Dual PID loops with motor encoder + scale encoder per axis

And yet, with zero command, you can rotate the motor by hand, right?

Craig

PCW wrote:

That's what I meant with:

"the inherent velocity feedback from the BEMF makes it so"

You do have velocity feedback (and damping) in a open loop PWM system
The gain in the loop is inversely proportional to the motor resistance.

If you slow the motor so its BEMF is less the the applied voltage, the current increases.
If you speed the motor so its BEMF is greater than the applied voltage, the current reverses
to oppose the motors motion.



 

A velocity loop maintains commanded velocity which can be anything including zero velocity. The drive responds to disturbances to maintain commanded velocity including zero velocity. The actual RPM makes no difference. Turning by hand when commanded velocity is zero, constitutes a load disturbance. IOW the velocity loop won't allow it.

Craig

Friday night and I have a date with a certain cask ale. This is fascinating stuff. Back in the am.

Cheers,

Craig

A velocity loop maintains commanded velocity which can be anything including zero velocity. The drive responds to disturbances to maintain commanded velocity including zero velocity. The actual RPM makes no difference. Turning by hand when commanded velocity is zero, constitutes a load disturbance. IOW the velocity loop won't allow it.

Actually it will, it will just drag with a velocity and loop gain dependent torque.
This is precisely what a shorted PM motor will do (Same as PWM with 0V applied)
This is because of the inherent velocity feedback. It will have lower equivalent gain
than a velocity mode servo but its still chiefly velocity mode.

Tinine wrote:

And yet, with zero command, you can rotate the motor by hand, right?

Correct.  0 PWM command feels no different than if the drive is unplugged.

It does occur to me that this could be an artifact of how the input to the drive works.  It's offset PWM on a differential line driver.  So 50% duty cycle = 0 output.  0% = -100% output, 100% = 100%.  However it does NOT have a separate enable signal, it uses the lack of PWM to shut itself off, which I assume is A=B on the differential inputs.  It only occurs to me now that maybe it interprets 50% duty as "off" as well.

0 PWM should be the same as 0V which means the motor should resist applied
motion (because its acting as a generator into a shorted load). This may not be apparent
for very slow motions but if you spin the motor, it should resist.

PCW wrote:

0 PWM should be the same as 0V which means the motor should resist applied
motion (because its acting as a generator into a shorted load). This may not be apparent
for very slow motions but if you spin the motor, it should resist.

OK, I just re-tested and I was mistaken.  I hooked the X motor belt back up to the table, which still has the manual handle.  It's much more obvious with trying to move it fast via the handle.  I can confirm that it does resist as if the motor is shorted when the drive is enabled and has a 0 PWM command, just as you said.  When I disable the output (which disables the diff line driver output), it goes slack again just like the motor is unplugged.
 

I'm sure it's obvious, but maybe it's worth stating that while it does resist being turned in the 0 value PWM open loop state, it's not in the same way that a closed loop system would fight back. It's more of a modest drag/braking action.

I would say that its the same (resistance proportional to K * velocity error),
but the loop gain (K) is much lower than a typical servo system velocity loop.

The way you can tell that it does not behave like a torque mode drive 
is that in a torque mode drive, the current at a given speed would not change
based on motor load.

BTW, you can test torque mode behaviour easily if you have a constant current
bench supply.

PCW wrote:

I would say that its the same (resistance proportional to K * velocity error),
but the loop gain (K) is much lower than a typical servo system velocity loop.
 

Fair point, it resists but it can only resist so much.

BTW, you can test torque mode behaviour easily if you have a constant current
bench supply.

I do have a bench supply that can do CC, I think I will try that this weekend just to gain some more intuitive understanding.

thanks,
Steve