Throttle scaling different?

[rput]

I am just getting my new system with the DIMU going but in the initial flights I noticed that the throttle response in manual mode was extremely sensitive, so much so that holding altitude or even landing was difficult.

Back home we did some probing into this and noticed that in the logfile the motor outputs (i.e. MOT_MOTOR0 etc.) are going up to 4000 on full throttle. On the analog board they would only go to 1900 or so. I probed the PWM outputs from the AQ board with a scope and indeed notice that on the DIMU version the PWM saturates at 1900 a little above mid-stick. Also, I noticed that there seemed to be little response at the lower end of the throttle, which means the mid-stick PWM values are about the same (50microsecs apart) for both systems but the DIMU system shoots up much faster.

Both systems have the throttle factor set at 0.7 currently. Any idea where these differences might come from?

I also noticed in the parameters that the CTRL_MAX is 300 on the analog system and 1440 on the DIMU. Not sure if this has anything to do with it.

The radio inputs were the same.

Will try to add some graphs with the data later today.

Cheers,
Richard

[rput]

throttle radio vs motor pwm on DIMU.jpg (18.74 KiB) Viewed 18407 times

Here is a graph from the scope measurements. Radio throttle input was measured from the telemetry. The I was put to 0 in the attitude PID settings and both systems were strapped to same table.

It doesnt actually seem so different, although you can see the slight S-curve in the DIMU system. Not sure if this explains the difficulty in controlling altitude in manual mode though.

btw, firmware versions are:

DIMU: 7.0-EXPR R399 B1758 - DIMU
Analog: 6.7-BETA R224 B1400

[Max]

Hi Richard,

The throttle scaling hasn't changed, but a lot of the other values, like attitute PIDs and motor output numbers, have (to account for greater resolution when using CAN).

What ESCs are you using? The unsteady altitude sounds like what I was getting on 2 systems with flashed Chinese ESCs. You could try the firmware I posted here: viewtopic.php?f=27&t=3269&start=30#p24486

-Max

[rput]

Thanks Max, I am using ESC32's from Viacopter.

[Max]

CAN or PWM?

-Max

[rput]

PWM, for now.

I seem to have found the issue. I was using very high values for Tilt Angle P (300) and D(20000) as well as Tilt Angle Rate D(25000)

I put the system 'on the ropes' and tuned it from scratch again. Now the values are Tilt Angle P (60) and D(5500) and Tilt Angle Rate (5000).

This really made a large difference and it now flies as I am used to in manual mode. Still some slight 'wobble' when flying against the wind, so I will try to continue tuning (any tips are welcome....

Anyway, I did not think the PID settings for Tilt would effect the throttle response in manual mode, but apparently it does... that is good to know.

Thanks for the help Max!

[Max]

Hi Richard,

The Tilt rate and angle settings do affect overall altitude stability, so that makes sense. I few others have also reported that reducing the Tilt PIDs improves performance with DIMU. Those are large changes though, which seems strange to me. Are you happy with the response to controls in manual mode (no lag/delay)?

If you could try out the other firmware I posted (), starting with default Tilt PIDs, I'd be very interested in your feedback regarding the difference.

Thanks,
-Max