Paul & Sabrina's cheap DIY 144v motor controller

[MPaulHolmes]

EDIT: WELL, MAYBE IT WAS THE CAPACITORS! HAHA!

Well, it sounds like it was the mosfets. Now, does everyone want to hear a secret??? I did a really bad job making sure the mosfets were really flat against the copper heat spreader when they were soldered. The clamps were clamping some much harder than others. I thought that might turn out to be a problem, but I wasn't sure. I've fixed that issue in the version 2 power section, but I'm still not sure that was the problem.

There is lots of good news, though! A bona-fide professional expert electrical engineer fixed up the schematic, using opto-isolation to separate various sections of the board, and did other stuff too. I believe A HUGE source of the problem with the current sensor was the fact that the power section ground was being shared with the weenie control section ground, causing terrible EMI issues. Even if all the mosfets are bad, there is a lot on that sucker that can be salvaged!

[MPaulHolmes]

Hey Ben! Now is the time I wish I would have made the control section separate, only plugged in, rather than soldered to the power section. Oh well! Version 2! ya!

Do you have like a couple pieces of wood, so that the controller can be suspended in the air by setting the bus bars on the wood, with the heat sink facing up?

Next, unscrew the 2 big screws from the top of the heat sink.

Now, the heatsink is screwed to a sheet of lexan. To remove that lexan (and the heatsink at the same time), just pull real hard on the heatsink! haha!

There aren't any screws attaching the lexan sheet (which is screwed into the heat sink) it to the rest of the controller body. It's just that clear caulking stuff keeping it stuck there.

Once the heat sink/ lexan sheet is off, you will see the capacitors/diodes/mosfets.

[taffy]

Bugger! Sorry to hear about its bang! Keep the updates and improvements coming.

[Wherewolf]

You weren't running around with a big metal hook on the roof of the car - trying to snag energy from the power grid were you? It's a metro - not a Delorean! It's a controller - not a flux-capacitor! Just Kidding!.....glad you're ok! Bet we will all learn something from this!

Wherewolf

[esoneson]

Ben,
Can you give a blow by blow of what was happening at the time of the big bang? Might help to figure the cause.
Eric

[MPaulHolmes]

I love debugging. This is so much fun! My wife gave the financial go-ahead for an immediate replacement. I only have the 3 ounce copper, but Wherewolf has been finding sources for 4-6 ounce copper. Maybe we should wait for that. 3 ounce might be OK, but might not. I know 4 is fine. I can't wait until we see the failure points!

[Roger Heuckeroth]

Paul,

Don't you want to the forensic examination on your controller. After all, nobody knows it better than you, no offense to Ben.

A couple thoughts:

One of the advantages to Mosfets over IGBTs is that they have a positive temperature coefficient. This means the hotter they get the larger their voltage drop, so it helps balance the current. The ones with a higher temperature get less of the current. IGBTs are prone to thermal runaway due to their negative temperature coefficient. That makes them hard to parallel, verses Mosfets. So, slight variations in the pressure against the heat sink may not be the problem. No guarantee, but I wouldn't jump to conclusions.

Regarding safety. You may recall that I suggested that the controller take care of engaging the precharge resistor, time delay and then the closing the main contactor. If the controller is in charge of closing the main contactor, it can also open it back up the instant it senses a fault. Failing in full throttle mode with no controller interlock (besides a manual disconnect) is not safe. We all are glad Ben is OK, but we have to redesign for safety first.

Its great that Fran is helping you redesign the control board. Remember my offer to pay for the first batch of control boards for you. Sounds like we may be getting closer.

I hope your not feeling to bad about this failure. We learn more from failure than we do from success.

Roger

[bennelson]

By the way, I hope that everyone realizes that I have always been JOKING about safety.

My car is wired up to just flip the key to kill the controller, and then the contactor.

There is also a main fuse and a manual disconnect.

The car was not in even in gear when the controller went. I was getting ready to leave a parking lot. When I turned on the controller, it just went full blast. I flipped the key off. That's it.

I looked under the hood and saw the plexiglass box was full of smoke.

Paul, what did you put in the heat sink screws with? An impact hammer?

[MPaulHolmes]

haha! No, Ben! No impact hammer. But I did use quite a bit of "non-permanent" loc-tite. Sorry about that. hehe. Interesting! It failed while turning it on. hmm....

Is it possible that the 12v control section was turned off when the contactor or pre-charger resistor was activated?

[MPaulHolmes]

When you shut off the car, the 12v power to the controller turns off before the contactor?