search

olliw42 / mLRS

2.4 GHz & 915/868 MHz & 433 MHz/70 cm LoRa based telemetry and radio link for remote controlled vehicles
GNU General Public License v3.0
295 stars 67 forks source link

rx, transmit diversity #105

Closed jlpoltrack closed 1 year ago

jlpoltrack commented 1 year ago

Works on the bench - but not sure if I'm missing anything here :)

olliw42 commented 1 year ago

ah, very interesting. I just happen to have started to continue my work on this, albeit I'm focusing for the moment on the tx side. I've just made me a logger to log the complete real-time transmitter stats data, to work out a scheme.

I can't claim for sure, but I think this is too simplistic. Especially I'm concerned about what's happening near connection loss (i.e. in the presence of substantial frame losses) as well as when disconnected. I also somehow feel that there should be some time constant, e.g. from some filtering, as well as some hysteresis. The point is that the momentary interference or loss of packet on a frequency does not typically mean that the over-the-air channel properties have changed, e.g. becasue the vehicle has changed it position oin space. So, the simplistic scheme may actually most of the time jump to the bad antenna. Just my thoughts on the matter.

BTW, I'm going to do my experiments for tezh tx transmit diversity with a 2dBi omni antenna on one port and a 7 dBi directional antenna on the other port. Would be a cool general configuration, I guess. (doesn't make sense on rx side, of course)

jlpoltrack commented 1 year ago

Fair enough - will be interesting to see your findings. I will try to do at least 1 flight with this basic implementation and share what I see.

Hardware wise - I seem to be moving towards single antenna (directional, e.g. moxon) Tx and diversity on the Rx side.

olliw42 commented 1 year ago

Fair enough - will be interesting to see your findings.

the findings will be unspectacular ... it's more the conclsuions i.e. model which on may draw from it. The idea is basically to record some logs and to then do the dev using these logs.

I will try to do at least 1 flight with this basic implementation and share what I see.

PLEASE do so :)

Hardware wise - I seem to be moving towards single antenna (directional, e.g. moxon) Tx and diversity on the Rx side.

this sure makes sense, given that it's teh vehicle which is moving

I think that the tx diversity however can make lots of sense when a combination of omni and directional antenna are used, since the omni allows position free operation in near range and the directional long range ... pretty much makes an antenna tracker superfluous. And probably not coincidentially it's used on "many" systems, like the herelink, or some folks do so on fpv googles

olliw42 commented 1 year ago

any technical reason for closing which would be worth knowing?

jlpoltrack commented 1 year ago

No technical reason, just tidying :)

olliw42 commented 1 year ago

ha, ok, thx!

you have seen this PR https://github.com/olliw42/mLRS/pull/108 ?