PDA

View Full Version : One axis looses steps when going fast under load - change accel setting


Greg J
Fri 28 November 2008, 09:24
Gerald or Mike or whomever,

I’m experiencing some random “missed” steps on the X-axis. The part is the 3D leaf that is a standard cut file in Vectric’s Cut3D. I was cutting a part in Ash when the X-axis missed a step in the middle of the part on the finish cut. The next missed step was right at the start on a piece of Walnut, also on the X-axis. Walnut and Ash are expensive test subjects, so I started doing “dry” runs and everything ran great. I next ran the same program on cheap pine and the parts came out great. Since the pine parts came out fine, I decided to lower my feed rates from 160 ipm to 100 ipm and try a part in Walnut.

Right at the start of the program, the X-axis missed / jumped. . I heard a strange motor/gear/whatever noise, but stopped the cut file in time to save the Walnut. I re-winded the cut file and started again. This time the Walnut part came out perfect. Nice finish, etc.

I’ve noticed that all 3 times, the missed step is in the positive X direction. So maybe on the first reversal of X direction, the motor cannot catch up. The X-axis is off by about ˝ to 1 inch in travel.

I’ve checked all mechanical parts (loose pinion, motor, swing bracket) and everything is good. I’ve looked at the speed-torque characteristics (PK296A1A-SG7.2 / 35 tooth pinion) and calculate 62.8 lbs of force per motor (more than enough). My acceleration setting in Mach is set at 25.

Can either of you two suggest area’s I should be looking at? Mechanical? … electrical? …. Software?

Gerald D
Fri 28 November 2008, 09:42
Without any work on the table, you can hold a gantry (or car) back by hand to feel how much force it could generate. Try this for a simple rectangle cutting program, or get Karen to push the cursor keys while you hold it. You should detect a dud motor, dud drive, or bad motor tuning with this test. No harm in holding a car/gantry until the motor/s loose steps.

Could you "rotate" the part 90 degrees so that the y-axis does the work that the x-axis was doing before - ie. see if the problem transfers to the y-motor, drive, accel setting etc.?

Greg J
Fri 28 November 2008, 10:23
Thanks Gerald,

I have tried holding the Gantry while dry cutting. Those two X axis motors have allot more force than I can stop.

I'll try transferring the work to the Y axis. Simple in Cut3D.

Richards
Fri 28 November 2008, 10:38
Greg,

Also, check your acceleration. Stepper motors are picky about a lot of things. Too slow acceleration is just as bad as too fast acceleration. The motors have a "harmonic" region where they have trouble. If acceleration is set improperly, that "harmonic" region can cause missed steps.

One of the problems with stepper motors on a CNC machine is that one axis might be required to move very slowly when the the other axes are moving more rapidly. Slow movement can translate into speeds in the "harmonic" region. Sometimes reprogramming the part to run faster or slower can help.

Most stepper motors are used to position an axis without requiring other axes to be synchronized, in other words, a command is given to move two or more axes to end up at point X, Y, Z, A, B. Each axis moves independent of the other axes. As long as the final destination is reached, the move is successful.

On a CNC machine, it is expected that all axes are co-ordinated. That can cause one or more axis to run in its "harmonic" zone and problems can be caused.

Experiment a little to see if faster speeds or slower speeds help.

For example, a move such as G01 X96 Y1 F450, might cause problems with the Y-Axis, since it is only moving 1/96th as fast as the X axis.

Gerald D
Fri 28 November 2008, 11:08
I see that Mike is also suspicious of your accel settings . . . . from this thread (http://www.mechmate.com/forums/showthread.php?t=875), you should have an accel setting of 20 to 25 inch/sec/sec.

If you aircut that particular tough file, can you induce the glitch by holding the problem axis back?

Greg J
Fri 28 November 2008, 12:58
Thanks Guys,

I made a video and have posted on Youtube. The video is poor quality, but I think the sound is more important. I was jogging the X axis at 300 ipm. My accel is currently set for 25 inch/sec/sec.

http://www.youtube.com/watch?v=JrXoq0mA8A4

Greg J
Fri 28 November 2008, 17:09
Gerald / Mike,

I've lowered the accel. from 25 to 20 in/sec/sec and it seems to have made a difference. With accel set at 20 in/sec/sec, I haven't been able to reproduce the bad performance.

Always something to learn. Now, back to work. :)

Greg J
Fri 28 November 2008, 20:15
Here's a video cutting Walnut after reducing the accel. from 25 to 20 in/sec/sec. Cut quality is great and working fine.

The video quality is very poor (using digital camera) so unless you don't have anything better to do on a Friday night (my wife says I need to get a life :)) view at your own risk of wasting time. I'll get a real digital camera for Christmas.

http://www.youtube.com/watch?v=BcZu9Ti4sXM

Gerald D
Fri 28 November 2008, 20:25
Glad to hear (not actually - don't have sound on my PC) that your problem is solved for now. Let's leave it for a few days before copying this to the troubleshooting thread.

Greg J
Sun 07 December 2008, 18:11
Gerald,

Lowering the Accel. from 25 to 20 (in/s/s) has made all the difference. The blue beast has not skipped a step since. :)

Gerald D
Sun 07 December 2008, 21:55
Glad to hear you have something reliable and predictable now.