Paul & Sabrina's cheap DIY 144v motor controller

[mike_lera]

Jack
thanks for info
are you telling me that you leave power on the board when you re-program it?
your way ahead of me i have never programed a chip not even a pic as for C thats going to be fun

[jackbauer]

yep the board stays powered up during programming. I just woundnt have it hooked up in your car when you reprogram as the pins may toggle! Just let me know if you need any help setting up pony prog.

[mike_lera]

Thanks
might tak a fue days Honey dues are next in line
Mike

[squiggles]

Originally Posted by MPaulHolmes

The gate resistor output will NOT look like a pwm signal because the software is wanting to get feedback from the LEM. You might see something weird if you are lucky, since I think you have Joe's code, so it's not purely throttle proportional to current.. The only way you will know if it's working is if you either put software on that is purely throttle proportional to pwmDuty, or build a whole power section and run a smallish motor. hehe. I'll mail you a chip asap.

Surely the LEM feedback can be emulated with a pot and a bit of thought. It is just giving a variable voltage signal.

[MPaulHolmes]

Hey squiggles is right! Or with crack and a bit of a snack. hehe. you can fake 0 amps by having 2.5v output of the "lem" (which is secretly your fake output). 500 amps is 3.54v. So, you need about 1 volt swing to simulate the 500 amps.

[MPaulHolmes]

So, I had finished another beta controller, but with the $8 super fancy schottky diodes that are rated at 120 amps instead of the other ones that are rated at 60 amps. It was going to be shipped out already, but never mind, it's a long story. I made a little circuit that Fran had suggested to measure voltage spikes. It's so dang freaking elegant and fancy and awesome and dang it was awesome and cool. Joe told me about it. Voltage spike magnitude is a function of current. It's independent of pack voltage. So we can come up with a graph of voltage spike magnitude as a function of current for current = 0 through maybe 509 or so. haha. Well, it might not be "voltage spike magnitude". It would be better to call it the "peak voltage that the controller ever sees from M- to B-", since the pack sags the more current you suck out of it! But who gives a dookie?! All that matters is HOW HIGH IS THE VOLTAGE GETTING??!! We just need to stay away from 200v. Well, I guess 200v is even OK a little every now and then, since they are "avalanche rated". I totally know what that means. I'm not lying.

Now to the results!

OK, First I made the awesome little circuit. Then I precharged the caps and watched on the dedicated volt meter what the max voltage was from Drain to Source. The volt meter started off at 78v. Then I drove around and finally managed to get the peak voltage all the way up to 85.0v I think. It was either 84.5 or 85. I have a bad memory. hehe. That was probably a few hundred motor amps, but I am not sure. I was in 2nd gear at low rpm on a small uphill. So, maybe figure for a spike of 7 volts above max pack voltage? Not bad. Joe's controller has M- B- and B+ all out the same side. I wonder if that makes any difference compared to how it is with the controller I was testing as far as spikes go.

If 7 is the spikes above max voltage, then it's officially safe to drive around at 193v. hahahaha. Just kidding.

[MPaulHolmes]

Originally Posted by jackbauer

yep the board stays powered up during programming. I just woundnt have it hooked up in your car when you reprogram as the pins may toggle! Just let me know if you need any help setting up pony prog.

I programmed it in the car today. I just made sure the main contactor was open. I wonder if you could reprogram it while driving? That would be pretty funny.

[jackbauer]

Thats interesting. I'd just be concerned as to whether the circuit can detect really fast transients that occur at switch on and switch off ? On a related matter my gate ringing seems to have disappeared with the new power section.

Best bet for on the move reprogramming is at a traffic light. Maximum current ramp. Hammer down. Let em eat dust

[MPaulHolmes]

I used an "ultrafast rectifier diode" and a cap with a very low ESR. The cap would definitely catch the electrons in the very fast transient spikes. And the good thing is, the capacitor stays charged, since it's so big. So it lets energy in, but basically doesn't let it out.

It's like a crab trap. The electrons get pushed by the higher voltage past the diode and into the cap, and then they say "Wait! I don't want to live inside a capacitor!" So they attempt to leave, but the diode won't let them go back. hehe. I love getting crab. I'm allergic to it though, but it's super fun to pull up the rope and make guesses as to how heavy it is, and then flip it into the boat before any of them fall out. I hate it when a big star fish hogged all the crab bait.

Oops, the "positive fake electrons get pushed past the diode" hehe. Stupid +/-!

Hey we can use this to find out the "stray inductance". I was just reading about it:

V_overshoot = L_stray * Di/Dt

So we can know Di/Dt without any trouble. Just look at the waveform on the oscilloscope. We can easily find V_overshoot, so then we can find L_stray!

[jyanof]

BTW,

I hit the 1000 mile mark the other day! I've been testing this controller since July and have since racked up just over 1000 miles. We've had a lot of software revisions (and I'm very pleased with the latest version), but the hardware has definitely held up.

The one glitch was with the current sensor connector while I was installing the extra diodes to the bottom of the control board. When I got the controller all back together, the current would just rise uncontrollably (way beyond the 500A limit if I let it). Somehow, the connector was damaged and the signal wasn't reaching the control board and the code I had just thought there were 0 amps, so it kept on increasing PWM. Now, the new code constantly checks to make sure the current sensor is connected and it'll shut down if it detects otherwise.

So, if anyone is writing their own software, I'd suggest including such a feature as well.

Someone else asked some questions about how the power board looks after a lot of use. I didn't take the controller completely out of the car when adding those diodes, so I didn't get an excellent look, but everything looked as good as new, from what I could tell.