Time for a new HV Battery

[Recell]

indeed

as with many things in life there are of course no guarantees. but that’s where applied mathematics and the wealth of data that we pump up to the cloud come into play. running that data through our software tools, we get to a minimum spec of less than two standard deviations that all our packs are required to meet (and in turn is used to certified a pack for a given range - 210+ miles, 240+ miles, 250+ miles - based on that distribution criteria).

the reality is that a failed brick is literally so pathological that it lies completely outside the rest of the capacity distribution. it’s not even a question of being long-tail - statistically, it’s out past Pluto when it fails. of course, you can then quantify any borderline or ‘at-risk’ cells and exclude them from the population if needed. it’s simply a matter of how tight you want that capacity distribution - our minimum spec is < 2 std dev, often we end up with < 1 std dev.

oh, and have we mentioned that we like data!

 

[Videomatic3]

@NV Ray I'm super interested in the Supercharger charging speeds with your new pack. If you'd be so kind (along with anyone else!) as to share your experiences (when you have some decent data under your belt) I'd appreciate it.

also note the temperature. one time i did charge and the temperature outside was 77 degrees, and it stayed above 100kw until 53% and reached a peak of 120kw, so i might have to make another video in the summer time to compare winter/summer
im curious what the weather was like when tlP85 was charging

so in comparison i say in 1 hour of charging on the old pack i got 2 hours of driving
with the new pack 1 hour of charging gets me 3 hours of driving, its a pretty significant difference

 

[Videomatic3]

This is the highest % I've seen it hold 100kw+, have to say I'm impressed, not sure why it doesn't do this every time

 

[howardc64]

oh, and have we mentioned that we like data!

haha yeah. Data is great in space that don't have much metrics. But of course I can only make guesses

Since you guys have seen 500 packs. Presumably all have failed and lets say average of 125k miles. Thats ~50k bricks with 1/~100 brick failure rate at an average of 8? years and 125k? miles. Interesting how you would calculating average rate for 2nd brick failure from this data. Seems like missing info.

If have data on large number of good packs that haven't failed @ similar age and miles, that would bring down the failure rate of course

 

[dark cloud]

we categorize water/moisture ingress as water/moisture ingress. the 85% we cited above does not include packs with either visible signs of water ingress or onboard detection of moisture by the BMS. the overriding factor here is calendar age, mileage, and chemical deterioration/dendrite growth. in short, life.

some portion of the board and voltage sensor failures are either moisture related or manufacturing defects, or both; and some portion of the 5% 'other' remaining in the pie is outright water ingress. As we've said before, water/moisture ingress isn't zero, but far outweighed by the aging factors noted above.

But, you are located in Texas. And I know you are going to say you get cars from all over the country, but I'm going to guess most of your cars are from low-moisture and salt-free geographic areas, and I would guess, (without any data to back up my claim), packs from a location such as Norway, or where I live, or the eastern US, would have a much higher percentage of moisture/salt related corrosion failures.

 

[Recell]

relatively few of the cars we see are from the south, and even fewer from Texas - that will change of course, but you have to remember that outside of a few key markets (CA, NY, NJ, CT, MA, FL), market penetration on the early Tesla’s was limited.

If it helps, our last 4 consecutive packs were from NJ and Massachusetts. srsly, go figure ¯\_(ツ)_/¯

makes it really hard to get the pack bolts off, and the road salt is certainly no friend to the main fuse cover, but otherwise no major skew in our brick data

 

[kirkusss]

I wonder if battery replacement prices will also go down following the current new car prices. For my 85D Tesla is giving a 21k CAD trade-in, less than what a replacement pack is, reman or new

 

[aerodyne]

relatively few of our cars are from the south, and even fewer from Texas - that will change of course, but you have to remember that market penetration on the early Tesla’s was limited.

If it helps, our last 4 consecutive packs were from NJ and Massachusetts. srsly, go figure ¯\_(ツ)_/¯

makes it really hard to get the pack bolts off, and the salt is no friend to the main fuse cover, but no major skew in our brick data

Asking for the new owner of my 2015 85D...

Based on your data, could you estimate the percentage of how many 85 packs are making it to either 150k and or 10 years, subtracting out those that have failed due to moisture intrusion?

The car has been in the rain less than 5 times, always garaged, SoCal based, factory ac drain installed.

 

[Recell]

unfortunately, that is harder than it might seem. first, there’s several variables/degrees of freedom involved in what you’re asking. but more importantly, remember that Recell lives in a world of failure () and our population is exclusively failed packs vs. the general population chugging along just fine. in stats jargon, our population isn’t a representative sample. it becomes difficult/tenuous to extrapolate percentages, mean time between failure and such to the larger population, at least based on the data that we track.

anecdotally, we see plenty of packs out past 150,000 miles doing just fine.

fwiw: @wk057 does track a segment of general population data that might be closer to what you’re looking to understand.

 

[aerodyne]

unfortunately, that is hard to do. first, there’s a lot of variables involved, but more importantly, remember that Recell lives in a world of failure and our population is exclusively failed packs vs. the entire population chugging along just fine. in stats jargon, our population isn’t a representative sample. it would be pure speculation to extrapolate percentages, mean time between failure and such to the larger population.

anecdotally, we see plenty of packs out past 150,000 miles doing just fine.

know that’s perhaps not the answer you were looking for.

Best answer I have to date, thanks.

The pack in question always had a weak string, module 2, string 2. Always 10-20 mv low per SMT, although pack level nominal capacity and range (75.5 kwh, 260 miles RR) was excellent per the fleet data on Teslafi.

Does having this data help at all in doing a preventative refurb?

 

[howardc64]

The car has been in the rain less than 5 times, always garaged, SoCal based, factory ac drain installed.

I wonder how much car wash contributes to the water ingress issue. My HV got disabled 2 days after car wash. Found gen1 DCDC HV connectors likely corroded (unplug replug fixed it haha)

My car wash method is multi-step with electric power washer (water to knock off major debris->spray foam soak->rince->soap wash mitten->rince->squeegy off water to avoid spots) A ton of water is going to go down the windshield and dump near the firewall/bulk area just above the battery.

2013 MS85 75k miles in rainy Pacific Northwest. Original owner. Always garaged. Sees rain on short trips. Door handles are all still original. Originally battery. Still charge > 240+mi (rarely do so don't have exact #, not that matters anyway with calibration fuzziness)

BTW, hybrids also have HV electronics in the front (my Prius's AC compressor is HV. Inverter also sits next to the ICE in front) but 2 major differences:

• Water runoff from windshield/hood seams is a more matured design than Model S
• ICE is a big heater to quickly dry off water intrusion

The pack in question always had a weak string, module 2, string 2. Always 10-20 mv low per SMT, although pack level nominal capacity and range (75.5 kwh, 260 miles RR) was excellent per the fleet data on Teslafi.

I have 20mV imbalance in module 10 (right behind the passenger front tire and closest to the fuse bay door) and 12 (2 modules down) is like 2/3 of that. All 6 bricks in every module are very close. imbalance is mainly between modules. Try charging > 93% to trigger re-balancing a couple of weeks ago. No effect on inter-module rebalancing (maybe intra-module got better? not sure, maybe placebo haha)

Talking to a Roadster battery expert and suggested sweeping SOC % to capture imbalance behavior. Put together a spreadsheet/chart (linked below) and now collecting data. Here is a rough summary of spreadsheet if anyone is interested to do collection

• In numbers (Apple's spreadsheet) Its what I'm familiar with and wanted a 3axis chart (SOC, brick #, mV imbalance). Unfortunately doesn't plot the SOC axis according to SOC # so just collect in increments of 10%. I plan to just drag charge % slider down 10% steps while collecting next couple of weeks.
• Excel export also attached
• Just capture SMT (on latest android version 1/16/22, can't find v#) BMS page (record start->10s->stop), find file in log and transport to computer, grab last row and dump into the spreadsheet. Change time stamp to SOC% from Battery page
• For now, 2 data point at 69.1% and 66.7% SOC. All other SOC is just copied and fake ready to accept real data once collected.
  • Sheet 2 is the raw SMC BMS data rows.
  • Sheet 1 collects the relevant columns and compute each brick voltage to pack average
• Not factoring in temperature for now

Numbers Sheet

brick imbalance SOC.numbers

drive.google.com

Excel Sheet

brick imbalance SOC.xlsx

Export Summary This document was exported from Numbers. Each table was converted to an Excel worksheet. All other objects on each Numbers sheet were placed on separate worksheets. Please be aware that formula calculations may differ in Excel. Numbers Sheet Name,Numbers Table Name,Excel Workshe...

docs.google.com

 

[3mp_kwh]

RE: Time for new HV battery
Not there yet, but what is the zeitgeist on Tesla's C-packs versus WK057's? For example, I'd rather a new 350v, 90KWh, pack, than a ~400v, 100KWh pack if it has ~100k on it from 2016? On the other hand, Is Tesla putting failed 100's, minus two modules, into its $23k C-batteries? IMO, there could be a huge value difference which begs more transparency.

I'd guess Jason H. could negotiate for a specific donor 100, but hard to pull the trigger if that jacks the price and it comes to pass all the Tesla C-packs are basically 100-shells with 90KWh of new 18650s. Lastly, I agree if Tesla can lower prices 20% elsewhere than we might expect (.8)$19k=$15,200. At least, one can hope.

 

[kirkusss]

The new C pack have a new manufacture date on them. All reman carry the original packs date, even if -01 is added for reman.
Anyone knows if before the 85kWh reman packs codes were visible in the parts catalogue? Was checking yesterday and no 85kWh reman was present, only a 90kWh for 2012-2016. Today I cannot log in to the catalog anymore....
In march, after my warranty expires will try requesting a pack change with a reman, hoping I'll get the new 350V one.

 

[aerodyne]

Lastly, I agree if Tesla can lower prices 20% elsewhere than we might expect (.8)$19k=$15,200. At least, one can hope.

If recent history is a guide, hope is not a strategy regarding Tesla promised features or service.

I'd go third party. Service has to be better than Tesla.

 

[3mp_kwh]

I'd go third party. Service has to be better than Tesla.

Prefer best-pack/decent-service, over best-service/decent-pack. I live ~7mi from an SC (FUSC) and am paying $.40/KWh; all the more reason for the best possible pack. RE: Tesla's 100-shell C-packs, I worry that WK057's earlier reporting that Tesla is aggressively bidding 100's in the used/salvage market may be how some are being sourced. Would they leave the original 100's tag (presuming these weren't "C")?

 

[ucmndd]

Prefer best-pack/decent-service, over best-service/decent-pack. I live ~7mi from an SC (FUSC) and am paying $.40/KWh; all the more reason for the best possible pack. RE: Tesla's 100-shell C-packs, I worry that WK057's earlier reporting that Tesla is aggressively bidding 100's in the used/salvage market may be how some are being sourced. Would they leave the original 100's tag (presuming these weren't "C"?

Parts marked new are new. Remanufactured packs are clearly designated as such. The law generally requires it.

 

[dark cloud]

The new C pack have a new manufacture date on them. All reman carry the original packs date, even if -01 is added for reman.
Anyone knows if before the 85kWh reman packs codes were visible in the parts catalogue? Was checking yesterday and no 85kWh reman was present, only a 90kWh for 2012-2016. Today I cannot log in to the catalog anymore....
In march, after my warranty expires will try requesting a pack change with a reman, hoping I'll get the new 350V one.

But there are some reman listed in the 2016-2021 catalog, and since my 2014 received the 1014116-00-C pack listed in that catalog, the catalog the pack comes from doesn't matter. Don't know if the old reman 85 is now called the reman 90 or not.

 

[kirkusss]

Cannot be called differently as it's different capacity. But I hope they will not provide any 85s reman, after seeing all issues with them

 

[tubaprde]

Cannot be called differently as it's different capacity. But I hope they will not provide any 85s reman, after seeing all issues with them

what's wrong with an 85 reman pack? if it fails within 4yrs you get another pack with another 4yr warranty, right?

 

[kirkusss]

If paid out of pocket, yes. But for those having a failure in the last months of warranty is not.