axis.5.motor-pos-cmd = -nan

No, it s a standard C-axis config. command comes from motion.

axis.5.motor-pos-cmd => pid => 7i77 analog-out...

C-axis movement is achieved by fiddling with the feedback signal in gearhob.comp.

Henk

#ing out the addf line for gearhob.comp doesn't solve the issue as suspected.

thanks
Henk

Henk wrote: No, it s a standard C-axis config. command comes from motion.
C-axis movement is achieved by fiddling with the feedback signal in gearhob.comp.

OK, and that output only goes to the PID, I assume? (This doesn't matter for investigating this problem, I am just curious about whether your way is better than my way as I do a fair bit of gear hobbing myself).

When does the NaN appear? I wonder if Halscope would be able to track it propogating through the system.

As soon as the gui loads, I check this in Hal configuration and it shows nan. this is before the e-stop is released on the gui and the machine still off. it seems as if it starts up that way.

Henk

As my next step I would comment-out everything related to the axis in HAL and see if the problem still occurs.
You can then run the HAL commands one-by-one in an interactive HAL session (halcmd -kf after LinuxCNC loads) to see if you can work out where the problem starts.

Are you using Trivkins?

Hi Andy
Yes, I'm using trivkins.

Ill try that, thanks

Henk