fdp2532 fets pulsing throttle prob

[solarbbq2003]

i have a customer in italy ( antonio ), he has modified a 20amp controller (36v/48v) crystalyte, he changed the standard fets to the 150v fdp ones,
he using a 408 motor with 4 lots of 24v lithiums in series ( cobalt oxide ) getting about 115volts fully charged. He's getting about 60km/hr top speed in 26" rim. But he is getting some pulsing of the throttle at certain speeds, where the motor will be revving up and down of its own accord.
Anyone got any theories on this one or have found similiar thing happening?

 

[The7]

The other components (diving transisiotrs and capacitors etc) in 36/48V controller are not able to stand 115V and the controller will behave abnormally.

 

[solarbbq2003]

oh I forgot to mention he changed the three large capacitors to 200v,
ok so drivers are the cause, thanks for that

 

[solarbbq2003]

it appears that the problem only occurs between 25-30km/hr, below and above that range its ok, anthony tells me its not too bad of a problem though, I hope he will put some posts up soon, heres a bit of a guess at what might cause it ( probably way off the mark): perhaps the pwm frequency at those speeds causing some sort of interference with drivers or something else, and some sort of feedback loop gets set up, and the pwm does a bit of up and down variation??? just a guess, might be solveable somehow?

 

[fechter]

That's pretty strange.

The only feedback loop I can think of would be in the current limiter circuit. He might try putting a capacitor between pin 1 of the KA3525 and ground to try and dampen any oscillation. Wild guess, around 1uf.

Another test would be to hang a voltmeter across the throttle power to see if the +5v is remaining steady during the oscillation.

Is the pulsing affected by load? (any different when climbing a slight grade or going downhill)

The only other thought I have is a possible 'cross talk' between wires, especially the hall and throttle wires. If the wires are bundled closely together, they act like a capacitor and can transfer a signal. Try unbundling any wires to get the signal wires (hall and throttle) away from the high power wires. Even a few millimeters of separation will greatly reduce crosstalk.

 

[solarbbq2003]

ok thanks for that will pass along not sure if different when under load will find out

 

[solarbbq2003]

sorry to report antonios controller died, pulsing problem got progressively worse, he will take apart see if can find what happened