Compensation for Battery voltage depletion

[Paulmann]

Hi,

when the battery voltage gets lower we have to compensate wither higher throttle.
Wouldnt it be nice to have the same flight / hovering points independent of battery voltage ?

-- the failsave decay with a fixed throttle value would be less battery sensitive.
-- the closed loop gain of PIDs isn t constant . It goes down when battery voltage falls.
-- for altitude hold the battery voltage change is one systematic error source.

I could be a nice feature to compensate.
Maybe to multiply the motor -outputs with a correction factor which is dependant from the VBAT - Measurement.

[Hamburger]

The third value of the governor feature does that already.
It was originally for helicopters but works for multis as well.

[copterrichie]

We really need a road map as to what can and can not work in conjunction with each other. I don't have a clue how to go about this this other than user input into some sort of database.

Code: Select all

 /* Governor: attempts to maintain rpm through pitch and voltage changes
     * predictive approach: observe input signals and voltage and guess appropriate corrections.
     * (the throttle curve must leave room for the governor, so 0-50-75-80-80 is ok, 0-50-95-100-100 is _not_ ok.
     * Can be toggled via aux switch.
     */
    //#define GOVERNOR_P 7     // (*) proportional factor. Higher value -> higher throttle increase. Must be >=1; 0 = turn off
    //#define GOVERNOR_D 4     // (*) decay timing. Higher value -> takes longer to return throttle to normal. Must be >=1;
    //#define GOVERNOR_R 10    // (*) voltage impact correction scale in 0.1 units. Higher value -> more compensation for voltage drops. normal is value 10 <=> 1.0; 0 is off

[copterrichie]

Suggestion, it would be very nice if the third define becomes part of the VBAT section of the config.h file.

[Hamburger]

Currently the entire governor thing can be switched via auxN

[jy0933]

why not just support i2c esc and use rpm feedback for it.. aka direct rpm control?

[copterrichie]

why not just support i2c esc and use rpm feedback for it.. aka direct rpm control?

Now that is something I REALLY would love to see happen however, many say the i2c esc are not a good way to go. Wish I had a few to experiment with.

[balrog]

why not just support i2c esc and use rpm feedback for it.. aka direct rpm control?

Now that is something I REALLY would love to see happen however, many say the i2c esc are not a good way to go. Wish I had a few to experiment with.

I've been doing that in my tricopter fc, the discussion is at https://github.com/sim-/tgy/issues/6 , I was basically reading the rpm every 4s and slightly adjusting the throttle gain every time. While I'm learning to fly and use no telemetry yet I decided to leave it alone though so I could see when the voltage was low.

The serial protocol I use (dumber than i2c) is not the prettiest but I think within the limitations of a generic solution that will work with every ESC that supports SimonK-firmware it's not terrible. It uses 4 wires instead of three for two-way comms + BEC power like usual.