Paul & Sabrina's cheap DIY 144v motor controller

[Christ]

Originally Posted by MPaulHolmes

I read the guy's process, but I don't know if I saw lapping. I still don't know what that is. boohoo! Makes me mad!

LOL - you're not the first person to say that, and I'm not really sure if it's ever been called anything different, but that's pretty much what I've done as well. It's basically polishing both surfaces to get better thermal conductivity.

Originally Posted by MPaulHolmes

Did you see the advertisements for the new Fast and the Furious?

*Facepalm...* No, I haven't. I'm sure there's a Jesse joke coming from somewhere... LOL.

[brihoo2k]

The pictures of the fets are great Paul keep up the good work

[MazdaMatt]

Paul, are those fets soldered through the board, or flat onto it? If flat, i'd be concerned about re-aligning them when you screw then into the heat spreader causing some pulling on the solder joints. pulling + vibration -> crack -> poof.

[MPaulHolmes]

Excellent point Matt! I secretly attached them to the heat spreader, soldered them in place, and removed the heat spreader again. Each one was perfectly flush against it when being soldered in place. ya! I'm so sneaky sometimes.

Another nice thing is, they won't be bolted to the heat spreader, only clamped against it, so it allows them up/down/left/right freedom when reattaching them.

[MazdaMatt]

Good stuff

[bennelson]

In terms of getting useful heat from the controller, I think it is almost a moot point.

One COULD put the controller inside the passenger compartment of the car the gain the little bit of heat, but then it's adding heat in the summer too!

My experience with controllers is that there is so little heat coming off of it, it's not worth messing around with.

My charger, on the other hand, does put out a bit of heat. That is installed inside the passenger compartment. In the grand scheme though, it's not putting out that much heat.

Paul, I know it's getting ahead, but have you thought about an enclosure yet? The latest e-mail from INSTRUCTABLES had a nice little bit about making boxes out of acrylic with T-bolts.
How to Make Anything (Using Acrylic and Machine Screws)

I have seen some other controllers with acrylic boxes that looked really nice.

Plus, your controller just looks so dang cool, it would be a shame to put it in a plain black metal box.

[motor_control]

Motor produce hit too and need to be cooled.

[esoneson]

Paul,

I do not wish to distract you from making progress on this fantastic project, but if you have time, I have another question.

If, per chance, I was thinking of using 156 Volts or even 192 volts and maybe thinking of drawing 1000 watts instead of 500-600 as this design can handle, would this simply (yes, I know, it is never simple) mean that there would be changes in the components of the high power module only? With perhaps, changes of software settings too?

If this becomes a nice 'black box' module system with software configurable low power module and plug-in high power module, then this could meet controller requirements from scooter to dragster.

Anyway, since I was thinking 156+volts and pushing 600+ amps, what changes would be necessary in order to guard against pushing the envelope like I am thinking of doing and not creating that dreaded magic smoke?

Thanks again.
Eric and the dogs

[MPaulHolmes]

Originally Posted by bennelson

Paul, I know it's getting ahead, but have you thought about an enclosure yet? The latest e-mail from INSTRUCTABLES had a nice little bit about making boxes out of acrylic with T-bolts.

Plus, your controller just looks so dang cool, it would be a shame to put it in a plain black metal box.

Thank you for the link to the instructables! That thing is awesome. I think I can do a simple, inexpensive box using their ideas. Thanks! It would be really nice to have the inside visible. Especially if you are going to show the car this summer at that alternative energy fair! I think most people have no idea how a controller is actually super simple!

[MPaulHolmes]

Originally Posted by esonson

If this becomes a nice 'black box' module system with software configurable low power module and plug-in high power module, then this could meet controller requirements from scooter to dragster.

Anyway, since I was thinking 156+volts and pushing 600+ amps, what changes would be necessary in order to guard against pushing the envelope like I am thinking of doing and not creating that dreaded magic smoke?

Going much over 156v is probably not a safe thing for the 200v mosfets. In fact, 144v is probably the safe peak. I looked it up, and there ARE higher voltage mosfets that are reasonably priced. However, their heat loss is about 3 times that of a really good 200v mosfet. They are only about $7 each.250v mosfets

Hmm... You'd need higher voltage diodes, which I don't think is any problem, and then you'd be just fine for 192v. I don't know the safe paralleling limit on the mosfets. I'm doing 10 for this first attempt. Ian is using 12 to get 600 amps. I bet you could do 18 if you used 2 or 3 different mosfet drivers. I don't see why it wouldn't work. The control board would be the same up to 900 amps. After that, you would need a different current sensor. I don't know if any exist above 900 amps. I haven't seen any on digikey or mouser.

The nice thing about this design is that if you change the mosfets/diodes, everything else doesn't care about the higher voltage. If you throw in some 300v mosfets, the controller will run on anything from 12v to 300v. The key is that I'm using an isolated DC-DC that powers the controller with the 12v car battery. Most controllers use DC-DC's that have limited inputs like 90v-156v or whatever. That's why most controllers have strange voltage input ranges.

Another option is IGBTs. Ebay often has some 1200v 600amp IGBTs for pretty cheap. I once bought 4 for $50. They have a fixed voltage drop, so are less efficient at lower voltages, but if you do a good liquid cooling cold plate, I think it would work really well. Then, you could have any voltage you want, and very high max current. That's what the Tesla's CTO did before working at Tesla. It's actually pretty simple (ish).

Actually, 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.