Simon's new Xantia V6 and Leaf blog

[Gibbo2286]

Are these inverters repairable?

Everything is repairable Chris fancy getting inside one of these? This is a Seimens example.

I'm up for having a go at repairing anything electronic (within reason) but even I'd balk at this one... The scope for epic fireworks and injury are very, very significant...

I say that, knowing the principles of operation of these things; the chances of a successful repair are low. Many try to repair switch-mode power supplies - they operate on similar principles - and invariably even the most skilled have to give up in frustration at the continual failures, almost fixes and lots spent on new components...

They injure their fingers too from the continual crossing of them they do...

Ah! but think about it Jim, you pop the cover off and staring you in the face is a blown resistor or a puffed up capacitor, you'd have a go wouldn't you.

[Mandrake]

Are these inverters repairable?

With the right service data, in principle yes.

As with most complex electronics these days, the problem is getting the service data out of the manufacturers, which can be like trying to get blood out of a stone.

Most failures are likely to be high power components like the IGBT's. Another possible failure mechanism for a drive inverter might be a coolant leak inside the case - since they're liquid cooled using the same glycol loop as the motor and front radiator, a leak of this coolant onto one of the boards would be pretty devastating and probably write it off. Without seeing how it is designed inside though it's not clear whether a leak of that type would be possible.

It would certainly help to know what fault codes it is giving at the moment!

Price on request for a second hand one. Ha!

https://www.proxyparts.com/car-parts-st ... d/7736095/

Might need coding to the EV ECU with a Lexia before it will start... maybe...

You can see the coolant loop spigots - one on the side and one on the bottom...

[Mandrake]

I'm up for having a go at repairing anything electronic (within reason) but even I'd balk at this one... The scope for epic fireworks and injury are very, very significant...

I'd baulk too Jim, unless it was my car that failed and the alternative was scrapping it. Then I'd have a look, but without circuit diagrams I'm not sure how far you could get on one of these!

I say that, knowing the principles of operation of these things; the chances of a successful repair are low. Many try to repair switch-mode power supplies - they operate on similar principles - and invariably even the most skilled have to give up in frustration at the continual failures, almost fixes and lots spent on new components...

To be fair EV drive inverters are built to a much higher standard than typical switch-mode supplies in say a TV or a PC. PC power supplies are one of the worst, I've repaired many over the years (not for a while now) and apart from a few high end brands all use the same basic design and the same shoddy quality components and cut the same corners.

Having repaired many and seeing how awful they are I'm actually amazed that some PC power supplies make it out of the factory still working let alone past the warranty period!

[Mandrake]

How much is a new inverter... car up to £1300, could be a few quid to be made in repairing it! I'm making a space in my garage.... a narrow Ion would hide in there no problem

See the link in a previous post - "price on request". In other words, we have one of the only scrapped ones in the world, how much will you pay us for it...

While I would attempt to fix my own drive inverter, I would not buy a non working car and a second hand inverter of unknown provenance and expect to get a working car at the end of it. There are too many things that could go wrong like hardware revision or firmware mismatches (and although I have a Lexia, we can't do firmware updates without a PSA account...) or perhaps it needs to be coded to the main EV ECU or BSI before it will operate.

There just aren't enough of these cars around for there to be many second hand parts for people to have had experience doing this. Since only a few thousand were sold world wide its quite possible that there are only a handful of failed drive inverters on these in the world!

Not a project I'd be willing to take on I think.

[Gibbo2286]

I'm too old now, turned 83 last Saturday, I think the C5 with the missing autobox was going to be my last 'project', ten years younger and I would have had a go at anything.

[Zelandeth]

I've tried repairing a couple of industrial motor variable frequency drive units. Unsuccessfuly.

On both occasions the sequence of events seemed to start with a high power semiconductor breaking down - in itself not a major issue, but the things it took out along with it were. I eventually gave up after replacing what felt like half the semiconductors on the board, testing pretty much all the passives and doing everything by way of "cold" testing, but being greeted by nothing other than a bright flash, loud bang and a blown mains fuse on applying power.

I suspect given the similarities in construction I guess the same may well be true with these inverters.

[CitroJim]

Zel, that's generally my experience...

Ah! but think about it Jim, you pop the cover off and staring you in the face is a blown resistor or a puffed up capacitor, you'd have a go wouldn't you.

The problem there is knowing what caused those components to fail...

In big SMPS/inverter type systems and the like the obvious failures are likely to be the symptom and not the root cause...

In the early days of SMPS being fitted to TVs and some professional gear, a critical part of the service information was what else to replace in addition to the obviously failed components to ensure they did not fail again...

The list was often extensive and could end up being expensive...

Also, some components in such a system need to be replaced with identical/approved components to maintain integrity and electrical safety...

You might successfully fix it using non-approved spares and then in an accident - even if not directly caused by what you did - might affect insurance...

Having repaired many [computer PSUs] and seeing how awful they are I'm actually amazed that some PC power supplies make it out of the factory still working let alone past the warranty period!

Same here Simon and that goes for high-end server PSUs too, not just ordinary PCs... Gosh, I've seen some of those server PSUs go off spectacularly with big flames shooting out of them

The general build and component quality is rubbish...

Good job the servers in question had multiple redundant PSUs!

I've also seen some brilliant fireworks from high power UPSs too - and those are closer perhaps to an automotive inverter than a PSU...

Having said that, if the car was free, or nearly so, there's be not much lost to have a go but I'd exercise great caution given the volts and amps involved and have a good fire extinguisher handy!

[Hell Razor5543]

I remember somebody telling me about a particular printer (this was a long time ago) that had a known issue. If a particular thyristor failed (I think that was what he called it) it would loudly take a resistor with it, while quietly failing a transistor. If you did not know about that transistor, and you replaced the other two components, next time you applied power to it 132 other transistors would fail! Normally you only had this happen ONCE to you.

[CitroJim]

I remember somebody telling me about a particular printer (this was a long time ago) that had a known issue. If a particular thyristor failed (I think that was what he called it) it would loudly take a resistor with it, while quietly failing a transistor. If you did not know about that transistor, and you replaced the other two components, next time you applied power to it 132 other transistors would fail! Normally you only had this happen ONCE to you.

Yep, that's exactly what can (and frequently does) happen James...

The manufacturers love it as it effectively writes off the item and a nice, shiny new one has to be bought...


Good for them and bloody terrible for the environment

[Hell Razor5543]

Those printers were too expensive to just throw them away. Anyway, my friend was chatting to the stores man when somebody came down and asked for this particular thyristor and resistor. My friend was not listening, but when this guy went away he suddenly realised what might have happened. He got the stores man to pick out all of the relevant parts, just in case (including that ruddy starting transistor). Five minutes later the guy was back, and was surprised to find all the bits he required waiting for him!

[Mandrake]

I've tried repairing a couple of industrial motor variable frequency drive units. Unsuccessfuly.

On both occasions the sequence of events seemed to start with a high power semiconductor breaking down - in itself not a major issue, but the things it took out along with it were. I eventually gave up after replacing what felt like half the semiconductors on the board, testing pretty much all the passives and doing everything by way of "cold" testing, but being greeted by nothing other than a bright flash, loud bang and a blown mains fuse on applying power.

A great technique passed down from my Dad - and the way I used to test something like a switch mode PSU that is likely to go KERSPLAT again (and also older TV sets that had similar tendencies!) if you haven't found every single faulty component the first time, is to power it through a suitably rated 240v incandescent bulb in series with the mains phase. Typically between about 25 watts and 100 watts depending on the size of PSU.

If there is no fault the bulb will briefly and dimly light during the capacitor charging inrush and then stay dim or nearly out, keeping a low resistance and allowing the PSU to start up and run normally (albeit only with no load or a light load applied) with very little voltage drop across the bulb.

However if there is any fault remaining of the "kersplat" variety that would blow the fuse and/or further semiconductors the bulb just glows brightly, increases resistance dramatically and thus limits the voltage and current available to cause damage to the PSU - and unless you choose a bulb of too high a wattage 9 times out of 10 times no damage occurs! It gives you time to make a few measurements without too much risk of damage to help further the diagnosis.

A large power resistor doesn't work because if you have a high enough resistance to limit the current below the level that will cause damage in the event of a fault, the voltage drop is too great when there is no fault and usually the PSU will hiccup or not start at all.

But the non linear resistance of the bulb which also has a bit of thermal inertia means that the resistance is very high during a fault, enough to protect the semiconductors but low enough absent a fault to let it run normally. The hot to cold resistance ratio of an incandescent bulb is something like a factor of 10 to 1... You can also start with a lower wattage bulb and work your way up if you're being extra careful. If the bulb wattage is too low it usually prevents the PSU's startup procedure from completing or it may hiccup.

Sorry if you guys already know and use this technique but I thought it was worth mentioning.

I suspect the same technique could be applied to testing a just-repaired and not yet trusted inverter to avoid expensive damage on first power up, although it is DC at the input and would probably require a much higher wattage bulb(s) as even the inverter in the tiny Ion is rated to 55kW !!

[CitroJim]

A great technique passed down from my Dad - and the way I used to test something like a switch mode PSU that is likely to go KERSPLAT again (and also older TV sets that had similar tendencies!) if you haven't found every single faulty component the first time, is to power it through a suitably rated 240v incandescent bulb in series with the mains phase. Typically between about 25 watts and 100 watts depending on the size of PSU.

Oh yes, the standard method of first-time power-up of any old and unknown-provenance vintage wireless stuff...

Known as a 'lamp-limiter'... I really must make one...

For the past 45 years I've just done basic checks with an AVO and visual and if all looks fair I'll power on and see what happens but always with a finger dangling over the off switch and at the first sign of distress, such as an escape of the magic smoke, off it goes...

That's been good for me thus far but no good for modern stuff with switch-mode power supplies... they go bang in picoseconds under fault conditions...

The other method is to bring it up slowly on a Variac - a variable transformer...

Sadly, now they're as rare as rocking horse poo...

[Hell Razor5543]

That same friend of mine managed to play a good April Fools' joke on a new guy in the workshop. This guy had just finished soldering, and popped out for a call of nature. My friend quickly laid some pipework between his bench and the other guys one (and primed it). When the guy returned my friend was smoking a cigarette (shows how long ago it was!). When the guy turned on the unit my friend blew hard down said pipe, resulting in an impressive plume of smoke. The 'victim' turned off the power very quickly, but omitted an important step. Several hours later, despite all of his testing, he could not find anything wrong, and was even more bemused that, when he turned it back on, it worked properly.

To anybody who is seriously into their electronics repair, the important step that was missed is fairly obvious. To those less knowledgeable what he should have done was to take a sniff of the smoke, as different items (when they fail) have obvious scents.

[CitroJim]

James, I love that Reminds me of some of the japes played on me in my early days and then some more I've played on others...

Still one of the best is a blown-up crisp packet you hide and then burst just as the unsuspecting person is turning on a bit of kit for the first time after a repair...

For some the call of nature then comes a little sooner than expected...

These days those large air pillows they use for packing work a treat for this sort of practical joke although you often need to stamp on them...

[white exec]

Are these inverters repairable?

For us, that's a No, then.
Sobering information. Wonder what the cost of a new one is?
I'm guessing that the inverter is the most complicated bit of the EV powertrain, and executes all the motor (acceleration, regen braking, etc) control functions . . . or are there other similarly complex units?