IRQ affinity on PC Debian Bookworm

I don't, but I ordered a Intel NIC expansion card so I can get the USB to NIC adapter off of the USB. I use the existing on-board NIC for the MESA connection. I am also getting a pair of memory stick so it can run in dual channel mode. Right now, there is just one 4GB stick in the machine.

Last night I started it up and homed the machine and then just left it sit. Later when I went down to check the pendant was disconnected. To get it to reconnect I have to exit LCNC, Unplug and replug the pendant and then start LCNC. When LCNC starts it must start a communication session that doesn't get started when plugging the pendant in. You would think that temporarily pulling the pendant USB plug would reset it, but it doesn't. Might be interesting to put a scope on that.

rodw wrote:

Correct, IRQ (Interrupt Requests) relate to  hardware which linuxcnc isnt. eg its the hardware requesting the kernel to allocate some CPU time.
Back in the DOS days, you would chain a task onto the timer interrupt but now you create a thread (which is just a timer based  process)
from  memory, there is a task called something like rtapiapp which is the servo thread that shows up on a kernel trace. if you use isolcpus=3, the RT thread will be isolated to that core.

I am trying to understand what causes rtapi_app to run on core 3.  On my Core i5 machine with 4 cores I have isolcpus=3 and it looks like rtapi_app is running on core 0.  If I am doing this right I believe that is what the "0" shown below is telling me.

john@X3CNC:~$  ps -efP | grep rtapi
root        1116       1   0  5 08:36 ?        00:03:18 /usr/bin/rtapi_app load tpmod

Is there something I can do to ensure/force it to use core 3?
 

I finally got this working. I was doing a lot of investigation and learning along the way, but I was headed down the wrong path. I saw in another thread where a member was having issue with a USB RS 485 connection and PCW and tommylight both suggested it was an EMI issue. So I thought hmm could that be it? I pushed the PC further from the main cabinet and rerouted the cable from the pendant to the PC and the problem is solved. I left the machine up all even and all night and the pendant was still working this morning. The motherboard of the old PC was using was inside the main cabinet. The new PC is a standard low profile PC it its' regular case. The old PC PSU ran off of the 240v inside the cabinet. The new PC is just plugged into a nearby 120v outlet. Would running the new PC off of the 240v from the cabinet be better EMI wise?

Thanks
John