Paul & Sabrina's cheap DIY 144v motor controller

[MPaulHolmes]

It works again! I de-soldered the 2 bad mosfets, and put 2 new ones in, and it works good now. I changed the software. I'm testing how fast current drops under different circumstances, and how fast the software responds to it. I'm getting fast at switching out bad mosfets for good ones! That's not a skill that you want to brag about.

[bennelson]

Originally Posted by tech2

Another random thought:
If you used ten 12 volt batteries to run the car...... , you should not need much in the way of a battery charger

That's right. There are ways of building basic battery chargers for cheap once you are around this voltage.

Flooded lead-acid batteries would be the best, as they are very forgiving with how they are charged.

With a higher voltage controller and more batteries, a person would also need a higher voltage charger (or lots of little chargers, which gets expensive too)

Controllers and chargers are two of the more expensive items on an electric car.

An Open Source controller AND charger would really help bring the cost down. There's plenty of info on "bad-boy" chargers on the web, but that's the start of a whole 'nuther thread.

[tech2]

Well hi bennelson!
I have been following your progress for quite some time, in fact thats how I wound up here on this forum. How's the metro doing? Broke anything lately, or got any more tickets HA HA LOL
Here is what I had in mind:

[MPaulHolmes]

Well heck, I was asking how to get a straight DC voltage source, and there it is! hehe.

I was testing the controller with a new program in the garage. I think I've been running the motor backwards all last year. I switched M- and B+ on the motor, and it sounds a lot smoother. I did the same moron mistake as earlier, but the voltage was only 12v, so I think that saved me from destroying them i'm guessing. Maybe I damaged them, and I only have 1 cheap one left, so I'm going to buy a few more cheap ones to work out the bugs with the software.

Sabrina is going to help me make a protocol for each time I start the controller and turn it off during this testing time. I keep forgetting to plug in the wires that power the control board each time I bring it back from the room where it get's programmed.

[MPaulHolmes]

ya, the mosfets are acting really weird. It still works, but it's acting strange. This is the last time I do that mistake. The wiring is really ugly out there, I'm not going to lie. I need to really clean things up, and we're going to draw up a step by step procedure each time I re-hook it up after programming it. I've not been as sharp lately, since I haven't gotten that much sleep! anger!

[MPaulHolmes]

I'm pretty sure I can now eliminate the problem of having to turn on controller low power section before pre-charge resistor. Just 1 15kOhm resistor for each mosfet, from gate to source. That will keep gate at 0v when the controller is off, rather than having it floating. I think that gate was floating, and allowed power to go from the capacitors from drain to source. The current must have been huge for an instant, killing them!

Mosfets are a funny thing. Sure, put 200 amps through them, sure, do that at 150 volts! No problem! But don't have gate to source voltage go over 20v for an instant, and don't allow a large current surge for even a few millionths of a second! They are so touchy! Geeze!

I have 1 mosfet left. I shall be super careful with it. My wife is buying me a couple more practice mosfets. They're 3 bucks each so it's all good.

[Intrigued]

Whew! Learning the hard way sure is a pain! I'm beginning to understand why the commercial controllers look so blasted complicated!!! :-(

GET SOME SLEEP, MAN!!!!! :-)

[MPaulHolmes]

I removed the 2 bad mosfets, and added the last mosfet for testing. I also added a 15kOhm resistor from gate to source. I also added a line to the software that initializes to 0v the PWM output pin on the micro-controller.

I now believe that you can turn the controller on in any order that you want. Either pre-charge resistor and contactor first, or 12v power supply first, and then pre-charge resistor and contactor.

One of the reason the controller was acting so weird before switching out the mosfets is because the battery was completely dead that was powering the controller! Totally dead! Under load, it was sagging to about 4 volts. yikes.

I like the idea of the voltage range of 12v-156v because it lets me test it at 72v, which is all I have.

[Intrigued]

<:Taps for the dead battery. May it rest in a recycle bin!:>

[MazdaMatt]

Paul, for surface mount soldering we use solder paste instead of solder wire. It is a syringe filled with a slurry of flux and tiny solder balls. You squirt a tiny drop on each copper pad and place the component down with tweezers. The paste holds it still and you can heat up the pads till the solder balls melt.