Hi Michael,
Since the encoder gain inside the feedback loop increased by 60X theoretically the PID gains should be decreased by 60X (all 3) to get the same dynamic response (same Bode plot, band width, and so forth). But I also wonder if other mechanical effects have been introduced in the measurement - backlash or resonances.
You should not be exceeding the Encoder count frequency which is 1 MHz.
Regards
TK
| Group: DynoMotion |
Message: 4723 |
From: Michael Rosenfield |
Date: 5/1/2012 |
| Subject: Re: new encoder |
No backlash, but possibly resonances. I also needed to increase the V, A, and jerk by 60X. It is working reasonably well now, at least in KMotion. However, moves in KMotionCNC are jerky. Jogs are not. Also, when I move the axis in 0.010" steps (using the jog buttons), it moves ok, but when I move it in 0.001" steps, it doesn't move sometimes until I've given it 6 or 8 steps. I have 20 counts of deadband set up, but at 100,000 counts/in, should that be causing this? Where should I be looking, do you think? Thanks, Michael
| Group: DynoMotion |
Message: 4724 |
From: Tom Kerekes |
Date: 5/2/2012 |
| Subject: Re: new encoder |
Hi Michael,
Yes those would all need to be increased by 60X. Did you increase your Z resolution in KMotionCNC | Tool Setup as well?
Are you displaying Encoder position or Commanded Destination? You can look at the Axis Screen in KMotion.exe to see which if any are changing appropriately.
Regards
TK
| Group: DynoMotion |
Message: 4725 |
From: Michael Rosenfield |
Date: 5/2/2012 |
| Subject: Re: new encoder |
Yup, I remembered to change KMotionCNC. I tried displaying both, and I had a dial indicator on the axis, as well.
| Group: DynoMotion |
Message: 4726 |
From: Tom Kerekes |
Date: 5/2/2012 |
| Subject: Re: new encoder |
Well, what were the results?
| Group: DynoMotion |
Message: 4728 |
From: Michael Rosenfield |
Date: 5/2/2012 |
| Subject: Re: new encoder |
They were what I said below - It is working reasonably well now, at least in KMotion. However, moves in KMotionCNC are jerky. Jogs are not. Also, when I move the axis in 0.010" steps (using the jog buttons), it moves ok, but when I move it in 0.001" steps, it doesn't move sometimes until I've given it 6 or 8 steps.
| Group: DynoMotion |
Message: 4729 |
From: Tom Kerekes |
Date: 5/2/2012 |
| Subject: Re: new encoder |
Hi Michael,
There are various possibilities why an axis may not move. The axis may not be commanded to move for some reason. Or the Axis may be commanded to move but the servo doesn't have enough gain to cause the motor/encoder to actually move. Or the encoder may actually move, but mechanically the axis doesn't physically move. It is easy to determine which of these is the case by observing the various screens while commanding the small moves (and reporting the results). With a resolution of 100,000 counts/inch a move of 0.001 inch should cause the commanded destination to advance by 100 encoder counts. Please command steps of 0.001 inches and after each step record:
#1 The KMotionCNC DRO
#2 The Axis Screen Destination
#3 The Axis Screen Position
#4 The dial indicator
Regards
TK
| Group: DynoMotion |
Message: 4734 |
From: Michael Rosenfield |
Date: 5/2/2012 |
| Subject: Re: new encoder |
Tom, I started checking these moves, and it was moving most of the way, (I was making 0.1" moves at the time) and then about 15 seconds later, moving much closer. So I tore the encoder mount apart again, and discovered about 20 degrees of play in the encoder! Guess that system isn't going to work... So I'm going to rebuild that part again, and find a way to get rid of the play, but have more that 1666 counts/inch. Sorry about that. Michael | | | | | | | | | | | | | | | |
