Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Take out HV batt pack / cut out bad cell? (S P85+)

This site may earn commission on affiliate links.
a *minor* ride height adjustment might be in order - the delta up front between a 14 module and a 16 module pack is around 110-120 lbs max - but that’s about it. nothing that can’t be handled by a knowledgeable tech - even only slightly knowledgeable tech.
 
Last edited:
You think third parties are going to send a pack to the garbage heap instead of selling it for recycling? :eek: (Not that there are many third parties dealing with HV packs at this point.) Third parties actually sell the good modules from the S&X packs to be used in other projects, so they get a second life before being recycled.

@wk057 @Recell what do you guys do with failed/bad modules? Are you able to sell them to recyclers to recover some value?

Shoot even if the recycler wouldn't pay for it, you at least wouldn't have to pay to dispose of it.
Good points but my point is Tesla CANNoT handle the volume of work that is coming their way in 2-3 years. Third party shops, many of them manned by prior Tesla mechanics, are popping up. Usually a core charge is assigned in rest of industry to facilitate that return channel to ensure costs are minimized as well as environmental efficiencies as a side effect. All Tesla would need to do is either internally or through a third party contract to take battery pack cores back and credit them. They could easily build a technical bulletin and update their toolbox 3 software to do a quick eval of the battery pack at a shop and have a calculation tool that toolbox calls through an API to determine present module market pricing based on outputs of diags from toolbox. That could lower a battery replacement from 12-25k to 5k and allow Tesla to have a datalake of battery repair info to augment their battery vehicle telemetry data.
 
As a general note, given the overall construction I'm not convinced that recovery of raw materials from these modules for recycling is really viable at any scale without some drastic (magic?) advances in that field. Recovering individual cells without damage is a lengthy and labor intense process in itself based on what I've seen from the third party noted above. Recovering raw materials would be worse all around. I don't see it happening.
I was always wondering about this.. when cell degrades or dies, do raw materials "die" too? or something else happens in the cell? any chemists here :)

It could be Tesla changed the design at some point to be more universal or perhaps cover a wider weight range. I'd expect the weight change from the old S85 to the new 90 kWh packs to be minimal, but I don't know for sure. Nothing close to that pulling out two modules or installing an original S60 pack in an S85.
I believe wk057 posted in other threads a single module from 100kWh n 90/85 kWh packs. 100kW ones use 2 extra rows of cells so they are heavier.
So my guess, 16 old modules would be about same weight as 14 new ones, just diff voltage...
Its when u go from 85/90 > 100 is when u run into weight/suspension issues...
 
BTW, been thinking if there are auto self graceful degradation solutions that would make cell failure more owner/cost friendly.

A design would need to include bypass for every module in the voltage series. Bypass would need handle the high current involved (contactor/module?) Its not some cheap low power low cost chip tech. And the more one divides the pack into more submodules, the more bypass circuit would be necessary. And for safety sake, probably don't even want the failed module staying in the pack...

So probably can't get out of the manual labor to remove/reconfigure an 1k lb item sealed for moisture. So we get to 10s of hours of labor rate pretty quickly every time a cell fails.

If there is no auto self graceful degrading solution, then EVs can't get out of this high battery replacement cost quandary.
 
Wow, you know the volume of work coming, and Tesla's expansion plans. o_O Are you from the future?


Tesla does charge a core charge if you want to keep your old pack, but them keeping your pack is already included in their pack replacement price. (I think they charge $10-15k if you want to keep your old pack.)


No, it wouldn't. Everyone is already assigning a core value to the old packs. For example, you can see @Recell's prices in this post. ($6-11k) You could always ask them how much more they would charge if you didn't provide a core HV battery pack. (Assuming they would even sell, and install, a battery without a core.)
Hey - appreciate your info . It's helpful. Don't appreciate the sarcasm.
 
  • Like
Reactions: Droschke
Wow, you know the volume of work coming, and Tesla's expansion plans. o_O Are you from the future?


Tesla does charge a core charge if you want to keep your old pack, but them keeping your pack is already included in their pack replacement price. (I think they charge $10-15k if you want to keep your old pack.)


No, it wouldn't. Everyone is already assigning a core value to the old packs. For example, you can see @Recell's prices in this post. ($6-11k) You could always ask them how much more they would charge if you didn't provide a core HV battery pack. (Assuming they would even sell, and install, a battery without a core.)
Couple of things - not all of us writing here have the same level of knowledge nor the happy ears that Tesla can do nothing wrong. It is simple to see Tesla cannot handle nor will they be able to handle the future repairs of batteries in volume that is needed because a) they have not publicly acknowledged it is an issue, b) they don't have data across the fleet over sufficient time for MY and M3 to know failure rates long term, although MS can be used to predict a bit, c) most of all, there is no infrastructure to manage long term battery maintenance.

The idea Tesla had to swap batteries instead of supercharging was in an interesting direction. Why not take that idea in to repair world? There is a company who absolutely understands this and they are a Vietnamese company I think called VIA cars. They lease the battery - that makes sense. Tesla should move to that model OR provide a simple way to pay 5k after warranty expires if the battery needs repair to swap to a refurb battery. Third party companies like car shield limit EV battery repair to 6k - that's not enough under current paradigm as the range is 5-20+k. Imagine you spend 50k on a model 3 and drive it 20k miles a year - in 7 years you have a single component fail on the battery but it happens ti be a component that causes the BMS to degrade the pack to be essentially unusable. Are you going to shell out 24k for a new battery? Can't repair it can't swap it can't put it under extended warranty. ICE vehicles don't have this issue. ICE vehicles at least can get coverage from a extended warranty To 200-300k miles and that same 6k cap will get almost a new engine on most common cars. Might have to shell out 3-4k to make up difference but not 15k difference like on a Tesla.

I think three people buy Tesla's - those that are car enthusiasts and like mods and driving, financially well to do people, and people who think EVs are next Gen but can't spend 60k every 5 years and expect the car to last 10 years without a repair being needed well in excess of the value of the car. A Camry ten years old if it did need a new engine, which is rare, would be reasonable to repair. Tesla can make that happen for their car powertrains too - but not the way it stands today.
 
  • Like
Reactions: Droschke
Imagine you spend 50k on a model 3 and drive it 20k miles a year - in 7 years you have a single component fail on the battery but it happens ti be a component that causes the BMS to degrade the pack to be essentially unusable. Are you going to shell out 24k for a new battery?
No, I wouldn't shell out $24k for a new battery, as last we heard non-warranty cost to replace a Model 3 LR battery was under $15k. (It might be as low as $12k now.) And that will likely continue to decline in the coming years.

They lease the battery - that makes sense. Tesla should move to that model OR provide a simple way to pay 5k after warranty expires if the battery needs repair to swap to a refurb battery.

People tend not to like to lease major required components for their vehicle. IMO, if they are worried like that they should just lease the car itself. And how did you come with $5k as being a reasonable cost to replace a battery pack? (Are you saying Tesla should lose money on every out of warranty pack replacement they do?) Elon was hoping to get the cost of replacing the modules in a failed/worn out pack down to the $5-7k price, but I don't think they are there yet. (And that probably isn't even possible on the current 4680 pack design.)
 
No, I wouldn't shell out $24k for a new battery, as last we heard non-warranty cost to replace a Model 3 LR battery was under $15k. (It might be as low as $12k now.) And that will likely continue to decline in the coming years.



People tend not to like to lease major required components for their vehicle. IMO, if they are worried like that they should just lease the car itself. And how did you come with $5k as being a reasonable cost to replace a battery pack? (Are you saying Tesla should lose money on every out of warranty pack replacement they do?) Elon was hoping to get the cost of replacing the modules in a failed/worn out pack down to the $5-7k price, but I don't think they are there yet. (And that probably isn't even possible on the current 4680 pack design.)
Yes - the 4680 pack, not the cells, is dumb. I actually was going to cancel my order if they switched me not because of the chemistry but the pack design. They put foam in it and made the whole pack non-serviceable. No shop can even attempt to repair them. 5k is reasonable because car warranty companies have set 6k as a common cap on any repair number 1, and number 2 it doesn't cost Tesla anything in profits IF they built a reliable repair and refurbishment system.

If they don't want packs field repairable, buying a new pack is 24k. That's absurd but there are simply no refurbished packs available and I suspect why - they can only come today from crashes - right? But in 5 years - enough cells will fail in battery packs across millions of vehicles that the good cells they can harvest should be vast majority.

In other words, there are three scenarios. Packs that fail with a handful of cells but 80-90% of cells remaining can still be used. Maybe a lot of these are still under warranty. These can easily be refurbished depending on the overall condition of the remaining good majority cells. Tesla could repair a failed pack like this for hundreds of dollars.

The second scenario is the situation of an abused battery. Lots of supercharges, running it to 5 perceht, etc. - I don't know enough about how Tesla can show abuse, but it is in everyone's best interest that they start thinking about how to track this and just stop worrying about peoples reactions to range loss. I think 95% of population owning Tesla's would rather baby the battery and lose 10-20% usable range in exjange for lower TCO and higher reliability. Abused packs of tracked should not be charged to Tesla or via-a-vi Teslas profits, us all as a whole but rather should incur some financial penalty to customers to pay refurbishment.

The third case is out of warranty packs where it may fail but the remaining cells are simply too worn to be refurbished in to another pack. This is the one that concerns me most. There are not enough failed units right now to harvest good cells that's why we have to get new packs. That may never change - it could be these packs for 95 percent of people last 200k miles - in that case Tesla may never care. But the 5% of folks who get unlucky - because of a bad batch or maybe an accident that created latent thermal or vibrational damage to the pack that is detectable. Think of spinning HDDs. One time my boss dropped a server full of spinning disks. He said "it still runs who cares" - I said "mark my words in 6 months that system will fail" - and in less than 2 months 4 out of 12 disks failed and 3 more failed in rest of year.

The electronics and circuit boards on a Tesla are second to none of any other car - therefore, even though they are solid state and don't have moving parts like old HDDs as per the analogy above, these statistically will fail and cause physical damage to inverters, drive units, bms systems, battery cells, you name it - it's all digitally managed now.

So the end result of this new world we are in is reliability should be higher for the masses, at the expense of the fewer who do have issues having to pay 4x the typical cost for an ICE repair. It's not acceptable for Tesla to think this way - because Tesla owners are right now very intellectual. If we expect the average person to want a Tesla, psychologically, they have to know that they are not facing a 5% or whatever risk of 24k fir a battery, or 4k for a steering rack because Tesla doesn't believe in component level repairs, batteries included.

Not to say the final issue here - what the heck happens to all of these cells at 200k miles or whenever the cars do get tossed? Is Tesla going to melt them down and build new cells? I think they need s separation process to do that and I heard a third party is doing it already - hope that matures too.
 
5k is reasonable because car warranty companies have set 6k as a common cap on any repair number 1, and number 2 it doesn't cost Tesla anything in profits IF they built a reliable repair and refurbishment system.
That is an interesting way to set prices. o_O What if it costs Tesla $7k to provide a refurbed pack? (Things cost what they cost, regardless of how much a warranty company wants to pay.)

In other words, there are three scenarios. Packs that fail with a handful of cells but 80-90% of cells remaining can still be used. Maybe a lot of these are still under warranty. These can easily be refurbished depending on the overall condition of the remaining good majority cells. Tesla could repair a failed pack like this for hundreds of dollars.
No, they can't.

The third case is out of warranty packs where it may fail but the remaining cells are simply too worn to be refurbished in to another pack. This is the one that concerns me most. There are not enough failed units right now to harvest good cells that's why we have to get new packs.
Not possible, you can't mix new and old cells. In fact you can't mix old cells from different packs.

Not to say the final issue here - what the heck happens to all of these cells at 200k miles or whenever the cars do get tossed? Is Tesla going to melt them down and build new cells? I think they need s separation process to do that and I heard a third party is doing it already - hope that matures too.
Tesla has said that they are already recycling ~50 packs per week. And that number has probably gone up since they announced it.

And every pack they currently make uses foam. Most are foamed at the module level, only the 4680 is foamed as an entire pack.
 
Last edited:
  • Like
Reactions: brainhouston
No, they can't.


Not possible, you can't mix new and old cells. In fact you can't mix old cells from different packs.


Tesla has said that they are already recycling ~50 packs per week. And that number has probably gone up since they announced it.

And every pack they currently make uses foam. Most are foamed at the module level, only the 4680 is foamed as an entire pack.
From what I know, I am not in automotive but I have worked with LIFe batteries, and used to work with LiPO, as long as the resistance matches you can mix match. Would like to see w057 post on this. if enough cells are being renewed pulls from other packs, they should start ti match up eventually within margins. Ie if I have 1000000 good refurb cells coming to my docs each week wouldn't enough of them start to converge and match? Guess we would need to see standard deviations on cell resistance. Are there other variables that would determine whether cells can be reused?
 
From what I know, I am not in automotive but I have worked with LIFe batteries, and used to work with LiPO, as long as the resistance matches you can mix match. Would like to see w057 post on this.
He has, repeatedly. For example:

Every battery pack has aged differently. From the factory, the cells are all a little tiny bit mismatched, but all from the same batch of cells. So they'll quickly find their place within the whole and stay there, with a little nudge from the BMS occasionally, for the life of the pack without issue.

When trying to mix and match, things like mileage don't matter. Cycles don't matter. Capacity doesn't matter. How hard or lightly it's been driven doesn't matter. It's every factor combined over the entire life of the pack that determines how it ages and how it will continue to perform. You might be able to get a module that has a very similar total capacity to the rest of a pack, or even get it to balance pretty well with the rest of the modules. But, it's impossible for a replacement module to perfectly match the rest of the pack. Things will always be slightly off, and because these bricks are in a large series set these differences always result in a feedback loop that amplifies the differences over time no matter how small the differences are... and that's just how it is. You don't mix and match cells in series because current under load/charge is constant in a series string by definition. With mismatched modules you end up with the same current at different loaded voltages, resulting in unequal power/energy distribution throughout the pack, resulting in an imbalance.

I could go on and on about the details, but suffice it to say many have tried (myself included), and all have failed. I've tried this several times, always thinking I've gotten close enough to beat the BMS (or, well, physics/chemistry)... and every time the result has eventually been failure. The difference between me and others seems to be that the folks who've been super public about their miracle battery fixes never follow up with the failure portion of the "repair"... 🙄
or

This is what Tesla's BMS tries to prevent by monitoring far more metrics than just voltage. They calculate out the capacity of every cell group individually. They calculate out the impedance of every cell group individually. They map impedance vs temperature for every cell group. They map the surface charge capacity of every cell group. They map impedance vs load/charge for every group. And way way more. In a normal pack, when some of these don't match up, that's an indication of a problem, and Tesla's BMS reacts accordingly. A mismatched module induces differences in these values. It may take some time for the BMS to become aware (some of these metrics are calculated slowly over hundreds or thousands of miles of use), but it will eventually realize something is screwy and react accordingly.
or

The cells in factory built original packs aren't perfectly matched, no of course not. That's impossible. The BMS has balancing circuitry for a reason.

BUT, they're matched well enough to be within the ability of the BMS to keep them in balance over an indefinite period of time under normal use. Unlike when mismatches are purposely introduced (replacing a module, cutting a cell, etc), the imbalances that happen "naturally", for lack of a better term, in a factory original pack are always going to be within the ability of the BMS to keep in check, especially since it's with the pack for the entire life of the battery to catch these imbalances early, learn which cell groups need the most work, and preemptively work on those as needed. It's quite complete, and not a "dumb" stateless system (the Roadster, for example, was closer to a "dumb" system, for example, which is why some more crude repair techniques may have some lasting functionality). Changing things like a whole module, or "remove" a cell, throws this completely out of whack, and the BMS might even make the situation worse for a while after because it thinks it's working on something that isn't the case anymore.
I'm sure there are many more related posts if you look.

Ie if I have 1000000 good refurb cells coming to my docs each week wouldn't enough of them start to converge and match?

There is also the issue that Tesla keeps changing the formulation. So all those 1M cells would have to be split out by formulation.
 
As a general note, given the overall construction I'm not convinced that recovery of raw materials from these modules for recycling is really viable at any scale without some drastic (magic?) advances in that field. Recovering individual cells without damage is a lengthy and labor intense process in itself based on what I've seen from the third party noted above. Recovering raw materials would be worse all around. I don't see it happening.

My limited understanding of what the big companies do in this space (e.g. Redwood Materials, JB Straubel’s new company) is essentially shred/pulverize the modules in totality and then apply more or less the same chemical separation/purification techniques that they would use on ore coming straight out of the ground.

DOE / National labs funding a lot of research in this area too:

 
..A design would need to include bypass for every module in the voltage series. Bypass would need handle the high current involved (contactor/module?) Its not some cheap low power low cost chip tech. And the more one divides the pack into more submodules, the more bypass circuit would be necessary. And for safety sake, probably don't even want the failed module staying in the pack...
While I think your idea is technically possible, it would dramatically lower the reliability and increase costs. Contactors are not cheap ($50-$150 each). The contactor's mechanical operation makes it the least reliable component in the battery pack. They are also heavy and large, so you'd reduce space for cells.
As Sandy Munro points out, you can design high reliability or high repairability, but not both. Tesla has moved to reliability with the structural pack. Other companies have copied Tesla's old approach with modules, with far more places for things to fail. And that's what is happening with some of these old Tesla packs. I expect other EV makers are going to find this out the hard way in 6-10 years. So Tesla has already come up with a new solution that should be dramatically more reliable. Of course, we won't know that for quite a few years yet!
 
  • Like
Reactions: MP3Mike
While I think your idea is technically possible, it would dramatically lower the reliability and increase costs. Contactors are not cheap ($50-$150 each). The contactor's mechanical operation makes it the least reliable component in the battery pack. They are also heavy and large, so you'd reduce space for cells.
As Sandy Munro points out, you can design high reliability or high repairability, but not both. Tesla has moved to reliability with the structural pack. Other companies have copied Tesla's old approach with modules, with far more places for things to fail. And that's what is happening with some of these old Tesla packs. I expect other EV makers are going to find this out the hard way in 6-10 years. So Tesla has already come up with a new solution that should be dramatically more reliable. Of course, we won't know that for quite a few years yet!

Wasn't suggesting bypass is a good idea. Just brainstorming through what would be involved. I think the reality is can't get out of human labor to pull pack and remove bad module with a bad cell.

I don't see how structural battery helps the key factory on battery reliability. Will probably improve macrostructural reliability and cost which are important of course (but likely increase the human labor significantly for recyclability. I'll skip that details here but not going to be possible to feed a 1k lb battery pack into a cruncher to turn into dust followed by chemical separation)The key battery failure issue is still a rechargeable cell's chemical and microstructural degradation internally in the cell over time and use. Rechargeable batteries are tiny chemical systems and failure modes have been largely similar since 100 years ago. Regardless how "advanced" it has become (mostly in energy density, PbAcid->NiCd->NiMH->LiON) In reality, increasing energy density results in ever smaller anode/cathode pitch and the ion repeated movement between them along with parasitic reactions (basic highs school chem lab teaches no reactions are 100%, you get unwanted reactions) that damages the desired microstructure. Lithium ions moving back and forth btw the anode/cathode form dendrites that eventually bridge the gap to cause a short.

So cell failures will occur with time+use and If has to be removed when it happens will results in the fundamental high cost > residual value of the car after say 10 year of usage via complete replacement of the high capacity battery pack. Used module demands are the only way to reduce the cost but really not a hugely scalable solution.
 
Last edited:
Yes - the 4680 pack, not the cells, is dumb. I actually was going to cancel my order if they switched me not because of the chemistry but the pack design. They put foam in it and made the whole pack non-serviceable. No shop can even attempt to repair them. 5k is reasonable because car warranty companies have set 6k as a common cap on any repair number 1, and number 2 it doesn't cost Tesla anything in profits IF they built a reliable repair and refurbishment system.
I think Structural/sealed pack could actually benefit in another way. U can ONLY use it for the designed car
If u look at S/X packs, bunch of ppl take them apart n use it for conversions/solar n that takes away from used market for ppl who actually need it for their car as a replacement..

If they don't want packs field repairable, buying a new pack is 24k. That's absurd but there are simply no refurbished packs available and I suspect why - they can only come today from crashes - right? But in 5 years - enough cells will fail in battery packs across millions of vehicles that the good cells they can harvest should be vast majority.
If u look on eBay right now, Model 3 packs are up for grabs pulled from low miles cars ranging 5k-14k
I think with the amount of Teslas on the road n being made we won't have a problem with finding good used ones from wrecked cars n such for very affordable price.

The second scenario is the situation of an abused battery. Lots of supercharges, running it to 5 perceht, etc. - I don't know enough about how Tesla can show abuse, but it is in everyone's best interest that they start thinking about how to track this and just stop worrying about peoples reactions to range loss. I think 95% of population owning Tesla's would rather baby the battery and lose 10-20% usable range in exjange for lower TCO and higher reliability. Abused packs of tracked should not be charged to Tesla or via-a-vi Teslas profits, us all as a whole but rather should incur some financial penalty to customers to pay refurbishment.
I don't think its abuse, packs were designed to go between 0-100% n SC. it was recommended to use 90% if full range not needed.
Thats why Tesla added buffer, SC limits n such.

In general, it seems early model S's starting to fail around 8-10yr mark n since then pack designs improved a lot so i think newer cars should easily go 10yrs
After that, with multitude of used packs i don't think its gonna be an issue anymore

I look at my Gas savings stats in the app n its around $400/mo, thats 48k in 10yrs.
U can buy new Tesla for that much n def can afford replacement pack n still be on top.