[bbutcher85710]

I have a strange encoder problem. This encoder is on the Z-axis of a Bridgeport milling machine converted to CNC. I am getting what I would describe a drift in the zero position. The servo motor behaves fine with no runaway or other obvious encoder issues.

If I start at Z=0 and move to Z= -0.1 inch, then back to Z=0 repeatedly for 100 repitions with a dial indicator measuring the motion of the spindle, after 100 moves the zero position reads about - 0.006 inch on the dial indicator. The zero continues drifting another 0.006 inches for each 100 cycles. The motion is 20,000 counts per inch so the drift is about 120 counts, so I may be missing one count per each motor reversal. Am I connecting something incorrectly here? Maybe an error on a data sheet?

I have a US digital E5D-500-250-HG encoder mounted on a brush type servo motor. The encoder is a differential output, and it is connected to a Galil DMC-1750 controller through an ICM-1900 interface with about a 6 foot twisted pair ribbon cable. I am only using the A+ (pin 6) and B+ (Pin 10) outputs from the encoder, connected to the ICM-1900 +MAZ (Pin 97) and +MBZ (Pin 99) terminals as recommended by Galil. The encoder has +5 and Ground connected to the correct pins. I have checked for mechanical slipage and end play in bearings, and all are good. The encoder wheel is tight on the motor shaft.

The X and Y axis have similar encoders connected the same way and seem to work correctly. I did have a very similar problem when I connected encoder A+ and A- signals to the ICM-1900 board by mistake previously.

[Jim Dawson]

Why are you not using the -A and -B signals since you have them available? The Galil card would be very happy with those connected and much more noise resistant.

[bbutcher85710]

Why are you not using the -A and -B signals since you have them available? The Galil card would be very happy with those connected and much more noise resistant.

Thanks, I will try that. Should I connect to the Galil -MAZ and -MBZ pins?

[Jim Dawson]

Yes, A+ to +MAZ, A- to -MAZ, B+ to +MBZ, and B- to -MBZ.

Note, when you do this you may have to change your CEZ value. Also, this should double the PPR. I recommend that when you make this change that you mechanically disconnect the servo from the axis, or set your TLZ to a low value, maybe 0.5 or so. Just to make sure the motor doesn't run away when you make this change. Once you are sure the motor is under control, then you can set the TLZ back to normal.

[bbutcher85710]

Yes, A+ to +MAZ, A- to -MAZ, B+ to +MBZ, and B- to -MBZ.

Note, when you do this you may have to change your CEZ value. Also, this should double the PPR. I recommend that when you make this change that you mechanically disconnect the servo from the axis, or set your TLZ to a low value, maybe 0.5 or so. Just to make sure the motor doesn't run away when you make this change. Once you are sure the motor is under control, then you can set the TLZ back to normal.

OK, I understand now that you want me to try differential mode. I will try that. I doubt this is a noise related issue since the error seems quite consistent from one repetition to the next. I am beginning to wonder if perhaps the pinout for the encoder is incorrect. I may need to put a scope on it to verify the waveforms. Thank you for your help, and I will be careful to prevent runaway of the motor. I do have limit switches on it connected to the motor controller to avoid serious damage. With that said, I managed to burn up a servo motor a couple of years ago when it jammed and I did not notice it in time. I have the torque limits set a bit below 10 Amps now to prevent overheating of my Copely motor controllers.

Bob

[Jim Dawson]

Yes, differential quadrature. Always use this where available.

When you are running your test, are you using an absolute position move or incremental? Error can creep into a series of incremental or relative moves.

[bbutcher85710]

I am using absolute position mode through Mach 3. I have used the same method to verify the X and Y axes, and they work correctly. Only the Z axis is causing trouble. I think I have the same encoders on all three axes, I will check that and use differential mode on all if it is available.

[Jim Dawson]

I did see this same thing happen on another machine that was using a Galil card and an A/B encoder, but not running Mach3. There was enough electrical noise introduced that the Z axis was a bit erratic, and only the Z axis. Going to a differential encoder fixed the problem, so maybe it will help in your case also.

[bbutcher85710]

I did connect in differential mode and it seems to work correctly now, except once in a while the z-axis runs away. I think I may have a bad cable or connector on the encoder. I plan to rebuild the controller in a new box with new wiring so maybe that will help.