ADC Offset

[leandro.martins]

I'm driving a voice-coil using the Power Brick LV 5A/15A and my goal is to control the force exerted. 

Measuring the current with a multimeter, when it's issued the command

#nout0

a current of -0.15A is measured, but when I look to the Motor[x].IqMeas  the value corresponds to approximately zero amps.

 

I could adjust this by the Motor[x].IaBias, but in the Power Brick LV it says there is an autocalibration routine and the bias should be left in zero

 

If I set some value in Motor[x].IaBias, in a manner to correct the observed bias, is it possible the autocalibration routine interfer in some way and the bias keep changing?

Is there any other problem in setting the bias?

 

Thanks in advance

 

 

[leandro.martins]

On 3/5/2022 at 3:51 PM, leandro.martins said:

If I set some value in Motor[x].IaBias, in a manner to correct the observed bias, is it possible the autocalibration routine interfer in some way and the bias keep changing?

 

I made some tests and this is really happening, is there some way to disable the auto-calibration routine (by software or by hardware)?

 

Thanks again

[Faraday MC - Tony]

Hi Leandro,

 

Try setting Motor[x].CurrentNullPeriod to a negative number (try -1000 to start with). It will measure the current with the motor energised  and remove any offset caused by the output stage.

[leandro.martins]

Thanks Tony,

I will try this approach.

Reading about Motor[x].CurrentNullPeriod in the manual, it seems that is applied only to external amplifiers. I don't know if I'm misinterpreting.

So, I guess that using Motor[x].CurrentNullPeriod we will be applying a correction that complements the corretion of the amplifiers, is that right?

Edited March 15, 2022 by leandro.martins

[Faraday MC - Tony]

The current sense firmware in the Brick attempts to null the offsets automatically so you don't normally need to use the PMAC offsets. This only affects the current feedback, not the output. External amplifiers do not always perform the offset measurement automatically so it needs to be done in the PMAC.

With a Brick there is nothing to stop you applying the PMAC offsets as well. If you use a negative number for CurrentNullPeriod it has the effect of compensating for the small offsets in the output caused by the switching transistor characteristics. It might not work well for you because it could cause the voice coil to move when the test current is turned on. In that case you could calibrate the system with the stage locked and then use fixed values for IaBias and IbBias.