best to make a 3:1 gearing to utilize the holding torque best. Or is it possible to get cheap ballscrews with shorter than 5mm lead?

I would go for 10mm pitch and direct drive.
On 500rpm it is 5000mm/min, more than enough.

PotatoMill
19-06-2016, 08:43 PM
Hi PotatoMill your math certainly is correct but it doesn't account for the full picture. I use Kollmorgen MOTIONEERING to calculate more precisely what kind of motors are required, it's a little tricky to understand in the beginning but sticking with it pays off massively! http://www.kollmorgen.com/en-us/service-and-support/technical/motioneering-online/ here's the link to their site. I use their older application at the bottom of that page, it looks like they have a new online tool available now.

Kollmorgen have a huge range of motors available and the MOTIONEERING app helps point you to which is suited for your machine, I just use the data from that motor to base my search for a cheap servo / stepper system off of.

Also always remember that steppers holding torque is not the same as a servo rated/max torque.

A good rule of thumb that I use (though I'm not entirely sure it's correct) is that a 3Nm stepper is roughly equal in power to a 100W Servo.

Hope this helps!

Fenza

Great tool, very helpful. I'm trying both the new and old now, the old one seems better actually.

PotatoMill
19-06-2016, 09:22 PM
I would go for 10mm pitch and direct drive.
On 500rpm it is 5000mm/min, more than enough.
Wouldn't I want shorter pitch? If not I would not get close to the corner speed at all. And waste a lot of usable torque.

remrendes
19-06-2016, 09:53 PM
Wouldn't I want shorter pitch? If not I would not get close to the corner speed at all. And waste a lot of usable torque.

With steppers? No, not really.
What do you mean with a corner speed?

Here you can see a torque curve for both. You can calculate with a 0-500rpm range.
18743
Unfortuantely it is Hungarian, but here you can see my calculation for my X axis (https://drive.google.com/file/d/0B5Q2OeXv4-eTUTBENmY3cjk5V2c/view?usp=sharing) with servo. This is the optimal for a dynamics, but I chosen 2.5:1.

PotatoMill
19-06-2016, 10:26 PM
Here is some info on corner speed. https://www.google.no/url?sa=t&source=web&rct=j&url=https://www.geckodrive.com/gecko/images/cms_files/Step%2520Motor%2520Basics%2520Guide.pdf&ved=0ahUKEwjXkNfe77TNAhXLjiwKHe4BBEIQFggaMAA&usg=AFQjCNFeiMr4qAhguhvFv3vH6od3AB08dQ&sig2=H-hItspQJAddXnAgnBfq5A it's more apparent on these speed torque curves http://www.sanyodenki.eu/IMG/pdf/SteppingWithEncoder_06.12.N_e.pdf

remrendes
19-06-2016, 10:47 PM
Here is some info on corner speed. https://www.google.no/url?sa=t&source=web&rct=j&url=https://www.geckodrive.com/gecko/images/cms_files/Step%2520Motor%2520Basics%2520Guide.pdf&ved=0ahUKEwjXkNfe77TNAhXLjiwKHe4BBEIQFggaMAA&usg=AFQjCNFeiMr4qAhguhvFv3vH6od3AB08dQ&sig2=H-hItspQJAddXnAgnBfq5A it's more apparent on these speed torque curves http://www.sanyodenki.eu/IMG/pdf/SteppingWithEncoder_06.12.N_e.pdf

Yes, it is on the speed/torque curve, where the motor starts to loose his torque. Approx 4-600rpm on a cheap steppers.

Fenza
19-06-2016, 11:54 PM
Hi PotatoMill, I see you specced 10m/s^2 acceleration. Now I'm not too experienced with routers and steppers (Mills and Servos is more my thing) but 1g is kind of a lot! I did a quick calculation and a 20mm ball screw 1,200mm length 5mm pitch with 40kg load will require over 3Nm of torque at the shaft to accelerate all the moving parts at 10m/s^2

Cutting forces are only small but the force required to overcome inertia and accelerate a 40kg load is great!

I believe 2.5m/s^2 would be a good target to try and aim for, though It would be nice if someone experienced with steppers and routers could chime in and advise. I believe 50in/s^2 is considered not too shabby in Mach3 which would be 1.27m/s^2

Hope this helps! That Kollmorgen app really is a great tool once you get used to it :beer:

Fenza

PotatoMill
21-06-2016, 08:35 AM
I think it's more clear to me now.

Requirements
Speed: standard 0-2000mm/min max 3000mm/min
RPM needed:
5mm lead; standard 0-400RPM, max 600RPM
10mm lead; standard 0-200RPM, max 300RPM

A stepper motor seems to lose it torque from 300-500RPM
A standard servo motor has full torque up to 3000RPM. However to utilize the whole range I would need a 10:1 or a 5:1 reduction.

5mm lead will put me over the speed where the stepper starts to loose torque but it will also give me twice the force of a 10mm lead. I'm not sure i'm going to machine over 2000mm/min but for rapids I need a higher speed.

The weight of the X and Y cart is 30Kg. The Z cart is 20Kg
Cutting force; max 150N

After reading more on servos the torque advantages are clear. If my machine was bigger it would be obvious to use them, with the size om my machine there is not so much of a difference.

Fenza
21-06-2016, 09:15 AM
You could always use more powerful Stepper motors to negate the loss of torque at higher rpm. Also you could simply run your machine with higher rapids using servos so less of a gear reduction is required. 10mm pitch 10m/min rapids and a 3:1 reduction doesn't sound too bad! Ball screws have a maximum rotation speed so make sure you're not going to be over rotating them. 1,000 rpm is a pretty conservative speed but just be careful when it comes to spinning long ball screws at higher rpm.

If you'll be happy running at 3m/min then Open Loop Steppers will be a good choice. Rpm and microstepping both change the available torque so it'll be best to work out what torque will be required and choose the motor based on your calculations.

Cheers,

Fenza

remrendes
21-06-2016, 09:42 AM
Requirements
Speed: standard 0-2000mm/min max 3000mm/min
RPM needed:
5mm lead; standard 0-400RPM, max 600RPM
10mm lead; standard 0-200RPM, max 300RPM

A stepper motor seems to lose it torque from 300-500RPM
A standard servo motor has full torque up to 3000RPM. However to utilize the whole range I would need a 10:1 or a 5:1 reduction.


If you buy a not Chinese stepper, probably it is higher, like the price :) But yes, you should calculate up to 500rpm.
With 5mm lead from 2:1 to 3:1 usually enough with servos. On 2.5:1 ratio I can get up 6000mm/s no-load speed on 3000rpm. But some motors could do 4000rpm.

PotatoMill
21-06-2016, 10:30 AM
You could always use more powerful Stepper motors to negate the loss of torque at higher rpm. Also you could simply run your machine with higher rapids using servos so less of a gear reduction is required. 10mm pitch 10m/min rapids and a 3:1 reduction doesn't sound too bad! Ball screws have a maximum rotation speed so make sure you're not going to be over rotating them. 1,000 rpm is a pretty conservative speed but just be careful when it comes to spinning long ball screws at higher rpm.

If you'll be happy running at 3m/min then Open Loop Steppers will be a good choice. Rpm and microstepping both change the available torque so it'll be best to work out what torque will be required and choose the motor based on your calculations.

Cheers,

Fenza

Why not closed loop?

Fenza
21-06-2016, 11:17 AM
Why not closed loop?

Personally I'm not a fan of closed loop steppers... Open loop steppers are cheap and Servos are great performers, closed loop steppers are neither...

I think it is possible for you to achieve your desired level of performance using the cheaper open loop steppers. 3Nm stepper should probably be around about right directly coupled to a 5mm ball screw.

Servos you'd be looking at 100-200W I believe.

Hope this helps!

Fenza

remrendes
23-06-2016, 08:19 AM
Personally I'm not a fan of closed loop steppers... Open loop steppers are cheap and Servos are great performers, closed loop steppers are neither...
Fenza

A bigger steppers is a good idea. You can get enough torque in a no-load operation. Just you will get higher inertia as well.

PotatoMill
23-06-2016, 10:15 AM
I have heard open loop steppers is not good for point to point 3D toolpaths. They loose step easy and you cant push them very well.

magicniner
23-06-2016, 02:27 PM
I have heard open loop steppers is not good for point to point 3D toolpaths. They loose step easy and you cant push them very well.

Closed loop steppers are a solution looking for a problem, correctly designed open loop systems don't lose steps, if that can happen on a machine then it's badly configured or broken.
Closing the loop doesn't make a bad machine good, using the correct motors & screws and running it within it's limits does.

Boyan Silyavski
23-06-2016, 04:48 PM
Closed loop steppers are a solution looking for a problem, correctly designed open loop systems don't lose steps, if that can happen on a machine then it's badly configured or broken.
Closing the loop doesn't make a bad machine good, using the correct motors & screws and running it within it's limits does.

While this is true generally, closing the loop with servos definitely makes machine faster. I also am not fan of the closed loop steppers price wise.

But if the machine will be machining Metal i dont see why at all there is a discussion for sth else than servos and closed loop. Given the price of aluminum material that will be machined and eventually a mistake made.

Jonathan
23-06-2016, 07:03 PM
Closed loop steppers are a solution looking for a problem, correctly designed open loop systems don't lose steps, if that can happen on a machine then it's badly configured or broken.
Closing the loop doesn't make a bad machine good, using the correct motors & screws and running it within it's limits does.

Well said.

magicniner
23-06-2016, 07:15 PM
closing the loop with servos definitely makes machine faster.

But is this a function of the closed loop or is it the fact that the drivers are part of the package and this stops cheapskates buying inadequate hardware and then blaming low speeds and step loss on open the loop?
;-)

- Nick

PotatoMill
23-06-2016, 11:07 PM
I thought selecting a stepper or servo should bee easy, but it seems like nothings easy here. So for each person I ask I get a different recommendation. Steppers, open loop, closed loop, no only servos is good enough. The common thing is that everyone recommends servos if I want to use the money. But they are waaaaaaaaaaay more expensive than normal steppers. And all the cheap servos and closed loop steppers is a pain to get information on, since the datasheets and torque curves is not easily available and I have to ask the sellers for it for every motor. But my logic is as follows, if open loop steppers is good enough for Tormach it's good enough for me. First to comment on how this is not a mill. Then I can use some higher quality steppers instead.

By the way can someone remember the good reason to not use overkill steppers like 8nm nema 34? It's not like they are a lot much more expensive. I read it somewhere but I can't find it.

Clive S
23-06-2016, 11:57 PM
With nema 34 you will need mains voltage drives to get any decent results You won't go far wrong with decent 23,s from Zapp or CNC4you

With AM882 or similar drives and 68V power supply non linear

Fenza
24-06-2016, 12:36 AM
By the way can someone remember the good reason to not use overkill steppers like 8nm nema 34? It's not like they are a lot much more expensive. I read it somewhere but I can't find it.

I think Nema 34 steppers would be too large for your machine. The Load/Motor Inertia Ratio would probably be below 1:1 and therefore you're using more energy accelerating the motor shaft than you are accelerating the 40kg load!

I'm not sure about Steppers but for Servos a 5:1 Load to Motor Inertia Ratio is pretty normal, with 2-3:1 being higher performance and even 1:1 is used for super speedy machines.

A belt reduction gives you flexibility in the gear ratio. If you happen to under size the motors you can lower the gearing and get more torque with less speed, but if you over size the motors you can get some free speed!

I'm sure you can see how this becomes a hobby for many people... Haha!

Cheers,

Fenza

magicniner
26-06-2016, 12:01 PM
How about fixed screws and rotating ballnuts?
No problems with whip from the screws and reduced rotating mass ;-)

Boyan Silyavski
26-06-2016, 12:58 PM
...or is it the fact that the drivers are part of the package and this stops cheapskates buying inadequate hardware and then blaming low speeds and step loss on open the loop?
;-)

- Nick

I couldn't have said it better.

PotatoMill
26-06-2016, 04:15 PM
How about fixed screws and rotating ballnuts?
No problems with whip from the screws and reduced rotating mass ;-)
I'll just need CNC router to make it first

routercnc
26-06-2016, 05:05 PM
I'll just need CNC router to make it first

Jonathan has made them for people in the past - worth an RFQ?

PotatoMill
28-07-2016, 08:10 AM
Do someone have some opinions on this stepper http://www.aliexpress.com/item/Nema-23-Stepper-Motor-3Nm-425oz-in-2-Phase-4-lead-4-2A-L112mm-for-CNC/32660247354.html?spm=2114.13010208.99999999.264.8p hZye? 1.8mh seems nice, but the seller dont looks so professional.

Boyan Silyavski
28-07-2016, 08:24 AM
Do someone have some opinions on this stepper http://www.aliexpress.com/item/Nema-23-Stepper-Motor-3Nm-425oz-in-2-Phase-4-lead-4-2A-L112mm-for-CNC/32660247354.html?spm=2114.13010208.99999999.264.8p hZye? 1.8mh seems nice, but the seller dont looks so professional.

https://www.cnc4you.co.uk/Stepper-Motor

https://www.poscope.com/nema23

PotatoMill
28-07-2016, 08:27 AM
But the price is higher and the inductance is higher

Boyan Silyavski
28-07-2016, 08:36 AM
The price is same or lower. You have shipping costs and eventually duty to pay. If you are saving on stepper motors, you will not go far with that build. The links i gave you are the cheapest in Europe

18935

PotatoMill
28-07-2016, 08:57 AM
Inductance is probably my biggest concern.

Boyan Silyavski
28-07-2016, 09:08 AM
This is the motor that everybody here uses for a step motor on their CNC. No surprises, works well at 70V. The motor you show has less inductance on paper, but how true is that...

PotatoMill
28-07-2016, 09:13 AM
This is the motor that everybody here uses for a step motor on their CNC. No surprises, works well at 70V. The motor you show has less inductance on paper, but how true is that...

That was kinda my question. Is there anyone who have tested this/looked into it before?

Edit. I see that cnc uk sells the steppers in a kit with a 36v supply. I asked if they would put on a 48v supply instead. According to the classic 32*sqrt(inductance) rule about 57v should be optimal. Why is they selling 36v supplies in their kits?

Boyan Silyavski
28-07-2016, 09:25 AM
Cause most people / not on this forum/ don't build their own power supplies. And a 70VDC non regulated 550W PSU for 4 motors is not exactly cheap or light to manage and ship. So they offer the typical stuff assuming if you are to build strong and fast CNC you are knowledgeable from where to buy and build the PSU