Yesterday, 07:00 AM
Thank you for the clear answers, that helps a lot.
So a lost press is not repeated and the HAv2 does not wait for an idle bus. Collisions can only be avoided on the
controller side, which is what we do: the F24 polls the TS only when the bus has been quiet, never within 10 s of a
button frame, and tries up to 3 times. Good to know that foreign frames do not disturb the HAv2 itself.
Update from our bench: 18 hours with the TS polled every 10 s on the HAv2 bus, using the patched kc868_ha.
6,438 polls, 6,436 replies (one retry, one poll in flight when we read), 0 failed cycles, no reboot. In production
we will poll only every 90 minutes.
The code change is here as a pull request: https://github.com/hzkincony/esphome-kc868-ha/pull/3
Thanks again.
So a lost press is not repeated and the HAv2 does not wait for an idle bus. Collisions can only be avoided on the
controller side, which is what we do: the F24 polls the TS only when the bus has been quiet, never within 10 s of a
button frame, and tries up to 3 times. Good to know that foreign frames do not disturb the HAv2 itself.
Update from our bench: 18 hours with the TS polled every 10 s on the HAv2 bus, using the patched kc868_ha.
6,438 polls, 6,436 replies (one retry, one poll in flight when we read), 0 failed cycles, no reboot. In production
we will poll only every 90 minutes.
The code change is here as a pull request: https://github.com/hzkincony/esphome-kc868-ha/pull/3
Thanks again.

