Cube Black Baro problems


in my Octo APR in Coax configuration I made many test flights and all went well without problems

a few days ago I flew about 150 km from the test airfield and the Coax had a serious altitude problem. No Crash but after the first few minutes of flight he starts doing a yo-yo.
After an analysis of the log it is noted that 2 out of 3 barometers are perfect (BARO 2 and BARO 3) but one (BARO) is very different from the others

How do I solve it? … is it always better to adjust the pressure at each flight ?? (I hope not !!)

There are only 2 internal baros in the Cube.
Can you tell which one is the external baro?

hi Alvin and thanks for the reply. As a GPS I have a Here V2 and via Uavcan I enabled Baro to have the external barometer located on GPS … correct?..

If you are using Here2 in serial mode, its baro will be detected by default. If using in CAN mode, you have to enable the baro.

the external Baro is enabled in Can (4-wire cable)
The problem is that I don’t know from the graph which is the cheater providing the false information

Baro = problems
Baro 2 = ok
Baro 3 = ok

so of these, which are the 2 internal Baros and which are the external Baros?
Logically, I would identify Baro and Baro 2 as internal and Baro 3 as external … is there a “map”?

Normally the ardupilot should assign external baro as primary baro, which is 0.

interesting … so it would be easy to disable it (Uavcan> Baro = 0)
Perhaps because the Baro was on value 1, that is, it had to use the learning of the pressure value through the procedure provided, i.e. set it to 2 and let the pressure be determined and then return to 1 to use the value read. Except that the procedure didn’t work and the value is always 0.
All this is interesting but I don’t understand why many flights went well and only the last two off-site went very badly …