Losing counts during quick acceleration on HBL2350

Hello everyone,

I am using an HBL2350 controller to drive a motor that is rotating a turret. I wrote a script to find the z pulse of the encoder by using the closed loop count position mode to increment count by count until the z pulse is found. When I drive the turret couterclockwise, everything works fine, but when I drive it CW, the turret occasionally jerks. I think this issue has something to do with the mechanical assembly of the turret, but this jerking action is always accompanied by a lot of lost counts, around 2000 counts lost in what should be 14,400 counts per revolution. I don't think the encoder is slipping relative to the turret because when I stop the turret on the z pulse, it is always in the same position. If the encoder were slipping, I would expect to see the location of the z pulse change. I probed the encoder signal going into the controller, and it looks fine, even during this period of rapid acceleration. I measured the maximum pulse frequency to be around 1kHz, which is well within the spec for the controller, so it isn't that the turret is moving too fast for the controller. The only other thing I can think of is that, for some reason, maybe the controller has difficulty capturing very rapid acceleration. According the data I captured with Roborun, these jerking motions are accompanied by accelerations of around 20,000RPM/s. Could it be that the controller has trouble accurately capturing counts for this kind of acceleration?

TL;DR: Is it possible for the controller to lose counts during periods of very high acceleration (20,000RPM/s)?

Thanks for your help.

Have you changed your min and max value counts in your encoder settings?

I haven't edited the min and max counts, but these counts are being dropped when the count is still well within the min and max range. I usually reset the counts to zero before I run the motor.

Make both motors are connected to controller, go to the run tab of Roborun+.

Set up 2 channels in your lower right hand area. Hall counter 1 and hall counter 2.
The number should be zero for now.
Slowly rotate the shaft of each motor so that you spin the shaft 1 full revolution.

Does the hall counter on both channels equal each other?

If you set up 2 more channels, this time for encoder 1 and 2 counters, and do the same, are they equal to each other as well?

It may be possible that your encoder is receiving noise on the lines. Move your encoder wires as far away from your motor as possible. Also you can try shielding your wires as well to see if this makes a difference.

Thank you for the responses. I have some new information. I turned the turret manually, in open loop mode, both clockwise and counterclockwise. When I turned it CCW, the number of counts for one revolution was 14,400, as expected, but when I turned one revolution clockwise, the count was around 12000. So it seems like counts are being dropped in the CW direction, which must have been causing the jerking motion I was observing since the lost counts would register as loop error. Given this new info, it seems unlikely that signal noise is the issue just because I can't think of a reason why it would only effect one direction of rotation, but I'll try shielding the wires. I also probed the quadrature signal coming into the controller in both directions, and it looked fine, but it is difficult to get a good look at the signal when the counts are being lost. I easily could have missed something.

The problem seems to have been resolved. The second encoder input appears to be faulty. When I plug into the first encoder input, the count is fine in both directions. When I plug into the second encoder input, the counts are being dropped in one direction. Thank you for your feedback.