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

[MPaulHolmes]

Let's face it, AC is better than DC! An AC motor will last forever, since there are no brushes to wear out. You get regenerative braking for free, which both recharges the batteries, and saves the brake pads! And the new lithium batteries can absorb huge currents. The main problem is cost. What a shocker!

With Field Oriented Control, you get to control the torque of an AC motor with a PI loop, like with a DC controller. This makes it very well suited to EVs. It's sort of confusing, but I've been reading about and I'm learning.

A high powered 3 phase AC inverter with field oriented control can cost $8000-$10,000. There already exists an open source inverter that has a parts cost of $2500. I would like to drive the price down further, and also it just sounds fun! hahaha. I've got very good help too. It will happen.

One thing I hope to do is take advantage of the surplus IGBT market once a process comes together, so people can make their own very cheaply. I would like to see the parts cost be under $500 for an inverter that will get a car on the freeway. Also it will have regen.

Also, the parts count isn't THAT much higher. There are no free-wheel diodes like in a series wound DC controller, but you need a 3 phase bridge, which is 6 IGBTs (or 6 banks of mosfets or whatever).

I'll first do a small one, that can drive a 1/2 HP 3 phase AC motor rated for maybe 220v/440v. Then I'll scale it up. I will probably do a more expensive scaled up version, then try to make it cheaper. My electronics background is not good, so I have to keep it simple.

Here's what I need:
Most of it is from Digikey/Mouser I think.

1) PICKIT 2
2) IRAM136-1060B-ND
3) dsPIC30F2010 microcontroller
4) U.S. Digital encoder, model E3-500-500-IHT. I don't know where to get this. Maybe I'll get a different one.

I should come up with a parts list of what I need to get moving! If anyone wants to donate to get those parts, I won't stop them. As usual, I get payed on the 10th, and am broke on the 11th. hahaha.

[Daox]

That would be AMAZING!

[MetroMPG]

I should come up with a parts list of what I need to get moving!

No, no, no. FIRST the project needs a nickname!

Just kidding!

Paul, you're an EV rock star.

[esoneson]

The one thing I want to know is.............where do you find the time?
You know, you can't do this in your sleep! :-)

[TheSGC]

Wow, another controller!

Don't even bother with MOSFETs, go all out IGBT's. MOSFETs are decent in controller up to 156 volts, anything above that you need IGBTs to handle the higher voltage, and I believe the pricing actually goes down when you get into that high voltage range.

[bennelson]

Well, here comes another 1000+ post thread....

A friend of mine is planning a built-from-scratch three-wheeler - all-electric, running AC or bust. This might be just the thing for him.

[NiHaoMike]

Originally Posted by TheSGC

Wow, another controller!

Don't even bother with MOSFETs, go all out IGBT's. MOSFETs are decent in controller up to 156 volts, anything above that you need IGBTs to handle the higher voltage, and I believe the pricing actually goes down when you get into that high voltage range.

For low powers, MOSFETs are cheaper than IGBTs. Only at high power does IGBT become cheaper.

A good start would be one to power a small 1/4HP motor for the power steering pump or a 1/2HP A/C compressor. Then move onto the main motor.

BTW, small motors generally do not need resolvers. If you use a common industrial induction motor, you can just lock the transmission into one gear (maybe even remove the unused gears to reduce friction) and use the speedometer signal as the resolver.

A lot of industrial motors can be wired in different ways. For a common 208-240v/480v motor, there are 4 ways to connect it - 208v Y, 208v Delta, 480v Y, and 480v Delta. If you're going for single speed, you'll probably want it wired for 208v Delta. Or use contactors to dynamically switch between the connections. The two sets of windings are wired in parallel for 208v, so it should be possible to drive the two sets from two power stages in order to facilitate current sharing.

[sawickm]

Hi Paul,

This sounds like another interesting project !!!

I read an article in Circuit Cellar Magazine (August 2008) that covered a similar project. Like you said it cost them $2000 - $3000 to build that AC Controller.

I would be willing to donate to the cause, give me a list and count of the parts you need to get started.

I would also love to see an open source group effort on a “3-phase Switched Reluctant (SR) Controller”. The architecture for that type of controller is not that much different from AC, the SR can also do regenerative braking. The neat thing about an SR system is that the EV motor is EXTREAMLY simple to make. The SR motor is a just a housing, shaft, end plates, laminated rotor, and laminated stator with coils, NO BRUSHES or MAGNETS !!!

An open source SR EV system could be made which would be a low cost combination - Controller and Motor.

www.freescale.com/files/product/doc/AN1932.pdf
http://www.itee.uq.edu.au/~aupec/aup...ll%20paper.pdf

- Mark

[NiHaoMike]

I thought switched reluctance is mostly good at high speed, low torque. One of my friends who works with HVAC has built a water cycle A/C that uses a compressor she made out of a turbocharger and a custom machined SR motor. It works very well (SEER rating around 21 in "hybrid" operation), but the compressor going at 75kRPMs makes a lot of noise.

[Janis]

Hello, Paul
I have read all DC controller thread, That was awsome, and last few month waiting for this thread to show up. AC controller is what I am looking for. I have worked with industrial AC frequency converters and as much as possible I will help tell their behaivior.

Only why you switch from Atmel to Microchip?


AVR Freaks

http://www.atmel.com/dyn/resources/p...ts/doc7546.pdf

We are on te way.......