Paul & Sabrina's cheap DIY 144v motor controller

[adamj12b]

Hi Paul,

I was looking at the schematics of the control board and comparing it to some the datasheets of some parts and noticed that the pins on the DC-DC converter you have dont match the datasheet. Why is this? They also make the converter in a surface mount which those machine I mentioned could place as well.

Also, What software do your use to design the schematics?

-Adam

[MazdaMatt]

Adam, one of our major design criteria is DIY-ability. Even people who CAN solder have a hard time with surface mount soldering unless they have the proper tools. SM has huge advantages in automation and board size, but that's just not the goal here.

[charlie_fd]

Originally Posted by esoneson

Paul and Adrian,

....
... I will email you some guidelines for dealing with bits-on bits-off so you don't have to do the 128+64+....stuff. It will all be clear.

...
Eric

What I did when using C for microcontroller projects was to use macros to be able to represent binary as binary. Using hex is another option that makes it more readable than the decimal sum

I've uploaded the header with those macros on svn /trunk/macro-binaryconstant.h - Open ReVolt - Trac

using the macro will allow you to write smth like:

#include "macro-binaryconstant.h"

int main()
{
int i = B8(1010); // 8 bit binary representation of the decimal number 10
int j = B8(10000000); // 8 bit binary representation of the decimal number 128


return 0;
}

[jyanof]

Originally Posted by bennelson

Yep, we gotta do some more testing before getting everything out to the general public.

It's amazing what can be missed until you actually try something out!

"

Is there anyone slated for testing that lives in a hot climate?

I'm in Phoenix and am about to go through my first summer with a 500A curtis 1231c. I have a thermal probe in the heatsink and, so far, the heatsink seems to level off around 30 degrees F above ambient when highway cruising and about 40F in the city. It reached 100F outside this past weekend and it began to limit current to 375A at the tail-end of some accelerations. Also, it often runs in its 'efficiency mode' at 1.5kHz, even while cruising at lower amps. There are others in AZ with Curtis', so maybe it'll be ok when it gets to 115+.

I know the Curtis is pretty inefficient too, especially when compared to the newer mosfets and upgraded design of this DIY model, so maybe it won't be a big deal.

Nevertheless, I'll be happy to help (and pay for parts) if testing in hot weather is desired. If everything is worked out soon, I can be driving it around in July and August which are the worst months.

-Joe

[MPaulHolmes]

Hey jyanof! Good idea for the hot weather testing. Overtemp shutdown hasn't been tested in the field yet. I can easily adjust the values and mail you a new chip if it comes to that, like if it turns on too soon. Right now, I have it start to limit max current when the internal heatspreader hits 75 degC, and there is complete thermal shutdown at 85 degC (internal heat spreader temp). We have Ben's 10" motor (man that's big) for worst case control of current at low RPM (like no inductance in that gall dern feisty motor. ), your hot weather testing, and Romanian testing? See how it stands up against Dracula! Ya!

The 3 limited edition Cougar controllers are officially taken.

.................................................. .................................................. ..
Good idea, MazdaMatt! I'll ask futurlec what you suggested.
.................................................. .................................................. ..
Adamjb, I checked the schematic. Thank you for attaching that! I'm using the high voltage isolation (3kV isolation for 1 minute) DC-DC, which has a different pin layout than the 500v isolation DC-DC layout. We got it right. Thank you for checking! By the way, I would love to do another version with surface mount components and have your friend do his super fancy soldering machine on it. May as well have both options available! That will be coming too! I was using ExpressSCH, but my friend that is helping me is using Protel. He converted my schematic to Protel, and added the new fancy schmancy stuff.
.................................................. .................................................. ..
I really like the idea of b(01011101) rather than 64 + 16 + ... . My way was really stupid. Fast and small, but hard to read and confusing. Or maybe I could do that & or | with ~(1 << ???) or who knows what.

OK boys, I'll order those PCBs! Git R Done! hahaha!

[jyanof]

Originally Posted by MPaulHolmes

I can easily adjust the values and mail you a new chip if it comes to that, like if it turns on too soon. Right now, I have it start to limit max current when the internal heatspreader hits 75 degC, and there is complete thermal shutdown at 85 degC (internal heat spreader temp).

This might be appropriate though for keeping the diodes/mosfets from getting too hot. You're still using the thermal pad to isolate all the components, correct? There could be a large temperature drop across that pad, especially if there's a lot of heat generation and the heat spreader (and thermistor) haven't had time to heat up. The silicon could be much hotter and having lots of margin might be desirable. (I'm having trouble with my charger design getting hot and blowing FETs. Those high voltage FETs aren't as efficient!)

If it happens that thermal shutdown occurs, it might be better to try to reduce the thermal resistance to keep things cooler rather than adjust the threshold higher.

Guess we can cross that bridge if it comes up - maybe everything will work fine and we won't have a problem! 120F is 49C - That leaves 26C above ambient for a controller that's gotta be more efficient than the Curtis.

[MPaulHolmes]

Originally Posted by jyanof

There could be a large temperature drop across that pad, especially if there's a lot of heat generation and the heat spreader (and thermistor) haven't had time to heat up. The silicon could be much hotter and having lots of margin might be desirable.

Well said. That heat spreader is like a big dang thermal capacitor. I asked the EV Tech list about this, and Lee Hart said it was one of the classical problems of control theory. Otmar drilled holes into the back plate of the mosfets and IGBTs to get right behind the silicon to get his thermal numbers. I don't have a drill bit that skinny. That guy's crazy. haha!

Originally Posted by jyanof

If it happens that thermal shutdown occurs, it might be better to try to reduce the thermal resistance to keep things cooler rather than adjust the threshold higher.

My thermal shutdown code is very gradual. There is an 8 stage process, so you will definitely start to notice it gradually, rather than all at once. That's an excellent idea about reducing the thermal resistance. It wouldn't be a quick fix on your controller, though. That would be a total redesign. Maybe bolting the backs of the mosfets/diodes DIRECTLY to their own respective heat spreaders, and then isolating the 2 electrically isolated heat spreaders from the heat sink. That would require some thought.

By the way, I have a cost tally for the controller parts. We still need an enclosure, though. That's the big unknown. I was going to do the new and hopefully improved power section layout for you and Adrian, which would require a 6"x8" power section, which would fit into the largest stock build-a-box that Wherewolf found, I think. I think a metal enclosure would be much better thermally speaking than a Lexan box with a heat sink on top. Can I get your email address? You can PM me if you want. Do I already have it? I forgot!

[Intrigued]

Originally Posted by SteveU

Intrigued: By reducing the field, the armature has to spin faster to generate a back EMF that matches the input. Some big AC motors use this for speed control, but if the field supply fails the motor goes crazy. Likewise, an automotive alternator is controlled by controlling the field current.

A basic concept is that the motor current is the difference between the input voltage and back EMF divided by the coil resistance, so as the torque load increases, the motor slows down a tiny bit, reduces the back EMF and the current increases, pulling harder. So the motor speed is a function of input voltage divided by the field current. Of course, there are practical limits like the field current must not be too low.

Steve

Whew! Got my paper turned in, and am catching up. I still actually have enough brain power to understand the most of that! I used to be an ASE Master Automotive Tech, so I can see how alternators relate, and can now understand how they can (sometimes at least) be used as a motor also.

I'm really glad to better understand the field weakening idea. It's frustrating to hear about things you don't understand... and you guys are light-years ahead of me...

[charlie_fd]

Take a look at these standardized (European) enclosures:
E-Case aluminium enclosures by Lincoln Binns UK: Aluminum enclosures & cases for electronic applications.

The D class will probably fit our current design of powerboard ...

[xtian999]

I have heard of using copper rivets for through-connectors on circuit boards. Looks like a good idea for small batches.