Always increasing current on a servo motor

Hi all,

Quick reminder I have a 7i43 + 7i52 + 8i20 and a used servo motor, as I stated in my last post I haven't the technical information of the motor.
After several tests I managed to get the following details about it:

• 6 poles
• encoder 2500 x 4

When I'm using the AXIS UI to jog a bit (ie. 5 mm) I can see that the current (as reported through hal meter) is around 0.7 Amps but also I noticed that the current is always increasing when the motor is in stand-bye up to ~5Amps and then AXIS disconnect it.

Current related parameters for the 8i20 are :

• current_maxlim 5.0
• current_minlim -5.0
• current_scalemax 5.0

I don't think that it's normal can you confirm ? If not normal what can/should I do

I'm attaching the .ini and the .hal files (there is a cleaned version that didn't contains all my comments)

if you have that much current and the motor is not spinning or stalled it means you have a commutation problem
if the motor current gradually rises when not moving, that suggests that you are losing/gaining encoder counts
from noise

Hi Peter,

I checked, there is no noise it seems, the rawcount counter (and the non raw one) are not really varying...
It takes from 90 to 150 seconds for the counter to decrement (or increment) of 1 unit.
Looking more closely, the motor does move but not very much but moves still.
Thinking twice it looks like a static error, as if the machine wasn't able to reach the destination.

Also I noticed that there is an angular play, it's easy to rotate a bit the shaft back and forth before feeling the torque of the motor.
Changing the deadband factor by 10x didn't seems to have any influence on this angular play.*

I'd like to remind that my control chain is not the usual one, I have a 7i43 with a custom firmware that I got from Mesa website.

If the current changes, it must be because the PID output has changed
The PID outputs can change because the encoder position changes, the commanded position changes
or the integral error builds up, that's about it. Again if you have 5A of current and the motor is not
working against a load and not spinning, you have a commutation issue, that is the current angle is such that you
are pulling radially on the rotor (so you can have lots of current but no torque) not tangentially as proper commutation
should.

Can you spin the motor equally fast in both directions for unlimited times?

Hello

I'm not sure to understand " that is the current angle is such that you
are pulling radially on the rotor (so you can have lots of current but no torque) not tangentially as proper commutation
should.", can you explain it in a different way please.

Nevertheless, I tried to spin it in both direction (I had to use the motor without load due the small range of the table), maybe 10 times in each direction. I didn't notice any difference.

I started to monitor the emcmot.00.pos-cmd and pid.x.error, I noticed that the error was diminishing very slowly when the current was increasing in absolute value (at what seems to be a plateau in the pos-cmd) I started to wonder if the deadband parameter was not too demanding for my configuration or my motor. When I reduced it to 0.05 the current was not increasing anymore soon after reaching the plateau for the command.
According to the documentation it seems that it should be at least 10 times smaller given the resolution of the encoder (if not more).

I'm attaching the scope of the current before and after reaching the plateau of the command, we can see that the current starts to increase (I'm not very skilled at halscope and I'm not able to capture a trace for a long period).

what is the motor current rating?
can you plot the following error at the same time?

Hello Peter,

I'll do it this afternoon.
In the meantime, I'm still not sure of the meaning of "that is the current angle is such that you
are pulling radially on the rotor (so you can have lots of current but no torque) not tangentially as proper commutation
should".
Can you re-explain ?

Thanks.

Yes you can have lots of current that does not translate to torque if the
magnetic field angle is not 90 degrees to the rotor angle, that is,
if the commutation is not right (so you could have full motor current but 0 torque)

Thank you it's clearer now.
But given the fact that at when jogging with AXIS the motor seems to move correctly at the right speed (even when changing the speed) don't you think it's not a problem magnetic field ?

Hello Peter,

Find attached the new graph with the error added.