Paul and Sabrina's Cheap 3 Phase Inverter (AC Controller) with Field Oriented Control

[thingstodo]

Originally Posted by MPaulHolmes

Or you could do the testing with the base plate sitting in water. It could get to 100 degC, and then would definitely hold there, because it takes a ton of energy to evaporate water.

How well would the controller do in a pan of boiling water?

I could try a pan of oil .. but the cat would likely want to BATHE in it .. makes a mess ...

[MPaulHolmes]

I have a more practical idea. Take the controller to the top of Mt. Everest. Then, put it in water. There is a lower boiling point on Mount Everest.

[e*clipse]

How 'bout this - I mean he is doing the testing in a pretty cold place already; use it to your advantage!

- bolt the stuff to a big aluminum plate - that battery pack one would be good.

- Set all of it out in the driveway

- Wait 'till about midnight, when it's nutty cold, then start testing. The dT between a worrysome 50*C component and an ambient of -40*C is pretty huge. That alone might work - you won't need much of a heat sink.

If it starts snowing, you may want to stop. -

- E*clipse

[e*clipse]

Paul - wiring questions for the gate drivers

- what IGBT's did you design the version one ACControlAndDriver board for?

- the connector for DESAT - on the physical board, it's just a small hole. Does this line up with a pin, or do you run a wire?

- Can I just run separate gate resistors to each individual mosfet from the control transistors (for example Q1 and Q9) ?

- For each mosfet, should there be a corresponding diode to the emmitter ( for example D1)

- How is the output voltage controlled?? How is the +15 , -8.2 V set? Does the Attiny/ transformer do this?

- E*clipse

[MPaulHolmes]

Originally Posted by e*clipse

Paul - wiring questions for the gate drivers

- what IGBT's did you design the version one ACControlAndDriver board for?

- the connector for DESAT - on the physical board, it's just a small hole. Does this line up with a pin, or do you run a wire?

- Can I just run separate gate resistors to each individual mosfet from the control transistors (for example Q1 and Q9) ?

- For each mosfet, should there be a corresponding diode to the emmitter ( for example D1)

- How is the output voltage controlled?? How is the +15 , -8.2 V set? Does the Attiny/ transformer do this?

- E*clipse

ACControlAndDriverBoard is designed to plug right into 600v 50amp igbt half bridges up to 1200v 600amp IGBT half bridges. The pin spacings are the same for all of them.
Powerex Product: CM600DY-12NF

So, no wiring is required for the gate and emitters. The collector is either the B+ tab, which requires a wire, or is the high side emitter tab (E1 = C2). So, a wire is intended to be used for C1 and C2 holes (for desat detection).

Yes, you can run multiple resistors. For this design, it's 2 gate resistors per IGBT. One for turn on and one for turn off. So, you would have 3 turn-on resistors going to the 3 separate mosfets and 3 turn-off resistors going to the 3 separate mosfets. There's around 8 amps (continuous!! lol) of drive current available for those npn and pnp package transistors, so I'm sure you could probably do 12amp peak.

The reason there is the NPN -PNP totem pole arrangement is because this was required to boost the 2.5amp from the SOIC driver, but still maintain the soft turn-off feature of the driver when desat is encountered. If we just dumped those, and put a regular old mosfet/igbt gate driver in place of the npn-pnp totem pole, a desat event would turn off the igbts in like 8nsec, which would possibly destroy them.

A separate back to back zener diode pair for each mosfet would work, but the 500amp controller just uses one TVS diode right at the gate driver (before the gate resistors). However, Otmar says the protection diodes should be right at the legs of the igbts or mosfets (if we are talking to-247).

The input voltage is 24v. The attiny pushes and pulls the gate drivers, which cause the output to be close to 24v too, after it gets rectified. Then, you do this:

+24v ------------- 3k resistor ----------< 8.2v zener ---------- 24v ground

Then, we call +24v "+15v", and we call the spot between the 3k resistor and the 8.2v zener cathode "emitter", and we call the 24v ground "-8.2v". The output voltage is limited to no more than a total of +24v by a 24v zener from +15v down to -8.2v (also known as +24v down to 24v ground). There is no regulation on the output voltage other than keeping it <= 24v. The voltage drops as the current increases, but in practice, the current just never goes all that high, so the voltage doesn't sag much on the output. Generally, the real voltages for switching are around -8.2v up to 15.8v or something like that. Sometimes 15.3v, etc...

[thingstodo]

Originally Posted by e*clipse

How 'bout this - I mean he is doing the testing in a pretty cold place already; use it to your advantage!

- bolt the stuff to a big aluminum plate - that battery pack one would be good.

- Set all of it out in the driveway

- Wait 'till about midnight, when it's nutty cold, then start testing. The dT between a worrysome 50*C component and an ambient of -40*C is pretty huge. That alone might work - you won't need much of a heat sink.

If it starts snowing, you may want to stop. -

- E*clipse

Good ideas.

There are some issues, though. When you put stuff outside to cool off, the humidity condenses on the surfaces - water right now and frost a bit later. No big deal for the battery base. Not so great for the controller. So I think I'll bolt the controller to the cold base after bringing the base in from my driveway.

It's not yet (thank goodness) -40F outside. More like 35F. But that's still enough for my first set of tests. So I'll try that out.

I'll likely funnel distilled water into the motor water cooling connections and let the load on the motor warm up the water as the motor is used. That should be enough cooling for 30 minutes. I will be monitoring, so if it isn't enough I'll change something.

[e*clipse]

Originally Posted by MPaulHolmes

ACControlAndDriverBoard is designed to plug right into 600v 50amp igbt half bridges up to 1200v 600amp IGBT half bridges. The pin spacings are the same for all of them.
Powerex Product: CM600DY-12NF

So, no wiring is required for the gate and emitters. The collector is either the B+ tab, which requires a wire, or is the high side emitter tab (E1 = C2). So, a wire is intended to be used for C1 and C2 holes (for desat detection).

Ok, I'll add another connector for each phase output and **at least** one for the B+ bus. Is the situation dynamic enough that local measurement for desat is necessary? Should I put that connector close to each group of TO-247's as well?

Yes, you can run multiple resistors. For this design, it's 2 gate resistors per IGBT. One for turn on and one for turn off. So, you would have 3 turn-on resistors going to the 3 separate high side mosfets and 3 turn-off resistors going to the 3 separate low side mosfets (if you were using 3 high side and 3 low side). There's around 8 amps (continuous!! lol) of drive current available for those npn and pnp package transistors, so I'm sure you could probably do 12amp peak.

I haven't looked yet, but 8 amps should be appropriate for 3 TO-247's rated at 100A, right?

The reason there is the NPN -PNP totem pole arrangement is because this was required to boost the 2.5amp from the SOIC driver, but still maintain the soft turn-off feature of the driver when desat is encountered. If we just dumped those, and put a regular old mosfet/igbt gate driver in place of the npn-pnp totem pole, a desat event would turn off the igbts in like 8nsec, which would possibly destroy them.

Gotcha, thanks!

A separate back to back zener diode pair for each mosfet would work, but the 500amp controller just uses one TVS diode right at the gate driver (before the gate resistors). However, Otmar says the protection diodes should be right at the legs of the igbts or mosfets (if we are talking to-247).

I'll try to use Otmar's suggestion on this; hopefully there's space.

The input voltage is 24v. The attiny pushes and pulls the gate drivers, which cause the output to be close to 24v too, after it gets rectified. Then, you do this:

+24v ------------- 3k resistor ----------< 8.2v zener ---------- 24v ground

Then, we call +24v "+15v", and we call the spot between the 3k resistor and the 8.2v zener cathode "emitter", and we call the 24v ground "-8.2v". The output voltage is limited to no more than a total of +24v by a 24v zener from +15v down to -8.2v (also known as +24v down to 24v ground). There is no regulation on the output voltage other than keeping it <= 24v. The voltage drops as the current increases, but in practice, the current just never goes all that high, so the voltage doesn't sag much on the output. Generally, the real voltages for switching are around -8.2v up to 15.8v or something like that. Sometimes 15.3v, etc...

Ok, I think I got it. So the 3K resistor just limits current; it doesn't have anything to do with the output voltage from it's voltage drop...
If the switches required different voltage - example: ( +20 turn on, -5V turn off )
http://www.wolfspeed.com/~/media/Fil...2M0025120D.pdf

Then one could change the zener diode if necessary. Also could the Attyiny modulate the pulses - I assume they're 50% - for something less than 24V?

Thank you very much for the info; I'll post a schematic for the gate drive here when I get done.

- E*clipse

[MPaulHolmes]

turning the IGBTs off and on only takes the 8 or 12 amps for less than a microsecond with a duty of maybe 1%. Those pnp and npn transistors are rated for 8 amps continuous, so I think you could do just about any old current you want for very short bursts. Probably 10 Ohm resistors would be fine, and that would be around 7.5amp peak for 3 of them. But I doubt 4.7Ohm gate resistors would cause a problem.

I don't think you need local checking for desat detection. It was just mechanically convenient for me to have 3 local spots connected to 3 separate B+'s. I'm not sure how well the desat detection converts to mosfets though. IT's designed for IGBTs. Do mosfets have a 5 or 10uS short circuit capability ever? And would a drain to source on voltage of like 6 or 7 volts destroy the part before it had a chance to shut off? Now that I think about it, I bet it would work OK. A lot of these parts can dissipate several hundred watts. So, 600 or 700 watts (7volts * 100amp) probably would be OK for 1 or 2 microseconds. crap... Let's say the RdsON was 10mOhm. I don't know what it would really be. Then, to get to the required 7v desat on voltage, it would take 700amps. That would be 4900 watts. haha. Maybe desat isn't practical in that case. I don't know. I do know I've tested it on accident with IGBTs, and it works very nicely. It shuts off before the hardware overcurrent circuit from the LEM Hass current sensor. So, it's quicker than 3uS.

Yes, if you want -5v, just use a 5v zener. But if you are push-pulling with 24v, it will only be around 19v down to -5v. And yes, I think you could reduce the duty so that you could control the output voltage. Right now it is 50%.

[e*clipse]

Awesome - thanks a bunch Paul! This actually goes into some questions I thought of later. I'm getting close to fitting the driver on the board with individual gate resistors and back to back protection diodes. It looks possible to put each gate drive on the top of a 2 layer board, which will help make soldering easier.

On the subject of saving some space, what's the reason for the two banks of four 22uF capacitors - the ones tied to +15V , IGBTemitter, -8.2V? Are these buffer capacitors or filter capacitors? Why not just have one capacitor each at about 48uF? Or if these are filtering capacitors, one 22uF and one 66uF capacitor each?

Originally Posted by MPaulHolmes

turning the IGBTs off and on only takes the 8 or 12 amps for less than a microsecond with a duty of maybe 0.001%. Those pnp and npn transistors are rated for 8 amps continuous, so I think you could do just about any old current you want for very short bursts. Probably 10 Ohm resistors would be fine, and that would be around 7.5amp peak for 3 of them. But I doubt 4.7Ohm gate resistors would cause a problem.

This brings up the question about the gate drive's power consumption. I checked into the driver that push-pulls the transformer, and it's rated at 4 amps. So, if everything were matched perfectly 24V * 4A = 96watts - - - Close enough?
This would bring some opportunities to downsize certain components a bit - for example, it sounds like the gate drive transistors could be smaller.
and I've found a transformer that is much smaller, but a bit more expensive. It's made by Pulse and can take 5A, with something like a 1500V Hi-Pot test. This actually is a deal because many of the transformers I looked at could only handle < 100V. There's also a similar sized one my Eaton that can handle 4A and is "rated" for 300V for about $1 more.

I don't think you need local checking for desat detection. It was just mechanically convenient for me to have 3 local spots connected to 3 separate B+'s.

That's good to know about not needing a local desat reference. That will save 4 connections because I can use the 10mm pins that are already there.

I'm not sure how well the desat detection converts to mosfets though. IT's designed for IGBTs. Do mosfets have a 5 or 10uS short circuit capability ever? And would a drain to source on voltage of like 6 or 7 volts destroy the part before it had a chance to shut off? Now that I think about it, I bet it would work OK. A lot of these parts can dissipate several hundred watts. So, 600 or 700 watts (7volts * 100amp) probably would be OK for 1 or 2 microseconds. crap... Let's say the RdsON was 10mOhm. I don't know what it would really be. Then, to get to the required 7v desat on voltage, it would take 700amps. That would be 4900 watts. haha. Maybe desat isn't practical in that case. I don't know. I do know I've tested it on accident with IGBTs, and it works very nicely. It shuts off before the hardware overcurrent circuit from the LEM Hass current sensor. So, it's quicker than 3uS.

About needing desat detection at all - VERY interesting point. Since mosfets are breaking into the high voltage realm ( with SiC Mosfets ) maybe just don't bother with IGBT's? The current is limited to what the legs of the TO-247 can handle, which is covered by Mosfets.

So desat really relies on a big RdsON to work? Very interesting - because I'm trying to get away from as many losses as possible with this design. Maybe I should just commit to Mosfets... Wow, that's got my head spinning.

Yes, if you want -5v, just use a 5v zener. But if you are push-pulling with 24v, it will only be around 19v down to -5v. And yes, I think you could reduce the duty so that you could control the output voltage. Right now it is 50%.

Very cool!
Thanks again, Paul - I'd like to add that you are an excellent teacher.

- E*clipse

[Hersbird]

Soooo, is there a condensed version of this thread with details of what the title says? I just can't do 230 pages when just knocking around the idea of electrifying my Forester. I know the "kits" out there are way to much for me, but the couple posts above mine hardly give the impression of DIY.