Paul & Sabrina's cheap DIY 144v motor controller

[isaac_alaska]

Originally Posted by ElectricZX2

I second not using scrs. I ran a 500amp model for a while. They fail in funny ways if you run voltages too high above their rating. Mine failed on as I was coming up on a red light with a long line of cars...like forgetting you had the cruise control set but your brakes don't do anything. Luckily there was a dirt shoulder and the ignition contractor still worked. Getting home was interesting using controller bypass...

the SCR's wouldn't be used as switches, they'd be connected as diodes. At least i'm pretty sure they only allow current to flow in one direction...from what i've read about them anyway. i was planning to connect them as always conducting, so they'd essentially be giant diodes. So they only way they could fail would be to fail open (what's the consequences of them failing open?) or they could fail in such a way that they always conducted current in the reverse direction....i guess i should read up on them.

would someone be so kind as to link the power board schematic on here? even the mosfet power board schematic would be appreciated, i think i could see how everything gets connected up based on that.

*edit*

these are 800V scr's, and the nominal voltage of my battery pack is going to be 355, with a maximum of about 405. i would think this should give me enough of a margin of safety to not worry about them failing, as long as i keep them cooled

[MPaulHolmes]

they need to go from nonconducting to conducting in a very short time. LIke 50-100 nSec or something. So, usually you want to look for "ultrafast diode" or something. If it takes a "long time" to go to conducting, there would be a large voltage spike that would hit the igbt, since the current needs to keep flowing somewhere.

It says the turn on time is 7us for the hockey puck ones. Am I understanding that right? That would be catastrophically too long I think. The igbt would be off in maybe 1uS, so that leaves 6us of the current that is still going to flow somewhere to pile-drive through the igbt or the scr. Maybe creating thousands of volts. Am I understanding that right?

[isaac_alaska]

Originally Posted by MPaulHolmes

they need to go from nonconducting to conducting in a very short time. LIke 50-100 nSec or something. So, usually you want to look for "ultrafast diode" or something. If it takes a "long time" to go to conducting, there would be a large voltage spike that would hit the igbt, since the current needs to keep flowing somewhere.

It says the turn on time is 7us for the hockey puck ones. Am I understanding that right? That would be catastrophically too long I think. The igbt would be off in maybe 1uS, so that leaves 6us of the current that is still going to flow somewhere to pile-drive through the igbt or the scr. Maybe creating thousands of volts. Am I understanding that right?

You're probably right. Do you have the schematic to the power board so i can see how it all gets set up? I was just tryin to use my SCR's cause they're big and pretty...but if i can't use them i'll have to find somethin else.

[jackbauer]

Diagram attached. If you go back and look at my early control experiment videos they may help. Please forget about those scrs. They are totally unsuitable in any form whatsoever for a modern pwm controller build. Remember the freewheel diode and the switch each do the SAME work and disipate the SAME power and so as much care must be taken in choosing the fwd as the switch. Fortunatly , the body diode in the high side igbt works very well and the interconnect between the two devices in a dual module is optimised for low inductance. Get that link ringing and its goodnight and good luck

[isaac_alaska]

Thanks jack. that makes a lot more sense now that i see how they all get connected together. Looks like i'll be getting a few more igbt's...and probably put these up for sale if anyone wants them. i grabbed them while they were available but it looks like it's much better to go with the dual package.

I think the only other question that's burning in my mind is where do the filter caps get placed? from the other motor controller power board schematics i've looked it, it seems like one set of caps goes directly in parallel with the motor connections, and the other set gets tied from motor + to ground. I'm not seeing how the capacitors get pre-charged though....and i know pre-charge is important because of what we did to our logisystems controller....it turned into a big hot rocket engine for about thirty seconds. not something i want to repeat

[MPaulHolmes]

haha, well, you want the capacitor bank + connected to B+, and the capacitor bank B- to the B- in jacks picture. Also, you want the sum of the paths from Capacitor + to Diode cathode (the output of the arrow), diode anode (input to diode) to IGBT collector, IGBT emitter to capacitor bank B- to be as SHORT AS POSSIBLE!!!!!!!

The precharge circuit isn't shown. YOu basically need to apply full voltage to the capacitor bank, through a resistor. So a separate relay that allows the caps to fill up with a resistor in series, and then once the bank is full, you close the main contactor. The control board I'm sending you has 2 relay controls, so you can have all that done automatically.

[isaac_alaska]

okay. that makes a lot more sense. it seems i'm learning a ton of stuff every day...hehe. so really all the capacitor is doing is supplementing the battery so that it can provide a much higher pulse current over a short duration? i was under the assumption the the switch was somehow cutting out the battery from the circuit and the motor was being run from the energy stored in the capacitors, but i think i was picturing a different circuit. maybe it was a buck converter? i'll have to look into it.

[isaac_alaska]

CM400DY-12H Powerex IGBT Module 600V 400A New | eBay

what do you think of this one? more expensive than i was hoping for, but it DOES take the place of the power mosfets, power diodes, clamps, screws, thermal interface material, and soldering time. now the question is....how much overkill would it be to run four in parallel? since i'm planning to run 400 volts in and then only 200 volts out....the max current out should be twice the current in. we're getting up into the zilla 4k range if i run four of them in parallel.

what are your (anyone who reads this) thoughts?

[harlequin2]

Originally Posted by isaac_alaska

CM400DY-12H Powerex IGBT Module 600V 400A New | eBay

what do you think of this one? more expensive than i was hoping for, but it DOES take the place of the power mosfets, power diodes, clamps, screws, thermal interface material, and soldering time. now the question is....how much overkill would it be to run four in parallel? since i'm planning to run 400 volts in and then only 200 volts out....the max current out should be twice the current in. we're getting up into the zilla 4k range if i run four of them in parallel.

what are your (anyone who reads this) thoughts?

What do you intend to drive with this?
If you are limiting the duty cycle to 50% - and I wonder why you would do that, how are you going to limit the current?

[isaac_alaska]

driving a netgain warp7 motor with it. i was planning to limit the duty cycle to 50% to limit the voltage to the motor to half the pack voltage. Current limit would be handled by the controller.

higher pack voltage smaller heating losses in cables, battery leads, and igbt's themself, but the motor doesn't like to see much more than about 200 volts. HIgher pack voltage also means we need to parallel fewer cells (since we will be paralleling the battery at the cell level, rather than simply paralleling each string.) Also, if we end up switching to a higher voltage motor, it would be a very simple matter to re-program the controller to allow a full 100% duty cycle.