Fuel pumpmotor should have simple control system?

Supposing I have a 700W back-up fuel pump in an aeroplane (my personal plane is a Cessna, but I am enquiring for friends who fly airliners)
Do you agree that the control of the pump motor (DC brushed motor) should not involve a dynamically complicated feedback loop. After all, all that is needed is to rotate the pump motor at 8000 rpm in order to pump the fuel at the required rate.
Therefore, would you agree that the control algorithm should be
1...Place 20V at the motor terminals and then 2000ms later read the motor speed...
2...If speed is greater than 8000 rpm then decrement the motor voltage (decrease motor speed)
3....If speed is less than 8000 rpm then increment the motor voltage (increase motor speed)
4....delay for 250ms
5...Go back to part 2.

Do you agree that this is all that is needed?...and that this could be done with an SMPS providing the motor voltage.....and by "Increment the motor voltage", all one does is increment the SMPS duty cycle.

Does this sound like the way forward?

I'd agree, assuming the SMPS can handle a reactive load like a motor.


I'd modify step 1 - place 20V at 100% PWM......................

"100% pwm"....sorry I don't understand...there is no drive...its just a brushed dc motor fed by a buckboost converter.
Why is a reactive load, in your view, a problem for an smps?

I think a better way would be to monitor the fuel pressure and use that for the pump speed control, that way worn out pump seals could be compensated for. The voltage required for the fuel pressure can be checked as a maintenance item to give an indication of pump wear.
Frank

I would agree with Frank. The deciding factor might be whether the back-up pump is in series with the main pump or fuel is diverted between one and the other if there is a failure. In series the constriction of the main pump would have to be taken into account and it's method of failure might influence it's flow capacity. Measuring pressure would compensate for all scenarios.

The algorithm seems OK though. As the variation in speed would be over a small range, I would consider a linear power control system.

Brian.