[3.5.0-rc.1] 1HCL board takes 1 second to respond

[nikscha]

Hi all,
I'm currently working with the constant torque feature of the 1HCl board, and I'd like to change the torque on the fly.
For that I compute the desired value in daemon.g and then send a M569.4 to the 1HCl.
This however takes a long (>1s) time. I'm wondering whether there's a way to reduce that. CAN-FD shouldn't be the bottleneck.
Is there a way to make daemon.g execute faster? (I'm already doing a while true loop with 1ms wait and it's still only running about 3-5 times a second)
This alone doesn't account for the >1s wait, so maybe the 1HCL board is to blame?

Normal (open or closed loop) movement commands are executed immediately on the 1HCL.

[nikscha]

@nikscha It seems like problem is related to the input channel. When the gcode that runs M569.4 is executed from file, the torque is applied instantly.

[nikscha]

@nikscha Okay I'm not so sure about my problem anymore. Please excuse the confusion.

I want to test a few more things before I make any more assumptions/ask meaningless questions ^^

[nikscha]

@dc42 how long does it take for the object model to reflect a change in movement?

[nikscha]

My problem wasn't caused by the 1HCl board.
Instead daemon.g simply doesn't run often enough to catch a changed value in the object model in time.