The motor used a Johnson SP540
The propellar used a folding type from graupner 7.5 x 4
The batteries used for this test: Sanyo 2000 NICD 7 cells, and KAN1050 8 cells
Here is my messed up table with the test setup.
The shunt resistor used for the current mesurements is a 1.000mOhm (one mili ohm)
The digital scope used was lended special for this task.
Sanyo 2000 Nicd HOT battery, Peak current = 84A, constant current after speedup = 24A
KAN1050 NIMH Cold battery, Peak current = 60A, constant current after speedup = 20A
KAN1050 NIMH HOT battery, Peak current = 70A, constant current after speedup = 25A
Here we zoom in on the AC voltage ripple, the ripple is about 1.5Khz
We zoom into the area with high frequency noise, the 4MHz is strong.
Lets zoom in even more.. See a 100Mhz almost as strong. nice.
With 100nF capacitor: The motor voltage ripple, note most high frequency noise is gone !
See the 100nF is soldered to the case and black ground wire,
the red wire on motor is directly connected to the metal case of the motor,
This sounds crasy to somebody but this is the best way to remove noise,
you see the black wire is the power signal that is PWM switched by the regulator later.
Then I added a Diode directly over the motor in reverse direction,
Now see the problem is gone !
Here is the same signal measured at the regulator end, no diode at motor.
Also at the regulator end, but now the Diode at the motor has been added.
See here is the diode mounted at the motor to solve all noise problems.
The diode can be any power type you find in your jumkpile I used a 1N5401.
Don't forget to connect the positive terminal to metal can of motor.