1. Home
  2. ⏣ General Metalworking
  3. Encoder output signal question

Encoder output signal question

I've seem to be having an encoder issue with an old Hurco KMB1 mill. I've had to replace the original 5 wire 200 line BEI encoders with 8 wire 1000 line encoders from EPC, as part of an control board upgrade. The old encoders put out 'A, 'B and 'Z, the new encoder puts out 'A, A, 'B, B, 'Z, and Z. The new encoders are the differential line driver, type , (ie. they are supposed to be able to source or sink 20ma ~ 5V). EPC said these units should electrically interchange with the old BEI.'s... The position display on the control indicate position correctly, but I'm not seeing the "dot" in front of the X, Y or Z position lines on the control. The "Dot" is supposed to signify the servo controller is seeing the marker input. Right now I have 'Z index signal from the encoder tied to the controller "Marker" input lead. I'm thinking about trying the Z output instead to see if this works. The Z output is supposed to be logic HIGH, except when the encoder is sitting at the marker. Anyone know for sure if the "Marker" signal input the Hurco wants is the same as the typical encoder "index" signal?? Without the "Marker" inputs working apparently the machine cannot be "Homed in" correctly so I'm kinda stuck until I get this fixed....

Reply to
oldjag
Loading thread data ...

Give it a whirl. Unless your readout is really odd you can't hurt anything.

Note that you may have to get all three signals lined up, and the BEI nomenclature may differ from the EPC. Some encoders are set up to have the index signal active over most of one cycle of the A and B outputs, which requires gating in the readout to find the "real" index pulse. Others do the gating internally, so that the index pulse is lined up with just one combination of A and B outputs. If both the readout and the encoder do this gating, then you may have to do more swapping than just the Z and 'Z lines.

(I'm probably borrowing trouble. Don't worry yet. First reverse the sense of the index, and if that doesn't work come ask again).

Reply to
Tim Wescott

I swapped the Z and 'Z lines on all of the encoders and now I'm getting the "Dot" in front of the X Y and Z on the position display. So that seems to have been the problem with the encoders. Now if I can figure out why the servo amps don't seem to want to tune correctly I'll be all set.

Reply to
oldjag

Have you actually moved the machine to see if the machine is moving the correct distance? I doubt it will without changing the parameters that deal with ballscrew pitch/encoder ratio

Just changing the encoder would have changed the servo gain considerably and I would expect the motors to be running rough and drawing high current due to the gain being so high.

Good Luck

Dave B

Reply to
Dave B

Dave;

At the same time the encoders were changed I put in an upgraded Servo control board and a new MPU board that allows loading of a Master for either the Hurco conversational, or G code operation. This board combo is sold as an upgrade for the older Hurco's with type B control, (.001 resolution), to allow counting to the values needed for .0001 resolution. It's supposed to be a more or less drop in replacement. In any case the issue at hand now is that with the recommended base pot settings for GAIN, Balance, TAC, SIG, AUX, & CLM the machine vibrates like crazy when the servos are turned on. All axis seem to have the same issue. Backing off the gain to minimum only helps slightly, only backing down the SIG a lot kills the vibration, but it also increases the deadband to the point of being ridiculous. The position display looks to be reading correctly and in the right direction when I move the axis by hand. So I'm still looking for the reason this thing won't tune. The servo controls were dead stable before the boards and encoder were replaced, with excellent circular interpolation. The ballscrews and bearings are tight with less than

0.0005 backlash.
Reply to
oldjag

Going from a 200 line encoder to a 1000 line encoder will give you 5 times as many counts per revolution. It is pretty much the equivalent of setting your tuning 5X higher. Do you know the details of how your system is wired? On a my CNC mill and lathe, the drives have tachometer feedback and the controller gets the encoder feedback.

It sounds like the only 2 tuning pots you have that could help would be gain and signal. If you position reads right the it sounds like you have the problem taken care of in the control. What signal does the drive get for feedback? It sounds like there is still something 5X too high getting into the control algorithm. Some servo drives take the encoder input and have an encoder output that goes to the control. If that's the case then the control might be set right due to your board change but the drive wouldn't be set right.

You could try turning your signal and gain down to where the oscillation stops, then increase one pot until it starts oscillating again, back it off until it becomes stable. Do the same for the other pot (using gain and signal) until you get both pots as high as you can and back them off until it's stable. See if that will get the machine positioning.

Perhaps there is parameters in the control, maybe for P I D control? If so, that would let you tune the system and all you'd need is the signal and gain pots on the drive. If that's the case, then the oscillations would be caused by the control instead of the drive. Anyway, if that is the case, set the P, I, and D values to 1/5th of their current values for a starting point.

Good Luck

RogerN

.
Reply to
RogerN

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.