[archsakas]

Hello,
I have built my own CNC router and it was working relatively fine.
Attachment 300460

It had rack and pinion drive with reducer ratio 3:1. I used nema 23 stepper motors 425 oz-in.
Attachment 300462Attachment 300464Attachment 300466

But now I have decided to change to ballscrew type drive, because I need more precision.
I used three RM2005-1500mm ballscrews
Attachment 300468

This is my redesigned router
Attachment 300470

And now I have a problem.
I cant move or jog at the same speed as I could with rack and pinion. Motors suddenly stops if I jog too fast even with low acceleration.
I dont understand why?
Is it just because motors are too weak?
It makes strong vibrations at some points of movement and then motor squeaks and stops.
Any suggestions how to get rid of vibrations?
I know one mistake I made in this build in X axis. I have placed ballscrew too far from linear rails. So I will have to fix that...
Attachment 300472

Will it solve the problem if I buy nema 34 1600 oz-in stepper motor?

Other question is what PSU is needed for nema 34 1600 oz-in motors?
I found this kit on ebay, but I dont nudestand if I realy need 3 PSU for 3 nema 34 motors?
Currently I use 720W 48V15A Power Supply with M542 Stepper Motor Driver 1.0A-4.5A and motors: x3 nema 23 425 oz-in 3.0A and x1 nema 23 263 oz-in 2.8A (z axis)

Thank you for any comments or remarks

[awerby]

You didn't give us the specs on your R&P system, but it sounds like it hit the "sweet spot" for those motors. Even geared down 3-1, you get a lot of travel per revolution with R&P. Those ball screws, on the other hand, look rather finely pitched - again, you didn't tell us exactly what the pitch is, but I'd guess it takes about 8 revolutions to go an inch. That means that to go at the same speed you went before, the motors have to turn faster, and they're not up to the task.

I don't think that going with those 34-frame motors is likely to help with this problem. They will have more torque, but that wasn't the issue. Ball screws don't require a lot of torque from a motor; that's the generally reason they're used. In this case, what you need is more speed, not more torque. For that, a 23-frame motor with less inductance, and less holding torque, will likely work better. You might be able to test that theory by swapping the smaller motor from your Z axis with one of the ones on X or Y and seeing if it works any better.

Also, I'm not sure you really got that much more precision with those ball screws; it depends on what grade you got. Even with a C7 ball screw, the permissible error is going to be plus/minus 50 microns per 300mm : http://www.tpa-us.com/pdf/sbc/SBC-ba...ew-catalog.pdf (that's equivalent to +/- .002"/ft; not bad for woodworking, but not a very high tolerance for metal machining.)

[ger21]

What you're seeing is the result of several things.
1) You're motors now need to spin about 5x faster with the ballscrew for the machine to move at the same rate it did with the rack and pinion.

2) Most 425oz motors have high inductance, and often need over 100Volts to reach their maximum rpm potential.

3) A large ballscrew has a lot of inertia, so you need more power to accelerate it.

Will it solve the problem if I buy nema 34 1600 oz-in stepper motor?

No, the problem will get worse.

Stepper motors have their maximum torque when they are not spinning. The faster they spin, the less torque they have. This becomes even more of an issue when you get into really large motors, like those 1600 oz motors. Once they get over 300rpm, they have about the same or less torque than the motors you have now.

There's also the issue that steppers have a fixed resolution of 200 steps/rev. Combine this with the loss of torque at higher rpm's, and you see that in order to gain more resolution, you need to spin the motor more times, which means faster. Which gives you less torque.
Bottom line, with a stepper, generally, if you want more resolution, then you'll need to go slower. IF you want to go faster, than you need to give up some resolution. If you want both, then you really need to look into servos.

I think your best bet is to find some very low inductance Nema 34 motors (around 2mH), in the 400-600oz-in range, with a current rating around 6-7 amps. You made need to use a 60-72V power supply to get the rpm that you need.

Power supply voltage should be 32 x sq rt of the motors rated inductance, for maximum motor performance.

For those motors you found on Ebay, with 22mH inductance, you'd need 150 volts to get the best possible performance from them. And they would still probably be much slower than your rack and pinion setup was.

[A_Camera]

You didn't give us the specs on your R&P system, but it sounds like it hit the "sweet spot" for those motors. Even geared down 3-1, you get a lot of travel per revolution with R&P. Those ball screws, on the other hand, look rather finely pitched - again, you didn't tell us exactly what the pitch is, but I'd guess it takes about 8 revolutions to go an inch.

He says that the ball screws are RM2005, which means 20 mm in diameter and 5 mm pitch per rev, so 5 rpm = 1" travel.

[A_Camera]

What you're seeing is the result of several things.
1) You're motors now need to spin about 5x faster with the ballscrew for the machine to move at the same rate it did with the rack and pinion.

If we assume he had the same 5 mm pitch then he should need 3x stepper rev to get the same moving rate as before since he geared up 3x.

2) Most 425oz motors have high inductance, and often need over 100Volts to reach their maximum rpm potential.

3) A large ballscrew has a lot of inertia, so you need more power to accelerate it.

Are you sure about this? I mean, he geared up 3:1 before, so the same stepper rev gave 3x moving rev and paid the price with force. I am not sure there will be much difference in acceleration with the same motors, but of course, assuming he needs 3x the speed now, this can be impossible if he already had maximum speed before.

The precision will increase since there will be less, or zero backlash. Geared rack and pinion drives have several points (in his case three) where the risks for backlash is increased and compared with ballnut/ballscrew solution is better. Perhaps in real life use that is not the case, but I am pretty sure a ball screw is better in this respect.

...or, am I wrong?

[A_Camera]

Hello,

Attachment 300464

Another thing I would change, apart from moving the ballscrew is the Z fixture. I think you have to move the spindle closer to the support. I don't understand why you added that extra and why not make the plate which is attached to the support sliders longer so that you could screw the spindle holder to it as well. In fact, for better precision, I think that the extension is just too long, it will generate a lot of vibrations and that in turn might cause stalling or precision issues.

[archsakas]

Another thing I would change, apart from moving the ballscrew is the Z fixture. I think you have to move the spindle closer to the support. I don't understand why you added that extra and why not make the plate which is attached to the support sliders longer so that you could screw the spindle holder to it as well. In fact, for better precision, I think that the extension is just too long, it will generate a lot of vibrations and that in turn might cause stalling or precision issues.

The reason for longer Z axis is to be able to do deeper carvings and I was thinking to upgrade this machine with the 4th axis on the same table, so I needed some hight as well as area.. but I did not have time for that yet...
In other words, it is "a big experiment" with a big working area..

[Khalid]

People converting from Lead screws to R&P, Here is the reverse...

[archsakas]

I did some calculations to compare.


I used to rough cut at Feed 100 inch/min and finish at 150 inch/min (that is ~65mm/sec).
But now I can't cut at half of that speed... Motor stalls at some point of the job and spoils it.
So I lowered speed even more to be able to cut.. for now...

It is a huge differance...

[ger21]

If we assume he had the same 5 mm pitch then he should need 3x stepper rev to get the same moving rate as before since he geared up 3x.

Most rack and pinion users use a 3:1 reduction, and still get about 1", or 25mm of travel per stepper revolution.

Which the OP confirmed in the previous post. So 5x the stepper rev, as I mentioned.

[UA_Iron]

I did some calculations to compare.


I used to rough cut at Feed 100 inch/min and finish at 150 inch/min (that is ~65mm/sec).
But now I can't cut at half of that speed... Motor stalls at some point of the job and spoils it.
So I lowered speed even more to be able to cut.. for now...

It is a huge differance...

Just expanding on what ger21 said, your motors are not properly sized for the application:

Are your stepper motors the nema23 from automationtechnologies?
http://www.automationtechnologiesinc...86-20-8BTC.pdf

150in/min @ ~1in/rotation = 150RPM that your stepper motors were running at. At 150 RPM your steppers generate around 2Nm of torque or 282 oz-in.

You need a stepper motor that generates at least 282oz-in at 5x the RPM to achieve the same cutting speeds. So your new sizing specs would be 282oz-in @ 750RPM.

These will achieve that at 70V, you may be able to squeeze out your cutting speeds at 48V even (considering a ballscrew is a more efficient drive unit). But they aren't cheap:
STP-MTRH-34097 | Stepper Motor: NEMA 34, 861 oz-in hold torque, 6.3 A/phase
torque curve: http://www.automationdirect.com/stat...stepmotors.pdf

Should have went with the 10mm pitch ballscrews and it would have cut your RPM requirements in half for the required torque.

[archsakas]

Thank you very much for explaining stuff about the stepper motors. I did not know a lot.
But I still fear about vibrations of ballscrew... and I begin to think that this set-up will cause more problems.

[UA_Iron]

You want to ensure that your ballscrew is perfectly aligned with your guides, use an indicator on one of the guides or tables and probe the ballscrew to ensure it is aligned.

As far as RPM and whipping go, you are probably more than safe staying under 1000RPM. You can always use a calculator to determine the max RPM for the screw though.