andyf Posted June 20, 2012 Posted June 20, 2012 We have a Kollmorgan DC brushless servo motor being controlled by PPMAC ACC24E3. When executing jogs, we occasionally see the motor hit its Fatal FE limit of 2000mu. I ran the plot utility to see what is going on and found that the Commanded Position and Actual Position are jumping 8 million units when this occurs (see attached plot). At first I thought it might be interference between motor and encoder cables. However, even with the motor killed, if I put ActPos in the watch window, I see it jumping between the correct position, and other invalid positions. I also watched the interrupt captured encoder values (PhaseCapt and ServoCapt) and they do not jitter. Also, we don't see this happening all the time. Sometimes we run the motor fine for hours, and then all of sudden it gets stuck with this problem. This is the first time I have encountered this problem. Any help would be greatly appreciated.
Sina.Sattari Posted June 20, 2012 Posted June 20, 2012 Andyf, I think the problem is related to noise on backplane data bus. Can you try moving the CPU to a different position in the rack as a test? Try placing the CPU in the middle of the rack and post your findings.
andyf Posted June 20, 2012 Author Posted June 20, 2012 Sina, The problem seems to occur around the point where PMAC thinks the encoder zero position is (i.e. position at reset). We were able to manually move the motor away from that point, and now it is working OK. I also noticed that indefinite jogs move through the point OK, but point to point moves fail. I will try your suggestion and let you know... Andy
andyf Posted June 21, 2012 Author Posted June 21, 2012 Sina, Update: We do in fact see the value of ServoCapt jittering. I did not see this in our testing yesterday, but today I did. We found this problem on Chan[0] originally, and now today we tried Chan[1] and it seems worse (more obvious). We are trying the other two channels now. Andy
andyf Posted June 21, 2012 Author Posted June 21, 2012 Sina, Problem solved. After moving to Chan[2] the problem went away. This indicates that the primary card on the Acc24e3 (Chan 0 and 1) is bad, but the secondary card (Chan 2 and 3) is good. Time to check our warranty... Andy
Sina.Sattari Posted June 21, 2012 Posted June 21, 2012 Andyf, I'm still curious to see the result of CPU position change test. If you do have time, please do this test. Regards,
andyf Posted June 22, 2012 Author Posted June 22, 2012 Sina, I have it on my list to perform that test. Unfortunately we are behind on some data gathering due to this issue, and now that things work we need to catch up. I will update this thread as soon as we are at a point where this test can be done... -Andy
andyf Posted June 26, 2012 Author Posted June 26, 2012 Sina, After some time, we started experiencing the problem on Channel 3 (Chan[2]) as well. Although not as bad, it was still causing the occasional FE limit trip. We moved the CPU card as you suggested, and found the problem to be worse. What we see now is jitter on the Motor input when viewing the Motor Setup Interactive Feedback. The Servo captured position looks OK, but the Motor Input is erratic. If you are available for a webex today or tomorrow that would be great. -Andy
leeho Posted June 27, 2012 Posted June 27, 2012 Here is a screen of the interactive feedback with motor at stand still, concerning the issue that Andy spoke of. We have seen the erratic behavior with multiple motors now on the same ACC24E3 card on different channels.
andyf Posted June 28, 2012 Author Posted June 28, 2012 Sina, We tried moving CPU card next to ACC24E3, and still encountered the problem. In fact, near the end of the day the CPU stopped detecting the ACC24E3 card all together, so we believe the card has gone bad. Coincidently, we received a new PMAC CPU, ACC24E3, and chassis today as part of another order. We placed the old card in the new chassis, and set the DIP switches, but it was still not detected. Then we put the new ACC24E3 in the old chassis and everything works great. I did note that the old card was stamped PROTO. We will be sending it back tomorrow for replacement. Cheers Andy
Recommended Posts