Brown Out after adding Airspeed sensor

Hi Jamie,

I don't think a brown out was your major cause of problems. As far as I can see the major issues were:

it looks like you have a MTK GPS, and it was having trouble retaining lock
your settings had raised AHRS_RP_P to 0.5, which increased the gain of the GPS in attitude correction
the GPS kept claiming it had lock when I think it had lost it, as evidenced by the very high AHRS.error_rp values and the small satellites_visible count

This sort of problem with MTK GPS modules is one of the reasons I released 2.73, with a check on the satellite count to stop using GPS velocity for attitude correction.

So what I think happened is this. The MTK GPS lost lock, but kept claiming lock. The AHRS DCM code kept using the (incorrect) GPS velocity numbers to correct the accelerometers. That led the DCM attitude solution to be way off. The APM thought it was level, when it fact it was rolled at 25 degrees. You had a quite low roll limit in FBWA, giving not much authority to correct the incorrect attitude.

Additional factors that made things work:

you had an airspeed sensor enabled, but not working (it always read close to zero airspeed)
you had a ALT_MIX of 0.6, so altitude came 40% from GPS, and GPS was misbehaving
compass learning was off, and you had lots of magnetic interference, and poor offsets (about 2/3 of the magnetic field was from your motor/esc/battery, not the earth)

Cheers, Tridge

PS: As I mentioned in a private msg to you, we can go over the log in more detail together if you would like

Replies

Developer
Andrew Tridgell May 20, 2013 at 5:14am
Hi Jamie,

I don't think a brown out was your major cause of problems. As far as I can see the major issues were:
- it looks like you have a MTK GPS, and it was having trouble retaining lock
- your settings had raised AHRS_RP_P to 0.5, which increased the gain of the GPS in attitude correction
- the GPS kept claiming it had lock when I think it had lost it, as evidenced by the very high AHRS.error_rp values and the small satellites_visible count
This sort of problem with MTK GPS modules is one of the reasons I released 2.73, with a check on the satellite count to stop using GPS velocity for attitude correction.

So what I think happened is this. The MTK GPS lost lock, but kept claiming lock. The AHRS DCM code kept using the (incorrect) GPS velocity numbers to correct the accelerometers. That led the DCM attitude solution to be way off. The APM thought it was level, when it fact it was rolled at 25 degrees. You had a quite low roll limit in FBWA, giving not much authority to correct the incorrect attitude.

Additional factors that made things work:
- you had an airspeed sensor enabled, but not working (it always read close to zero airspeed)
- you had a ALT_MIX of 0.6, so altitude came 40% from GPS, and GPS was misbehaving
- compass learning was off, and you had lots of magnetic interference, and poor offsets (about 2/3 of the magnetic field was from your motor/esc/battery, not the earth)
Cheers, Tridge

PS: As I mentioned in a private msg to you, we can go over the log in more detail together if you would like

This reply was deleted.

Discussions

Brown Out after adding Airspeed sensor

Replies

Activity