Paul & Sabrina's cheap DIY 144v motor controller

[apowers]

What if you blipped the accelerater to a certain duty cycle and let off the pedal. This would open the contacter and no voltage to power section.
But what if the controller did not return to 0 duty , for whatever reason,
and then you pressed the pedal closing contacter sending all the pack voltage into the modulating mosfets. Would that be too much of a surge?

[adamj12b]

Originally Posted by apowers

What if you blipped the accelerater to a certain duty cycle and let off the pedal. This would open the contacter and no voltage to power section.
But what if the controller did not return to 0 duty , for whatever reason,
and then you pressed the pedal closing contacter sending all the pack voltage into the modulating mosfets. Would that be too much of a surge?

I dont think so. The caps would suck it up so fast that it would hardly reach the fets. The fets dont care about voltage either. as long as its within their range, and thats all they would be seeing unless there is a short.

-Adam

[apowers]

Ok. I was just thinking that if the controller was stuck at half or full duty and you closed the contactor on a stalled motor that would a big amp draw.

[Dwffy]

Hi Paul,
The loose wire was on the throttle pot, and yes for a while I had the ground connected to the vehicle ground and not the high voltage ground, but never took it for a ride like that, although I did run the motor in neutral to see if the throttle controller was working. All the gate resistors are open but no physical damage to look at. They look fine, test dud. There was also no chance that moisture got in. Diodes all test out OK. Removed one cap and it tests out ok too so I presume that the others will be ok as well. As I mentioned before all was working 100 percent until I tried to run it on 120 volts. I did use heavy jumper cables to connect up the second bank of batteries because they are mounted behind the seats in the cab. The cables are in good condition. Sorry that some of the pics are fuzzy.
Thanks again,
Doug.

[MPaulHolmes]

Did you resolder those gate resistors, or did I do that? I can't remember now. If I did it, then I must have removed the control board for some reason while I was building the controller. There were a couple control boards that I had removed in the past so as to fix one thing or another. I wonder if your control board had "walking wounded" components on it then. Recently someone told me that after something like a mosfet failure, a voltage spike through the gate resistor or through the source leg can partially damage something like the mosfet driver or the hcpl-4504 optocoupler. I wonder if the driver was only partially functional. I don't remember anything wrong with yours, but the board has been removed (probably by me). 120v shouldn't have been a problem if every 200v component was working as it should.

For all the mosfets to fail in unison is interesting. The only other failure I have seen is with only 1 or 2 mosfets blowing up. Ben's first controller, and Hondo's controller, which had a gate resistor break off from vibration. For them all to fail simultaneously to me means they must have all seen a voltage spike over 200v, or the gates all saw a spike over 30v, or something else weird. Is it possible that B+ and B- leads got reversed on the last time the 120v was hooked up? That would cause a short circuit which would blow them all up at the same time.

I think I have most of the parts to make another one. It would take a few weeks, though. If that's OK, I can get busy, and I'll try to be more careful this time.

[MPaulHolmes]

Hey! Fran suggested that the current sensor may not have been working. That would make a lot of sense, but I don't know why it wouldn't be working though...

[MPaulHolmes]

Adam has an awesome explanation, he's just taking forever to type it out. I just wanted the world to know that.

[adamj12b]

Well I think I've figured it out.

You said that your contactor is switched by the microswitch on the throttle pedal. What will happen when you apply throttle with no power is the duty cycle will rise to increase current. Current wont increase though because there is no voltage coming into the controller. When you get to the point of the throttle that closes the contactor, you are at 100% duty which means only the mosfets are conducting. So, now the ugly. When that contactor closed, 100% of the battery was connected to the motor. this means huge current and it happend super fast. If the current sensor was working, it would have tried to limit current. Also, Paul and I think that the first time it happend the fets failed shorted. The second time the contactor closed and finished them off.

This all explanes the fets being blown and the rest ok. If it was running the old firmware, that doesnt check for current sensor for safety. The new firmware > 1yr old, would have disabled the controller before it could even do anything.

Another thing. If you could take out the contactor and take it apart and look at the contacts, you will most likely find some serious pitting on it. If there are pits there, my philosophy is true.

Sorry to be the bearer of bad news

-Adam

[jyanof]

Originally Posted by MPaulHolmes

Hey! Fran suggested that the current sensor may not have been working. That would make a lot of sense, but I don't know why it wouldn't be working though...

There was that episode I had last summer. I had taken the controller apart to add those extra diodes and somehow screwed up the current sensor connector to the board. One of the pins wasn't connecting properly.

Thus, even a small throttle command would ask motor current. BUT, the control board would sense ZERO current (since the sensor is not connected) and would increase the power (PWM) endlessly until you took your foot off the pedal. Fortunately, I had only 'blipped' the throttle and noticed the motor current rising ridiculously fast for that blip. Like, near 1000A in less than a sec.

This is when we added the software check to make sure there is a current sensor signal. If not, it shuts down. If an earlier version of the software was running, it wouldn't have this feature...

It's pretty reasonable that a bad connection could vibrate loose. If the sensor is still connected, you could check for continuity with an ohm meter to make sure there is a connection.

If the fault happened at the lower voltages, the motor current wouldn't rise that quickly, or would only rise so high and could still be in a reasonable range. Perhaps 120V pushed it over the edge.

[apowers]

Adam thats what I said. My thinking was that the potentiometer was loose.
Tight enough to accelerate but came loose and did not return.