Oscillation during loiter mode when any amount of roll input is given causing it to crash

Using Tarrot 5008 with 18 inch Props and 60A esc Tflame esc. Everything is fine in stabilize and alt hold mode but as soon as i switch to loiter and give even a little bit of roll the whole drone starts to oscillate side to side as if trying to stabilize itself but failing to do so and sending too much throttle input to the motors than required. I am also attaching the drive link to the logs. Kindly help me to solve this issue.
https://drive.google.com/drive/folders/1yCLH06c2mYeDCWVMjRuRUJ3TGnodsSex?usp=sharing

you have got serious of vibration issue on your flight controller.
please show us how to mount the FC on your drone?

Yes, I noticed that too. The cube orange + is hard mounted to the body using the provided screws.
Although the vibration is under 30m/s/s which is the acceptable range of vibration according to copter documentation. Before using Tflame esc I was using 40A hobbywing xrotor and it had no problem during loiter(setup is completely identical and nothing is changed except ESC) but as soon as I switched to Tflame esc the problem started to occur.
PS the reason I switched to Tflame is that during one flight when I was landing one of the motors suddenly lost power( my conclusion was that esc desynced) and crashed that’s why I Switched to Tflame.
I also changed the mot_thst_expo from 0.7 to 0.2 ( according to mission planner initial parameter setup while selecting Tflame esc ). Later I changed it to zero because I read in one of the arducopter discourse that for linear esc it should be set to 0(also found out Tflame is linear esc, maybe wrong have not went too deep into it, correct me if I am wrong) and still the issue remained.

Vibrations are too high. Anything over about 15 or 20 is getting into a grey area, and up around 30 and more is going to give you trouble.
You should definitely do something to improve the vibrations before flying more. It’s OK to mount the Cube carrier on good quality antivibration foam or whatever works. The good thing in this case is the vibrations are reasonably equal across all 3 IMUs indicating there is NOT any strange issue with the inbuilt IMU damping and harmonics.

MOT_THST_EXPO of 0 would definitely be wrong.
For T-Motor Flame I would use:

MOT_THST_EXPO,0.40
MOT_PWM_MIN,1100
MOT_PWM_MAX,1940

and do the semi-automatic ESC calibration just to be sure:
https://ardupilot.org/copter/docs/esc-calibration.html#semi-automatic-esc-by-esc-calibration

then I suspect these should be more like:
MOT_SPIN_ARM,0.10
MOT_SPIN_MIN,0.13
but use MissionPlanner motor test to verify that MOT_SPIN_ARM is using the lowest possible reliable start up percentage. Then add a bit for MOT_SPIN_MIN.

Make sure these battery failsafe levels really are correct or you will damage your battery pack.
The normal values for LiIon 6S are:

BATT_CRT_VOLT,18.00
BATT_LOW_VOLT,18.60

Also set these just to get all the basics ticked off:

BATT_FS_CRT_ACT,1
BATT_FS_LOW_ACT,2  //  or 3
INS_ACCEL_FILTER,10
INS_HNTCH_BW,25
INS_HNTCH_ENABLE,1  //  set this then refresh params to see the rest
INS_HNTCH_FM_RAT,0.7
INS_HNTCH_FREQ,50
INS_HNTCH_MODE,1
INS_HNTCH_REF,0.17
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4
LOG_BITMASK,180222
PILOT_THR_BHV,7  //  I think you have a spring-centered throttle

Do another test flight just with Stabilise mode and AltHold if that works OK. Don’t try Loiter yet.
Hover thrust might take a while time to relearn after changing those MOT values so dont be surprised if hover needs some stick input at first.
Send that .bin log file.

Ok.Will do that and get back to you. And I am using solid state lion battery that’s why the batt volt is set to lower than default lion battery.

Yes, I make the Li-Ion SS parameters to be:

BATT_ARM_VOLT,18.50
BATT_CRT_VOLT,17.40
BATT_LOW_VOLT,18.00
MOT_BAT_VOLT_MAX,25.20
MOT_BAT_VOLT_MIN,16.20

I think those values might be a bit too low based on my experience with the Ares Semi-Solid State packs. I usually run 20.1V and 20.7V w/ Sag Compensation turned on.

Hi Manav - Check my spreadsheet.
The Lipo and LiIon values are long-standing, tried and true values.
The LionSS and LipoSS values I’ve found from their specs of common brands - they could easily be a bit wrong, but then why would you buy them instead of standard LiPo or LiIon.

You will see I list the High (fully charged) voltage and the Critical voltage for each cell type, and use some simple formulas to derive the parameters. Someone would be free to put their own values in that table when setting params for their battery pack.

image

There are a good amount of data-sets in here which show a higher value would be better for low and critical: Foxtech 30000mah 6S Battery, any one try it? - #28 by mtfunder. I target around 20% remaining and 10% remaining for low/critical (which I think is fairly standard).

The main reason behind buying the LiPoSS (from my point of view) is the high energy density paired with a fairly high (5C) c-rating. However, I don’t have enough flight time on them to fully trust for some of my more commercial projects (so take this all with a grain of salt).

I’ve read through that discussion multiple times too and checked over the graphs.
I came up with 2.7 volts critical for LiIonSS based on the foxtech value plus eye-balling those graphs. It could well need to be 2.8v and the advantage would still be the lighter weight/higher efficiency of the LiIonSS overf a similar-sized conventional LiIon pack.

Ideally a UAV in regular use should have the capacity values set correctly, and the voltage failsafe levels would be there for when something goes wrong or capacity has unexpectedly changed.

Yeah - I just worry that the user sets it too low based on a recommendation.

Here’s the bin file after using above parameters. I also switched to new FC and use damping pads. Also switched to default PID even though I don’t think tuning is the issue cause I have done one 50 min flight time(in loiter) and one 30 min flight with different ESC(hobbywing xrotor 40A) and during that time there was no issue everything was smooth.
So did a flight in stabilize mode and same issue.
https://drive.google.com/drive/folders/1OAp9QhRY45PdIsbccQFg84C2eIo8C8OF

There’s quite a large yaw bias like twisted motor mounts.
Z axis vibrations are still too high.
Plus there’s a few parameters that you need to put back in place.

Ok. Kindly tell me the parameters that I need to put back. Also which parameter did you inspect in the bin file that helped you to find the yaw imbalance. I want to know so that I can also be capable to find my problems in the future.

Thye RCout values show that motors1 and 2 (CCW) have to work harder than 3 and 4 (CW) , which requires a physical fix.


image

Your battery failsafe voltage levels and actions need to be put back, and adjust these

INS_ACCEL_FILTER,10
INS_HNTCH_FREQ,42
INS_HNTCH_BW,21

but we cant tell much more until you fix the physical issues, vibrations and yaw, then do a stable hover.

1 Like