BMS-029 - Tesla Must Do Better

JVL · Jun 5, 2023

NV Ray said:
Did they give you a free loaner?

Assuming that is not a joke....

No loaner.

tesluv108 · Jun 5, 2023

JVL said:
Assuming that is not a joke....

No loaner.

The loaners and friendliness expire along with the warranty.

ucmndd · Jun 5, 2023

Crutonius said:
Okay I think I see where you're coming from. Since Tesla didn't make the lithium ion cells then it's not their problem. And therefore it's not our place to pressure them to mitigate any cell failures. And really our beef is with Panasonic. So we should be writing angry letters to them for making batteries that eventually fail.

It's an interesting perspective. I'm not so sure I agree with it.

I see it slightly differently. Writing angry letters to anyone complaining about the laws of physics and the current electrochemical reality of lithium batteries is not likely to be particularly fruitful (although maybe it will make you feel better?).

The right way to solve a problem is to start with a problem statement and develop potential solutions. In this case, the problem is "we need to develop safe, reliable, cost-effective batteries for electric vehicles that are reasonably fault-tolerant and economical to replace when they do break or reach the end of their expected service life."

The wrong way is is to start with a solution and back into the actual problem. "An electric vehicle battery must be able to isolate any one arbitrary cell from a ~7,000 cell pack to ensure absolute longevity and service life of the remaining pack/cells" is a great example of the wrong way to solve a problem.

Is the latter example possible? Yes. Is it possible at an acceptable level of cost, complexity, and reliability such that the juice is worth the squeeze? I strongly suspect not.

Crutonius · Jun 5, 2023

ucmndd said:
I see it slightly differently. Writing angry letters to anyone complaining about the laws of physics and the current electrochemical reality of lithium batteries is not likely to be particularly fruitful (although maybe it will make you feel better?).

The right way to solve a problem is to start with a problem statement and develop potential solutions. In this case, the problem is "we need to develop safe, reliable, cost-effective batteries for electric vehicles that are reasonably fault-tolerant and economical to replace when they do break or reach the end of their expected service life."

The wrong way is is to start with a solution and back into the actual problem. "An electric vehicle battery must be able to isolate any one arbitrary cell from a ~7,000 cell pack to ensure absolute longevity and service life of the remaining pack/cells" is a great example of the wrong way to solve a problem.

Is the latter example possible? Yes. Is it possible at an acceptable level of cost, complexity, and reliability such that the juice is worth the squeeze? I strongly suspect not.

They wouldn't necessarily have to be able to isolate every single cell. That would be nice but probably costly. Instead what they could do is make a way to isolate the module with the failed cell in it. Instead of dealing with 7000 devices they'd only have to deal with 16. In the context of a 20k pack this seems feasible to me.

MP3Mike · Jun 5, 2023

Crutonius said:
They wouldn't necessarily have to be able to isolate every single cell. That would be nice but probably costly. Instead what they could do is make a way to isolate the module with the failed cell in it. Instead of dealing with 7000 devices they'd only have to deal with 16. In the context of a 20k pack this seems feasible to me.

Even to do that would probably require adding 32 contactors, you are look at ~$5,000 extra just for that, and replacing 16 short modules jumpers with 48 short and 16 long bus bars, all of which take space, add weight, an energy, so you might have to drop a module to account for all of that extra equipment. (And efficiency would likely take a small hit as well.) Cooling of those contactors could even be an issue. By the time you get done with all of the extra stuff to make the pack more resilient, you have made it worse overall, and cost so much extra that you could probably have paid for half of a replacement.

And are you really OK with a single cell failing causing you a loss of 6.25% of your range, which would also drop your pack voltage by ~50v, resulting in slower charging and less performance. (Well it might be possible to keep the performance, but it would increase the amperage demands on the remaining cells, which would likely make them fail quicker.) When the second module fails you would be down another 50v, and would be at the bottom of the supported voltage range for CCS charging. (I don't know how low of a voltage the Superchargers, or the onboard AC charger, support.)

To give you an idea, @wk057 thinks that Tesla removing two modules from a 400v 100kWh pack to make a 350v 85kWh is pushing things, and likely to cause more failures of the battery and inverter(s)/driveunit(s).

Bypassing two modules on a 400v "85kWh" pack with ~12% degradation to make a 350v ~60kWh, would give you a pretty rotten experience, especially when one more cell failure would then put you out of business.

gt2690b · Jun 5, 2023

MP3Mike said:
Even to do that would probably require adding 32 contactors, you are look at ~$5,000 extra just for that, and replacing 16 short modules jumpers with 48 short and 16 long bus bars, all of which take space, add weight, an energy, so you might have to drop a module to account for all of that extra equipment. (And efficiency would likely take a small hit as well.) Cooling of those contactors could even be an issue.

And are you really OK with a single cell failing causing you a loss of 6.25% of your range, which would also drop your pack voltage by ~50v, resulting in slower charging and less performance. (Well it might be possible to keep the performance, but it would increase the amperage demands on the remaining cells, which would likely make them fail quicker.) When the second module fails you would be down another 50v, and would be at the bottom of the supported voltage range for CCS charging. (I don't know how low of a voltage the Superchargers, or the onboard AC charger, support.)

To give you an idea, @wk057 thinks that Tesla removing two modules from a 400v 100kWh pack to make a 350v 85kWh is pushing things, and likely to cause more failures of the battery and inverter(s)/driveunit(s).

Bypassing two modules on a 400v "85kWh" pack with ~12% degradation to make a 350v ~60kWh, would give you a pretty rotten experience, especially when one more cell failure would then put you out of business.

Not to mention adding that much hardware would likely decrease reliability making the whole thing counterproductive and more expensive

Crutonius · Jun 5, 2023

MP3Mike said:
Even to do that would probably require adding 32 contactors, you are look at ~$5,000 extra just for that, and replacing 16 short modules jumpers with 48 short and 16 long bus bars, all of which take space, add weight, an energy, so you might have to drop a module to account for all of that extra equipment. (And efficiency would likely take a small hit as well.) Cooling of those contactors could even be an issue. By the time you get done with all of the extra stuff to make the pack more resilient, you have made it worse overall, and cost so much extra that you could probably have paid for half of a replacement.

And are you really OK with a single cell failing causing you a loss of 6.25% of your range, which would also drop your pack voltage by ~50v, resulting in slower charging and less performance. (Well it might be possible to keep the performance, but it would increase the amperage demands on the remaining cells, which would likely make them fail quicker.) When the second module fails you would be down another 50v, and would be at the bottom of the supported voltage range for CCS charging. (I don't know how low of a voltage the Superchargers, or the onboard AC charger, support.)

To give you an idea, @wk057 thinks that Tesla removing two modules from a 400v 100kWh pack to make a 350v 85kWh is pushing things, and likely to cause more failures of the battery and inverter(s)/driveunit(s).

Bypassing two modules on a 400v "85kWh" pack with ~12% degradation to make a 350v ~60kWh, would give you a pretty rotten experience, especially when one more cell failure would then put you out of business.

Where are you buying your contactors for that much? And why 32? And why contactors? I feel that you've reached for the most maximalist way to implement this solution in order to strawman it.

I'm not okay with losing 6.25% of my range. You know what's worse? Losing the entire battery. If I can get away with driving another 100k miles until the next failure I'll take it.

MP3Mike · Jun 5, 2023

Crutonius said:
Where are you buying your contactors for that much? And why 32? And why contactors? I feel that you've reached for the most maximalist way to implement this solution in order to strawman it.

You need two contactors to either connect to module n to module n+1 or module n+2. (Shoot you might even need another 16 contactors to disconnect the other end of the module to fully isolate it.) From what I could tell ~$150/contactor is really cheap. The actual price I was quoted was $300/each for contactors rated to 500v/1000A, but I figured Tesla would have a pretty good, 50%, volume discount. On the other hand you could use cheap, underrated, contactors like Ford did in the Mach-E where they weld shut, and they have to limit WOT duration, and put the car in limp mode to prevent outright failures, but I didn't figure you wanted to make it that much more unreliable.

Maybe there is a better choice than contactors for switching 1000A around, but I don't know what that is, do you?

Frankly I think the whole idea is ludicrous. The only feasible approach would be to design the pack to be opened, and sealed, fairly easily with the intent that you take the bad module out and install a "dummy" bypass module. Then maybe you figure there is a $2-3k charge for each module that needs removed. (Of course you would be pre-paying for the design, in higher initial vehicle costs, to be more serviceable even if you didn't ever need it.)

David29 · Jun 5, 2023

gt2690b said:
I don't know.. I know people that work at Dell that expect (engineer) their hard drives to fail very shortly after the warranty is up.. not because they are trying to screw customers but because they are making them just reliable enough to fulfill their warranty period... some would call that good engineering

on a related note.. how is your S85 2013 holding up!?

I had forgotten until you wrote this, that a former manager of mine worked for awhile for the large appliance division of Westinghouse (washers, dryers, that sort of thing), and described their engineering practice as something like that. Engineer the machines to last a certain period, consistently, so that you do not build a lot more longevity into one part than into another. I do not think he tied it to the warranty, but as I recall he talked of not over-engineering so -- he said they called it "value engineering. " In a way, it makes sense -- don't build long term value (and cost) into parts that are likely to be discarded when the rest of the machine is no longer worth saving or fixing. I am not sure how practical that is, in practice, or how safety requirements come into it. You don't want parts that carry natural gas in an oven or dryer to be failing prematurely, certainly.
Can any practicing machine/appliance designers comment on this?

Crutonius · Jun 5, 2023

MP3Mike said:
You need two contactors to either connect to module n to module n+1 or module n+2. (Shoot you might even need another 16 contactors to disconnect the other end of the module to fully isolate it.) From what I could tell ~$150/contactor is really cheap. The actual price I was quoted was $300/each for contactors rated to 500v/1000A, but I figured Tesla would have a pretty good, 50%, volume discount. On the other hand you could use cheap, underrated, contactors like Ford did in the Mach-E where they weld shut, and they have to limit WOT duration, and put the car in limp mode to prevent outright failures, but I didn't figure you wanted to make it that much more unreliable.

Maybe there is a better choice than contactors for switching 1000A around, but I don't know what that is, do you?

Frankly I think the whole idea is ludicrous. The only feasible approach would be to design the pack to be opened, and sealed, fairly easily with the intent that you take the bad module out and install a "dummy" bypass module. Then maybe you figure there is a $2-3k charge for each module that needs removed. (Of course you would be pre-paying for the design, in higher initial vehicle costs, to be more serviceable even if you didn't ever need it.)

Circuit breakers, fusible links. Probably some other methods if I looked into it more. If you want to get low tech about it then maybe even an access panel a tech could reach on the top of the battery where the output lines of the modules can be reached and cut if necessary.

Contactors are used where they are in the battery because there's a need to open and close the circuit a lot. Isolating a module is something that only needs to be done once and I believe only 2 lines need to be cut to effectively isolate a module. Maybe even 1.

David29 · Jun 5, 2023

MP3Mike said:
They have. They learned from the failures of the old packs and have made design changes so that future packs don't suffer the same failures. (It sounds like most of the failures are for pre-2016 packs, and that 2016+ packs should be much more reliable.)

And shortened the warranty..

MP3Mike · Jun 5, 2023

Crutonius said:
Circuit breakers, fusible links. Probably some other methods if I looked into it more.

Both good for disconnecting on an over-current event, but that isn't what people are asking to be done here. (Either one would disable the pack if triggered.)

Crutonius said:
Contactors are used where they are in the battery because there's a need to open and close the circuit a lot.

Or where you need to remotely switch something. (I assumed that they would use latching contactors, so they stayed in the last commanded postion, so essentially they are remote switches.)

Crutonius said:
Isolating a module is something that only needs to be done once and I believe only 2 lines need to be cut to effectively isolate a module. Maybe even 1.

Modules are connected in series, if you cut either connection to it you just disabled your entire pack.

ucmndd · Jun 5, 2023

David29 said:
And shortened the warranty..

For some, although I suspect the large majority of Teslas sold do less than 150,000 miles in 8 years, so the change to a warranty with a mileage limit was inconsequential for most.

"Infinite mile" warranties carry a lot of risk for the warrantor. A risk worth taking when you're a new company selling new technology and trying to convince people to buy by lowering their overall risk - and I must admit a deciding factor in my own decision to buy as a high-mileage driver. But probably unsustainable in the long run.

gt2690b · Jun 5, 2023

MP3Mike said:
Both good for disconnecting on an over-current event, but that isn't what people are asking to be done here. (Either one would disable the pack if triggered.)

Or where you need to remotely switch something. (I assumed that they would use latching contactors, so they stayed in the last commanded postion, so essentially they are remote switches.)

Modules are connected in series, if you cut either connection to it you just disabled your entire pack.

Here are some breakers you could use instead of contactors:

Square D HJ060 HJA36060U33X PowerPact Circuit Breaker T197523

Square D HJ060 HJA36060U33X PowerPact Circuit Breaker in stock, at a great price, with a 30-day warranty and fast USA shipping.

teamequip.com

I'm not seeing the savings though.. 36 of these is more than an entire pack

Crutonius · Jun 5, 2023

MP3Mike said:
Both good for disconnecting on an over-current event, but that isn't what people are asking to be done here. (Either one would disable the pack if triggered.)

Or where you need to remotely switch something. (I assumed that they would use latching contactors, so they stayed in the last commanded postion, so essentially they are remote switches.)

Modules are connected in series, if you cut either connection to it you just disabled your entire pack.

Ah. So they are. I was under the impression they were wired in parallel.

Different plan then. Patch panel on the battery where a tech could rearranged the wires to jump out a module. It'll take some extra copper but it shouldn't be that expensive.

brainhouston · Jun 5, 2023

father_of_6 said:
Why would losing one cell create added stress on the remainder? This has to be a solvable problem.

Its not stress per se, u cut one out, u now have one brick with 73 parallel, this is now ur weak link n will limit rest of ur bricks to 73p worth of capacity
On top of that this brick will charge/discharge faster now that its smaller, n voltage gain/drop will be faster n not matching other 74p bricks.
Only way theoretically this could work if u cut out single cell from all 96 bricks but thats a lot of work...

father_of_6 said:
Why can't an EV battery just keep the voltage where it needs to be and let the amperage reduce as cells die off?

Cause physics of series circuit

Crutonius said:
They wouldn't necessarily have to be able to isolate every single cell. That would be nice but probably costly. Instead what they could do is make a way to isolate the module with the failed cell in it. Instead of dealing with 7000 devices they'd only have to deal with 16. In the context of a 20k pack this seems feasible to me.

This function already there, many ppl posted brick voltages n imbalance, u can see which brick is out of population

Crutonius said:
Ah. So they are. I was under the impression they were wired in parallel.

Different plan then. Patch panel on the battery where a tech could rearranged the wires to jump out a module. It'll take some extra copper but it shouldn't be that expensive.

As was said before 96 in series, 6 series ones hiding in one module.
Any extra panels will add moisture ingress possibility...
Only remote way possible would be if all 96 bricks were separate, then u can remove one n jump it but you'll need software support for it...

Also, Tesla already went different direction, 3/Y's have only 4 modules, n Plaids/4680s have all cells glued in foam so no way any module replacements happening in the future... These S/X packs will die off eventually n only full pack swaps will be the future. my 2c.

father_of_6 · Jun 5, 2023

Crutonius said:
Fundamentally I think the problem is that Tesla's battery management system is just inadequate.

ucmndd said:
I see it slightly differently. Writing angry letters to anyone complaining about the laws of physics and the current electrochemical reality of lithium batteries is not likely to be particularly fruitful (although maybe it will make you feel better?).

Yeah, I'm not seeing this as a Tesla specific problem or even anyone's problem that we can assign blame or fault to, it's just a challenge of physics and economics.

ucmndd said:
Would you rather spend $5-8k on a new battery in 8-10 years, or a much greater amount in fueling costs over the life of the vehicle? I know my answer. Trying to mimic the exact economic model and value proposition of ICE cars is not a logical path forward.

I do think this doesn't work for most people though. While I agree with you and rationalized the purchase of the most expensive vehicle I've ever owned (M3LR) at least partially because of decreased fuel costs, I don't think the economics work for most people.

I read your question as "Would you rather spend $8k upfront or $14k over 10 years?". It's obvious which is the better choice, but with the amount of credit card debt and people living paycheck to paycheck in the U.S., I think most people would choose to pay the higher amount over a longer period of time.

From what I've read, half of those with 5 figure salaries live paycheck to paycheck. Most people don't have $8k sitting in the bank to buy a new battery. Currently, lenders don't commonly issue loans for car repairs, but perhaps that will just become a thing... replacement battery financing.

If battery recycling is in full swing, and the economics of replacing a battery is worked out, then I guess this is already a solved problem and not too wasteful. It just *feels* that way to me. It's also a bit hard to stomach that 99% of my battery cells are fine, but I've got to pay $15k for a new pack. That doesn't sit well with my frugality.

I *still*, *still* think someone can overcome this problem. Perhaps it will require an entirely different battery technology, or just a very different perspective on the problem.

MP3Mike · Jun 5, 2023

father_of_6 said:
It's also a bit hard to stomach that 99% of my battery cells are fine, but I've got to pay $15k for a new pack.

But you don't know they are fine. There could be a bunch more just about to fall in the same way. (At least one person who had the snip surgery done had a cell in a different module go out shortly after. Which isn't unexpected given they were made at the same time, originally matched and were subjected to the exact same usage.)

jmcdono362 · Jun 5, 2023

ucmndd said:
I’m at 170k and counting, I’ll let you know. So far so good.

It appears those with high mileage have better luck with the old Model S. It's the ones that have low miles for its age that are getting the BMS errors. My guess is they sit too long, allowing moisture to corrode inside the battery packs. Where the high mileage ones heat up the battery enough to evaporate moisture build ups.

David_Cary · Jun 6, 2023

jmcdono362 said:
It appears those with high mileage have better luck with the old Model S. It's the ones that have low miles for its age that are getting the BMS errors. My guess is they sit too long, allowing moisture to corrode inside the battery packs. Where the high mileage ones heat up the battery enough to evaporate moisture build ups.

Do we have real data on that or are you stringing some anecdotes together?
Someone with 50k and a failed battery is more likely to complain (on a forum) than someone with 150k and a failed battery. Also the 150k car is more likely to be in a car accident and totalled first (and it is worth less so more likely to get totalled with the same accident). Or the person driving that much doesn't have the time to come here and post about it.
Someone who buys a $80k car and then doesn't drive it much maybe someone more likely to tell Tesla to keep the car and moves on.
I would be very surprised if these and a lot of other factors that I haven't thought of skew the anecdotes so we really have no idea. Now Tesla probably does so I will eat my hat if we actually have some leaked real data. I would take WK057 as another source but also a skewed one - knowledge differences, frugality difference etc.
I also would be less likely to go to WK if I had 200k rather than 50k because I would figure there was less salvageable cells.

It is really awful to learn that packs are becoming harder to service. Makes the whole sustainability argument go completely out the window.

BMS-029 - Tesla Must Do Better

Member

2023 M3 SR LFP, (plaid something, someday?)

Well-Known Member

Member

Well-Known Member

Active Member

Member

Well-Known Member

Active Member

Member

Active Member

Well-Known Member

Well-Known Member

Active Member

Member

Active Member

Membler

Well-Known Member

Member

Active Member

Similar threads