Paul & Sabrina's cheap DIY 144v motor controller

[jyanof]

I was thinking about ways to make this controller adaptable to a variety of applications - seems like each rig is different based on motor size and battery voltage...

Seems like possible candidates of things to customize might be rise rate (that current rise constant at the beginning of the code) and maybe this 'high end sensitivity' thing recently discussed in Ben's thread. Maybe max current, too.

These could be little pots connected to ADC's (if there are enough available- they could be read once at the beginning) that are adjusted with a small screwdriver (kinda like the curtis), but maybe you have to take the cover off - unless you want to drill access holes and have rubber plugs just like the curtis.

If something like this is added to the next version of the control board, it could allow these throttle response parameters to be customized by the user without swapping micros or using a bootloader...

[taffy]

I really would like some kind of feedback to know what the setting is. When adjsuting the curtis you are blindly adjust things and not sure if you have set at 60% or 75%.
A remote settings changer would be a plus, so you dont have to pull it apart when testing. Do 1 run, ramp up current limit, run again, etc.

Paul: 180V and 700Amp...!!! Wahooba! Sadly my budget wont stretch a 180V thundersky pack though.

[MPaulHolmes]

Hi Taffy! Fran is going to help with making the controller settings programmable through the serial port through any ordinary laptop or whatever. It will then save the settings in the EEProm. I sent him a control board today.

[Sutitan]

Finally got myself an ecomodder account. I started following this thread somewhere mid page 150 (but went back and "read" the rest) I found you through a link posted a while back on diyelectriccar. I gotta say congratulations Paul on the progress so far.

anyways, as my build begins to approach, I have a few questions to see if your controller could work (summer '10) I am considering this controller and the zilla 1k lv. You mentioned way back that there would eventually be a more powerful version of this controller, and on the last page, you made it seem as if it were just scalable. Anyways, I was sort of looking for a controller that could handle 1000amps with not to much fuss (few seconds. It would be watercooled as well) No need for extreme voltages just 144v.

I know its hard to compare these 2 controllers, but I have high hopes for it. cant wait to see what you manage to accomplish in the next few months.

[MPaulHolmes]

As soon as I get a little free time, I'm going to try multiple gate drives and learn the AT90PWM3b micro, which was designed for synchronous rectification. Then eventually I'm going to take about 12-15 of the new fancier 230amp 200v mosfets, and do a synchronous rectification version, so the freewheel mosfets will have way more current carrying capacity than what I'm using now. I think that one will do 1000 amps. I'll need maybe 6 or 8 ounce PCB, and I'll need to do something tricky with the LEM current sensor, since it can only go up to 900 amps. That one may only be safe at 132v or so, since the inductive voltage spikes get bigger with large di/dt, and it's already cutting it sort of close to 200v with the 144v nominal inductive voltage spikes.

[taffy]

Paul, that sounds perfect! Cant wait to see what comes from that. I am awaiting a pile of parts to turn up so i can build my 144V 500Amp edition. Yay!

Does that serial port give real time feedback? Cause that would make for some awesome logging & display capabilities. (Not a requirement! So dont run out and do it and not see your kid or anything!!!!)

Sorry to say but i think a Zilla is better then a revolt. The Revolt seems better then a curtis, kelly and logisystems but possibly not as good Zilla and possibly the Solution1.
Someone on here made a 1000Amp version didnt they?

[MPaulHolmes]

It's really easy to have real time serial updating. I have done it from time to time when I was tuning the PI loop. So, YES! HEHE.

Yes, a zilla is the best brushed DC controller ever made (at Least that's what I think). It's pretty much bulletproof.

Adam has some really really high power stuff in the works. Also some really cool user interface things too. And a sports car and lithium batteries to test it all in.

[jackbauer]

Now THATS a transistor:
IGBT Modul FUJI 1MBI800U4B-120 1200V 800A on eBay (end time 05-Oct-09 16:41:24 BST)

[jyanof]

Originally Posted by MPaulHolmes

Then eventually I'm going to take about 12-15 of the new fancier 230amp 200v mosfets, and do a synchronous rectification version, so the freewheel mosfets will have way more current carrying capacity than what I'm using now. I think that one will do 1000 amps.

I'm trying to follow how you'd get more current with less switching devices... It boils down to heat, right? Keep it cool, push more current. So, with multiple gate drivers and sync-rect you can switch faster to minimize switching losses (at the risk of higher voltage spikes). Also, the rectifying mosfets should be a bit more effecient than their diode counterparts, particularly at lower currents.

That sound right?

[MPaulHolmes]

The freewheel mosfets are much more efficient at lower currents. Around 500 amps or so, there's a break even point, and then after that, the diodes are more efficient.

But the new mosfets' legs are rated for 160 amps. The inside is rated for 230 amps. The diodes I have been using are rated for 60 amps. So, even though the mosfets are less efficient at really high currents, they can just brute force take more current at higher powers. And for normal around town 100-200amp driving, the freewheel mosfets are quite a bit more efficient.

Let's say you are cruising around town at low power. look at 10 amps per freewheel mosfet:
The voltage drop is about...
1.8*0.0075*10 = 0.135v

And the heat loss:
10*0.135 = 1.35 watts

(the 1.8 comes from the graph on the datasheet assuming 100degC operational temperature)

The freewheel diode at 10 amps:

1v (or so).

Heat loss:
1*10 = 10 watts.


Now, let's say a short burst of 100 amps per freewheel mosfet:
Voltage Drop is about...
1.8*0.0075*100 = 1.35v
Heat loss:
100*1.35 = 135 watts! hahaha (VERY BRIEFLY!!!!)

Now, 100 amps per freewheel diode!
IMPOSSIBLE! SINCE IT WILL BLOW UP! This is WAY over the max rating.