Instability resolution for low thrust (very light vehicle)

ArduCopter Copter 4.3
, , ,
1

Instability resolution for low thrust (very light vehicle)

Total weight of the vehicle is 3.72 kg
Vehicle Type: Quadcopter
Vehicle driving: (ESC) Air 40A / (Motor) U8 Lite (Prop) T-moter G30*10.5
Vehicle Flight Control: (FC) orange tube 4.3.7 version

Until now, it has been designed to match the Hovering output of 40 to 50% of the ESC and motor of the vehicle’s drive, but the vehicle I inquired about is very light compared to the output that can be hovering even at about 20% of the drive.

I adjusted some parameters to prevent light vehicles from soaring upward.

MOT_SPIN_MAX : 0.7
MOT_SPIN_MIN : 0.15
MOT_PWM_MIN: 1,000 (adjusted to a minimum PWM value to allow the vehicle to land)
MOT_PWM_MAX: 1,600 (maximum PWM value considering vehicle drive)
MOT_THST_HOVER: 0.2 (MOT_HOVER_LEARN is learned and stored, so the final value changed and stored when checking the log after the flight is 0.554091)

However, there are difficulties in testing to reduce output and have horizontal stability during flight for this test.

The difficulties encountered during the test are as follows.

  1. When you take off after ARM, the aircraft gradually begins to lose its level significantly in place and eventually loses its balance and crashes. MODE: LOITER / STABILIZE (Predicted from ESC)
  2. It is predicted that it will be difficult to cope with special situations by lowering the PWM maximum.
2

Please refer to the log file above.

3

isnt mot_spin_min way too high? thought that was supposed to be the lower limit commanded to motors.

is using lower battery voltage / higher batt cap an option? (so that hover throttle can be brought closer to before)

4

I’m sorry
There was a part I wrote wrong.
Please note that we have updated the post again.

There is no exact reason why MOT_SPIN_MIN was designated as 0.15, but I designated it as 0.15 because the motor was operated at 0.15 when the Mission Planner tested it.

  • In my opinion, there is a risk that the motor will stop when adjusted below 0.15.

The battery-related option you mentioned is a function I haven’t used yet, so I didn’t test it because I couldn’t trust it.
Lordvon, why did you ask about this function??

5

Your method of tuning is incorrect. The proper way to tune is to use the correct pwm value and change the mot_thr* parameters afterwards. Not the other way around.

And using mot_thr_max to limit the climb rate is also incorrect. There is a different parameter for that.

Do not be creative, follow the existing documentation and tuning methods and order.
It will save you time and money

6

Thank you for your advice.

The parameters were changed again according to the documentation and documentation.

But the same problem recurred.
But I think there’s a clue.

Predicting the cause of the problem is as follows.

  1. It is determined that the weight of the vehicle is very light (almost bone only) compared to the drive, causing the motor to stop rotating outside the minimum dead zone of the ESC and motor (the weight is evenly distributed and the weight of the vehicle is added to test)
  2. A re-measurement of the centre of gravity predicted that approximately 340g of weight would be more concentrated rearward, causing forward and rear shaking during vehicle flight
7

TEST VIDEO

LOG

8

It’s good that you have the emergency stop switch set up. That is essential with these bigger copters.
Do all these steps exactly

1. set up your battery monitor for at least voltage, and current too if possible. It is not safe, and sub-standard tuning, to ignore voltages with these large props. Check it’s accuracy at low voltage like 20.0 volts.
Reboot after you get that working.

2. set all these by copy/paste into notepad and Save As “updated_parameters.param”

ATC_ACCEL_P_MAX,22000
ATC_ACCEL_R_MAX,22000
ATC_ANG_PIT_P,6
ATC_ANG_RLL_P,6
ATC_INPUT_TC,0.2
ATC_RAT_PIT_D,0.003
ATC_RAT_PIT_FLTD,10
ATC_RAT_PIT_FLTT,10
ATC_RAT_PIT_I,0.110
ATC_RAT_PIT_P,0.110
ATC_RAT_RLL_D,0.003
ATC_RAT_RLL_FLTD,10
ATC_RAT_RLL_FLTT,10
ATC_RAT_RLL_I,0.110
ATC_RAT_RLL_P,0.110
ATC_RAT_YAW_FLTE,2
ATC_RAT_YAW_FLTT,10
BATT_ARM_VOLT,22.10
BATT_CRT_VOLT,21.00
BATT_FS_CRT_ACT,1
BATT_FS_LOW_ACT,2
BATT_LOW_VOLT,21.60
BRD_BOOT_DELAY,3000
FENCE_ENABLE,1
GPS_GNSS_MODE,65
INS_ACCEL_FILTER,10
INS_HNTCH_ENABLE,1 
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4
LOG_BITMASK,180222
MOT_PWM_MAX,2000
MOT_SPIN_MAX,0.9
MOT_THST_EXPO,0.77
MOT_THST_HOVER,0.15
PSC_ACCZ_I,0.4
PSC_ACCZ_P,0.2

then in Full Parameter List, load from file and select that file you just saved from notepad.
It will come up with a comparison window - accept everything it offers and write them to the flight controller.
You can verify they all got written by trying to load them again, or using the Compare Params function.

3. reboot the flight controller then do the semi-automatic ESC calibration as per
https://ardupilot.org/copter/docs/esc-calibration.html#semi-automatic-esc-by-esc-calibration

4. use MissionPlanner motor test to verify that MOT_SPIN_ARM and MOT_SPIN_MIN are still valid, and adjust them if necessary. MOT_SPIN_ARM should be as low as possible to ensure reliable/smooth startup of motors, then MOT_SPIN_MIN = MOT_SPIN_ARM + 0.03

Do not lower MOT_SPIN_MAX or MOT_PWM_MAX that I’ve specified within that big list of params → we should be able to get this flying with normal parameters.
You might have to add a bit of dummy payload. Please reply with what the take off weight is now, and what you expect it to be in actual use.

5. Do a test flight of just hover, start in Stabilise mode, and change to AltHold for a while if everything is going OK.
If there is instability land immediately and reduce these:

ATC_ANG_PIT_P,6 -> 5 or 4
ATC_ANG_RLL_P,6 -> 5 or 4

Let’s see that log file!

1 Like
9

I’m sorry for the delay in answering.

When I proceeded in the way Shawn told me, I found it to be very stable compared to the previous “the phenomenon of pitch and roll movement gradually increasing and falling during flight.”

The test was not conducted in Stabilize mode, but in Alt Hold mode, and it was stable apart from the concerns.

There is an additional question.

  1. When testing a motor in stabilization mode, can the speed of each motor vary depending on the posture angle of the aircraft? (Prop X)

  2. I’d like to see the logs and videos below and see if there’s anything else I need to supplement.
    https://drive.google.com/file/d/1q1V3IKlNLUnUVaQorZBqvi-RYleg_map/view?usp=sharing

10

  1. Yes, if you can find somebody brave enough or crazy enough to hold this copter over their head then moving it around in Stabilise mode should make lower motors speed up so the copter tries to level itself. It’s risky and probably not worth doing, especially since you’ve been able to launch this copter already.

  2. There was only a log in that link.

But anyway, this is my first option for Harmonic Notch Filter:

INS_GYRO_FILTER,15
INS_HNTCH_BW,8
INS_HNTCH_FREQ,20
INS_HNTCH_FM_RAT,0.7
INS_HNTCH_HMNCS,15
INS_HNTCH_REF,0.15

should give this:

image
image1230×443 47 KB

If there’s adverse affects from those settings, you will have to use this:

INS_GYRO_FILTER,15
INS_HNTCH_BW,16
INS_HNTCH_FREQ,38
INS_HNTCH_FM_RAT,0.7
INS_HNTCH_HMNCS,3
INS_HNTCH_REF,0.15

image
image1223×442 46 KB

Once you have relatively stable hover with those Harmonic Notch Filter settings, I would try these PIDs

ATC_RAT_PIT_D,0.004
ATC_RAT_PIT_I,0.12
ATC_RAT_PIT_P,0.12
ATC_RAT_RLL_D,0.004
ATC_RAT_RLL_I,0.12
ATC_RAT_RLL_P,0.12

but from then on you will probably need to read about and use transmitter based tuning, unless you are Ok with using Autotune. Probably set:
AUTOTUNE_AGGR,0.08
and do one axis at a time.

11

Thank you for your answer.

First of all, thank you for your reply to the above request.

I told you that the semi-automatic ESC motor calibration did not seem to proceed normally in the case of No. 1.

I will share the link again for number two.

Also, as you answered, we will test by applying parameters for Harmonic Notch Filter. Thank you.

As a result of flying the aircraft today, it was noticeable that the Loiter mode can hover normally even with small output, but it moves even with weak winds (2 to 3 m/s).

  • I suspect FC’s response is a little slow because it’s a small output. (I’m going to try more various tests.)

2023-08-16 15-03-58.bin - Google Drive, 2023-08-16 13-45-46.bin - Google Drive,

12

uhhh why not just disconnect all other motors and, in stabilize mode while keeping hands on throttle, lift one side of the copter to observe motor change

1 Like
13

Yeah it can be done easily and safely with props OFF too

1 Like