Updating el axis encoder offsets to get more flight-ready
Make a gyro data validity check more lenient:
acs.c checks the number of valid gyro packets recv'd since the last loop increment against an expected number
If this check fails more than twice, it sets the gyro fault mask to ignore readings from the gyros
The gyros + an inclinometer are now on a USB hub (c.f. #62)
Although USB has more than enough bandwidth, it seems we marginally fail this check on a regular basis, i.e. the number of gyro packets received between 200 Hz loop updates is less consistent than BLAST
Inspection of gyro data packets shows that all data is unadulterated, so data validity is not a problem, it's a latency/jitter issue, probably from USB interrupt latency?
Making this check only fail gyros if the count fails thrice restores the expected behavior of the gyro fault mask, i.e. very infrequent masking of one or the other gyro.
Make a gyro data validity check more lenient:
acs.cchecks the number of valid gyro packets recv'd since the last loop increment against an expected number