Help making sense of Megohmmeter results from transaxle and inverter

I wish you had asked first.

The highest voltage used at the motors in Gen 2 is 500 V.

The test at 1000 V may have been within the safety factor of the internal insulation.

It may not have survived the tests at 2500 V and 5000 V.

 

Yeah, that's nine volts, of course, but the voltage settings on the meter aren't kidding around; a boost converter inside the meter takes the nine volts up to the selected test voltage.

The tester doesn't use a lot of current. But you might need to do some reading on what could have happened to the insulation in there while exposed to 5,000 volts.

Here are a couple articles I found in some quick googling:

The Professor: Megohmmeters for Preventive Maintenance | 2011-12-05 | ACHRNEWS | ACHR News

"Any megger with a higher voltage output than 500 volts DC should be used by an experienced technician. A high voltage for too long of a time may further weaken or fail motor windings and the winding insulation could be damaged by the testing procedure."

https://protecequip.com/blog/can-using-a-megger-meter-be-destructive/

"It is always a good practice to be familiar with the designed limitations of the insulation along with the proper voltage at which to perform the test. This knowledge will help to ensure that the insulation is not damaged or compromised as a result of this type of testing,"

You can also find references that say it's no big deal and the insulation isn't lastingly degraded as long as it didn't actually break down during the test. So there is some diversity of opinion out there, and I can't really tell you what to make of it, because I've really never been in this situation before.

Part of the story might be in another quote from that protec blog;

"While there are ways to electrically stress insulation, insulation resistance testing is typically focused on characterizing the insulation rather than stressing it. Therefore, these type tests are normally performed at voltage levels that are lower than the rating of the insulation and generally much less than if the intent was to stress the insulation."

In other words, some of the opinions out there that megger testing doesn't do lasting damage may be based on a (stated or unstated) "because people choose their test voltage settings to avoid that". So you're kind of in uncharted waters here, and I can't help much with the decision on whether it's a good idea to put those units back in service.

As for your actual numbers, they're not so easy for me to read, as I'm guessing you wrote them all as something at-sign something, so for me they all show up as

[email protected] 296 and rising
[email protected] 273 and rising

and so on. Maybe if you edited to put spaces around the at-signs or something, that wouldn't happen. Or maybe just say whether you've ordered them with the 500 V tests at the top or the 5000 V tests at the top.

The "and rising"s probably reflect the capacitance and dielectric absorption effects you might have read about in A stitch in time.

As far as interpretation, Toyota seems to have given pretty much a go / no-go criterion; 10 MΩ or higher good, less than 10 MΩ bad. On that basis, your MG2 results look ok, MG1 or its cable, not so much.

Your second set of readings also look low. But that may involve a weirdness in that edition of the repair manual that somebody else pointed out here only recently. (I'd look that thread up for you but the member content search seems to be broken at the moment.) It was something like, step 17 and step 18 in the P0AA6 troubleshooting are really the same test—you arrive at them by different paths, chosen up around step 2—only the second one has the arrows in the illustration pointing the wrong way. The text beside the illustration still says "cables on the transaxle side" just like the earlier test, and the go / no-go criteria are, of course, the same.

But if you followed the flipped arrows in the drawing and applied the megger voltage into the inverter itself, that could explain the difference in readings. For the interpretation, you might be able to search the manual for some other troubleshooting step where they really do say to test the inverter side, and see what go / no-go numbers are given there. I don't remember for sure, but I think there might be such a test in there somewhere, and the expected resistances might be lower.

I also don't know what the higher test voltages might have done to the inverter.

 

Clearly something was botched in the editing, so some interpretation by the reader is needed to work out what they intended, and whichever one of us is right, your interpretation is the reverse of mine. I'll stick with pictures bad description good.

A short would be a value lower than the 10 MΩ, and that's the NG result. 10 MΩ or higher is the OK result, in both cases.

But notice that the positions of "OK" and "NG" have also been swapped in step 18 compared to 17. In 17, NG is to the right and OK is down. In 18, OK is to the right and NG is down.

In contrast, there is no change in what OK and NG tell you to do. NG tells you to replace the transaxle (because it's the thing you just tested, and the result was NG, so it's at fault). OK in both 17 and 18 tells you to replace the inverter (because the transaxle is what you just tested, and the result was OK, and the inverter is what's left).

Notice also that there's no sequence where you would do both step 17 and step 18 (which would be redundant, as they're the same test). If you reach step 17 (via an OK result from step 16), you will replace one thing or the other based on the step 17 result, and be done. You only get to step 18 if you jumped there from step 2 (C).

Studying this whole troubleshooting tree, it turns out there is no step where they want you to touch the inverter itself with the megger. That kind of makes sense, as the inverter's a complex device with semiconductors and not the same simple resistance behavior as a motor or a wire (to say nothing of what test voltages might do to it). So also in step 19, for example, if they want to decide between replacing the frame wire or the inverter, they have you megger the frame wire. Replacing the inverter is what you do by elimination if the frame wire checked out OK; what they don't have you do is megger the inverter.

Going back in time to the Gen 1 manual, there they do have a step where you megger the inverter. That's probably the step I was remembering. It gave a go / no-go criterion of 0.9 MΩ for the inverter circuitry, compared to the 10 MΩ specified for motors and wires.

The Gen 1 manual also specified the test voltage to use, 500 V. Gen 1 ran the inverter and motors at 274 volts, so they specified a test voltage not quite twice the working voltage, which seems like a typical choice. Because Gen 2 added a boost circuit up to 500 V, it's maybe not outlandish to use a 1000 V megger setting on its motors and motor cables, again roughly twice the working voltage. Still, I'd rather have the manual just say what voltage to use, as it did for Gen 1.

 

When you do step 18, you've disconnected the MG1 and MG2 cables from the inverter, but left the other ends connected to the transaxle. Whatever you read between a U, V, or W phase lead and body ground is a test of both that cable and the winding of the corresponding MG.

I assume that all the reading numbers you've given are in MΩ. Anything 10 MΩ or higher is fine, per the instructions. All of your MG2 ("Hybrid Vehicle Motor") readings are handily above 10 MΩ, so those are fine.

Your MG1 ("Hybrid Vehicle Generator") readings are severely under 10 MΩ, and therefore not fine. The 5000 volt readings are different, which is probably a limitation of your meter. Ordinarily if you find an unacceptably low resistance at a lower voltage (like the 10 kΩ at 500 V), there is no point proceeding to test at higher voltages. Even a 500 volt test into a 10 kΩ resistance would require the megger to source a twentieth of an amp, and at 500 volts that's 25 watts, quite a lot of work for a six-AA-powered megger. Realistically, it was not reaching 500 volts for that test, and didn't need to; it simply divided what voltage it did produce by the current it saw, and showed you the quotient. The higher-voltage tests into that low a resistance are surely just reporting "0.01" as underflow for an unrealistic test, and at 5000 volts it was so unrealistic some bogus arithmetic happened.

All of which is only an effort to explain the weird pattern of readings where the ones at realistic voltages are all showing 10 kΩ. Those details aren't necessary to see that 10 kΩ is a thousand times too low to be a passing result for MG1.

So unless you have the good luck to isolate MG1's cable and find the fault to be just in it (which would be an inexpensive fix), you have found out that the fault is in MG1.

 

Thanks for the update!

For others reading along, the reason a megger shows rising readings on a piece of big equipment like a motor is that the megger is putting energy in, at a slow rate (the megger uses a low current to test with), and the equipment is storing that energy.

It's usual to wait for the rising to taper off, and take down the reading at that point, after the megger has stored energy in the equipment, up to its test voltage. 30 seconds or a minute may be enough.

And yes, that does mean at the end of a test, say at 500 volts, that motor is holding a charge up to 500 volts, and that can hurt you. The megger likely has a function built in that slowly bleeds the charge back off the equipment after you release the test button; sometimes the display will change to voltage and you can watch the voltage declining until it is safe to take off your test leads. Do not skip that part and just blithely pop your test leads off right at the end of the test.

A large motor has two different ways of storing energy like that: the capacitance of its windings, and polarization of its insulation. The capacitance part accounts for the rapidly rising reading at first, and then the polarization (or dielectric absorption) is why the reading continues to rise more slowly over the next half minute or so.