Can't Kill A Toyota Inverter

[Bigpie]

Thanks, I've not got that far with my inverter yet, but I'll add to the wiki page as I get to it.

I guess this makes the jumper on the board redundant as the DCDC is now under control of the Atmega.

[Jack Bauer]

The atmega does not control the dcdc converter. Just the buck boost stage. Jumper is still required.

[celeron55]

So what do you think, should we have a default non-charger program for the atmega328 that maintains about 200V at the dc-dc side so that the dc-dc will work without modifications no matter what the battery voltage is?

[Bigpie]

The atmega does not control the dcdc converter. Just the buck boost stage. Jumper is still required.

It'll not work unless the Atmega turns on the high side of the buck boost though, unless I'm not following?

The jumper in the current v1c still requiring connecting, but it can be turned on or off via the Atmega

[Jack Bauer]

So what do you think, should we have a default non-charger program for the atmega328 that maintains about 200V at the dc-dc side so that the dc-dc will work without modifications no matter what the battery voltage is?

That would be the ideal solution.

[Jack Bauer]

The atmega does not control the dcdc converter. Just the buck boost stage. Jumper is still required.

It'll not work unless the Atmega turns on the high side of the buck boost though, unless I'm not following?

The jumper in the current v1c still requiring connecting, but it can be turned on or off via the Atmega

Yeah you always need the enable line on the V1c to be either jumped or wired externally. You then have two choices to supply hv to the dcdc side. Either just jump it with a wire or run the little code snippet above on the atmega.

[Bigpie]

Cheers. I like the idea of celeron55, it would be nice to not have to change the resistors.

Do you know if the buck/boost has the unkillable Toyota spirit? If I inadvertently enabled high side and low side while playing with the atmega, would it ignore me and laugh or self destruct?

[Jack Bauer]

Do you really need to ask that question:)

[johu]

So in other words you cannot connect the battery to the original 201V input and run the inverter via the boost inductor. Like "B+ -> inductor -> diode -> inverter" AND expect this to maintain good (>30kW) power output? Or does 30kW just refer to the boost converter actually operating?
And that would mean you can not use the buck/boost stage as a battery charger while using the inverter as inverter.

[celeron55]

If the boost converter's limiting factor is something else than raw current carrying ability, then maybe you could just connect it like it originally was intended and just drive the boost high side continuously on. If someone has a spare inverter to blow up, that's definitely worth testing for cases where one has Toyota's nice HV input connector handy. Altough, you might as well just add an insulated bus bar link to do the same thing...

But options are nice. It's a software solution to a hardware problem!

I'm going to bet Toyota's hardware will limit this in some annoying way.

[ZooKeeper]

post deleted

[xp677]

Do you know if the buck/boost has the unkillable Toyota spirit? If I inadvertently enabled high side and low side while playing with the atmega, would it ignore me and laugh or self destruct?

From experience with the GS unit, which I imagine is similar, the IPM should prevent this occurring. However I've had a GS converter self destruct with signs of a HV - Ground fault, after several hours (probably 20 or so hours) of testing. So I'm not sure how robust that protection is.

Unless I completely mis-understood the documented features of a Prius Inverter/Converter.... The boost feature both increases MG1/2 output during "regen" to get volts high enough to recharge the battery AND boost B+ to allow higher than pack voltage to drive MG1/2 according to power demand.

The converter bucks under regen to charge the battery with a "safe" voltage.
The converter boosts under load to drive the motors.
The generator (MG1) can be used to power the motor (MG2) directly through the DC bus.

So in other words you cannot connect the battery to the original 201V input and run the inverter via the boost inductor. Like "B+ -> inductor -> diode -> inverter" AND expect this to maintain good (>30kW) power output? Or does 30kW just refer to the boost converter actually operating?
And that would mean you can not use the buck/boost stage as a battery charger while using the inverter as inverter.

The 30kW rating is down to the max current the converter IGBTs can withstand, and how effective their cooling is. If you wanted to run the car solely through the converter, you'd need to suitably fuse it to protect the converter. This is one of the reasons why you can drive in battery mode at low speeds, but as soon as you try to accelerate, the engine starts up.

You could use the converter as an interface to a battery charger, if it suits the project, and if the voltages are suitable, but you'd be limited to 30kW again.

IIRC, some Prius converters are smaller than 30kW.

[celeron55]

I think the most useful way to think about what toyota is doing in these things is to primarily think of them as e-CVTs where simply mg1 makes electric power from ICE output and mg2 consumes it, and the pwm ratio creates the virtual gear ratio. It does not need a battery to function and operates at its own independent voltage.

Then, just to have some electric mobility, a battery at its independent voltage and other characteristics is added via a buck/boost converter to suck out and inject power into the e-CVT. It's an added secondary energy reserve. Any voltages internal to the e-CVT are independent from the battery.

Now, it just happens that you can drive using battery power only using the original system, and you can plug a battery into the e-CVT bus bars and it can utilize as much power from that as it could from the ICE. But trying to think what toyota is doing from that perspective puts you in the wrong mindset. It's not how the toyota engineers thought about it.

[clanger9]

Perfect explanation, Celeron55! It's an electronic CVT first and foremost. That's what the Toyota designers were originally trying to build (and is why the DC bus flaps about so radically). The buck/boost converter and battery are "extras" that came later.

Speaking of which: does anyone know what the maximum power capability of a Prius 4 buck/boost? Or do I need to buy one and try to blow it up? I have a cunning plan (and I need ~35kW..).

[edit] Answering my own post, the 2017 Prius inverter is designed for 53kW / 600V.

[arber333]

Lets continue the charger debate in gen3 charger thread. I may have an idea on how to implement both traction and charger in the same housing.

[Jack Bauer]

So when I posted some videos of the Prius inverter cap failure in the E39 I was surprised by some comments. The short version seemed to be that even though I had admitted to doing it wrong, I was still doing it wrong in my theory of the failure. So while waiting for the replacement inverter for the E39 I continued with experiments on the Range rover. The new encoder worked wonders and I could finally drive somewhat normally using just the rear motor on the MG2 power stage of the IS300H inverter. Drifting off to sleep on the previous night a thought popped into my head. Could I use one encoder and one Johannes brain to drive the two enova induction motors. Reasoning was they are the same pole count etc, same gearbox and same ratio to the road. Well it seems to work just fine:) I started building an obstacle course in my drive way and set about some rock crawling. Working great. Then I hear the pop. Jumped out and lifted the bonnet to see a thin whisp of smoke from the inverter. Cut the power, removed and dismantled the inverter to see....yep same failure mode!

Why? and why now? Well the Rover is only running 240vdc hv so the caps had more headroom them in the E39 at 360v. And why now because two motors running on the two power stages for twice the inductive pleasure when moron features here....you guessed it! had the HV connected to the WRONG TERMINALS...............

[clanger9]

Never mind a few minutes: with all this practice at doing it wrong, you'll soon be able to destroy Toyota inverters in *seconds*!
I look forward to the videos.

[clanger9]

PS: would you be able to get an oscilloscope trace of the spikes on the DC bus when "doing it wrong"? Just for the LOLs, of course.

[Mouse]

PS: would you be able to get an oscilloscope trace of the spikes on the DC bus when "doing it wrong"? Just for the LOLs, of course.

Not just a oscilloscope trace but a Before and After trace to show the difference of attaching the power to the right place despite the unintuitivly short connecting bus bars. It would be very educational.

[Jack Bauer]

Decided to do a little autopsy on the cap. Found two failure points. Now, the is300h inverter does not have any snubbers annnnd it seems to feed the main HV from roughly the same points I was using.

[RetroZero]

On the one hand I'm glad this happened to you and not me, but on the other, even though we learn MOSTLY from our mistakes, I hope it has not caused too much damage...It's not dead but "wounded"...
But Damien, you're doing it WRONG

[Mouse]

Thanks for the autopsy.
There's a saying that goes along the lines of... If you're not breaking something you're not trying hard enough.

[celeron55]

So the facts are: (are they?)

1) There are voltage spikes, and
2) the battery is connected to the same location the boost converter output goes to, and
3) the inverter was very unlikely to be broken when you started testing.

So what can this even mean?:

1) The caps are incorrectly rated, or
2) The power stage is run incorrectly, maybe there has to be a specific deadtime that the power stage doesn't enforce or maybe they have to be run somehow synchronously compared to each other so that one motor always snubs the inductance spikes caused by switching the other motor (a motor is an inductor though? doesn't make sense), or
3) you can't feed power to these things directly using long wires from a battery, because maybe any wires or batteries have too much inductance compared to the buck/boost converter coil (I'm sure that's not the case... an inductor has more inductance than cables? right? maybe the main cap is tuned for _more_ inductance than a cable has and it only then doesn't explode? is that even possible?), or
4) the boost converter has to be controlled somehow in order for this not to happen. (This seems very unlikely, it should be unable to affect this in any way positively no matter how you ran it, as towards the main cap it either looks like a coil or an open circuit)

Maybe there is a better explanation.

Notably the Yaris inverter is similar to this in that it seems to have only one(?) snubber compared to 3(?) snubbers in the Prius inverter.

[Jack Bauer]

Found the culprit. celeron55 is very very close in one of his points above. Video later.

[Jack Bauer]