Welcome to LinuxCNC.org

ArcEye wrote:The symptoms are the same, but the cause is almost certainly different.
This is only going to hit people using the pico-systems ppmc. The two chaps you linked to are using stepper-mm.ini and sim-axis.ini which look like the sample configs. Neither make likely candidates for conversion to a PPMC system, but users do odd things.

I think that you get the "waiting for s.axes" problem any time that the realtime part of a HAL component crashes the realtime thread.
It might be worth adding a diagnostic to check for that cause.

andypugh wrote:OK, I hope that VNR can test this, as I don't have an old enough board to test it here. But,
the problem was quite obvious, the velocity pin did not get created, and then the encoder
module tried to write a value to it.

I did test it with a USC and a UPC board, and it works fine in both cases, so it didn't
break anything (I think). If the lines were moved to the right place to satisfy C syntax,
I don't see how this code could go wrong.

Jon

andypugh wrote:Right, and write a much more informative message. Not exactly sure how you detect a crashed
RT module in a thread, but hopefully there is a simple way to do so.

This particular bug did not cause the driver to crash during initialization, which would
cause LinuxCNC to shut down immediately with a slightly more informative message.
It only caused it to crash on the first servo thread execution.

Jon

I instaled the update and it is working OK with my USC card (by the way is an old card Rev 2.2 2005-9-1 Number 0046)
Thanks Jon for the patch.

VNR wrote:Excellent! Thanks so much for FINDING the cause, and then testing the fix!

The board rev # is not the same as the firmware rev #. But, a huge change was
made at the rev 3.0 board going to a newer FPGA chip, so older boards
couldn't accept newer firmware.

Jon