Linux Mint - PlasmaC upgrade confusion

well try 100 and 100...

PCW wrote: well try 100 and 100...

That works!

PCW wrote: If you still get the error on joint 3 it may be that joint 3 is being asked to do something it cannot.
That is, the commanded positions violate the acceleration or speed constraints of the
stepgen.

That is EXACTLY what I think is going on here. We both only get this error on Z and only on torch retract (which causes some offset issues on my machine at least). See my latest post under the Feature Request thread: here

Seems like there's something going on in the code that controls the Z movements. I'm sure Phill will get it sorted out eventually.

OK that suggests a glitch in the commanded position, maybe when changing
modes from THC to direct Z or vice versa.

This should be pretty easy to plot by plotting commanded position, feedback position
and f-error in halscope single trace mode triggered by f-errored

I'm not very good with hal scope - are the large spikes in commanded a smoking gun?

I couldn't find the error signal.

These are what I would look at:
joint.3.f-error
joint.3.motor-pos-cmd
joint.3.motor-pos-fb

I was trying to catch the bit where it was tripping - I might need to read the halscope documentation, but the buffer seems quite small. Any use?

You can capture it with normal trigger mode perhaps trigger on "joint.3.f-errored"

(assuming you set the following error limits small enough that you trigger a following error)

You could also set the to trigger on "joint.3.f-error" though I'm not sure whether the
large error is positive or negative. Make sure the scale for joint.3.f-error is at least
.1mm/ division, your first plot has the scale too small so tiny glitches are magnified
The second plot shows a 60 u error (.06 mm) which would have triggered a FE
with the 0.05 ferror limit

That 60 micron spike does seem to be the culprit for both the conversational circle and my scribe test code.

I'm going to stop for the day now, but I'm tempted to stick with the 2ms IPO - was it correct to have halved the P value?

Thanks for the Sunday support!