Roadster PEM failure

[Rolf68]

Hello

What are exactly those insulation strip you're ordering? It's the grey part looking like modeling paste on the last picture? They cut that with laser :^O

 

[Jonathanm]

They look like this and go between the heat sink (the aluminium plate) and the IPGT's....
The OEM insulation appears to be a silicon looking type of material. This is what is failing if you look at my pictures...

 

[Scene]

But if I had my PEM apart anyway to replace a few IGBTs, I would replace all of them with the newer low saturation voltage versions. I confess I didn't think there would be any issues with the driver.

Any suggestions for alternative IGBTs (to save me doing the research)?

I didn’t even get a chance to use my new Can before the PEM went!

 

[DanielFriederich]

I thought about doing this pro-actively to get less heat overall and more efficient operation - including more range. But when I ran the numbers it didn't seem like the benefit was worth all the work taking my PEM apart and soldering. That is especially true given that my PEM is working fine.

But if I had my PEM apart anyway to replace a few IGBTs, I would replace all of them with the newer low saturation voltage versions. I confess I didn't think there would be any issues with the driver.

My guess is that even with a lower saturation voltage and with better heat transfer (ceramic stripes) it won't affect efficiency so much. It maybe compensate once or twice hitting the pedal to metal...
The effect might be more on the overall lifetime of the IGBTs (the hotter electronic parts get the faster they age) but I'm gonna ask the wise guy (electronics engineer) about his opinion and will post it here

and another thing: soldering the IGBTs is a pain in the a... and when soldering (new ones) more or less solder could lead to different (slightly!) resistances and maybe inductive loads at the pins of the IGBTs. My electrical engineer advised me if doing soldering to use only as much soldering lead as needed to make a solid connection between the pins and the circuit board...

also keep in mind that the 1.5 PEMs use a different type of IGBT with a slightly higher max amperage!

 

[DanielFriederich]

But if I had my PEM apart anyway to replace a few IGBTs, I would replace all of them with the newer low saturation voltage versions. I confess I didn't think there would be any issues with the driver.

Any suggestions for alternative IGBTs (to save me doing the research)?

I didn’t even get a chance to use my new Can before the PEM went!

inside PEM 2.0 and probably PEM 2.5 the Infineon IKW75N60T is used
inside PEM 1.5 its the IXYS IXGK120N60B (not 100% sure, please correct if I'm wrong!)

Besides Infineon and IXYS there's Fairchild and some other companies producing IGBTs with (almost) same specs (max amperage 75 or 80 amps e.g.)
All those companies advertise with their special type of technic making it the best of all in comparison to other technics... well... thanks marketing :/

There are sites on the web where you can provide the needed specs and get the types and producers of IGBTs suitable.
e.g. here (I added some of the needed values for convenience):
600 V TO-247-3 1.5 V Through Hole +/- 20 V IGBT-Transistoren | Mouser Deutschland

Keep in mind that our engine is not running at 50 or US 60 Hz but likely at a higher frequency for smoother operation (I know this from garden trains being digitally run, the motors get a frequency up to some kHz so the trains run smoother!)
I once read in the forum that the driver boards in the PEM 1.5 might run up to 32 kHz! and the 2.x might be in the same range. But then the IKW75N60T wouldn't make sense anymore if you take a look on the operation curve corresponding to driver frequency.

Here https://www.infineon.com/dgdl/Infin...n.pdf?fileId=db3a304412b407950112b42890113e25 the Infineon data sheet of the IKW75N60T for your convenience

 

[DanielFriederich]

concerning lower saturation voltage for the IGBTs, here the reply of the electronics engineer guy:

Lower saturation voltage also means other transistor parameters according to Barkhausen (S x D x Ri = 1).
One of these is the steepness S (roughly comparable to the current gain B in bipolar transistors)
If you take other transistors, then of course you MUST look into the data sheets, (before / after) and compare other values as well.
Maybe then the driver stages must be changed too if the control/driving for the IGBTs is not well defined, otherwise there's possibly a parasitic oscillation, (self-excitation) at the end 100% deadly for the circuit.

In other words... we will have to stay with the 1.5V saturation voltage and only the heat dissipation is a parameter which when lower would be a advantage to the megapole boards. IKW75N60T has 428W max of heat dissipation!

 

[Scene]

Thanks Daniel,

So it seems best to stick with the IKW75B60T given the unknown effect of changing any patameters - even Vsat with regards to the drivers etc.

Luckily I'm good at soldering and am also a director of an excellent contract electronics manufacturing facility here in the UK!
www.printsys.co.uk

My only reservation about all of this is not being able to easily diagnose if anything else is wrong with the PEM.

 

[jeremyz]

If somebody is feeling adventurous, it might be possible to switch to Cree silicon carbide mosfets. They claim to have much lower switching losses than IGBT's. This would likely be a much more involved project though.

Wolfspeed Introduces New SiC MOSFET for EV Drive Trains

 

[Jonathanm]

Thanks Daniel,

So it seems best to stick with the IKW75B60T given the unknown effect of changing any patameters - even Vsat with regards to the drivers etc.

Luckily I'm good at soldering and am also a director of an excellent contract electronics manufacturing facility here in the UK!
www.printsys.co.uk

My only reservation about all of this is not being able to easily diagnose if anything else is wrong with the PEM.

Well if you are not able to diagnose any problems with the PEM, given what I see you do at your company - then nobody can!

 

[hcsharp]

If somebody is feeling adventurous, it might be possible to switch to Cree silicon carbide mosfets. They claim to have much lower switching losses than IGBT's. This would likely be a much more involved project though.

Wolfspeed Introduces New SiC MOSFET for EV Drive Trains

Those look promising but they have a higher gate voltage threshold. It would be tempting to try them if you were designing from scratch but I think you'd run into driver compatability issues in the PEM.

 

[dpeilow]

I’m seriously considering repairing the PEM myself because Tesla (UK) can’t tell me if their expensive PEM refurbs have the insulation issues addressed

It would be great if you have the facilities and do this, because sooner or later I'm sure several UK owners will be interested in the same job.

 

[Scene]

It would be great if you have the facilities and do this, because sooner or later I'm sure several UK owners will be interested in the same job.

Well actually my car is in at Tesla right now for inspection. They reckon there might be other things wrong judging by the error messages.

I would only attempt the PEM repair if I knew it was just the transistors.

Will keep you posted.

 

[thefortunes]

Unfortunately, my turn.

Car is undrivable after hearing a "pop" and having the dreaded PhaseB Low Side Desat fault (among others).

I'm assuming the PEM is dead - it will go to Tesla this afternoon for diagnosis.

Weighing options:
1) Engage Gruber for a repair/upgrade
2) Remanufactured PEM from Tesla
3) 3.0 battery - trying to work this angle. Since the upgrade comes with a modified PEM, maybe we can sneak this in the $29k cost?
4) Others?

Thoughts?

 

[Jonathanm]

Sorry to hear about your PEM failure. My guess it is the insulation gone on yours as well with an IGBT shorting to the heat sink on 1 of the output boards.
Can you smell anything burnt around your PEM?

If you feel like attacking this yourself it is not that complicated.
I go for point 4)
If you can wait until the weekend 1st December I will be doing the full repair that weekend. The new insulation (aluminium/ceramic strips) are about to be posted out to me now. They cost me €250. The PEM is already opened. Cleaning the old insulation is dead easy - it all just falls off. I need to buy myself a little bit of Loctite for the IPGT nuts on the clamps + I have purchased a torque limiting screwdriver.
The IPGT's need to be tightened to 0.8 N/Metres.

Sealey Torque Screwdriver Digital Readout 0.05 - 5Nm 1/4" Hex Drive STS103 | eBay

Before mounting I will test the IPGT's only for short circuit failure. This can be done with a Multimetre with a "diode" symbol. Anymore detailed testing needs them to be removed from the board. (not that keen on doing that) I see nothing burnt at all in my boards so if none are shorted I will go with it as they are.

I am pretty confident that this will fix my problem.

In about 2.5 weeks I will know a little more....

Hopefully this will be an inexpensive solution... I will photograph the reassembly as well and post later on...will need a bit of luck as well...

 

[ion_1]

1) Engage Gruber for a repair/upgrade
2) Remanufactured PEM from Tesla
3) 3.0 battery - trying to work this angle. Since the upgrade comes with a modified PEM, maybe we can sneak this in the $29k cost?
4) Others?

Sorry to hear about your possible PEM issue.

Is this still your original PEM?

One major question I have re remanufactured Tesla PEMs is what is a "remanufactured" PEM?
It comes with a short warranty, but I worry the "remanufactured" PEM may have not have insulators that are in great condition.

Does Tesla inspect/ replace as necessary the insulators as part of the remanufacturing process or could you end up with
someone else's remanfactured PEM (that possibly failed for a completely different reason) that has insulators 50% degraded on their way to failure?

It would be great if someone could extract this answer from Tesla of what consists of the "remanufacturing" process,
It was one thing when you received a remanufactured PEM under warranty, another when you are dropping $10k.

If you need a battery, I would go with the 3.0 if you could get the PEM thrown in. Besides spending the $30k, this is an easy call.

A more permanent solution may be the Jonathanm route. If it works, that would be fantastic. If it doesn't, as long as you can still use the PEM as a core exchange for a remanufactured PEM, you lose very little except time (but you gain knowledge and the rebuild experience).

Companies such as Gruber may be the go to as the years go by. In fact, they may even be the go to by Tesla one day to keep us owners happy who knows.

Let us know what you decide and outcome.

 

[thefortunes]

Sorry to hear about your possible PEM issue.

Is this still your original PEM? It is the original PEM. I have ~78k miles on the odometer

One major question I have re remanufactured Tesla PEMs is what is a "remanufactured" PEM?
It comes with a short warranty, but I worry the "remanufactured" PEM may have not have insulators that are in great condition.

Does Tesla inspect/ replace as necessary the insulators as part of the remanufacturing process or could you end up with
someone else's remanfactured PEM (that possibly failed for a completely different reason) that has insulators 50% degraded on their way to failure? This is probably my biggest hesitation to go with a Tesla PEM

It would be great if someone could extract this answer from Tesla of what consists of the "remanufacturing" process,
It was one thing when you received a remanufactured PEM under warranty, another when you are dropping $10k.

If you need a battery, I would go with the 3.0 if you could get the PEM thrown in. Besides spending the $30k, this is an easy call. I don't need a battery at this time, but since I plan to keep the Roadster as my daily driver (~12k miles per year) I certainly will at some point. Currently I was thinking in about 5 years. This could change tomorrow if I had a sheet or two go bad and then I would have lost the opportunity to get the PEM thrown in. To be clear, I have not yet investigated this with the SC so it may not truly be an option.

A more permanent solution may be the Jonathanm route. If it works, that would be fantastic. If it doesn't, as long as you can still use the PEM as a core exchange for a remanufactured PEM, you lose very little except time (but you gain knowledge and the rebuild experience). Yep, I'm a little intrigued by this. I have little/no electrical engineering experience so I am a little hesitant (although I did put together and solder the OpenEVSE kit and have done the PEM cleaning ). Not sure I can wait 3-4 weeks though.

Companies such as Gruber may be the go to as the years go by. In fact, they may even be the go to by Tesla one day to keep us owners happy who knows.

Let us know what you decide and outcome.

 

[Scene]

My car is still with Tesla for investigation. If they confirm it’s just the PEM then I have a choice:
1. Remanufactured PEM
2. Self or 3rd-party repair

Currently I’m struggling to get any sense from Tesla after recovering the car to them several weeks ago.

I don’t mind dropping the cash for a fully decent warranted repair from Tesla; BUT only if they can tell me that they have corrected the original design fault and used improved insulation that will last. However, they cannot seem to answer this simple and perfectly reasonable question. Especially annoying in my case after having already paid 10k GBP for a full extended warranty that has expired just before the latest failure...

Personally I don’t want to experience a failed PEM for the third time. And let’s be clear about this - it’s ****ing dangerous, potentially lethal, to have power instantly cut without prior warning. Their firmware does not even kill regen when this occurs. I’m sure Tesla do not want this publicised....

But, if I fix the transistors and insulation and something else is wrong I’m back to square one. Will Tesla accept back a PEM that has been worked on by anyone else?

Perhaps I’ll enquire about the 3.0 battery/PEM. Had not considered that before.

Looking forward to see how Jonathan gets on.

 

[ion_1]

1) Engage Gruber for a repair/upgrade
2) Remanufactured PEM from Tesla
3) 3.0 battery - trying to work this angle. Since the upgrade comes with a modified PEM, maybe we can sneak this in the $29k cost?
4) Others?

I know this can be a controversial subject around here for other reasons, but companies like Gruber may become an interesting option.
It seems Gruber , when performing a PEM rebuild, will tell you what they have replaced and upgraded (electrolytics/insulation) etc.
They offer a couple of rebuild - upgrade options.

I'm just not sure if anyone has had their PEM repaired through them yet. If anyone did, it would be very valuable to hear a review.

 

[Scene]

I know this can be a controversial subject around here for other reasons, but companies like Gruber may become an interesting option.
It seems Gruber , when performing a PEM rebuild, will tell you what they have replaced and upgraded (electrolytics/insulation) etc.
They offer a couple of rebuild - upgrade options.

I'm just not sure if anyone has had their PEM repaired through them yet. If anyone did, it would be very valuable to hear a review.

I've had a very good discussion with Peter Gruber so far. He has provided some excellent assistance. Gruber have some associates in Europe they are working with to provide PEM repairs over here.

I've not gone any further with Gruber just yet due to the current dialogue with Tesla.

 

[hcsharp]

I'm just not sure if anyone has had their PEM repaired through them yet. If anyone did, it would be very valuable to hear a review.

I've spoken with two people who had their PEMs repaired or replaced by Gruber. Both of them were happy.