Reason for Taycan heater failures?

[whitex]

I found this by accident browsing for other things, for those interested in what's inside the Taycan heater. Apparently in this case what failed was some digital programmable part lost its content and just needed to be reflashed. I wonder if it's flash memory that could not handle the heat (literally) - NAND flash is definitely susceptible to corruption due high temperatures, especially if left hot while unpowered (should be well known fact by engineers designing automotive embedded systems that use NAND flash storage). Could it be that the whole Taycan heater recall saga was caused by some fresh engineers not knowing this (not reading specs, or using a part which did not spec this)? ?

Sponsored

 

[alexsas]

I think Porsche just believed Webasto who are meant to be "experts" in these heating systems. Who would have thought that a heater may get hot - scandalous!

 

[chun]

I think they just worked with what they knew: heaters from other Porsche models, from a trusted supplier. And they are simply not adequate for the amount / density of heating cycles the taycan does for the battery. Doubt they actually did their own testing of the module, and just truested the supplier.

They did the same with the charging cables, using the ones from their hybrid cars.

They also used battery pouches without much reserch into failure rate, trusting the supplier yet again, choosing to use LG E66A ( NCM 712 chemistry); which have been proven to have high failure rates in bolts and hyundais. They chose to avoid tarifs on chinese quality batteries, and used LG ones made in EU, with proven high risk failure rate. For J1.2 / facelift, they droped these batteries in favour of newer chemistry LG batteries, designed in china however, and produced still in poland i belive - so we'll see if these prove to have better failure rate.

It's all a risk / reward balance. And they took a lot of risks with j1, even though the car is 200.000+|... they could have simply just use purpose specific parts that are known for their quality

 

[McgR]

For what is worth: one of the sales guys in Brussels said it was because of the chip shortage they used inferior electrical parts to be able to sell the cars anyway.

 

[chun]

For what is worth: one of the sales guys in Brussels said it was because of the chip shortage they used inferior electrical parts to be able to sell the cars anyway.

I have also heard reports here in Switzerland of them switching the heaters / repairing / upgrading them without informing the customer / mentioning it on the maintenance paper. I'd say rumors, but you never know, Switzerland doesn't really have customer protection laws.

 

[KenR]

A service advisor at my local dealership told me that Porsche fixed the problem by adding a fan to the heater to keep it cool. So a fan to keep the heater cool ?

 

[chun]

A service advisor at my local dealership told me that Porsche fixed the problem by adding a fan to the heater to keep it cool. So a fan to keep the heater cool ?

He for sure was pulling your leg. There is no space for a fan, it's a closed enclosure of metal, as can be seen in the video. And a fan outside of it will do 0 to cool it inside.

I don't think porsche themselves have actually repaired any of the units that broke down. They just replaced them. So broken / cracked / erased chips due to heat sounds plausible. Only way to know is to look at BOM difference between new and old model of heater.

 

[daveo4EV]

I found this by accident browsing for other things, for those interested in what's inside the Taycan heater. Apparently in this case what failed was some digital programmable part lost its content and just needed to be reflashed. I wonder if it's flash memory that could not handle the heat (literally) - NAND flash is definitely susceptible to corruption due high temperatures, especially if left hot while unpowered (should be well known fact by engineers designing automotive embedded systems that use NAND flash storage). Could it be that the whole Taycan heater recall saga was caused by some fresh engineers not knowing this (not reading specs, or using a part which did not spec this)? ?

this is a great find! thanks for sharing.

 

[whitex]

So my speculative theory here is that there is NAND flash onboard for some microcontroller, which the designers didn't account for corner cases. One such corner case is content decay which is highly accelerated if the NAND device is left at high temperature unpowered (there may be other factors, such as how quickly it cools, etc). This can speed up the decay from 10 years to as little as 3 months on some parts I have worked on in the past (it might be even more extreme in this case as this is a heater which likely heats up quite high, the scenarios I have analyzed before was parking a car in the hot desert for few months).

The ironic part is that this can be mitigated by refreshing the NAND content periodically, so a software fix. Had Taycan supported OTA updates for the heater, they could have done step one mitigation by simply deploying the same update periodically (to refresh the part), followed by step 2 which would automatically refresh the content of NAND on its own (without OTA). Those refreshes of course count towards expected program-erase cycles, so they would increase wear on the NAND part, but maybe doing this once a month would not be an issue at all. They could even do smart refresh, vary the refresh frequency based on heater usage pattern too. Other mitigations could have been cooling the heater off before shutting off its power, but that might affect car's range (and require an OTA for quick deployment).

These are all my speculations, but this is the first piece of data I came across which allows me to even speculate what might have happened - Porsche never did release any details AFAIK. Given how simple the heater is inside, there are not that many things that could break inside (the video lists the main 3 - power FETs, heating wire, microcontroller).

 

[WasserGKuehlt]

I found this by accident browsing for other things, for those interested in what's inside the Taycan heater. Apparently in this case what failed was some digital programmable part lost its content and just needed to be reflashed. I wonder if it's flash memory that could not handle the heat (literally) - NAND flash is definitely susceptible to corruption due high temperatures, especially if left hot while unpowered (should be well known fact by engineers designing automotive embedded systems that use NAND flash storage). Could it be that the whole Taycan heater recall saga was caused by some fresh engineers not knowing this (not reading specs, or using a part which did not spec this)? ?

Interesting theory, and definitely matches the circumstances of my heater failure - which I’ve observed live while charging at a HPDC station: longish charging session, above 80 SoC, high batt temp, fans came on to cool it off, followed immediately by a sudden stop and a charging error message.

The part number of the replacement heater is PAD-963-507-A, which seems to be unchanged.

 

[violuma]

NAND flash is definitely susceptible to corruption due high temperatures, especially if left hot while unpowered

I have no idea what ever happened to this tech, but it seems that Macronix tried to make lemonade out of this particular lemon about a decade ago, if I'm reading that whitepaper correctly. What you call "corruption" or "content decay" they tried to treat as "self-healing" that could purportedly extend the life of the flash RAM by periodically turning it into a tabula rasa using targeted heat and then rewriting it.

 

[whitex]

Interesting theory, and definitely matches the circumstances of my heater failure - which I’ve observed live while charging at a HPDC station: longish charging session, above 80 SoC, high batt temp, fans came on to cool it off, followed immediately by a sudden stop and a charging error message.

The part number of the replacement heater is PAD-963-507-A, which seems to be unchanged.

Perhaps they put in a more robust part, or reconfigured existing part into pSLC mode instead of MLC (1 bit per cell vs. 2), or adjusted refresh algorithm to be aggressive. Any of those are possible, or something completely different. My hope however is that my speculated cause for the heater issues is not a new failure mode that they haven't discovered yet, i.e. has been masked by other problems and will only start showing up on cars only after few years.

 

[bluesky]

This video was posted before.

The original NHTSA doc on the heater failure says something to the effect of “improper welds” or bad solder joints. There is probably more than one failure mode, since there were successive models of the heater.

 

[whitex]

This video was posted before.

The original NHTSA doc on the heater failure says something to the effect of “improper welds” or bad solder joints. There is probably more than one failure mode, since there were successive models of the heater.

In other words a complete clusterfuck of a design? While I haven't seen this video until recently (it's 6 months old), it does show that the heater is a very simple device - small microcontroller turning current on or off to a resistance wire to create heat. It doesn't get much simpler than that. I'm surprised Porsche still chose to continue to use them for new designs. If what you're saying is true, it seems the heater manufacturer has issues across the board, from design to manufacturing, and maintenance functionality (lack of ability to update its firmware in the field for example - perhaps they thought they could get such simple software right on their first go).

 

[KenR]

I believe I recall someone in a recent review video saying that all the cabin heating in the new Taycan will be provided by a heat pump. Presumably this doesn’t rely on a hot wire.

Sponsored