Pretty cool. I made a comment on the SSW version of this thread. I admit I have been thinking similar thoughts on how to add rpm--not just keep them the same. When I look at my airspeed indicator on my E-Nobler Arf, I find than my minimum airspeed is just about the climbing 45 degree point on the reverse wingover for example---and this is with the rpm actually holding constant.  What you need to do is actually add power (=increasing rpm setpoint for example) during the initial part of the climb, and then begin to throttle back sometime after passing the 45 degree point.
How to do that, either by monitoring where the nose is pointed, or g-forces, or adding positive feedback to the power as you mention on the SSW thread sounds like a trial and error effort. I don't really have the programming tools to make my own throttle control (which would effectively change the setpoint depending on some readback), so for now I am basically in "imagineering" mode!
What I think is important is to first understand what a well-tuned power setup is actually doing (even IC).

Hello Igor,
So do I understand correctly? You have used an accelerometer in the nose-to-tail direction to add some throttle when the airplane decellerates, and vice versa? Cool!
If it is something else, then what?

bravo!
    Dean P.

Hi Dennis,
Yes, a pitch rate gyro will add power any time a "G" loading is being adde.
On the other hand, a simple circuit that adds power any time the elevator is being moved quickly will add some power that leads the highest drag load.
Fore and aft acceleration represents yet another different kind of information.
My bet is that some combination will be the eventual result ...
Most importantly, though I belive that the most robust way of implementing this is with something like the existing governor. The new inputs will be used to adjust the RPM setpoint, rather than as the primary throttle feedback. This is because the governor is still the fastset control loop possible.

This is interesting stuff!

Dean

I asked one of my "Dark Arts" e-flier friends about trying to get the equivlent of a 4-2-4 break in an electric.  He speculated linking the throttle input to a gyro...

Yes, it will work, angular speed equal to lap speed can be base for speed regulaton, but it needs to solve two things.

1/ you must separate yawing (from presecssion, wind, line swinging) from lap time, it needs to make som low-pass filter .. it will be critical point of the system, because speed changes in corners are sometimes quicker then yawing. I do not say it is impossible, just I think it will need lot os time and effirt to make that software. On opposite side, that other componet (blocked yawing) can be used for rudder inputs to damp yawing. That is also possible. I had gyro on my model (look picture), I did some tests and may be later I will continue, for present time I think system on base air speed is better than speed constant to groung. But I am not very sure what we really want - constant speed to ground or cosnatnt speed to air.

2/ the sensitivity of actual gyro units is ot very stable. It will probably need to change gain very often, may be I had very simple unit, but it really needed to change it permanently and biggest problem is, that you must do a test flight, because you do not see its effect on groung (like RPM which you can measure by tacho)

It is possible to have a runaway that is very much analogous to the classic 4-2-4  setup when things are just a teeny bit off setting.

Yes, I think it happens if the setting is more than respodn to slippage. Or by othe words this system cannot add more rpm then what the setup loses because of slippage.

Do you limit the change in RPM setpoint to some small amount, no matter how large the change in absorbed power?

I do not have implemented any limits. This is first version and right now I see several points I must solve. The device now is constructed for range 0-50A and does not work with voltage. It has two points - for 0A and for 50A and that make the response sensitivity. Model on video uses curren only from 6A to 10A and those two end points are made to have proper number 4100 for 6A in level and 4600 for 10A. The only problem is, that if I want make proper slope for regulation, I must measure motor, prop and battery and then calculate proper numbers. I will need more experience to make it better. I will see. The first point will be implementing of voltage measuring, becaut failed with old worn-ot battery - the slope was too strong for low and especially falling voltage under load and it started to oscillate strongly.

Do you guarantee stability by limiting the slope of RPM versus Wattage consumed?

As I wrote - no limits yet.

Is there a corresponding RPM setpoint reduction when the power needs are lower than average?

Still - no limits. The only real limit is, that the esc does not brake in this mode, means minimal power is 0 power, prop must brake itself.