Overheated Inverter - Need Expert Opinion

Overheating is not good for electronics and can cause premature failure. You already know that or you wouldn't be posting this question. The real question is whether your inverter has been damaged. If it isn't throwing a code, there is no way to prove damage. I can't see a dealer or Toyota agreeing to replace an expensive part that has not failed. That said, if you don't have any codes, your inverter is probably fine.

Tom

 

Hi PD,

I doubt your going to get them to replace the inverter. Possibly you could get them to extend the waranty on the inverter, based on the possible effects of a waranteed item that failed early.

 

if there is no evidence of damage to the inverter and you choose to insist it is replaced, you will be paying for the replacement.

 

The inverter is covered by the 100K mile hybrid system warranty. So you've got coverage on that component for a reasonable amount of additional time and mileage.

 

Make sure that the problems and potential long term impact on the inverter are documented by the dealer. It doesn't make sense to replace it at this point. However, it if fails later due to the other warranted part that failed, you will have a pretty good basis for replacement. Of course, as Patrick noted, you already have a very long warranty on the inverter, but it doesn't hurt to have something to reference back to if they try to blame something on you later...

Do they do a flush and replacement of the inverter coolant system when they replace the pump? If not, then it really should be done. What about the transaxle fluid? Would it get a bit toasted since MG1/MG2 lost coolant and ended up in lockout? Seems like it would be prudent to flush it as well.

Seems like Toyota needs to give us some decent gauges so that we have some warning when parts like the coolant pump are failing. It's silly to have more ECU's than you can even list, and virtually nothing in the way of an interface for them. Even their dealer shops have trouble with stuff like this. A simple temp. readout on the inverter coolant would eliminate a lot of problems and guesswork.

 

So I suppose Scangauge II would not read the inverter coolant T?

 

If it had the proper X-gauge set up for it, it could, but to my knowledge no one has figured out the X-gauge code for it from the CAN data stream.

 

Ask the dealer whether SSC 50P needs to be performed or repeated.

 

I realize this is about 6 months too late, but here goes...

You're fine. The behavior of the system (all of the warning lights) tells you that there are temperature sensors in the inverter that cause the computer to shut it down before there is actual failure of the system.

The high-power parts in the inverter are IGBTs (Insulated Gate Bipolar Transistors). These parts can function well above 100 degrees C - common testing temperatures of these parts for qualification testing is on the order of 140 degrees C, so there's quite a bit of margin.

If the IGBTs themselves overheated, they would short (catastrophic breakdown) causing the bond wires attached to them to blow open.

Because your inverter is still working, you didn't damage anything. You might have marginally reduced the component reliability, but not significantly so.

If it were my car, I wouldn't bother having them replace the inverter.

(I work in the extremely high reliability electronics business - stuff we put out has to run for 15 years straight with NO possibility of repairing it, so we're pretty anal about this sort of stuff).

 

It should be pretty easy to measure the gate current. If there is any damage at all then this will show as a gate current increase, problem is you need a reference or the IGBT datasheet.

I wonder if it is actually measured by the ECU.

 

I don't think so. The IGBT has an insulated gate (meaning that there is effectively no DC gate current). The Field Effect Transistor (FET) structure of the IBGT isn't the thing that breaks down during overheating - in fact, FET structures are well-behaved during overheat, and tend to decrease in current (because they are majority carrier devices).

During overheating of a bipolar transistor, the current increases (because the resistance of a minority carrier device has a negative temperature coefficient). Unless the circuit is otherwise current limited, this causes more power, more heat, and even more current, until the transistor blows.

When the transistor is overheated and overcurrented, the emitter region punches through to the collector region, and the transistor goes into secondary breakdown. At this point, the transistor fully conducts and can't be shut off. Unless the collector current is otherwise limited, the bond wires of the transistor will blow (spectacularly I might add).

None of this happened though - the thermal sensors in the inverter told the computer to shut things down way before this happened. Don't worry, be happy, drive the car.

 

Note to self, don't post before breakfast and stop posting crap.

 

Where are these temperatures measured? I imagine there is a large temperature gradient from the active region to the surface of the device packaging and on to the coolant.

 

The temperatures that I mentioned would be junction temperatures. Those are the temperatures within the devices themselves.

There will be some thermal resistance between the part and the cooling plate, but I'd be surprised if it was anywhere near one degree per Watt.

The neat thing about the design of these switching systems is they waste very little power. Either the transistors are off and the current is ~ 0 Amps, or the transistors are turned strongly on, and the Voltage across them is small. Either way (since the power consumed in the transistir is Current x Voltage) the transistors themselves burn very little power while controlling lots of juice.

But the key point is my original one: The inverter's fine - drive it and be happy. I would.

 

This is good info to know. Thanks again!