Machine locks up while travelling at high speed - Tuning the motor speeds and accels
 

I finally got around to doing some rapid traverse tests today. My machine has Vampire Geckos, PK296A2A-SG7.2 motors wired half coil with 35 tooth pinions.

I started with the pulse rate set at 35,000. From a dead stop, she traveled six feet in an average of 5.825 seconds which is 12.36 inches per second. Then I increased the pulse rate to 45,000 and retuned the motors. Under the same conditions, she averaged 4.64 seconds which is 15.51 inches per second.

Something strange happened during the tests though. Sometimes, she would travel 66 inches and then lock up. If she didn't lock up on the first pass, then on the return pass, she would lock up in the first 6-12 inches. I gradually slowed the acceleration rate down and when I reached 25, she quit locking up. Now, I wish I had waited to special order the gantry tubes in a thinner wall size. I used 1/8 inch.

I traversed at least a dozen more times without a lockup. However, I refuse to believe that it was still accelerating after 66 inches and I'm wondering if she was binding up or something. I haven't had a chance to set the Y axis square to the X axis and when I turn her on she clunks so maybe I need to drop the motor and spin the pinion until she starts more smoothly. Maybe the clunk gets her out of square and creates a high torque condition. Maybe the rails are slightly wider at the 66 inch point on my machine and the additional torque required to drive the gantry assembly through that point is enough to stall the motors out. Anybody else have a theory?

You make no mention of the Velocity setting in the motor tuning. This is the place where you set the max speed. The computer frequency has to be adequate, but shouldn't be the limitation on the speed.

That start-up clunk is never totally eliminated, but it can be reduced by squaring and engaging the pinion to rack at a different pair of teeth.

Gerald,

I left the velocity setting on the highest possible which was 1030. I think that's inches per minute. I believe the computer is a 2.8 Ghz Pentium with 1 Gig of RAM. After I reset to the highest pulse rate, I ran the driver test like Art recommends and it appeared to be fairly stable. About as stable as his computer was in the tutorial. If the gantry had stopped at random locations, I would have suspected the pulse rate was too high but it kept stopping at the same two locations. Tomorrow, I think I'll raise the acceleration rate to 50 and try the test from the other end of the table. If it stops at the same location on the gantry, that would seem to indicate a binding/high torque condition at that point. However, if it traverses approximately 66 inches and locks, that would seem to indicate a problem with the computer keeping up. I'll let everyone know what I discover.

Apparently, the higher the set frequency of the pulse engine on the PC, the more chance for problems. The recommendation appears to be to set the PC frequency as low as possible. That means you have to calculate the frequency needed to move at your desired top speed. What is your calculated frequency for say 15 inches/sec? (By the way, the speeds you are mentioning already sound quite respectable :))

On the acceleration and effect of mass, suggest you do a little test; clamp 20 pounds of iron on top of the gantry and see how much difference that makes. That will give an indication of what subtracting 20 pounds will do.

The pulse engine of Mach is not very smooth. It gets worse with bad PC's, or with multiple axes running at the same time. One day you will be able to plug a "Smoothstepper" between your PC and your BOB, but it is still under Beta test.