Paul & Sabrina's cheap DIY 144v motor controller

[sawickm]

Hi Paul,

I came across a few low cost IGBT gate drive IC’s, the ST Micro TD350, TD351, TD352 series. The IC series was designed especially for cost-sensitive motor control applications. The TD35x series are advanced IGBT drivers with integrated control and protection functions.

They might be a good solution for the IGBT power PCB that you are looking at designing, check them out.
FYI:

STMicroelectronics | Part Number Search

Power Electronics : Celeritous Technical Services, Corp., The Prototype Specialists

[MPaulHolmes]

Hey Matt! If you want to go Alvin's route, and get some of the annoying drilling done yourself and ordering from digikey and mouser and stuff, then you could save yourself a few hundred dollars. It's just annoying and time consuming.

[jackbauer]

Paul , am i right in thinking that your software looks at the throttle input then ramps the pwm until the lem senses the set current? From looking at the scope today thats what seems to be happening.

got some 1600v 800a igbts on order!

[MPaulHolmes]

Hi Jack! (Oh no, don't ride my airplane)... It was just a PI loop with current proportional to throttle, and Joe added a few lines to incorporate pwm duty and throttle sensitivity. It made his controller feel a little more like his Curtis I think. I just sent you that one since he had been happy with it.

[jackbauer]

thanks paul i'll let you all know how i get on with the igbts. At least you get the "24" jokes

[apowers]

Hi Paul
Yes I was trying to build this EV on the cheap. When I ordered my stuff you
were only selling the boards and micrcontroller. I still spent $650. I bought more
than one of each. Just in case. My motor came from a forklift 11 inch GE $50.
I bought low amp batteries maybe not a good idea. Been charging all week
one at a time. I need a charger. Has anyone out there built a bonn charger and
used it? Let me know.
Thanks again Paul
Alvin

[drbobwoolery]

you wrote:
Hey Bob! I'm using 5/16" holes in the bus bars. If it is a thin manganin
sheet, how does the mounting usually work? Could it just replace the
section where the LEM usually is? I've never seen a shunt. Is it sturdy
for bolting?
I am supposed to get a 4x4 inch or so piece of .060 annealed manganin (Cu, Mn, Ni), which should saw and drill easily, and might just answer to my biggest tinsnips.
as I work it out, a 1 inch wide strip, with the current uniformly distributed, has an area of .387 cm squared. Manganin has a specific resistance of 38 microohms . cm, and a temperature coefficient of plus minus 15 ppm. That is the reason this alloy is used for metal resistors. Anyway, dividing 38 by .387 gives 98 microohms per centimeter.

So for a 0.1 milliohm shunt, the voltage taps need to be 1 cm apart. The bolt holes for connection to the busbar and cable can be at the diagonal corners of a square an inch on a side. I'll either bolt or silver solder one or two pairs of sense leads near the centerline, spaced 1 cm apart. Filing or grinding the opposite corners should allow calibration at around 200 amps with standards traceable to a wastebasket in Guandong, China. No lab standards around my shop these days.

The resistance was chosen to allow for 1000 A full scale, with 100 watts dissipated in the portion between the sense points. Think you can fit such a piece inside the controller, to replace the noisy Hall sensor? I'll put on two pairs of leads, to allow for a 200 mv panel meter and the current feedback to be separately hooked up. Check my figures and reasoning if it comes handy.
We will know about two productive hours after the stuff comes in the mail.
Better allow a week.
bob

[jackbauer]

Following takes place between 3pm and 4pm:

Wired up a 12v nissan starter motor to the igbt and upgraded the wiring. It works! I bypassed the solenoid so it just runs the motor. Takes 30 amps to get it moving then ramps up to 60amps and full bore. My 7ah gell batt suffers a bit of "pack sag" at that current dropping to just under 10v. After about 30 seconds of running the igbt base plate was warm. About 30-40c i would guess but its only a 150 amps device , got no heatsink and sits on a wooden bench. So not bad for a first run.

next moves : 4 igbts in parallel , caps , 24v , two 12v starters in series for some white zombie type action!

[MPaulHolmes]

Jack, that is one crazy mad scientist lab you have going there! That was a very productive 3-4pm. Do you have a freewheel diode anywhere, or are you not worrying about it since the igbt can handle really really high voltages?

Hey Bob! There's plenty of room where the current sensor usually is. I would have to change the control board to include some sort of op-amp and isolation for the current A/D input, and also change the hardware overcurrent circuit, but that's all doable. The hardware overcurrent circuit triggers in about 2-3 microseconds right now. Are there any opto-isolators with super short propogation delay? Alternatively, how would we get current A/D input isolation from pack ground

[jackbauer]

The igbt i'm using is a 2 pack push pull type. I've tied the gate to the emitter on the high side and use the intrinsic diode as the freewheel. Seems to work ok. Does anyone know if starters are series wound?