Mb2hal speed irregular timing of reversing of spindle

I am having trouble with fwd/rev spindle delays 0-2sec variance when using modbus 2 hal userspace progam.
I have baud rate as high as i can for the x200 vfd. But it seems like the internal timing of the user space mb2hal is too slow.

Is there a way to get linuxcnc to output fwd/reverse over rs485 at consistant sub second speed. Rigid tap dose not like delays in reversing.

Have you tried checking the debugging messages for mb2hal while Linuxcnc is running? Large delays can be caused by communication errors (I have seen this myself, but haven't been able to do much about it.)

Havnt seen any debug messages. I look into it. I have started to look at the mb2hal file to try understand how it works.

I checked ther mb2hal.num_error pins in hal and they all change 0 to 1. So i guess this means illegal function.

You set the debuging level in the mb2hal.ini file, then the messages are printed out in the terminal when you start Linuxcnc from the command prompt.

For maximum speed check this:
SLOWDOWN=0.0
MAX_UPDATE_RATE=0.0

also for serial link check this
SERIAL_BAUD= highest as posible
SERIAL_DELAY_MS=0

set the debug levels to see information
INIT_DEBUG
DEBUG

Charlieslasher wrote:

But it seems like the internal timing of the user space mb2hal is too slow.

The problem is not mb2hal, it is as fast as your link throughput or VFD response.

Can you explain in more detail how to run linux from comand prompt . I tried opening terminal and running linux cnc from there but didnt get any error log messages.

I changed serial delay to 10ms this made it faster and more consistant but i think im still geting modbus errors.

Overtravel rigid tapping can be upto 8mm at 600 rpm due to the delay. Crashing tools.

just type linuxcnc in a terminal window. The error messges are just printed to the console in that window, not to an error log.

Thanks go it to work. Was using a shortcut to linuxcnc.

With some testing update rate is 6hz. And no errors idle.

I think i might be losing data from interference. Possible reason for random timing behaviour.
I dont have terminating resistors at the vfd. So i will work on this and see if things change. No oscilloscope to test.

All fixed and running very nice! Main problem was data com errors caused by vfd noise.

Grounding the 485 cable at the serial converter on the pc fixed all the errors. Grounding the cable at the vfd had little to noneffect. I used cat 6 sheilded cable 3m length with no terminating resistors.

Response is 10hz and repeatable timing plenty fast enough for rigid tapping spindle control.

Thanks for all the directions to fix the problems.