Paul & Sabrina's cheap DIY 144v motor controller

[MazdaMatt]

Paul, any idea what the big guys are using for control? Pruis is somewhere around 300v, right? I assume they are driving that with IGBTs (and using an AC motor)... liquid cooled? Looking at Digikey, i can see IGBT modules rated for 600V, 400A for 120 bucks (cdn). I have not checked ebay.

[MPaulHolmes]

I think that anything in the 300v range is using IGBTs for high power applications. The fixed voltage drop just becomes less loss at that point. New IGBTs are way expensive! And you should only (according to Lee Hart) run an IGBT continuously (or was it for a very short time like 1 minute?) at about 1/3 of it's current rating. I think you would need like 4 600 amp devices to get the 1000 amp controller, and liquid cooling.

However, if it pans out, a company that spoke to Rod Hower I believe (the head designer at Ametek) told him about their company's breakthrough high voltage mosfets with low RdsOn (resistance from drain to source while gate is high) that would make IGBTs obsolete. They would be really expensive at first, though.

[MazdaMatt]

I believe (90%) that you can't parallel IGBT's. You can only buy the biggest module you can. However, that would not be the case for multi-phase AC - each phase could be driven by a module.

I'm starting to think DIY here, but time/budget aren't there for me right now. I would love to develop a DIY 3-phase AC igbt-based controller.

[MPaulHolmes]

I also thought you can't parallel them, but you can. I spoke to several people that actually did it. They said the hard part is IGBTs have a "negative temperature coefficient", so are subject to thermal runaway. You can be careful to keep them cool, and watch their temperature closely, and safely parallel them, it's just harder than mosfets. Mosfets tend to balance themselves in parallel because the ones that get more current increase their RdsON, lowering their current draw.

I love your idea about an AC controller! I would like to learn more about it too.

[MazdaMatt]

Interesting. I'll have to look up application notes on them... or just go with 3-phase AC. I was perusing datasheets this morning and i noticed some IGBT's, i believe by IXYS, that advertised positive temperature coefficient. Either way, they all seem to have built-in thermistor for temp monitoring. They also have large copper or ceramic isolated heat sinks attached, making cooling a snap.

[MazdaMatt]

FYI - unrelated to this controller - article on VFD (aka AC controller) using IGBTs. This is the basics and it is pretty clear. It looks quite simple to me.

http://www.kilowattclassroom.com/Archive/VFDarticle.pdf

[esoneson]

I have been following a DIYer building an AC controller with IGBTs on diyelectriccar.com. Here is his web site:

Tischer's Inverter for the Siemens motor built for Ford

He has a novel approach which fully supports the idea of the 'black box' approach.

Eric

[MazdaMatt]

Thanks!

[motor_control]

Originally Posted by MPaulHolmes

for higher current sensing, you could use 2 current sensors in parallel. You could have 2 separate M-'s! That's be sort of weird, but at least you could monitor currents up to 1800 amps.

I think 1 LEM is anough. Russian way



Split you copper bar into n patrs and throgh any part flow current I/n with good presize. n=2 or 3 èç very simple designinig and wiering I think.

You can get 1 wire from power cable and LEM need to be for I/(number of wires).

[MPaulHolmes]

Beautiful Russian elegance!!! Thank you motor_control! I love it! It makes perfect sense!

Did you ever hear how the american astronauts couldn't get a pen to work in space, because of no gravity, so they spent a ton of money on developing a complicated pen. Later they asked russia how they overcame the problem, and russia said they used a pencil.