Nan - voltage

There is something I have read somewhere about a bug generating a nan in a NML message. Maybe it was somewhere here. read on a bit.
forum.linuxcnc.org/41-guis/36920-labview...cnc?start=100#182307

dosent bother me what is displays it could be -500,000,000 if it still performs as expected its fine


okay, so lowpass = 0 results in, 0 instances of nan


but lowpass (for me) stops crazy reactions to tiny variances in voltage.

whata i do?


machine at the moment, if it dosent load with a nan, runs PERFECT.

but it'd be nice if i could open the machine, and every time 100% be confindent that someone else running the machine wont run into a Nan.

Tesremos wrote: but lowpass (for me) stops crazy reactions to tiny variances in voltage.

Have you tried with a higher lowpass frequency?

You can change the value using setp while it is running .

rodw wrote: There is something I have read somewhere about a bug generating a nan in a NML message. Maybe it was somewhere here. read on a bit.
forum.linuxcnc.org/41-guis/36920-labview...cnc?start=100#182307

The test comps I posted ensure all inputs were not nan and yet the output still created a nan.

The thing is I cannot get it misbehave here

Actually, didn't the low pass code just change? Maybe there is a bug.
August last year there was discussion about an update. Did it go into the 2.8 release?

rodw wrote: Actually, didn't the low pass code just change? Maybe there is a bug.
August last year there was discussion about an update. Did it go into the 2.8 release?

plasmacs lowpass is internal. It was based on the lowpass comp but changed so it used the cutoff frequency rather than the gain.

So maybe it has the same bug...

rodw wrote: Actually, didn't the low pass code just change? Maybe there is a bug.
August last year there was discussion about an update. Did it go into the 2.8 release?

phill, emailed some stuff from the developers mailing list.

BUt if an NML message was bad, I wonder if that could impact at the display only?

rodw wrote:

rodw wrote: Actually, didn't the low pass code just change? Maybe there is a bug.
August last year there was discussion about an update. Did it go into the 2.8 release?

phill, emailed some stuff from the developers mailing list.

BUt if an NML message was bad, I wonder if that could impact at the display only?

Thanks Rod, I got it.

The display is a direct read of the calculated voltage HAL pin.

I wish I could get it to happen here...

Tesremos wrote: dosent bother me what is displays it could be -500,000,000 if it still performs as expected its fine

okay, so lowpass = 0 results in, 0 instances of nan

but lowpass (for me) stops crazy reactions to tiny variances in voltage.

whata i do?

machine at the moment, if it dosent load with a nan, runs PERFECT.

but it'd be nice if i could open the machine, and every time 100% be confindent that someone else running the machine wont run into a Nan.

Another option that may be worth trying is the stock LinuxCNC lowpass component:
linuxcnc.org/docs/2.8/html/man/man9/lowpass.9.html
You could stick it between your encoder output and plasmac.arc-voltage-in in the connections.hal file

I'm a bit confused about the lowpass value of 25.

The PlasmaC manual says that this value is measured in Hz. So, at a value of 25Hz, only frequencies below 25Hz are allowed to pass. The THCAD works in the ~140000-1000000Hz range. I'm not sure how the voltage measurement would work at all with everything above 25Hz being ignored?

I've not had any coffee yet, getting ready to drive to work. Maybe my head's a pretzel, it's 05:06am here.

EDIT: Fixed the THCAD frequency range based off of my THCAD 5 values.
EDIT2: I drank coffee and fixed the frequencies for real this time.