High Rates of SoH degradation in Taycan

[ciaranob]

Data Source Thread:
https://www.taycanforum.com/forum/threads/baseline-for-hv-battery-soh-performance.12815/

I have a thread (above) dedicated to tracking SoH (State of Health) of Forum member Taycan's (with some external additions) - we currently have data from 50 cars and 73 readings of SoH taken directly from the PCM via OBDII tools. You can purchase these tools very cheaply and retrieve a huge amount of data that is readable with the free app CarScanner - highly recommended for all owners.

The reason I started the original thread, assuming we had sufficient participants to be statistically meaningful, was to see how SoH in our new Taycans compared with other studies usually heavily biased towards Teslas which of course have some of the highest road miles of any current EVs.

So to cut to the chase, below is an externally published plot of Tesla Model S/X SoH (labeled (Battery) Remaining Capacity) vs Mileage (in km) supposedly from hundreds of cars over multiple years reported in 2020 for example in these articles:

https://electrek.co/2020/06/12/tesla-data-battery-degradation-limited-mileage-packs-equal/
https://electrek.co/2020/06/12/tesla-data-battery-degradation-limited-mileage-packs-equal/
https://electrek.co/2020/06/06/tesla-battery-degradation-replacement/

I scaled their chart (which has some peculiarities to it but can discuss later!) to my chart from the BaselIne SoH thread.

Clearly the data 'cloud' for the Tesla cars is showing significantly better SoH performance of Tesla HV batteries than what we are recording for the 50 cars in our dataset. Arguably the worst Tesla performing cars are essentially equivalent to the best Taycan SoH performance.

Now whether we are truly comparing apples to apples here is up for discussion/debate (all sorts of variables effect the SoH value including how buffering effects readouts, battery chemistries, cooling systems, dominant charging, climate (although clearly all 4 climate data zones in our data sit below the Tesla cloud!) etc.).

Interesting plot!

So if we take this at face value, not only does the Taycan battery degrade (a lot) faster early on but longer term projection curves look like Taycan's performance will sit anywhere from 5-7% lower than the Tesla equivalent.

EDIT - Possibly a big variable here when comparing to Tesla cars might be the buffering issue and how exactly SoH is read in each system - is it conceivable if Porsche's measure is more conservative (bigger buffers and say even reporting SoH from the worst vs best performing cells etc) one could possibly shift the average SoH curve for each population to be a lot closer than suggested in this graph? Wild speculation on my part and clearly not an expert in how these measurements are performed.

Here is the plot in original form I used as an underlay (a zoomed in version like my own) from teh above listed sites (also my original Forum plot above in the Baseline SoH thread):

 

[Hirschaj]

What I find interesting is that people are not reporting reduced range on this forum as far as I can tell. There are multiple people reporting about 250 miles range at 100% charge regardless of SoH reading. Am I perceiving that wrong? Is SoH reading irrelevant until a certain percentage reduction is reached?

 

[npx]

What I find interesting is that people are not reporting reduced range on this forum as far as I can tell. There are multiple people reporting about 250 miles range at 100% charge regardless of SoH reading. Am I perceiving that wrong? Is SoH reading irrelevant until a certain percentage reduction is reached?

this has always been my understanding, possibly anecdotal. that some of the battery capacity was held in reserve to replenish cells that lost capacity. Tesla does this too, but after a decade of knowledge about their deg rate, they are more confident to have a smaller kwh buffer.

 

[ciaranob]

this has always been my understanding, possibly anecdotal. that some of the battery capacity was held in reserve to replenish cells that lost capacity. Tesla does this too, but after a decade of knowledge about their deg rate, they are more confident to have a smaller kwh buffer.

And this is where I suspect we just don't have enough data/insight in the comparison between brands in respect how the buffers are managed, and together with other metrics, how the respective calculations are made in respect backing out a valid comparison.

There's enough expertise and knowledge in this Forum I further suspect that we might still be able to determine how valid that initial chart comparison is.

 

[ciaranob]

What I find interesting is that people are not reporting reduced range on this forum as far as I can tell. There are multiple people reporting about 250 miles range at 100% charge regardless of SoH reading. Am I perceiving that wrong? Is SoH reading irrelevant until a certain percentage reduction is reached?

Or are the metrics being 'manipulated' or just 'measured' differently enough that predicted range impacts are yet to be seen - they are after all predicted ranges based on inputs/ouptuts controlled by the PCM - now I sound like a conspiracy nut !! But I do think the actual values used by proprietary algorithms are potentially somewhat unique to a brand.

 

[jvincent]

SOH provided by a car's own battery management system via OBD is ballpark useless. SOH presented by another manufacturer's BMS is also not directly comparable. A proper SOH check takes many, many hours and involves a full charge and discharge under controlled conditions. Heat loss and initial buffer size also need to be considered.

Respectfully, the data does not support the thread title assumption. The assumption might be correct, but the data isn't there.

 

[ciaranob]

SOH provided by a car's own battery management system via OBD is ballpark useless. SOH presented by another manufacturer's BMS is also not directly comparable. A proper SOH check takes many, many hours and involves a full charge and discharge under controlled conditions. Heat loss and initial buffer size also need to be considered.

Respectfully, the data does not support the thread title assumption. The assumption might be correct, but the data isn't there.

Again I have basically stated that we are comparing like to like I.e. what the OBDII tool is reading as a ‘measure’ of SoH as reported. I never said anywhere that a full factory SW measure of SoH will be the same - so with respect you’re point fairly pointless and redundant as the data we have is ‘there’ regardless of your opinion as to it being a proxy or an accurate or even an inaccurate measure of SoH.

What we ARE comparing and again as stated, all we are comparing, is a like to like measure from a consistent tool we have available and these comparisons are as such entirely valid.

If you actually read my post. I emphasize repeatedly that we of course don’t know what exactly the PCM is offering as a measure of SoH, but is consistent between cars, and further stated clearly that we cannot assume that how the data were measured for Tesla and other brands were equivalent to what is measured by Porsche - hence the offer for those more knowledgeable to discuss and clarify.

 

[jvincent]

If you actually read my post. I emphasize repeatedly that we of course don’t know what exactly the PCM is offering as a measure of SoH, but is consistent between cars, and further stated clearly that we cannot assume that how the data were measured for Tesla and other brands were equivalent to what is measured by Porsche - hence the offer for those more knowledgeable to discuss and clarify.

Actually read the whole thing and the "EDIT" information was on point. My only issue is forming a conclusion based on data you yourself rightfully call into question. Like for like OBD measurements would be comparing the delta between two Taycans both with calibrated battery management systems. Any other comparisons are not credible.

Bjorn Nyland has a rambling but informative video on battery degradation measurements. He touches on the efficacy of OBD comparisons and how manufacturers handle initial degradation masking.

 

[JP_21]

As Porsche gathers more data on battery performance/ degradation do we think there is a chance they may look to reduce the inbuilt buffer and ultimately offer more battery capacity through future software updates?

 

[AndiL]

The SOH data provided via OBDII ist not a measurement but a very rough calculation. This is what the Porsche Taycan boss Mr. Giek said at a meetup.

 

[TurboInCold]

I suspect that the reserve in between gross and net capacity are being used while the battery is degrading.

My MY22 Turbo reports SoH 91%, which should mean that out from the net capacity of 83.7kWh only 76.2kWh (83.7 * 91%) should be remaining.

However my last long trip with moderate speeds (90km/h roads) from last week was 386km (240miles) on a charge, car reported consumption of 19.4kWh/100km (3.20 mi/kWh) and I begun the trip with 100% and ended with 11%.

This means that 74.884 kWh was used during the trip and the usable battery capacity as today would then be 84.1 kWh ((19.4 * 3.86) / 0.89).

Based on this calculation is seems that car has 100% of the net capacity still available, even the SoH from OBD reports 91%…

 

[DirtyPop]

Data Source Thread:
https://www.taycanforum.com/forum/threads/baseline-for-hv-battery-soh-performance.12815/

I have a thread (above) dedicated to tracking SoH (State of Health) of Forum member Taycan's (with some external additions) - we currently have data from 50 cars and 73 readings of SoH taken directly from the PCM via OBDII tools. You can purchase these tools very cheaply and retrieve a huge amount of data that is readable with the free app CarScanner - highly recommended for all owners.

The reason I started the original thread, assuming we had sufficient participants to be statistically meaningful, was to see how SoH in our new Taycans compared with other studies usually heavily biased towards Teslas which of course have some of the highest road miles of any current EVs.

So to cut to the chase, below is an externally published plot of Tesla Model S/X SoH (labeled (Battery) Remaining Capacity) vs Mileage (in km) supposedly from hundreds of cars over multiple years reported in 2020 for example in these articles:

https://electrek.co/2020/06/12/tesla-data-battery-degradation-limited-mileage-packs-equal/
https://electrek.co/2020/06/12/tesla-data-battery-degradation-limited-mileage-packs-equal/
https://electrek.co/2020/06/06/tesla-battery-degradation-replacement/

I scaled their chart (which has some peculiarities to it but can discuss later!) to my chart from the BaselIne SoH thread.

Clearly the data 'cloud' for the Tesla cars is showing significantly better SoH performance of Tesla HV batteries than what we are recording for the 50 cars in our dataset. Arguably the worst Tesla performing cars are essentially equivalent to the best Taycan SoH performance.

Now whether we are truly comparing apples to apples here is up for discussion/debate (all sorts of variables effect the SoH value including how buffering effects readouts, battery chemistries, cooling systems, dominant charging, climate (although clearly all 4 climate data zones in our data sit below the Tesla cloud!) etc.).

Interesting plot!

So if we take this at face value, not only does the Taycan battery degrade (a lot) faster early on but longer term projection curves look like Taycan's performance will sit anywhere from 5-7% lower than the Tesla equivalent.

EDIT - Possibly a big variable here when comparing to Tesla cars might be the buffering issue and how exactly SoH is read in each system - is it conceivable if Porsche's measure is more conservative (bigger buffers and say even reporting SoH from the worst vs best performing cells etc) one could possibly shift the average SoH curve for each population to be a lot closer than suggested in this graph? Wild speculation on my part and clearly not an expert in how these measurements are performed.

Here is the plot in original form I used as an underlay (a zoomed in version like my own) from teh above listed sites (also my original Forum plot above in the Baseline SoH thread):

Great info. Where can I purchase the tools need to pull this information?

 

[TurboInCold]

I suspect that the reserve in between gross and net capacity are being used while the battery is degrading.

.......

Based on this calculation is seems that car has 100% of the net capacity still available, even the SoH from OBD reports 91%…

I did further analysis on all the chargings I have made. I have charged the car for 187 times, and I selected only such where I have charged 60% or more (77 different sessions).

I calculated the capacity of the battery using the the statistics I have using this formula :
Capacity kWh = ((Consumption kWh/km * Distance driven km) / Consumed capacity %)

To create a trend line, I selected the median value of each month from the data to the graph below.

The end result is interesting, it looks like the car has had net capacity of 89 kWh in the beginning and now it is closer to the spec of net capacity at around 83 kWh.

Also interestingly, the trend follows quite well the SoH measurements I have taken during the ownership (red dots in the graph)

 

[ciaranob]

Actually read the whole thing and the "EDIT" information was on point. My only issue is forming a conclusion based on data you yourself rightfully call into question. Like for like OBD measurements would be comparing the delta between two Taycans both with calibrated battery management systems. Any other comparisons are not credible.

Bjorn Nyland has a rambling but informative video on battery degradation measurements. He touches on the efficacy of OBD comparisons and how manufacturers handle initial degradation masking.

I hear you and again no one and certainly not me has ever suggested that the OBDII tools are giving us a precise and absolute value of SoH - just not the case but also not the point.

And also, as others are pointing out, there are examples of variances between observed Kw used and consumptive range observations against OBDII SoH readings but many of these also involve calculations and assumptions - easy to dismiss one data set for another.

So the point again is that whilst the car's (assumed) calculation (vs a deeper measurement) of SoH clearly might be sub-optimal it is nonetheless a consistent measure (regardless of accuracy) delivered to us via OBDII from a decent population of cars to date - what are the error bars around that measure we as yet have no idea (a deeper consumptive use with assoc calculations might provide guidelines) and frankly I hope that the current OBDII data represent a pessimistic determination (perhaps related again to cell state issues etc etc) that will allow us, should we ever determine what the PCM calculation is, to bulk shift the curves from the OBDII data to a lesser degradation state!

What I can say is that whatever and however the SoH value is determined (and again, despite all the chat here no one here actually knows how this is done and all comments are entirely speculative) it is clear that the outputs in respect for example behaviour of SoH in different climate zones would appear to match exactly that as reported from industry and Porches's own comments in respect very hot climates seeing worse degradation.

The latter point is actually one of the primary reasons I started tracking this OBDII data at all as clearly I am squarely in a Hot Climate zone so of def interest (zones as defined in the baseline thread).

So good discussion and I still don't buy that the OBDII data is effectively 'useless' - it is a consistent 'means' of measurement that shows individual and gross degradation rates that are not a million miles off other EV studies and do represent an albeit likely crude true measure of SoH through time/usage/miles driven.

My goal here was to drag any (preferably industry) experts out with that provocative title to provide ACTUAL hard SoH data to offset these OBDII SoH readings and they're apparent negative implications for the Taycan's battery performance to date. I would love to see proper SoH test data and equivalent OBDII readings from the same vehicles (if I ever get mine done I will def try to produce this) but also for multiple cars and preferably multiple battery cycles etc. - but this data undoubtedly locked up in Brand proprietary reports, so no doubt dreaming on!

 

[andb]

Is the OBDII reader able to tell the percentage/kWh amount of AC and DC charging and how many charging sessions were done to 100%? Some data should be stored somewhere.

I'm asking because in two weeks I should collect a used Turbo from Porsche dealer with 2 years warranty and 45k miles. Would be good to know more about how it was charged by previous owner.

Sponsored