help with stepper motor tuning

[Dale gribble]

Hi

I need help tuning my motors, I am running a dell p4 2.8ghz. Installing mach
went very well on a fresh install of win xp. The driver test had very small
spikes. my set up is running routout 10amp driver cards microstepping 3200
steps / rev. these are running 12nm (1700oz/in) motors. I am using a C11 BB
from cnc4pc. I've set mach to run at 45000khz and with a lead screw pitch of
5mm max speed is 4218 mm/m it runs nice at 4000 mm/m (157 in/m ) no problems
but sound rough at 400 rpm and 1400 mm/m which sometimes stall the motor's.
I then moved onto the x axis which has the same setup but has a 16mm lead
screw which ran happily at 12500 mm/m/m (492 in/m )but again sounded rough
around 500 (19in/m) and 1500 mm/m (59in/m) and if held at these speeds would
stall.

My driver cards needs a min of 1.0us on step
and the bb needs min of 3.0us

I altered the timing in incremental steps to 5.0us and 3.0u seems about the
best but the motors are still rough at the above feedrates.
The driver card asks to connect step and direction and also a terminal
return to pin 25 of the pc printer port. I have connected these to the c11
BB com and move the jumper so these are ground. Is this my problem? I have
not earthed anything in my control box as I am unsure of what to earth and
where to. I've uploaded a photo here

http://farm3.static.flickr.com/2798/...f94a8505_o.jpg

I am very weak when it comes to electronics. I have one side of the BB
powered from the pc usb cable and the outputs side from a switched power
wall plug @5v 2amp. I tried switching this with a regulated wall plug but
made no difference. I did remove the terminal returns and this had a big difference, motors would only just move in one direction. I am using shielded
cable on all motors.

Also tried reducing driver amps from 6 down to 4 but this did not make any difference.

Any help or wisdom you can shed on this matter would be gratefully received.

Kind regards

Dale

[ger21]

If it's rough at specific rpm's, it's a very good chance that it's resonance. An easy way to get rid of it is with dampers. See this thread.
http://www.cnczone.com/forums/showth...639#post256639

[Dale gribble]

Thanks for the reply, I've checked out the thread and I am currently making a rattler for trials. I'll post results.

Regards

Dale

[Dale gribble]

Just an update,I finished making the rattler / balancer and these things do work, I couldn't get my motor to stall, but from a range of 350mm/m (70rpm) to 550mm/m(110rpm) it did sound noisy and with the rattler on it sound like someone was digging up the road outside with a jack hammer.

I noticed something strange once I had the rattler on I moved the motor by hand and with no power connect to the drive. I could feel it was very lumpy and had to spin past at around the above rpm range. It's fine if I turn it slowly and if I can get past the lumpy speed it's OK. Does this look like I have a broken motor ? I have checked the other three motor by hand and they are smooth.

Any help or wisdom on this matter would be gratefully received

Alan

[wildwestpat]

Hi Dale

There are a number of contributaory causes to stepper resonance some mechanical and other electrical or electronic.

Starting with your bread board construction of your controller electronics. I am going to number the points to help:-

1. Use a common earth point and ground every thing to it by short thick wires. Include the frame work of your machine and the safety earth from the public mains supply. Think of this as a star point to which all earth wired are connected.

2. Connect the computer case to the star in 1. above with a thick wire. The LPT 25 way cable has a screen that will probably be connected to the chassis of the computer and hence to the zero volts of the mother board.

3. Use shielded cable for the limit switches and earth the cable screen to the breakout board zero volts which in turn has been grounded to the star earth point. Don't earth at the machine end and use twin core cable so that the switch contacts are brought back as fully screened to the breakout board. This will avoid most of the problems with induced voltage from the high motor currents.

4. Place capacitors accross the amplifier feeds as close as possible to each amplifier. The leads you have from the rectifier / capacitor are too long and need decoupling local to each amp. Suggest 100mfd electrolytic a 0.01 polyester and a 1000pf ceramic between the positive and negative power supply wires as close as practical to each amplifier. The voltage of the capacitors should be rated for more than the power supply voltage.

5. Try and use half step or micro step as this will improve the resonance stability.

6. The motors need to be attached to the mill mechanics for all testing after the initial do the motors spin.

7. It is stating the obvious but attaching an unbalanced load to the motor will put the bearings under stress and the rattler action can be very vicious as the weights move hence the noise. The rattler technology is that used to smooth out crank torsion in internal combustion engines and I would suggest that other than for the final tweak rattlers only be used after all the other sources of resonance have been addressed. For example are the gibs adequately tight - consider adding drag dampers if using low friction trolley type axis carts. Are the screws mounted correctly at the drive end? Are the screws coupled to the motors with a ridgid coupling - a coupling with some elasticity can help? Is the motor coupling concentric with the screw shaft being driven? Can the motor mount resonate with respect to the screw? You get the drift you need to look at each part of the drive chain to see that it is not acting as a contributer to resonance at any speed.

8. Stepper motors have a coging feel when rotated by hand in the power off condition. Test for play in the bearings to see if they are shot as a result of the rattler. Stepper motors can exhibit alsorts of resonances if spun under powere with no load. Beaware that it is not a good idea to open stepper motors unless you know how to prevent the loss of magnetism in the type you are dealing with. Some are ok others need special magnetic shunts.

Hope this helps.

Pat

[Dale gribble]

Hi, pat

Thanks for your insight, on the mechanical side I've got linear rails and zero backlash ballscrews connected to one fixed bearing and one floating end. The motors are connected via a flexible connector. I took all the motors off there mounting just to check if the resonance was still there it was. so I think I can rule out mechanical.

Next I'll add the earth's as you have mentioned. Do I add an earth to the negative side of my drivers power supply? I have shielded cable on all stepper motors, but have just twisted pairs cable for limit switches, I have routed these on the opposite side on my machine from my motor cables.

I am very weak with electronic knowledge and am struggling with understanding fully point 4

4. Place capacitors across the amplifier feeds as close as possible to each amplifier. The leads you have from the rectifier / capacitor are too long and need decoupling local to each amp. Suggest 100mfd electrolytic a 0.01 polyester and a 1000pf ceramic between the positive and negative power supply wires as close as practical to each amplifier. The voltage of the capacitors should be rated for more than the power supply voltage.

Do you mean bridge the negative and positive with a capacitor close to the driver power input? Also I am unsure on what decoupling each amp means.

They say a picture paints a thousand words, would it be possible to do a diagram for point 4 above.

Gratefully

Dale

P.S. I think your right with regards to the rattler I think all that energy has damaged the bearing on my z axis.

[Dale gribble]

Hi,

I've read up a bit more on wiki, on decoupling and understand it a bit more.

http://en.wikipedia.org/wiki/Decoupling_capacitor

Are these capacitor the right ones?

http://www.maplin.co.uk/module.aspx?moduleno=318

Kind regards

Dale

[wildwestpat]

Hi Dale

To give you my answers to your questions and add some observations:-



Thanks for your insight, on the mechanical side I've got linear rails and zero backlash ballscrews connected to one fixed bearing and one floating end.

This is one of the occasions when some friction helps as it helps dissipate oscillatory power that might build up due to resonances being excited.

The motors are connected via a flexible connector. I took all the motors off there mounting just to check if the resonance was still there it was. so I think I can rule out mechanical.

The flexible connectors if of the Lovejoy type with an elastic middle element are also going to help kill resonances as power will be lost in the elastic element. The middle element can be changed in these couplings for harder or softer material. Taking all the load off a stepper motor will invite resonances and it is possible to get the acceleration slope so high that the motor will just buzz and not rotate. They need some load to give inertia against which to work. Use a fly wheel.


Next I'll add the earth's as you have mentioned. Do I add an earth to the negative side of my drivers power supply? I have shielded cable on all stepper motors, but have just twisted pairs cable for limit switches, I have routed these on the opposite side on my machine from my motor cables.

Yes all earth wires go to that common star point. Good to keep the motor wires separated from the switch wires. However I would recommend using twin screened wire rather than plain twisted pairs. Take the screens to the common earth point. You need to check that the wire you have on the motors is OK for the current. Normally screened multi core cable would not be flexible enough to stand the constant bending need for the axis drives. The braid can fracture and the fractured ends penetrate the cores causing a short circuit. Not good for the drive electronics so I would recommend using flexible four core cable and it is worth spending a bit extra to get the special stuff that is OK with constant bending. It is good that you have fused the drives as this may help protect against accidental short circuits.


I am very weak with electronic knowledge and am struggling with understanding fully point 4

OK we all had to start from some where. The capacitor trio is to decouple the length of wire between the big capacitor used in your power supply and the drive electronics. It is possible that the drive electronics have such decoupling built into their PCB. Due to the long length of wire and the high pulse currents used to drive steppers the 9 inches of wire is enough to cause problems. If you look at a commercial packaged driver these are recommended and are regarded as good practice. The aim is to connect the capacitors directly across the power input to the electronics with as short a leads as possible. i.e. don't wire back to the star earth but to the 0 volts of the drive.

Sorry to have lapsed into electronic jargon. Decoupling just means placing capacitors in this case with as short a leads as practical across the source - in this case the electronics driving the stepper. In more viccious cases an inductor capacitor network might be used but should not be necessary here as we only have relatively low frequencies not radio stuff.

Hope this helps as I am reluctant to start drawing as the mose is a no no as my RSI is playing up.

If you need more ask and I will try and help.

PS on dampers. Placing a finger on the parts is a good thing to do provided it does not get choped off! Pressing on the part can help identify where a bit more friction can help. I use an old doctors stethescope to listen to bearings and slides to see if they are having a jig.

Regards

Pat

[Dale gribble]

Thanks Pat for explaining it for lay people like myself.

I'll get the capacitors and screened cable on order and then re-design my control box making the power leads as short as possible. With regard to the screened motor cable, it came off an industrial cnc using servo motors, it wasn't cheap but have seen on forums many people having noise issues.

Kind regards

Dale

[Dale gribble]

Hi,

I've had a Brain fart and have designed a flywheel with some balancer action as well, but with no rattle. I doubt I'm the first to think of the design but thought I would upload jpegs for some constructive criticism. basically the body is made from nylon or PTFE and steel weights added to the outside of the body these are glued in eliminating play. The thin 1mm webs allow some movement to dampen resonance.

Your thoughts on the design are welcome

Regards

Dale

[wildwestpat]

Hi Dale

Glad you have sourced that special cable for the motor feeds it can be difficult to get and using one of the multi core screened cable used for general electronics is to be avoided. Yes screening is good as it will help reduce interference levels and might also enable you to listen to the radio whilst making chips. My own choice is to use a cable guide that also acts as an eletrical screen ever since I put a heavy rotary table down with out thinking and cut a cable blowing the driver. An expensive mistake as I did not spot the problem until I had cut too much off the work being done at the time. You might like to ensure your cable runs are to the back of the machine and thus avoid the temptation to rest a heavy item on the same bench as the mill - assuming you are CNCing a bench job.

Good luck with the mods come back if there are further problems as there is a wealth of expertise on this forum and a lot of willing contributers.

Regards

Pat

[wildwestpat]

Hi Dale

No comment on the flywheel other than it looks sexy. Might be easy to get caught in the fingers but then the stepper should be shielded as they get too hot to touch. In my book it is best to reserve flywheels and their derivatives for the final stage of tune up when hopefully they won't be needed. At least that is my view! Others are ardent disciples of active balance shifters - I guess there is no one size fits all solution.

Regards

Pat

PS There has been a recently active thread on rattlers which you might like to read just incase you need to quell the odd troubling resonance problem.