Paul and Sabrina's Cheap 3 Phase Inverter (AC Controller) with Field Oriented Control

[MPaulHolmes]

The circuit I have in mind can measure accurately down to around 10femtoAmp, which is WAY below tingle. haha.

[Astro]

Originally Posted by MPaulHolmes

That thar looks like a sine wave, e*clipse!!

I was thinking of adding a leakage current monitor circuit, with a known large resistance between pack ground and chassis. And if the resistance drops below a particular value, then it would flag an alarm.

Paul, would this circuit be able to measure how much leakage current was flowing or would it just be triggered above a certain current?
Reason i ask is because one of the tests that the authorities here require is a measurement of the leakage current between the pack and the chassis.
Would be handy to know what the value is. Would also be nice to know if it changes in the future.

[MPaulHolmes]

It is on a scale. A resistance of any range you like. For example, 0-10TOhm, 0-10GOhm, 0-10Meg, etc.

I'm done PI loop tuning the PMSM motor at 12 volts. A 5 amp step converges in around 0.5msec-1msec.

Now I'm going to figure out what the offset is from index for which Id = NONZERO, Iq = 0 doesn't cause the motor to turn. Then, save that offset for the life of the motor/encoder pair. PMSM is so much easier than ACIM. it will be very easy to include support for both types of code. I can't wait to see it spin tomorrow.

[MPaulHolmes]

Well, the PMSM motor spins. Good torque control in both directions. I love that you just set Id to zero, and let the magnets do the work. The power factor is almost 1. It's nice and smooth too, and makes a cool sound. I haven't messed with the field weakening yet, but I'll try it.

I also need to get a video of trying to find the rotor flux position for the first time. It worked beautifully. I set Id to 5 amps, and Iq to 0. Since Iq is what is supposed to cause the rotation, and Id is supposed to cause the field strength, the motor shouldn't turn at all. But if you don't know the true rotor flux angle (which for all eternity is a fixed offset), the motor will spin if you guessed wrong. So, I let the guesses run from 0 up to 1023 "degrees" (one complete electrical revolution), and you actually see the dang motor spin one way, gradually slow down, stop (that's the sweet spot!), and turn around and go the other way. So, then I just read what offset it was when the motor was in the middle of its stopping, and then from that point on, commanded Id to be zero, and Iq to be throttle position. Then the dang thing just works! And the response is fast! Bang reverse! Bang forward! Now move super slow! I command you, Torque! haha. It doesn't lose sync like was able to happen with the ACIM at first.

[dave koller]

NICE !! And as they say in the old country --- Nice!

[MPaulHolmes]

Here's a video of testing a brushless DC motor for Rod's motor with field oriented control:

https://www.youtube.com/watch?v=mzLyc_yJI2Q

[MPaulHolmes]

I'm making a new board that's going to be primarily surface mount. I found out that there are places that will populate it for like $100 per board in quantity 10 down to $17 per board in quantity 500. And surface mount parts are cheaper.

The new board is going to have 8 pwm channels. 4 high and 4 low. You can run the controller as 3 phase controller with a boost stage on the front end, a DC controller that's rated for like 1600-2000 amps??? (if I can get all 4 IGBTs to current share well), or a regular old 3 phase controller without the boost stage, or a high powered solar MPPT tracker/charger, or a grid tie inverter, or non-grid-tie inverter for a house (3 phase or single phase). Some of those need either 1 or 3 inductors though. I want to figure out how to wind my own very high current inductors. The inductor I have is made of laminated steel sheets, and I think a roll of copper sheet wrapped around it. That's rated for 200amp. I would need a 400amp inductor though. That way a 72v or 144v car could be boosted up to 288v or whatever, and you could run the millions of cheap 20HP 230VAC motors on ebay. The dimensions will probably be around 14" x 10" x 5".

I'm also going to try doing a water cooled version. I had never heard of those ball nose mill bits, but eclipse told me about them, and so I'm going to get a base plate later (in the future someday when I save up enough money) and mill those round channels, and epoxy some 1/2" tubing inside the channels under the 4 600v (or 1200v) 600amp IGBTs.

The microcontroller is the dsPIC33ep512MU810 which is good to 70MHz, and can do 12 pwms, and 8 simultaneous A/D conversions. It has 100 pins, so I'm going to include some optional future use inputs and outputs.

[Astro]

Originally Posted by MPaulHolmes

I'm making a new board that's going to be primarily surface mount. I found out that there are places that will populate it for like $100 per board in quantity 10 down to $17 per board in quantity 500. And surface mount parts are cheaper. ...

Paul are you going to populate the older boards for the beta test team or are you going to hold off for the revised board?

Will you be keeping the driver circuits the same and mostly making changes to take advantage of the micro controllers extra capacity?

[MPaulHolmes]

The thru hole boards are populated already. The new one is really done just with the 4th igbt in mind, so you can do a boost input stage. Since your motor is already wound for a lower voltage, it would sort of be a waste. There would be an empty part of the board for no reason. To be honest, I would prefer just to keep things mostly as the current board, since I prefer thru hole , and like the diy model. But there are only surface mount chips with 8 pwms. So, the people who want to use those 700v motors need a boost stage.

[freebeard]

The take-away message for me is:

The dimensions will probably be around 14" x 10" x 5".

Questions:

1. Will the water cooling work with oil cooling as well?
2. What controls the controller? CANBus? Bluetooth?

What I like right now is E.V.I.C.