Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register
  • The final cut of the 8th episode of the Tesla Motors Club Podcast, featuring Balazs Biro, of the prominent Hungarian EV channel Villanyautósok, is now available. You can watch it now on YouTube or listen to it on all major podcast networks.

Failed Model 3 HV Battery - Out of Warranty - DIY

I am currently facing an interesting problem that I thought might be worth sharing with the community.

A few weeks ago, I bought a clean title Model 3 LRRWD with FSD and 143k miles for $23k. The main catch was a HV battery with significant imbalance (one brick was 3.50V with all others at around 3.73V). Therefore, the maximum range was about 60 miles, and deteriorating quickly.

Tesla wants $17.5k to replace the pack (which they keep). I wanted to see if I could do it for cheaper, and I also wanted to learn something new.

Removing the pack was not too difficult, just time consuming. I can explain this process in more detail, if desired.

20211216_021952.jpg

20211216_022318.jpg


Removing the necessary penthouse components before pack lid removal was time consuming, mainly because I wanted to be careful because of HV and because I hadn't done this before. I can also attempt to explain this in more detail if desired.

20211216_025459.jpg
20211216_025518.jpg


Lid removal itself is relatively easy, there is only glue around the outside edge (unlike the Model S/X), and it can likely be reused.

20211218_001455.jpg


Measuring the voltage of each module showed that module 4 (far right) had a lower voltage by around .23V, so that's the one I removed. This module has 23 bricks in series containing 46 cells in parallel.

Once removed, finding the bad brick location was possible by measuring the voltage at pads at the top of the module. The tops of the modules are not really meant to be removed and wouldn't be very useful anyway, so I rotated the module to the bottom side.

The bottom side is covered by a form of expanding foam/potting compound. I am not sure if it is for fire resistance, vibration resistance, or both, but it flakes away relatively easy with some plastic spudgers.

20211221_212353.jpg


The brick is losing around .01V per day, and my main theory is a small internal short in one of the cells. I was hoping that any bad cell would be revealed through FLIR imaging, but I had no luck with this. My secondary theory was a stuck bleed resistor or FET on the battery management board, but I saw nothing on the FLIR there (and my knowledge on this topic is nothing compared to some of the greats here, I don't even know if that's possible).

20211225_233012.jpg


Any burnt-looking areas are from soldering tests. I had run out of ideas for finding a bad cell except to isolate them by cutting the necessary fusible link, checking the cell voltage, and then soldering the link back together. This is my current roadblock, because soldering doesn't work.

As far as I know, the fusible links are copper with nickel plating. They are very soft and fragile. I cut this one in the middle in an attempt to potentially make it easier to solder back together, but the issue is that the collector plate and cell conduct my soldering iron's heat away too well. I cannot get a good solder joint because of this, and for obvious reasons it is undesirable to continue to heat the cell much with a soldering iron. Tesla uses ultrasonic welding to create very localized, brief heating that does not damage the cells. I do not have that capability.

At this point, I have two main options:

1. Find a way to repair broken fusible links and to find the bad cell (suggestions very welcome).​
  • If I find the bad cell, I can cut the links to that and then the brick will only have around 1/46th of a capacity loss. This will almost certainly go out of balance after a few months, so I plan on running leads (attachment to be decided) to the penthouse for easier access to manually bring it into balance every few months or so, I guess. I would then likely have a Model 3 LRRWD with FSD for around $25k plus weeks of my time.

  • I don't really know of a great way to find the bad cell (if it even is a bad cell and not a BMB issue, in which case I would really have to go to option 2). If anyone knows a good way to measure current flow without breaking the circuit (maybe like a current clamp that I've used for measuring AC current, but scaled down?), that would be great to know. Otherwise, I guess I need to find a way to fix these links (suggestions very welcome) and continue breaking and fixing until I find the one.

2. Give up, try to sell the good modules, and try to find a good, used pack (1104424-00-_ or 1104422-00-_ ) for ~$10k. and do something dumb with the 1 bad module?
  • A good, used pack is not easy to come by. I would expect them to be more plentiful due to parted out salvage cars, but eBay only has 1 or 2 suitable packs available. I believe a software update could essentially serve as the necessary firmware redeploy once the pack is installed, but who knows. Total cost would be around $35k, but the pack should be good for another few years.
I have already tried reaching out to Gruber Motors and wk057 with no response, but that's not really their fault. At this point, a repair is highly unlikely, and I doubt they have Model 3 LR packs available. The previous owner contacted Electrified Garage who then suggested a module replacement, but I am not sure about it.

"But AlphaTango11, why don't you just replace the one bad module? Even Elon says Model 3s could have a module replaced for $5-7k."


Good question. As far as I know (especially by reading from experts like @wk057, who would likely consider this whole post to be insanity), there is no long term viability for such a repair, if it even works at all. There will always be an Amp-hour capacity difference between cells of different ages, and that will almost always result in an uncorrectable imbalance.​
Why Elon tweeted that, I'm not quite sure. I don't think that was even possible on the Model S/X, which would theoretically be even cheaper (1 of 16 modules vs 1 of 4).​

"AlphaTango11, you're wasting your time to try to save a few grand, and I wasted my time reading this."

I chose to attempt this out of curiosity (and the fact that option #2 had always existed and would be always cheaper than replacement through Tesla). There was a chance that this could have been a cheap, easy fix if a bad cell was more obvious (thermally, visually, anything really).​
Sorry if you felt this post was a waste of time, I just ran out of solutions and also figured that this could be useful if someone was considering doing something similar.​
Right now, there are not many Model 3s with a bad imbalance like this, and even fewer that are out of warranty (which goes away after 100k or 120k miles for Model 3). I am hoping to continue updating this thread as I go, but I am a college student and this may take some time.

Would love any input (positive or negative), suggestions, or offers for buying my 3 good modules or selling a good LR pack.

Thanks everyone.
 

wk057

Vendor & Senior Tinkerer
Feb 23, 2014
5,925
13,394
Hickory, NC, USA
Pretty sure I spotted your ticket at 057 when my staff posted some internal questions (I remember the one photo), and I made notes for my staff to basically note our normal stance on module replacements (ie: don't do it), and that we could likely sell you a pack if needed. Not sure why you wouldn't have gotten that, although probably not ultra helpful.

It's pretty unlikely It's an issue with a single cell, also. Generally a group degrades pretty evenly. The exception would be an internal short, which would be clearly visible on a FLIR. I'd likely investigate a sense issue first (corrosion, bad BMB) by manually checking the voltages and if possible checking if the BMB is drawing any current from that group (FLIR the BMB's bleeders).

Even if it were a cell, removing that cell isn't a long term solution. Your external balancing idea would probably buy you some time, but the BMS would actually end up more confused in the end because it would appear that the group is self discharging or charging outside its control, which would be a red flag also.

Your best bet is definitely a full pack replacement. If you can do the work (physical + software), then don't bother wasting money on Tesla's 17k+core "deal"... which is insane.

As for soldering, you can't. You need an ultrasonic welder since the metals are dissimilar.

Edit: Looks like we actually have a LR Model 3 pack with 520 miles on it (odometer, not range obviously) in stock at the moment, although we can't beat the eBay chop shop prices generally.
 
Last edited:
Pretty sure I spotted your ticket at 057 when my staff posted some internal questions (I remember the one photo), and I made notes for my staff to basically note our normal stance on module replacements (ie: don't do it), and that we could likely sell you a pack if needed. Not sure why you wouldn't have gotten that, although probably not ultra helpful.

Yeah, not sure what happened, but I only have the initial default response. No worries though! At this point all I'd need to know anyway is if you have any packs you'd be willing to sell (which we could discuss privately if so).

It's pretty unlikely It's an issue with a single cell, also. Generally a group degrades pretty evenly. The exception would be an internal short, which would be clearly visible on a FLIR. I'd likely investigate a sense issue first (corrosion, bad BMB) by manually checking the voltages and if possible checking if the BMB is drawing any current from that group (FLIR the BMB's bleeders).

Right. That makes sense. I'm using a FLIR One Pro and there's absolutely nothing I can tell on the bottom of the module. The little dot in the mid-left part is a reflection of my body heat.

IMG_8049.JPG


The top of the BMB also appears fine. Voltages are in agreement from multiple measurement points and definitely drop by .01V every day or so while other bricks stay at ~3.73V.

IMG_8042.JPG


I guess it's possible my FLIR isn't sensitive enough, since the heat has to be going somewhere, but I'm not sure what else I could do (and apparently finding the bad component wouldn't do me much good anyway, at least for being able to use the pack in the car again)

Even if it were a cell, removing that cell isn't a long term solution. Your external balancing idea would probably buy you some time, but the BMS would actually end up more confused in the end because it would appear that the group is self discharging or charging outside its control, which would be a red flag also.

Makes sense, thanks for the information. Wasn't sure here.

Your best bet is definitely a full pack replacement. If you can do the work (physical + software), then don't bother wasting money on Tesla's 17k+core "deal"... which is insane.

Yeah, I agree. I know I can do the physical work, not sure what software work I would need besides a redeploy and coolant refilling. Was just going to try a software update and try priming the coolant pumps manually over time. Would love to buy a pack from you for the right price, if you have any.

As for soldering, you can't. You need an ultrasonic welder since the metals are dissimilar.

Makes sense, thanks again.

Edit: Looks like we actually have a LR Model 3 pack with 520 miles on it (odometer, not range obviously) in stock at the moment, although we can't beat the eBay chop shop prices generally.

What's the best way to contact you? I checked all of my inboxes but couldn't find an 057Tech response recently (had one from a while ago for an unrelated question).

Thanks for everything you do around here.
 
Reading this post and most of wk057's previous posts on Model S batteries and BMS, is it fair to say that the Model 3 battery architecture and BMS is pretty similar to the Model S battery architecture and that we can expect similar long term degradation and reliability for the Model 3? Or does Model S battery tech have additional failure modes that make Model 3 long term results hard to predict?

For that matter is there any reason to think other battery EV architectures that have active thermal management will do any better or worse than the Model 3/S. I'm assuming that the Leaf is a bit of an edge case because of passive cooling and that LFP is a whole other subject.

Side Note: wk057's posts on the operation of the BMS are fascinating and made me want to dive in deeper. Is it possible to get enough BMS data from a 2021 model 3 to start examining what's going on in there and hacking some code to examine the results? Any pointers to understand this stuff better? Only interested in monitoring and modelling, I don't have any hardware skills to actually mod anything.
 
Last edited:
Electrified Garage in Ocala, FL (also in the Northeast) does this. They have a way to test and find the bad cell. they said they charge about $6 to $7k start to finish. Nice guys, they may be willing to give you some tips or fix it for you depending on where you’re located.
 
I for one do not consider reading this thread a waste of my time. When there are thousands of Teslas out of warranty, it will be nice to have several repair alternatives so the cars can remain road worthy. I would probably never do what the author of this thread did, but I sure can learn for his and others experiences! I believe Gruber Motor Company has repaired several HV batteries that had a fuse-able wires that did not open on a individual cells, causing a problem with the entire pack. Maybe Tesla has solved this fuse-able link not opening problem in the newer pack designs? In the repair video they mentioned it is fairly rare event but it does happen. Be interesting to see if Gruber has or will repair M3/MY batteries as many M3's are coming out of warranty. The Model 3 HV battery top lid might be easier to remove than the older Model S vehicles but the battery is much easier to remove from the bottom on the latter. Interesting that repairing the batteries by swapping out a bad module for a good one is no longer recommended by Tesla or anyone else. I guess if you have enough modules (ie.. Tesla) they can match a replacement module with the good modules. Remains to be seen if the new structural batteries can be repaired/replaced or used in another application (ie.. fix storage etc) after they are longer useful in a vehicle.

THANKS for the thread, tribal knowledge is invaluable!
 
Last edited:

wk057

Vendor & Senior Tinkerer
Feb 23, 2014
5,925
13,394
Hickory, NC, USA
Eventually I'll have to start a thread on this, instead of correcting this misinformation every time it's posted, but the "fix" they claim to do is not a long term solution. Removing cells from a group causes an imbalance the BMS was not designed to correct (all except the original Roadster)... only detect and notify. Gruber's success in losing dozens of customer Roadsters in multiple fires... er, I mean, "fixing" Roadster batteries this way doesn't translate to the newer models.
 
Eventually I'll have to start a thread on this, instead of correcting this misinformation every time it's posted, but the "fix" they claim to do is not a long term solution. Removing cells from a group causes an imbalance the BMS was not designed to correct (all except the original Roadster)... only detect and notify. Gruber's success in losing dozens of customer Roadsters in multiple fires... er, I mean, "fixing" Roadster batteries this way doesn't translate to the newer models.
Please do because I thought the cars were designed to allow a cell fuse wire to blow/open and continue on operating normally. If true, from a laymen point of view how does this differ from cutting the wire to a bad cell that did not blow on its own? Would seem somewhat logical that out of thousands of cells in a battery pack one or more cells could become defective on many packs over time? There probably are many reasons why a cell could go bad that does not involve the fuse wire blowing and therefore cutting the wire is not the fix for these cases? You are probably 100% correct in your post, maybe I have just not found/read the correct explanation. But that is one reason why this tread is important.

Thanks Very Much for you post and time!
Regards
 
  • Like
Reactions: IdaX
Eventually I'll have to start a thread on this, instead of correcting this misinformation every time it's posted, but the "fix" they claim to do is not a long term solution. Removing cells from a group causes an imbalance the BMS was not designed to correct (all except the original Roadster)... only detect and notify. Gruber's success in losing dozens of customer Roadsters in multiple fires... er, I mean, "fixing" Roadster batteries this way doesn't translate to the newer models.

This would be great, especially as more and more Model 3 and Y have battery issues out of warranty in the next few years. Rich/EG, Gruber, and even Elon's tweets have led people to think that module replacement or repairs are viable alternatives to a pack replacement, when it really doesn't seem to be the case.

It's unfortunate that there's not a better option than replacing the pack, as it seems that even a single failure out of a few thousand cells will eventually doom the car to a $17,500 pack replacement, but I guess that's life with an EV like these.

The brick continues to drain ~.01V per day. It's been about a month, so I feel like even if I did magically find a bad cell and clip the leads, the entire brick has aged compared to the others because of the constant load. My other idea was to cut the leads of one cell from each brick in each module in an attempt to preserve balancing, but this also seems unwise, as it results in an immediate ~2% degradation and there's still one slightly older brick regardless.


By the way @wk057, definitely still interested in discussing with you about purchasing any of your company's available LR packs (especially the ~520mile odometer one). Haven't received any emails, but I'm sure 057Tech is busy nowadays. Maybe there's a Gmail issue, as only I've only received my confirmations (#057_8147679 from a few weeks ago, for example).
 

wk057

Vendor & Senior Tinkerer
Feb 23, 2014
5,925
13,394
Hickory, NC, USA
By the way @wk057, definitely still interested in discussing with you about purchasing any of your company's available LR packs (especially the ~520mile odometer one). Haven't received any emails, but I'm sure 057Tech is busy nowadays. Maybe there's a Gmail issue, as only I've only received my confirmations (#057_8147679 from a few weeks ago, for example).
We're pretty busy, and local weather has kept most of my crew from the office this week so far. Hoping to get mostly caught up today. Pretty sure we're not weeks behind on tickets, so will check into that.

This would be great, especially as more and more Model 3 and Y have battery issues out of warranty in the next few years. Rich/EG, Gruber, and even Elon's tweets have led people to think that module replacement or repairs are viable alternatives to a pack replacement, when it really doesn't seem to be the case.
In the case of Elon, giving the benefit of the doubt, I think he's misinformed, mixed up, or talking about some future iteration where modules are handled similarly to the powerwall/powerpacks where mismatches would actually only cause a drop in capacity to the weakest module with no other issues. The drawback is complexity (each module in a powerpack has a bidirectional DCDC) and a limit to max current. I don't see this being viable in a car any time soon.

The others are well aware at this point, including public failures of such "repairs", but claiming to have quick and cheap fixes for Tesla packs drives clicks and views, so..... yeah.

Edit: (missed this)

I thought the cars were designed to allow a cell fuse wire to blow/open and continue on operating normally.
That's the root of the misconception. The cell fuses are a safety mechanism designed to prevent a catastrophic failure, among other things. They can blow due to a number of issues, but the BMS detects this imbalance over time and the long term solution is pack replacement from the failure. It was never intended to be a mechanism to allow indefinite use after a fuse failure.
 
Last edited:
Here is an excerpt from a web link article published July 8th 2018 from CleanTechnica.


Tesla has employed a large number of parallel cells (46?), each with its own fusible link, allowing cells to fail independently without affecting the rest of the pack and assuring high lifetime functionality and protection against individual cell failure. The operational energy is limited to about 75 kWh, but the computed capability is about 80 kWh. The report says there is 78.27 kWh usable capacity out of 80.5 kWh total capacity.

Follows are couple of links that seems to debunk the statement above as wk057 points out. From a practical engineering perspective, it would seem to be a rational design approach.



Excerpt from the link above:
On the other hand, if for some reason the weak cell can not reach full charge, perhaps due to a very high self discharge, or in an extreme case, a short circuited cell, then the good cells, rather than the weaker cell, could possibly become overcharged.

I did not find any discussion relating to the above statement if the fuse-able link blew or was cut on a short circuited cell but it probably is addressed somewhere in the technical discussion.

If you had 46 50 ohm resistors in parallel and removed one from the circuit it would result in very little change to parallel resistance or current flow.

I am still reading/digesting the considerable information on the web site but it appears removing even a single cell via a fuse-able basically dooms the entire pack! Don't know why that seems scary to me but it does the only consultation it is supposedly a rare event.

Hope this is of some interest to the technical audience in this thread, otherwise feel free to ignore.
 
Last edited:

wk057

Vendor & Senior Tinkerer
Feb 23, 2014
5,925
13,394
Hickory, NC, USA
Correct: the loss of a cell fuse is supposed to be a very rare event, and if it does happen, the module is no longer viable in the rest of the chain of groups in series.

Under normal use, the variations brought about by initial manufacturing are pretty well evened out with dozens of cells in parallel or better, even when used in large series groups such as in an EV pack. This allows the BMS to keep everything well in balance easily provided there's no issues. Once a cell fuse is lost, either from a cell failure of some kind or a shop that doesn't know what they're doing breaking it on purpose, the groups are no longer able to be balanced effectively by the BMS, and the now-weak group will always be a weak link. It will function until the imbalance is too great to be safe (as in, under load or charging the BMS calculates a chance of being pulled under a safe voltage or over a max voltage, vs the rest of the pack). At this point the BMS has to shutdown use of the pack.

The cell fuses are a safety mechanism, and an effective one. But any use after one or more is lost is just borrowed time.

Put another way, imbalance in a pack is caused by capacity deltas. A 20mV imbalance, which is the most a Tesla pack BMS allows at rest for an extended period of time, can be caused by as little as a 0.05% delta in capacity between cell groups (using a Tesla 85 pack as the example). Removing a single cells causes an instant 1.35% capacity delta.... 27x worse than the BMS considers OK.
 
Last edited:

miimura

Well-Known Member
Aug 21, 2013
7,178
7,129
Los Altos, CA
It seems to me that any replacement module would have to be selected to precisely match the per cell group Ah capacity and cell chemistry of the rest of the pack. The independent garages that have done this so far probably don't have the equipment or selection of modules to find that precise match.
 
Just wanted to give a quick update on this.

@wk057's company had a battery that was cheaper than Tesla, but I ultimately went with a LRRWD pack I found on eBay for $9,000. Using dimensions from the old pack and some wood, we made a battery stand that raises the pack above the wheel wells of our truck, and it somehow made it the 10 hours each way.

It was that junkyard's first time working with a Tesla battery, but it seems okay except for some things mentioned below.
Here's the video of the somewhat sketchy loading process.

The replacement pack has ~9,000 miles on it and it is from a 2019 vehicle, so I felt pretty safe about it, even though it was from a rollover accident with airbag deployment. It looks a little dirty, but I'm hoping that's because it sat in their shop for a few months.

Status of the replacement pack:
  1. The pyrofuse is blown due to the airbag deployment.
    • This is not necessarily a bad thing, actually. It keeps the contactors from being energized and draining the pack, which is nice if the car sits for months on a salvage lot. I am pretty sure I can replace it with my good one.
  2. The module voltages seem very well balanced, and are not decreasing. Each half is 178.0V and 178.1V, which should be "within spec".
    • This is around 3.70V per cell, which is a fairly ideal SOC of ~42%. No complaints here so far.
  3. One of the coolant nubs at the front of the pack is pushed in, and one of the mounting bracket bolts at the front is broken off. :(
    • There's no good way for me to safely fix this without opening the pack and ensuring there's no other damage.
What's next:
  1. Since I'm not sure of the condition of the various coolant inlets / outlets and also not a fan of some scratches on the bottom of the pack, I'm planning on opening the replacement pack and transferring the modules to my original pack (that is still in great physical condition).
    • To do this, I have to take almost everything out of the penthouse (just like last time), carefully remove the lid, and carefully remove each module. (This is where I am now)
      • I plan to use heavy-duty (200lbs+) suction cups and straps, especially for the center modules (which require some maneuvering).
  2. Once the modules are transferred, I'll have to find some good sealant/adhesive to use for the lid. Suggestions here are welcome, as I don't know specifically what Tesla uses.
    • Before I seal the pack, I might try to find a good way to attach some silica packets in the pack in an attempt to absorb some humidity. Not sure if it's necessary, but I don't see how it can hurt as long as they are well-secured.
  3. Once the main lid is on, I have to put the penthouse components back in. Here is where things get a little more interesting:
    • I want to use the replacement pack parts whenever possible, since components like contactors will have 9k miles instead of 143k miles. Even if they can handle more use, I don't see a reason not to, except:
      • As far as I know, the BMS stores fault codes related to crashes / airbag deployments. I do not know if these are somewhere in the HVC or in each individual BMB.
        • Because of this, I don't know if I should swap HVC's (and/or other components like the PCS) or keep them separate.
        • The service manual for HVC replacement mentions backing up and restoring using Toolbox's Autodiag, but I really don't have that capability
          • Worst case scenario, I can trailer it to Tesla and have them replace the HVC or try to convince them to let me use my old one. (I don't think this Autodiag can be done remotely, unfortunately)
  4. After that, the rest is not so bad. The main concern here is finding a good way to get air out of the coolant lines, but it could just involve some repetitive pumping and venting if I don't want to pay up for Toolbox. As far as the necessary software redeploy goes, I am hoping that performing a software update will serve the same purpose. If not, I'll have to figure that out, too.
Essentially, the pack replacement seems at least plausible from this point forward. I need to find a good sealant, figure out what to do about airbag-related errors in the replacement pack's HVC (or figure out how to reuse my old one), and then put it all together and pray to any deities listening.

Of course, it's very likely that @wk057 or other battery experts will shoot a ton of holes in my current plan, which I will gladly welcome.

I'll hopefully have another update in a week or two instead of a month. So far, I will have around $35k in the car after this, which probably isn't that good of a deal after all. Regardless, it's definitely been a learning experience and I hope I can potentially be a resource for others in similar situations in the future.
 

Attachments

  • TestLid.jpg
    TestLid.jpg
    590.2 KB · Views: 116
  • ReplacementPack.jpg
    ReplacementPack.jpg
    783.2 KB · Views: 120
Last edited:
Here is where things get a little more interesting:
  • I want to use the replacement pack parts whenever possible, since components like contactors will have 9k miles instead of 143k miles. Even if they can handle more use, I don't see a reason not to, except:
    • As far as I know, the BMS stores fault codes related to crashes / airbag deployments. I do not know if these are somewhere in the HVC or in each individual BMB.
      • Because of this, I don't know if I should swap HVC's (and/or other components like the PCS) or keep them separate.
One thought would be to use your original parts and consider the ones from the replacement pack to be spares. In general it just seems less risky to me. Of course the experts might have much better insight into this.

Did your $35k net included parting out the old pack?
 
This is the first documented instance I am seeing of a Model 3 battery imbalance issue. I was under the impression the "new battery chemistry" (if that's even a real thing) in model 3 would alleviate these types of problems.. Perhaps a naive/ill-informed assumption.

Should we expect similar life in our batteries as much as earlier Model S/Model X? Maybe 250-300k on our batteries if we're lucky?
 

Products we're discussing on TMC...

About Us

Formed in 2006, Tesla Motors Club (TMC) was the first independent online Tesla community. Today it remains the largest and most dynamic community of Tesla enthusiasts. Learn more.

Do you value your experience at TMC? Consider becoming a Supporting Member of Tesla Motors Club. As a thank you for your contribution, you'll get nearly no ads in the Community and Groups sections. Additional perks are available depending on the level of contribution. Please visit the Account Upgrades page for more details.


SUPPORT TMC
Top