Paul & Sabrina's cheap DIY 144v motor controller

[squiggles]

Originally Posted by MPaulHolmes

Then eventually I'm going to take about 12-15 of the new fancier 230amp 200v mosfets, and do a synchronous rectification version, so the freewheel mosfets will have way more current carrying capacity than what I'm using now.

Paul, do you have any references for the design of the synchronous rectification circuit?

[MPaulHolmes]

Hi Squiggles! I've had quite a few discussions with people on the EV Tech list. Here's what I've learned:

Replace the freewheel diodes with mosfets. The diodes' ANODE (input) is replaced with the mosfets' SOURCE leg. The diodes' CATHODE (output) is replaced with the mosfets' DRAIN leg. Let's call the freewheel mosfets gates "GATE2", and the regular mosfets gates "GATE1".

You need a microcontroller that has 2 separate PWM outputs with programmable deadtime. The AT90PWM3B was made just for this purpose. Let's say PWM1 is connected to GATE1, and PWM2 is connected to GATE2.

Let's also assume we are starting with GATE1 High, and GATE2 low. So, the regular mosfets are in conducting state, and the freewheel mosfets are in the BLOCKING state.

Now, when you want to switch to the state where the regular mosfets are off and the freewheel mosfets are conducting, do this:

Bring PWM1 output low. The chip waits for a specified dead time (like 1/2000000 sec), and then PWM2 output goes HIGH.

When you want to switch to the state where the regular mosfets are ON and the freewheel mosfets are BLOCKING, do this:

Bring PWM2 output low. The chip waits for a specified dead time longer than the trr of the intrinsic diode (like 1/2000000 sec), and then PWM1 output goes HIGH.

Repeat this process each cycle. So, I don't think there is much external circuitry that needs to be added. Just an extra driver circuit and gate resistors. The power section would have to be modified so that there would be mosfets on both sides of the heat spreader.

[squiggles]

So it is a half bridge with the motor in parallel with the high side switch (FET).
Do you need a special gate driver for the high side device?

There are half bridge drivers specifically for the purpose that have the dead band built in and a high side gate driver. Just a bit concerned about forcing the high side FET on with the Source voltage going to a higher potential. Hopefully someone can tell me I am wrong

[MPaulHolmes]

I feel worried about that too. I hadn't thought about it before!

EDIT: Wait just a darn minute... You need another isolated 12v supply. crap.

[squiggles]

Get a copy of the data sheet for the IR2184 from International Rectifier.
It is a half bridge driver that can switch the high side device in a system with 600V supply. Problem is these expect the load to be between the high and low switches. In our system the inductive load is in parallel with the high switch and I feel (without proof) that that creates problems as the high potential appears on the source of the high side device.

Hopefully someone in our little community can tell us that it will work anyway....

I don't think an extra power supply is required.

[jackbauer]

Parts kit arrived today and i've started stuffing the control board. Very impressed with the layout and quality of the pcb. I've seen boards layed out by so called professionals look much worse! Well done Paul. Only problem i've came across so far is the silk screen for R12 is obscured by a via. Tiny issue but worth noting for the next rev. Hope to get it built fully tomorrow then start driving some serious igbts

[drbobwoolery]

I just got off the phone with a salesman from Vista Metals. He is sending me enough annealed Manganin sheet to make 3 or 4 1000 A shunts. What size bolts are you using on the main current connections on the controller? I will drill the outer holes to match that, and the cable will bolt up with the same hardware, to either the battery or the motor side.
I didn't get a price on more of the sheet, and it is an unusual alloy to find in this big a pieces. The temperature coefficient is specified at 15 ppm/deg C, either way.

I'll try to locate the voltage taps for 110 mV at 1000 A, so that a small divider can readily bring the output to 100 mV, once someone turns up a standard to calibrate it against.
BTW, at one point there was an issue with noisy current signal from the Hall sensor. A gain of 50 differential amplifier looking at this shunt should be much quieter.

I'll try for 110, but will accept 50 to 200 mv for 1000 A. This will be large current in a thin sheet, monitored at two points. An analytical solution is certainly possible, but would involve resurrecting too much mathematics I thought I understood in 1962.

If necessary, I can silver solder the taps, but I think that 1/4 - 20 brass bolts with crimp lugs should do fine. We'll see when the mail comes.

[MPaulHolmes]

Awesome Jack Bauer! Hurry, you only have "24" hours to complete it before the end of the world. hahaha. Actually, I didn't do the layout. It appears professional because a professional did it for me. I can't wait to see how the IGBT approach works out! I want to try that too.

Hey Bob! I'm using 5/16" holes in the bus bars. If it is a thin manganin sheet, how does the mounting usually work? Could it just replace the section where the LEM usually is? I've never seen a shunt. Is it sturdy for bolting?

[jackbauer]

The following takes place between 2pm and 3pm :

http://www.youtube.com/watch?v=PiND9UajAi8

[MazdaMatt]

Paul - It has been a long time since I've been around. I had a look at your website. 600 bucks for a 144V 500A (96 fricken electric horsepower!) is not to shabby! Good work!