Paul & Sabrina's cheap DIY 144v motor controller

[HaroldinCR]

Maybe it's just me, but, why are y'all trying to make a bullet proof enclosure ???

Had a business for 20 years, using 16 & 14 gauge sheet metal. SS, Alum, and steel, are ALL very durable in a box as small as the controller needs. I would bet the Zilla uses 18-20 gauge, or thinner.

Mount the controller up high in the motor compartment, and mount it securely. Run a flexible (vacuum or pool vacuum hose), from a catch box in the grill or just under the front end. Screen it, put a small hole in a low spot on the hose, in case water does get into the hose, and let the movement of the car cool the controller. Easy to keep clean. Just look at the screen and wipe off any trash ???

Faster you go, more air to pass over or through the enclosure ???

Good work coming from around the world on this project.

[MPaulHolmes]

I think I'm going to do an aluminum plate instead of a heatsink. I can mill the edges of the plate so that it has extrusions with holes that can be bolted to any old heat sink if you want, or it may not be needed. I think that's a much cleaner and professional (a la curtis, et al) way of doing it. I'm thinking of 3/8" plate. It will also keep the dimensions to a reasonable level. Not everyone could fit a controller that is like 6 or 7 inches tall (counting the heat sink) inside their car! Ben had a terrible time getting version 1 in his car.

I think HaroldinCR is right. All the support will be coming from the base plate, so it wouldn't matter if the sides are thinner. They would be there to keep out the elements, so I bet 18 or 20 gauge would be fine, Hondo. I'll PM you with more exact dimensions.

[blackpanther-st]

Originally Posted by MPaulHolmes

I think I'm going to do an aluminum plate instead of a heatsink. I can mill the edges of the plate so that it has extrusions with holes that can be bolted to any old heat sink if you want, or it may not be needed. I think that's a much cleaner and professional (a la curtis, et al) way of doing it. I'm thinking of 3/8" plate. It will also keep the dimensions to a reasonable level. Not everyone could fit a controller that is like 6 or 7 inches tall (counting the heat sink) inside their car! Ben had a terrible time getting version 1 in his car.

I think HaroldinCR is right. All the support will be coming from the base plate, so it wouldn't matter if the sides are thinner. They would be there to keep out the elements, so I bet 18 or 20 gauge would be fine, Hondo. I'll PM you with more exact dimensions.

Paul this is my first post here, but I have been lurking for almost a month now and have read the entire thread.

In my opinion I think that using the heat sink as the base is a good idea as long as it is sizable enough, the sheet metal enclosure shown earlier looks good to me but the aluminum version would be better, just bolt it to the heat sink. as for the end caps a flap on both ends for EMI reduction, even with generous cut outs for your bus bars will be nearly as effective as a full metal end.

now the details; make the end flap a little longer and bend back into the cover 1/4" and then down, so that it is inset the 1/4" . now after the controller is assembled, and the cover is bolted on, dip the ends just over the 1/4' in depth in a good potting compound.

this will give you good EMI control , good cooling and seal the unit from contamination. and it is relatively easy to make.

If you want the cover easily removable another option would be to skip the flap and instead use an aluminum end plate with an L at the bottom for mounting to the heat sink. again, pot the ends, but this time apply some Vaseline to the inside of the cover where the potting compound will be. the cover will then be removable.

because of the bending concerns mentioned in an earlier post, this second method with the removable cover, may actually be the easier version.

A little side note, I am also starting on my own DIY controller but in the very beginning stages, and yes it will be open source if I am successful with it. more on that later.

I wish you the best with this, and I am impressed with what I have seen here so far

[jyanof]

Since a lot of discussion has been about the case and heat dissipation, I was curious how the efficiency losses added up for this DIY controller vs. the Curtis 1231c I have.

I pulled up the component data sheets and calculated the mosfet losses according to:
Power Supply Engineer's Guide to Calculate Dissipation for MOSFETs in High-Power Supplies - Maxim

You can see the results here:
http://spreadsheets.google.com/ccc?k...KxRc7qBx_aacCw

It also breaks out the losses from mosfet switching, mosfet resistance, and diode voltage drop (all in watts). You can see that the curtis has a very high switching loss! (hence the need to drop the switching frequency to 1.5khz as thermal protection)

In Summary, the DIY controller is calculated to be roughly 4 times more efficient (as in, 4 times less heat dissipation) than the Curtis.

My heatsink runs 15C above ambient with the Curtis when I cruise at 200A. If we assume the heatsink to work just as well with the DIY controller (which, admittedly the curtis has a nice finned Al case to help it), 4 times less heat generation should result in about 4 times less thermal rise, so that puts us at roughly 4C above ambient...

That number might be a bit higher without the finned case, but that's something I can check for once we start testing.

[MPaulHolmes]

jyanof, "The more I think about th..."

Man, Now I'm going to wonder the rest of the day what you were going to say! Thank you for that spreadsheet. Very interesting! Man, they parallel 19 mosfets? Geeze.

[jyanof]

In thinking more about the case, I'd say the only reason for going metal would be for EMI shielding.

The curtis case helps convect heat away because it's a 1 piece extrusion. The components are on the bottom and heat conducts through the metal at the bottom to the sides and top of the case where it then dissipates to the air (in addition to conducting to any added heatsink on the bottom). I doubt very much heat makes it from component to inside-air to case to outside-air. Maybe some will radiate, but I bet there's just too much thermal resistance. It's way easier for heat to go through the heat spreader and heatsink, so most of it will.

I'm not sure how to comment on EMI though... If it ain't broke, do we need to fix it?

Perhaps cost and manufacturability are more important design drivers for a case. Cheap and easy!

[jyanof]

Originally Posted by MPaulHolmes

jyanof, "The more I think about th..."

Man, Now I'm going to wonder the rest of the day what you were going to say! Thank you for that spreadsheet. Very interesting! Man, they parallel 19 mosfets? Geeze.

yeah, i got ahead of myself... my bad!

but ya, check out Ian's write up (it's where I found the details):

Zero Emission Vehicles Australia

[MissileStick]

At those dimensions you gave me MPaulHolmes, the case would dissipate a little less at 25C temperature difference. Of course, now that seems extraordinarily high.

25C - 142 watts
20C - 113 watts
15C - 85 watts
10C - 57 watts
5C - 28 watts

This assumes 1/4 inch Lexan – 1/2 inch would dissipate half as much heat.

To me the main point in favor of a lexan (or anything else nonmetallic) case would be the fact that it would be electrically insulative. Of course there's other ways to do that – for instance painting the aluminum case with something hard/nonconductive, like epoxy.

Is the emergency cutoff relay still present? Would grounding out to the case be detectable, leading to a disconnect?

[xtian999]

Heat rises, so why put all the aluminum "base" on the bottom? Let the heat go up and out.

[MPaulHolmes]

I think that because the copper heat spreader is in direct contact (being pressed with like a 1000 pounds or whatever because of five 1/4 inch bolts torqued down) with the heatsink, the heat transfer from the mosfets/diodes to the base is pretty efficient. I think heat only rises in fluids, because the cold sinks, because it's more dense. With an aluminum case, thanks to Hondo, I'm guessing that if the case is bolted to the base plate, that the plate should transfer heat pretty efficiently to the case, sort of like the Curtis.