I’ve reported pieces of this somewhere before, but it might not have been noticed. I’ve spend now a bit more on it, and teh issue persistet, so, next try.
config: AUAV-X2, one “normal” gps+mag on uart and i2c, one uc4h uavcan gps+mag on can, AC3.5-rc11, latest MissonnPlanner
I’ve tried these configurations:
GSP_AUTO_CONFiG = 0 or 1
GPS_AUTO_SWITCH = 0
GPS_TYPE = 1 or 2
GPS_TYPE2 = 0 or 9
(so, in total 8 configurations, each tried several times)
I’ve noted this:
(i) With a dual config, i.e. GPS_TYPE2 <> 0, the first gps often has problems to get its fix, sometimes it can take quite long
it is easy to check that it is indeed the gps on the uart, since disconnecting the uavcan gps makes the gps2 data to reset
(ii) even though gps1 is still not fixed, the MP HUD displays GPS=3DFIX. It is obviously triggered by gps2 reaching a fix. This I think should not (must not) happen when GPS_AUTO_SWITCH = 0. The MP view next to it shows hdop: 100, Sats: 0, and does not show the vehicle, as for no fix. Something goes wrong here.
(ii) with only gps1, i.e. GPS_TYPE2 = 0, the gps on the uart gets its fix within seconds (as I’m used to see it with that gps)
(iii) it’s not clear which or if any of the GPS_TYPE2 = 9 configs worked “best”, I couldn’t really determine a pattern. My feeling is that there is a sort of race condition, if gps1 fixes before gps2, when everything is fine, otherwise not, but that is just a feeling.
(iv) With a dual config, GPS_TYPE2 = 9, and with both gpses fixed: When I disconnect gps2 from the can connector, its values go to zero, as they should. If I when connect it again, then its values reappear, as they should, but very briefly after that the values of gps1 go to zero for some seconds, and reappear that few seconds later. This I think should not (must not) happen, and I would take that as a further indication of the problem in (i)
I think this should be easy to reproduce.
Cheers, Olli