Excessive motor RPM on drops

[coyote_dc5]

So I've had my green hornet now for a week and it keeps doing this motor speed change during low RPM drops on its own. I have a video here of me just doing a quick test so I can blackbox it to see if anyone can help identify. The blackbox file (2/3) is located below and corresponds with the video timing exactly so the issue starts from about 1:07. The quick blipping on the throttle is not me, it does this on its own as i drop down which makes it hard to control.

BB Datalog file
Video

Mods - Is it worth having a new section in the forum for PID Tuning?

PID and Rates I have are as follows and filtering set to 1.2 for both Gyro and D Term

 

[L0stB1t]

Hi coyote, interesting problem. In the logs I see that around 1:06 motor 4 goes completely to 0%, that's not normal. After that I see the yaw I-term freak out quite a bit which is to be expected when only three motors were running.

What is your Motor Idle Throttle Value (dshot_idle_value) set to? Make sure it's high enough so the motors RPM doesn't go too low and they can respond fast enough.

If you have Bidirectional DShot you may also want to set up dynamic idle (betaflight/betaflight). That way you can prevent any motor to go below a specific RPM. With Bidirectional DShot I'd also recommend to use "set debug_mode = DSHOT_RPM_TELEMETRY" in your config so your blackbox will log the RPM of the motors.

I also noticed that during the first part of your flight where you're flying smoothly the logs show a fairly constant pitch I-term around 5, and motor 1 and 3 run slightly higher than the others. This tells me you have a bit too much weight in the back of your kwad. Try to move the battery forward to keep the center of gravity perfectly in the middle of the 4 motors.

 

[coyote_dc5]

Thanks for your response and help with this i think you've hit the nail on the head with idle control as I've spent the last 24hrs with my head in books so to speak learning tuning, filters and dynamic idle. I checked min value and it was set to 0. I've run the motors now and stall is around 600rpm so I've now changed this to 65 which is 650rpm. Havnt had a chance to arm again today as weather has been terrible so hopefully I'll have a play again tomorrow and run a new log . Thanks again for spending the time on this

 

[coyote_dc5]

What confused me with the 0 value was that if it really is stalling the motor on zero throttle then this should be causing a death roll dsync but it's not

 

[L0stB1t]

Ok great so the idle throttle of 0 was definitely the problem. I think it should be somewhere between 3 and 5.5 (the Betaflight default) for this kwad.

Regarding dynamic idle, if the motors stall at 600 RPM then your minimum RPM should be at least 1800, and the idle_min_rpm setting should be 18. Your setting of 65 would correspond to 6500 RPM which is dangerously high.

I actually recently had some interesting questions about dynamic idle so this may also help you: Determining idle_min_rpm in betaflight | IntoFPV Forum

 

[coyote_dc5]

So the BF GUI had 5.5 set but when i typed the get cli command it came back with 0

Thanks for heads up on the idle value, think I meant 6.5 for 650rpm

Why would you push min idle out to 1800? Trust you know what your doing , just curious

 

[coyote_dc5]

Knowing my stall rpm, what do you suggest the below should be set to?

 

[L0stB1t]

If the GUI had 5.5 % then the "get dshot_idle_value" command should return "dshot_idle_value = 550". A "diff all" won't show "set dshot_idle_value = 550" because it's the default but a "dump" command should.

If you use dynamic idle the minimum RPM should be quite a bit higher than the stall RPM because some reverse airflow (as you get when descending) would easily slow down the motors below the stall speed. Someone suggested the minimum RPM should be at least three times higher than the stall RPM but I can't find the reference anymore. In any case this value is still WAY below the RPM you get at 5.5 % throttle.

I suggest you make a list with your motor RPM at 1, 2, 3, 4, 5, and 5.5 % throttle. With that information it's easier to determine a good minimum RPM.

 

[coyote_dc5]

Ok so I decided to fly with profile 3 today which is BF4.1 defaults and just added the following idle settings

dshot idle 550
idle min 18

It definitely flies smoother but the motor jerks are still there during drops. Here is a quick vid I make from an earlier flight.

Next I will do some aggressive moves and log to file if your ok passing your eye over again. Be good to tune out this motor issues totally as it'll mean the moves on cam feel more fluids so I'm not fighting it etc

 

[coyote_dc5]

I'm thinking the residual drop issue now is simply prop wash as I'm descending into turbulent air. Do you think if I reduce RPM filtering on the sliders this might help eliminate? Any other setting I can adjust to help with wash?

 

[L0stB1t]

Would be great to have another log of a descent, especially with debug_mode = DSHOT_RPM_TELEMETRY. Logs of aggressive flips and rolls can help with general PID tuning.

To handle propwash you need reduced filtering so there's less delay and the FC can react to movement faster. But if you have too little filtering the D-term tends to go crazy which results in vibrations, hot motors and shorter flight times. You also need a good PID tune with a correct PD balance and enough P and D gain to battle the oscillations. I generally follow the guide from UAV Tech (UAV Tech - PID Tuning Principles.pdf) but I've always found if difficult to put all the theory in practice.

I'm a little afraid the big foam ducts of the Green Hornet are pretty bad for aerodynamics and it might be impossible to completely eliminate the weird wobbles. Still, a good tune always makes flying more enjoyable so it's definitely worth spending some time on it.

 

[coyote_dc5]

Great thanks , I'll push sliders one at a time tomorrow and keep eye on things. I'll wizz another BB log over too