T-drone M1000, post autotune the quad is oscillating

sanjay_pandit · March 21, 2022, 10:58am

As suggested by @manavgandhi17, I had packed the pixhawk and had given sponge vents at the bottom (for air pressure to equalise and for the turbulence to die). I flew it at 4m/s but still saw a real dip of over a meter dip in altitude of the quad while the HUD showed no dip. Please suggest any parameters change or any other change that would help me to overcome this.

The autotune has generated some value of yaw that is messing up with controls… When left in POS Hold mode the quad starts to oscillate in diagonal axis after 8-10 secs… When given an input the behavior goes off… But when left uncommanded the thing repeats. Anyone have experienced such behavior? . And what is the best method to go about solving this.

Please feel free to suggest… I am ready to even experiment with it.

amilcarlucas · March 21, 2022, 11:47am

no .bin files, no cookie!!

juzzle1 · March 21, 2022, 12:11pm

This may have nothing to do with it, but you’ve set:

INS_ACCEL_FILTER 20
INS_GYRO_FILTER 10

According to the setup guides, the recommended values for a machine this size are:

INS_ACCEL_FILTER 10
INS_GYRO_FILTER 20

sanjay_pandit · March 21, 2022, 12:44pm

Apologies, had forgot to upload it. I have uploaded it with the file name ‘where we started’.

Moreover, we took some more flights by changing some params. The file name shows the changes that we had made.

There is a video, in which oscillations in yaw axis can be observed

https://drive.google.com/drive/folders/1VUy_7I5RG20LOdYZGAS9p60SUWPE32Cj?usp=sharing

sanjay_pandit · March 21, 2022, 12:45pm

Thanks Justin for pointing this out. I don’t know how we missed such important thing. Will try a flight with the suggested params tomorrow.

Thanks again

sanjay_pandit · March 22, 2022, 5:01am

For the altitude dip during movements of around 4m/s, can the ATC_THR_MIX parameter be responsible?

xfacta · March 23, 2022, 12:23am

A .bin log file would really help to diagnose issues, otherwise we just have to guess

sanjay_pandit · March 23, 2022, 1:33am

@dkemxr it’s already there in this folder. Have given a video too.

https://drive.google.com/drive/folders/1VUy_7I5RG20LOdYZGAS9p60SUWPE32Cj?usp=sharing

xfacta · March 23, 2022, 3:45am

I would see if you can get MOT_SPIN_ARM 0.1 and MOT_SPIN_MIN 0.3 to be a bit lower, for example your hover thrust is around 0.24. But this doesnt appear to be causing an issue directly.

I would set these and retest.
ATC_THR_MIX_MAN,0.5
INS_ACCEL_FILTER,20
INS_GYRO_FILTER,20
INS_HNTCH_REF,0.198
INS_HNTCH_FREQ,40
INS_HNTCH_BW,20

You might want to try MOT_THST_EXPO,0.4 or even higher in 0.1 steps because pitch and roll do go a bit unstable as throttle rises for certain manoeuvres.

set too high you can see instability at low throttle
set too low you can see instability at high throttle

sanjay_pandit · March 23, 2022, 7:33am

Thanks @xfacta for sharing your observations and suggestions. But can you please tell why INS_HNTCH_FREQ, 40? coz when I had done the notch setup as per manual… I had got peak at 78Hz and had set everything according to it.
I had set MOT_SPIN_ARM TO 0.15 and MOT_SPIN_MIN 0.2 the takeoff wasn’t smooth and had to struggle a lot to lift it. Have made MOT_SPIN_MIN 0.23 and it feels better.
I will do MOT_THST_EXPO, setting and share the logs.
Thanks

sanjay_pandit · March 23, 2022, 7:36am

I did a small test today in windy conditions. The results were horrible and seemed like I would loose the copter. I am sharing the logs.

I request you all to please check and suggest changes, the systems is becoming harder to tame as I progress to make it worth a survey

https://drive.google.com/file/d/1i-tyn-nuty7uKb6OQcQGqJnDBa43uQen/view?usp=sharing

MindProbe · March 23, 2022, 2:43pm

@sanjay_pandit
Although this might not solve your problems but can make a difference sometimes
Your IMU target temperature is set to 45C
But you performed IMU calibration at 50C and from your latest log, IMU temps are around 55

Did you perform IMU temperature Calibration?
IMU Temperature Calibration — Copter documentation (ardupilot.org)

sanjay_pandit · March 23, 2022, 5:33pm

@MindProbe Thanks for the observation. Was never aware of this feature until you posted it here. Will surely do this asap. But here the temperature difference that I expect during a flight won’t be considerable.

How were you able to check my current IMU temp that you checked (is it the temperature that I see under Heat.temp?)

Also, can you please verify my logic that the Baro[0] is my external barometer on Here2 and rest two are from Cube? and is it the same with Acc, Gyro and Compass?

Thanks

MindProbe · March 23, 2022, 7:27pm

Hi @sanjay_pandit

Judging from your posts, you are preparing this quad for surveying purposes which could be performed in bright sunny conditions. This can very possibly increase your IMU temperatures and introduce minor gyro drift if not compensated for

Yes. Heat.Temp is the Current IMU Temperature

asdad
It seems like BARO1 is the external compass, which should be BARO[0] in the log (not completely sure though)

Check out this ardupilot/Tools/scripts/decode_devid.py for more clarity

bus_type=devid & 0x07
bus=(devid>>3) & 0x1F
address=(devid>>8)&0xFF
devtype=(devid>>16)

sanjay_pandit · March 23, 2022, 8:20pm

Very true. Thanks for understanding my needs and suggesting the right direction to achieve it. I will certainly try it.

Thanks

I am very noob to this, can you provide some link/source (to read and understand) like you did for IMU temperature calibration. That will help me understand it better. Currently, I am not getting a single thing in this direction and all I am doing is a guess work. I would like to make it a planned on rather than a guess work.

Thanks

dkemxr · March 23, 2022, 9:05pm

Just type the device ID into Mission Planners Decode HWID’s on the Ctrl-F menu and it will tell you the bus they are on.

Such as:

UAVCAN is obviously external as would I2C be if it shows that. SPI is internal.

sanjay_pandit · March 23, 2022, 9:23pm

Cool, that solves some of my problems. One more thing, I noticed the latest log that the mag1 and mag2 varies a lot.

Is it safe to disable internal compass and go with only one external compass(I doubt the what will happen if ekf rejects that)
Is is okay if I have that much variation in those two compass

Thanks https://drive.google.com/file/d/1i-tyn-nuty7uKb6OQcQGqJnDBa43uQen/view?usp=sharing 2

dkemxr · March 23, 2022, 11:42pm

Yes, this is commonly done.