Hi guys, this is Freeman. My Trex800 was flying well, but sometimes it will go into a toilet bowl in the Position Hold mode. Furthermore, when I was trying to tune the VFF for Pitch and Roll, it seems there is no way to make the Roll match the desRoll perfectly. I need some advice from you to make it better, many thanks!
Freeman,
Took a quick look at your log. Doesn’t look like you completed the tuning in pitch and roll. Did you follow the tuning guide to tune the pitch and roll rate P and rate D gains?
You have not set the ATC_RAT_RLL_I or ATC_RAT_PIT_I gains. These should be the same as your VFF parameter as a good first guess. Also the ATC_RAT_RLL_ILMI and ATC_RAT_PIT_ILMI should be 0.08.
These adjustments to your rate gains should make the pitch and roll attitudes follow the desired much more closely. Anytime you see the signals following each other but there is an offset, then that means you don’t have integrator working well (I gain).
Lastly, in order for loiter to work well, the attitude controller must be tuned correctly where the pitch and roll response of the aircraft closely follows the desired values. So if the corrections fix the disparity between the desired and actual response of the attitudes, then see how it responds when you put it into loiter. If it still seems like it there is a a toilet bowl effect, then you can try increasing the PSC_VELXY_D gain. Put this on a tuning knob if possible. only increase this by 0.1 increments and I wouldn’t go any higher than 0.5 on this parameter.
By the way, nice job on your vibrations.
Still appears that you have some residual vibrations or oscillations happening.
Set LOG_BITMASK to 131071 for your next flight. You will then be able to determine whether these are rotor vibrations. You could use the Notch filter enabled by setting INS_NOTCH_ENABLE to 1. Use the wiki to help you set up the filter.
I would recommend
INS_NOTCH_BW =10
INS_NOTCH_ATT = 30
and set INS_NOTCH_FREQ to the frequency of the oscillation in hz. So if your rotor speed was 1500 rpm. You would count 25 peaks in one second on the rate.R trace that I’ve shown. If you are counting less than 10 peaks in one second (or less than the peaks you would expect for the rotor speed) then it is unlikely that the oscillation is due to the rotor vibrations. I wouldn’t recommend using the NOTCH filter. These may be due to high rate P and D gains.
Bill
That’s Very nice of you shared your experience for my setup correction. I will follow your advice to improve it. Will share the flight report after the next flight.
Hi Bill, it’s been raining for almost 10days…But today I did a trial flight according to your advice in the rain. Luckily, it really worked! The toilet bowl disappeared, also the rate matches the Desrate pretty well as I think. I’m sharing you the log again and if you are free could you take a look on it?