Dynomotion

Group: DynoMotion Message: 9498 From: Hugh Sontag Date: 5/15/2014
Subject: SnapAmp faulting, test boards
Hi Tom,

I put the test boards in our system and tried running our X motors by themselves, and the motors ran fine.

Then I attached the Y axis motor to the second SnapAmp as before, with no power or ground connections, and the faulting started once the X0 and X1 axes were enabled. They weren't even commanded to move, just enabled, and the Y axis wasn't even enabled.

So the test boards have the same symptoms as our set of boards.

The system did work in the past, with unshielded 14-gauge wires for the motors. On a lark, we decided to disconnect the shields on the motor wires from earth ground.

I was surprised, but after doing that, the faulting did not occur.

Then I swapped the test boards for our set of boards and our set of boards worked as well. I've been running the system for a day, doing software development, not cutting parts, and so far, it's working fine.

I'd like to keep the test boards for a couple more days, just to be sure that things are A-OK, but it looks like the "fix" is to not have shields on the motor wires connected to ground. I should emphasize that the *only* connection to those shields was to earth ground, and that the faulting occurred *with and without* a ground connection for the +5V supply.

The wires still have shields, but now the shields aren't connected to anything.

I am not aware of an easy explanation for this behavior. In my experience, the parasitic capacitance of the shield should have a negligible effect on the high-power FET outputs to the motors. Perhaps the current-sensing A/Ds are negatively affected for some reason.

Hugh
Group: DynoMotion Message: 9502 From: studleylee Date: 5/18/2014
Subject: Re: SnapAmp faulting, test boards
Hi Hugh,
Just throwing my insight in here: Grounds and coupled noise can be very trying.  There's a pdf I put in the files section from a combat robot site that has some good insights on noise and shielding.

https://groups.yahoo.com/neo/groups/DynoMotion/files/Dealing%20with%20NOISE%20sources/

 

You probably know alot of this as you seem to be a sharp guy. Some connection impedence must have changed over the time of your good operation and the period it started the weird faulting.


The motor drives are usually PWM driven as rectangular waveforms of adjust able duty cycle, are rich in noise from RF being the  fundamentals and higher harmonics. The sharpness of the "squared" edges of the waveforms are the sum of the fundamental( sine ) and many higher harmonics that added together form the sharper edges ( vs the just the fundamental sine ).


This being said: this creates tons of noise radiated into the shielding. This needs to be shunted away from the control electronics and its ground. That's why many people recommend a "start" ground point vs a daisy chain

like ground scheme. Noise coupled into a ground that is common but of differing impedance of the controls and limit sensing can move the external signals relative feeding back to the control's 0v0(gnd) and VCC. These can be fast spikes and so the caps in the control supplies keep the control happy, but the external world seems to be jumping around like and earthquake signal wise. So the thresholds of high or low get hit.  


Frustratingly its usually a case per case situation with a few rules of thumb learned. If you are having weird faults or signals: Play with the grounds and shielding connections. I should have thought of mentioning that

awhile back, but it didn't click then as an issue since it had worked for a time.


Hope that pdf was helpful in some way.

Lee Studley
Group: DynoMotion Message: 9504 From: studleylee Date: 5/19/2014
Subject: Re: SnapAmp faulting, test boards
"start" ground point vs a daisy chain-like ground scheme

Should have been "Star" :: "star" ground point vs a daisy chain-like ground scheme.
Was tired when I typed last night.