Probably killed SDU inverter!

Topics concerning the Tesla front and rear drive unit drop-in board
Post Reply
Isaac96
Posts: 656
Joined: Sat Oct 05, 2019 6:50 pm
Location: Northern California, USA
Been thanked: 2 times
Contact:

Probably killed SDU inverter!

Post by Isaac96 »

I think I blew up my inverter today!

After testing some Chademo quick charging I went for a drive to discharge the battery some more.
As usual, full throttle made an appearance, and after that an overcurrent (which happens often, need to tune more).

But this time the inverter did not start back up! (autoresume tripmode)
I pulled over and tried the usual; power cycle and start -- but no response!
Connecting to wifi showed a continual stream of "OVERCURRENT" errors, one every few milliseconds.
il1 and il2 were low (0 and 0.65), and all inputs showed normally.
So I think I somehow defeated Tesla's protection hardware and destroyed multiple IGBTs -- this behaviour is consistent with what happened without the HVIL jumper shorted when bench testing, so it's probably the Tesla fault output.

I can't disassemble the car until I get it back here (only a couple miles, but will need to tow) so I will verify that the inverter is toast at that point.

Are there any other possibilities besides a toasted inverter? I suppose it could be the replacement logic board but that's unlikely...
Looks like it's time to go shopping again. It'll probably be tough to find just the inverter. Ah well.

-Isaac
User avatar
Boxster EV
Posts: 481
Joined: Fri Jul 26, 2019 9:32 pm
Location: UK
Has thanked: 56 times
Been thanked: 48 times

Re: Probably killed SDU inverter!

Post by Boxster EV »

😞

I’m interested in your findings. What we both have in common is that it seems both our batteries were fully charged during the fatal OC event. In the early days of my conversion, I must have hit OC 50 times(I live on a steep hill and struggled to find a working tune), however the battery was never fully charged.

Damien very kindly sent me a new power stage for my inverter and I’m now able to correctly see a voltage diode drop across all transistor banks, however I’m still seeing an OC fault.

IC10 reports a clean 5v across each pin.

In my case I’ve tried my logic board in a replacement drive unit and it’s fine.

I do however plan to try and continue repairing my old DU, I’m just not sure where to start diagnosing without just swapping out parts.
Porsche 986 powered by a Tesla large drive unit. Backwards. Build documented here and Instagram @tesla_porsche here.
arber333
Posts: 3557
Joined: Mon Dec 24, 2018 1:37 pm
Location: Slovenia
Has thanked: 133 times
Been thanked: 331 times
Contact:

Re: Probably killed SDU inverter!

Post by arber333 »

Isaac96 wrote: Sun Apr 18, 2021 2:14 am I think I blew up my inverter today!
......
Are there any other possibilities besides a toasted inverter? I suppose it could be the replacement logic board but that's unlikely...
Looks like it's time to go shopping again. It'll probably be tough to find just the inverter. Ah well.

-Isaac
That is a damn shame!
Can you try using longer deadtime + less/more boost maybe? Would that make a difference?
Also try loading firmware and bootloader again.
Hm... maybe you only have one of the transistors or drivers tripping desat signal. This would than propagate through the whole system.
You should take inverter off and test each phase separately with a light bulb load and some 48V. And then increase voltage in steps of 24V.
I remember i saw strange spikes from PWM signal pins when i restarted the brain on desk. That was with signals configured for active LOW.

Also I had a strange problem happening to me. Above certain phase current inverter would allways trip desat and before i used Volt inverter i would burn IGBT. But on my second car that same inverter would work fine. When i exhausted all inverter options i went and took the motor apart. What i saw shocked me. Rotor and part of stator were showing heavy water damage and i guess lacquered wire would short only above 150A. So try to inspect the motor too.

While you are searching for another SDU drive there is one thing you can do even if your inverter is toast. Connect your Volt inverter and test how much output you get from the motor that way. You may at least have a moving car instead of a realestate 8-).
User avatar
Jack Bauer
Posts: 3640
Joined: Wed Dec 12, 2018 5:24 pm
Location: Ireland
Has thanked: 9 times
Been thanked: 282 times
Contact:

Re: Probably killed SDU inverter!

Post by Jack Bauer »

Did the contactors drop out when you had the overcurrent?
I'm going to need a hacksaw
arber333
Posts: 3557
Joined: Mon Dec 24, 2018 1:37 pm
Location: Slovenia
Has thanked: 133 times
Been thanked: 331 times
Contact:

Re: Probably killed SDU inverter!

Post by arber333 »

Jack Bauer wrote: Sun Apr 18, 2021 9:53 am Did the contactors drop out when you had the overcurrent?
He said he had it set to autoresume. Contactor will allways be on then wont it?
jon volk
Posts: 574
Joined: Wed Apr 10, 2019 7:47 pm
Location: Connecticut
Been thanked: 2 times

Re: Probably killed SDU inverter!

Post by jon volk »

Someone one here once mentioned that while there’s hardware desat fault shutdown, the ability to deal with such events is not unlimited. Perhaps that event was the final roll of the dice.
Testing from low voltage and progressively increasing isn’t a bad idea to gain data.

Another variable we don’t have control over is that many of these drives come from salvage vehicles that may have had a high voltage disconnect under load due to an accident. In such cases we’re already in a compromised position before inducing further stresses.

For a future project I have some interest in attempting to build a replacement Tesla inverter using Infineon IGBT modules. It’ll likely be poorly executed but perhaps yield a functional high power option. Still just an idea at this point.
Formerly 92 E30 BMW Cabrio with Tesla power
User avatar
Jack Bauer
Posts: 3640
Joined: Wed Dec 12, 2018 5:24 pm
Location: Ireland
Has thanked: 9 times
Been thanked: 282 times
Contact:

Re: Probably killed SDU inverter!

Post by Jack Bauer »

Not wanting to hijack the thread but my 2 cents on a replacement inverter. I have been picking up chevy volt / ampera and Gen4 prius inverters whenever they show up for cheap. Have 3 on the shelf now. If/when I blow the inverter in the E31 it will get replaced by one of those with its main power stages ganged. These days I wouldn't even try building an inverter but that's me. The Toyota power electronics are in a class of their own.
I'm going to need a hacksaw
User avatar
johu
Site Admin
Posts: 6628
Joined: Thu Nov 08, 2018 10:52 pm
Location: Kassel/Germany
Has thanked: 346 times
Been thanked: 1489 times
Contact:

Re: Probably killed SDU inverter!

Post by johu »

arber333 wrote: Sun Apr 18, 2021 10:13 am He said he had it set to autoresume. Contactor will allways be on then wont it?
Yes indeed.
Support R/D and forum on Patreon: https://patreon.com/openinverter - Subscribe on odysee: https://odysee.com/@openinverter:9
Isaac96
Posts: 656
Joined: Sat Oct 05, 2019 6:50 pm
Location: Northern California, USA
Been thanked: 2 times
Contact:

Re: Probably killed SDU inverter!

Post by Isaac96 »

arber333 wrote: Sun Apr 18, 2021 9:48 am That is a damn shame!
Can you try using longer deadtime + less/more boost maybe? Would that make a difference?
Also try loading firmware and bootloader again.
Hm... maybe you only have one of the transistors or drivers tripping desat signal. This would than propagate through the whole system.
You should take inverter off and test each phase separately with a light bulb load and some 48V. And then increase voltage in steps of 24V.
I remember i saw strange spikes from PWM signal pins when i restarted the brain on desk. That was with signals configured for active LOW.

Also I had a strange problem happening to me. Above certain phase current inverter would allways trip desat and before i used Volt inverter i would burn IGBT. But on my second car that same inverter would work fine. When i exhausted all inverter options i went and took the motor apart. What i saw shocked me. Rotor and part of stator were showing heavy water damage and i guess lacquered wire would short only above 150A. So try to inspect the motor too.

While you are searching for another SDU drive there is one thing you can do even if your inverter is toast. Connect your Volt inverter and test how much output you get from the motor that way. You may at least have a moving car instead of a realestate 8-).

Hmm... I could just try not having overcurrents all the time :D
I think I had fslipmax a bit too high; fweak was raised a good bit to match the higher boost, which may have been making things unhappy. I optimized for low speed smoothness first though. Deadtime... not sure, haven't heard of any other SDUs blowing.
I will try playing with the Volt inverter. That thing was a total pain in the arse but at least it works.
Jack Bauer wrote: Sun Apr 18, 2021 9:53 am Did the contactors drop out when you had the overcurrent?
Nope, autoresume as Johannes stated.

Also I am quite interested in the Gen4 Prius idea; it is less space efficient and the single unit is very nice but bulletproof electronics seem to be necessary. Will check that out.

Anyways, it's off to the great eBay to try to find a new one. Are there any known sources, say from a unit with a dead motor/gearbox?

-Isaac
arber333
Posts: 3557
Joined: Mon Dec 24, 2018 1:37 pm
Location: Slovenia
Has thanked: 133 times
Been thanked: 331 times
Contact:

Re: Probably killed SDU inverter!

Post by arber333 »

Jack Bauer wrote: Sun Apr 18, 2021 12:52 pm Not wanting to hijack the thread but my 2 cents on a replacement inverter. I have been picking up chevy volt / ampera and Gen4 prius inverters whenever they show up for cheap. Have 3 on the shelf now. If/when I blow the inverter in the E31 it will get replaced by one of those with its main power stages ganged. These days I wouldn't even try building an inverter but that's me. The Toyota power electronics are in a class of their own.
Ah! That is why all my sources have dried up. :D.
Do you have any sources worth mentioning?
I would need at least one more Volt unit to have a spare, because i am about to test my parallel power units concept.
Isaac96
Posts: 656
Joined: Sat Oct 05, 2019 6:50 pm
Location: Northern California, USA
Been thanked: 2 times
Contact:

Re: Probably killed SDU inverter!

Post by Isaac96 »

Well it has been one month now since the inverter died and in celebration I have finally taken it out of the car.
The Johannes board looks fine.
Underneath it I don't see any smoke yet, but I have no idea how to get the plastic cover off the main driver board so that I can check the IGBTs for obvious damage. Do I just start breaking it? The conformal goo stuff is quite sticky.
User avatar
cloudy
Posts: 208
Joined: Wed Oct 02, 2019 12:15 am
Location: UK
Been thanked: 6 times

Re: Probably killed SDU inverter!

Post by cloudy »

Out of interest do you know if you were using 8.8khz or 17.6khz switching at the time?
Isaac96
Posts: 656
Joined: Sat Oct 05, 2019 6:50 pm
Location: Northern California, USA
Been thanked: 2 times
Contact:

Re: Probably killed SDU inverter!

Post by Isaac96 »

8.8 methinks as that is default.
User avatar
cloudy
Posts: 208
Joined: Wed Oct 02, 2019 12:15 am
Location: UK
Been thanked: 6 times

Re: Probably killed SDU inverter!

Post by cloudy »

Would be interesting to check next time you power up the johannes board... I guess you've read this since (and I know it's no use now!) But I think Damien found the root cause of the SDU issues as follows:

"Root cause of the drive unit inverter failures discovered and fixed. NOT caused by overcurrent through the igbts but rather the combination of using 8.8kHz switching frequency through a very low inductance stator and the sudden release of the main contactor on an overcurrent detection. The small drive unit inverter works perfectly at 17.6khz switching frequency and setting tripmode to precharge ensures the pre charge resistor can provide a path back to the battery during release of the main contactor. This was tested to the Nth degree on my recent trip to Romania"
Isaac96
Posts: 656
Joined: Sat Oct 05, 2019 6:50 pm
Location: Northern California, USA
Been thanked: 2 times
Contact:

Re: Probably killed SDU inverter!

Post by Isaac96 »

Interesting! I remember reading that a while ago but certainly forgot all about it. I shall check the settings and see what it's set to, I do remember the inverter ran super cool and the motor was a bit hot so raising PWM frequency would probably even the temperatures out. The default parameter set is 17.6 but not sure what I ended up with.
Isaac96
Posts: 656
Joined: Sat Oct 05, 2019 6:50 pm
Location: Northern California, USA
Been thanked: 2 times
Contact:

Re: Probably killed SDU inverter!

Post by Isaac96 »

My inverter is absolutely dead. It measures .354v between all phases and +/- except for one low side which is 0v. So that's a dead IGBT. I'm looking for a replacement now, should have the car running as soon as I get my hand on one. Anyone know where I can get just the inverter?
User avatar
EV_Builder
Posts: 1205
Joined: Tue Apr 28, 2020 3:50 pm
Location: The Netherlands
Has thanked: 18 times
Been thanked: 37 times
Contact:

Re: Probably killed SDU inverter!

Post by EV_Builder »

Isaac96 wrote: Tue Jun 01, 2021 8:56 pm My inverter is absolutely dead. It measures .354v between all phases and +/- except for one low side which is 0v. So that's a dead IGBT. I'm looking for a replacement now, should have the car running as soon as I get my hand on one. Anyone know where I can get just the inverter?
first try Ebay have them seen there. The complete part.
There is also a tesla breaker in the USA which sells you anything. if memory serves me right he is called 'autobahn'.

EDIT: Dug it up for you: http://www.autobahnparts.com/carmodels/tesla

No clue about the pricing...
Converting an Porsche Panamera
see http://www.wdrautomatisering.nl for bespoke BMS modules.
Post Reply