Is this saying that the keypress lag when wireless to host using the Dongle should be pretty much fixed but for wireless to host using BLE is not?
Depends, since the cause of ble hid lags is unknown.
All I can say is that the code that sets latency params is trying to force the minimum for all connections including ble hid.
So Dongle — yes definitely.
BLE HID — let’s see. Will give it a whirl.
@kareltucek Even 7ms may make a difference when timings of maybe 40ms or so are used to distinguish some homerow mod use cases. I am assuming correctly that you are not timestamping the messages from a synchronised clock source when the keystroke is detected, but instead the arrival time of the messages on the right half is used to calculate timings?
I notice that on the UHK 80, I more frequently get some wrong (non-)activations of my homerow mods; which worked absolutely fine on the UHK 60 with firmware 11.2. That’s for use cases where I press the mod on one half and the final key on the other half. Timing is pretty crucial for those HRM to work reliably.
Correct.
I have a plan to do timestamping and synchronise them in the right half (which will in turn allow us to decrease polling rates and thus conserve battery), but we are far from there as my backlog is constantly growing rather than getting shorter :-(.
I notice that on the UHK 80, I more frequently get some wrong (non-)activations of my homerow mods; which worked absolutely fine on the UHK 60 with firmware 11.2
I am wondering if the issue is really the latency, since uhk60 has similar latencies over its i2c as far as I know.