Wide switch mode always 6 bit, boost to 1:2 #3295
Conversation
|
Personally I think that changing the interpretation of the OTA data based on the TLM ratio is. a very bad idea, so I'm in support of 6-bit all the time in wide mode. w.r.t the TLMBoost, I'm ok with keeping the way it is now (i.e. with the new code). We could calculate where we are in the hopping sequence and if we are less that half way to a sync, then wait for 2 syncs otherwise wait for 1 sync. That way we boost for 1/2 to 1-1/2 periods. But this would introduce more complexity. |
|
Thanks, PK. You're definitely correct that changing the OTA format based on the TLM ratio was a bad idea right from the get-go; I should have recognized the problem with it when I created it. We can optimize the boost duration after 4.0 if we feel like it is excessive using something like you described if we find that it is not meeting our performance needs. Just gotta lock in that always 6-bit OTA format right now. |
mha1
left a comment
There was a problem hiding this comment.
I'm all for it. Reduces complexity, brings back 1:2 boost and gets rid of the 1:8 workaround. And if there's wide mode users in need of more resolution there's always the full modes.
SunjunKim
left a comment
There was a problem hiding this comment.
+1 for this, for the sake of simplicity.
ps. Who use non-Full modes for control surfaces? They stuck on Full mode (including ch5!) |
Removes the 7-bit mode for Wide switches when running at low telemetry ratios, making the Wide switches always 6-bit. This also returns TLMBoost to the original 1:2, removing the triage fix from #3210 that limited it to 1:8 to prevent it from crossing the boundary.
I feel like this is less confusing and returns the telemetry boost performance lost. REGRESSION: users may be upset about the lower resolution of the channels, making them maybe unsuitable for some control surfaces or head tracking.
Question
Should I also revert the TLMBoost request/active code that was added in #3222? The advantage to keeping it is slight and it does add minor code complexity.
Also tagged For Discussion because if we want to keep the terrible TLMBoost bandwidth and keep 7-bit Wide switches, I'm cool with that and just closing this PR.