22KW OBC FAILURE INSIGHT & POSSIBLE SOLUTION

[ze_shark]

It

Thanks. It is odd, however, that this is so random. I've AC charged 3x per week for three years with no problem. I wonder if it was a bad batch of MOSFETs.

It is a typical trap to assume that there is only one failure mode and root cause.
These dead MOSFETs, whatever makes them fail, could be the dominant root cause, or just one out of several.

Sponsored

 

[ct14garage]

It
It is a typical trap to assume that there is only one failure mode and root cause.
These dead MOSFETs, whatever makes them fail, could be the dominant root cause, or just one out of several.

But his car continues charging at lower current just not above 11KW it means the OBC HASN'T FAILED. Because it's still charging.

It could be an isolation problem from somewhere else triggering a shutdown emergency of HV system because the isolation deteriorates a bit as things get hotter with the higher charging current.

If the OBC charges at 9KW but not at 11KW. I'm willing to bet the OBC is perfectly fine.

Problem from charger, or from somewhere else in the car.

 

[F1Ruaraidh]

I can totally relate to this. When I use an AC charger with more than 11kW, it triggers a red error and fails to charge in less than 30 minutes. However, if I reduce the power to 9kW, the charging completes successfully up to 100%.

This issue has been bothering me for a while, especially since Tesla allows users to adjust the AC charging power directly from the screen, while the Taycan always draws the maximum available AC power.

That sounds thermal.

 

[F1Ruaraidh]

No scientist has PROOF that climate change is real and happening right now. If anything they say it’s gonna happen in 50 years which they’ve been saying since the 1930s.

I have a customer born in Bangkok in the 1950s, he used to hear when he was a kid that Bangkok would be underwater by the 1980s… it’s 2025 we’re fine.

Now the say Bangkok will be underwater by 2050… yeah right ???

Whatever. ???

 

[ct14garage]

Sunday..... repairing OBCs.... got a whole stack of broken ones to repair (62 pcs) I believe.....

Really tiring.......

 

[DannyD]

Sunday..... repairing OBCs.... got a whole stack of broken ones to repair (62 pcs) I believe.....

Really tiring.......

Where are you getting these from? Just Thailand or all over the world?

 

[ct14garage]

Where are you getting these from? Just Thailand or all over the world?

Most failed parts I get them from Porsche dealers in China. If I reckon I can repair it, I'll get it, repair it and then sell it working. It's a good business.....


Porsche is going down extremely hard in China. In the last 10 years they invested massively in the Chinese market (which until very recently, last year) was their biggest market!!! Far larger than Germany or United States. This is why Porsche developed many things exclusive for the Chinese market which are not available anywhere else, not even in Germany!!!

However since very recently they can't compete anymore with Chinese cars on price/value alone and are going down very harshly.

Dealer owners running away, without paying salaries etc....

As a matter of fact the PIWIS 4 tablet I use now, I bought if off a Porsche dealer employee in Jiangsu, China. Boss disappeared, cars disappeared, no salaries paid to employees, but most of the equipment remained behind so employees sold whatever they coould get their hands on.

He sold the entire PIWIS 4 TESTER to me for $2,000.... when it's worth well over $20K. Best purchase of my life LOL.

 

[ct14garage]

In spite of the expert opinion of haters such as @ze_shark, I can now publish some further findings which together with the fact that none of the repaired OBCs have blown up.... confirms that

The MOSFETS are the CAUSE and not the SYMPTOM. They are underrated for their task!

The MOSFETS that fail belong to the AC/DC part of the circuitry, also called BOOST circuit.

Charging at 22KW.... the current on the DC Bus will more often than not reach 36.7A (as directly measured).


Initial inrush peaks hard (~75 A) and drops off within ~50 ms.
Ripple persists on top of a high nominal load current (~36.7 A), keeping the total current well above the 40N120AG’s 30 A max rating.
Even without inrush, steady-state ripple often exceeds 30 A, risking long-term overcurrent damage.

Green is the nominal current... Red is the measured current.. Blue is the MOSFET max rating


WHAT DOES THIS MEAN?

1. Lower current charging will very much extend the life of the OBC. At 11kw it's much less likely to fail. Even more so at 3.6kw. (As evidenced by Porsche's patch which is to reduce single phase current)

2. Using high quality proper chargers will help. The cheaper the charger the more ripple. Ripple makes this existing problem way worse (pushing the mosfets further towards self destruction).

3. It's not a thermal problem at all... the PCB which holds the MOSFETs is mounted onto a heatsink the size of the motherboard.

This is by design, they've tried to compensate operating the MOSFETs beyond their spec with more than enough cooling. But they still fail of course!!!

At the time when this OBC was designed, there was no MOSFET with a higher amperage rating that would do the job. This was the only choice.

 

[snoopy]

So changing MOSFET eliminates this issue? Do you recommend we change them even on boards that are working ok not to have issues in the future?

 

[ct14garage]

if it aint broke dont fix it

 

[snoopy]

Question is if the failed mosfet would damage anything else that could maybe be avoided? In your usual fixes this was the only component that needed to be replaced?

 

[ct14garage]

Yeah

 

[arthur]

PS: It’s not so much that they don’t care about it, but all the Electrical system of the Taycan wasn’t designed nor built by them. They relied on LG, MetaSystems, HITACHI and plenty others. Those are the ones really responsible so Porsche doesnt bother with it. It’s not their job to fix their mess.

When the OBC fails and Porsche repairs it for you for free, Porsche then claims this cost back from Metasystems. Every single time.

Makes sense. But in a normal world Porsche should care about the fact that they are making their customers unhappy with an inconvenient issue that makes the car doesn't move and need to be towed. So beyond claiming the cost from Metsystems, the minimum is to make Metasystem accountable to investigate the cause and fix future OBCs sent to Porsche.
And if Porsche is so bad, I at least expect Metsystems investigating this issue after being hammered for months/years with Porsche after service claims (this is a huge cost for Metsystems).

Anyway, thanks @ct14garage for your work, your insights actually give me some hope that by now Metsystems figured this out as well and hopefully fixed it.
I had this issue once and it was a nightmare. If this happens to me a 2nd time I am really going to be pissed off.

 

[Fish Fingers]

Anyway, thanks @ct14garage for your work, your insights actually give me some hope that by now Metsystems figured this out as well and hopefully fixed it.
I had this issue once and it was a nightmare. If this happens to me a 2nd time I am really going to be pissed off.

I agree. @ct14garage from what i see, is single handedly providing the only real world insight into some of the workings of the Taycan in the whole world (given how tight lipped the dealer network obviously is).

I take my hat off to him, for all the pioneering work he is doing.
I find many of his posts on here remarkable.

Is it because there is no dealer network in Thailand, so it has grown into a DIY industry?

I would go as far as to say Porsche could actually pay him for his acquired knowledge, for future product improvements.
But somehow, they don't seem to be that bothered about reputational damage from all these issues.

I also see it now being reported that they are actually thinking of quitting the EV market in China altogether.

 

[snoopy]

Anyone got hands on this report? https://www.ltec-biz.com/en/report/obcmetasystem-porsche-taycan-22kw-teardown-report/

Sponsored