When Battery Rebuilds Go Wrong: Understanding BMSs, Spot Welders, And Safety

Batteries are amazing. Batteries are horrible. Batteries are a necessary evil in today’s world of portable everything. If you’re reading this sentence, even if it’s not on a mobile device, somewhere there is a battery involved. They’re that ubiquitous. There’s another thing batteries are: Expensive! And at $350 each for a specialized battery, [Linus] of Linus Tech Tips decided to take battery repair into his own hands.

Rather than do a quick how-to video about putting new cells in an old enclosure, [Linus] does a deep dive into the equipment, skills, and safety measures needed when dealing with Lithium Ion cells. And if you watch the video through, you’ll even get to see those safety measures put to good use!

The real meat of the video comes toward the end however, with its explanation of the different Battery Management Systems (BMS), and a discussion of the difficulty of doing battery repair correctly and safely. Lastly, the video covers something a bit more sinister: Batteries that are made to resist being repaired with new cells; DRM for batteries, so to speak.

Overall we found the video informative, and we hope you do too. You might also enjoy this peek into the chemistry behind your favorite battery types.

40 thoughts on “When Battery Rebuilds Go Wrong: Understanding BMSs, Spot Welders, And Safety

  1. Most notebook-grade BMS chips feature several thresholds which will trigger a permanent failure if exceeded. A common (and IMO necessary) permanent failure mode is the “safety undervoltage” protection, which will lock the pack out if the cell voltage drops below some critical level (usually 1.5V – 2.5V for NMC/LCO cells).

    This is not the standard overdischarge protection, which usually kicks in around 2.8V depending on PCM. This happens AFTER that, if the cells are getting discharged through a side channel or self-discharging in storage.

    Most people who have locked out their pack are triggering this protection. It’s not a tamper-resistance feature. It’s designed to prevent thermal runaway. When cell voltage drops to this level, the internal resistance permanently increases, which will cause an “event” if the pack is then fast-charged at its usual rate like nothing happened.

    Fortunately, these BMS chips all have configuration tools and software that can be used to reset permanent failures, so for someone with the know-how, these protections are not an impediment to replacing the cells.

    1. One would hope for a single-pin jumperable reset for these resets rather than software. I wouldn’t want to have to find some arbitrary binary blob that likely requires a paid license and some kind of serial bridge to reset it.

    2. Opened my laptop battery some months ago (it was swelling, though still holds a charge) and there’s a fuse inside, blow-able by the BMS chip. I imagine it’d take more than a bit of software if this thing goes into permanent failure.

      1. I don’t know the reality for laptop manufacturers, but the safety undervoltage is usually a soft permanent failure and does not blow the chemfuse, since when the cell voltage is that low you might not have enough available power to blow the fuse.

    3. The problem is unless it’s easy to reset it drives people towards potentially dangerous work-arounds or throwing the battery in the e-waste.

      An example work-around would be to connect a PSU across the batteries and trying to match the voltage, so they can be disconnected and replaced without triggering the cut-off. It’s common with small batteries for backup memories like old pre-flash memory cards and “suicide batteries” in arcade systems, but quite dangerous with large lithium packs.

      1. This is insufficient, since these types of packs are almost always 2S+ and the voltage of each cell is monitored. In order to trick the BMS, you additionally need voltage dividers to spoof the voltage of each individual cell. Or isolated PSUs for each cell, if you’re going all out.

  2. “Batteries that are made to resist being repaired with new cells; DRM for batteries, so to speak.”

    Laptop batteries you mean, with the built in selfdestruct fuse and firmware to activate in in case of deep discharge. With the persistance bit set. Jep i’ve got a few of those. At least my lenovo has it.

    1. We’re not likely to ever see that for power tools because the manufacturers use their battery designs to lock people into the brand. If you already have a couple of power tools from the same company, you’re likely to buy additional tools from the same company because the new one will be able to share batteries with the old ones. The only way we’ll ever get that is government intervention.

      1. TBH I’m surprised there’s not a bigger move to open-sourced 3D printed / OHW tool batteries that sticks two fingers up at all these manufacturers and allows interchange of packs and custom home-brew packs.

        It’s the sort of thing a hacking website could take a lead on, maybe offer some sort of prize as an incentive…

        1. Thingiverse already has a 3D battery holder with swappable components to fit various tool systems, some assembly required. Just add your cells, regulation circuitry, and contact tabs.

    2. In certain industries, there is more or less “standard” batteries.
      Be it an “old” (actually still produced) Nokia battery, or the NP-F battery form factor that Sony used for their cameras that then everyone and their dog started copying for other cinematography gear.

  3. I watched it and they don’t really understand how to use their tab welder.

    They discarded a dozen cells and gave up because they thought the sparks coming off due to bad technique, were due to cells equalising. Actually that’s just what happens when you flow hundreds of amps with crappy contact or dirty probes.
    Not to mention they equalised the cells to start with, so short of shorting out the pack, the only source for the sparks could be welder.

    Also didn’t even attempt to keep the BMS system energized which is a common way to circumvent the failsafe. Tried painting it as evil manufacturer intent when it’s usually the opposite: after a lithium cell has dropped below 2 volts, even once, it’s a fire hazard every time it’s charged in the future. The controller permanently shutting down is a safety feature in this case.

    1. You are not the only one to have noticed these two things. I thought the same thing about the sparking cells. The same maybe true with the failsafe in this particular instance, but they definitely have a point about the dark side of BMSs from certain manufacturers. I thought there was plenty of truth in this video, even if they weren’t 100% correct technically on a couple of parts. They did a good job of showing how hard it is for an average person to attack this problem. Especially since their primary target audience is gamers.

      1. Yes. This video was filled with what seemed like amateur grade disinformation presented with an expensive studio set-up. Just because you have the capital to make a confident sounding video on something doesn’t mean you should, especially if you don’t have the requisite skills and didn’t consult someone to troubleshoot.

        I’m kinda sad to see hackaday uncritically link it and promote it. This was not a deep dive and selling disinformation out of ignorance around safety doesn’t really make anyone safer. They threw away many perfectly good cells — completely ignoring their previous proclamation of reducing e waste — and this is something the article writer here noted as a positive example of safety measures.

        It’s of no use to try to simply scare someone and not actually pragmatically address WHERE risks are in a process. And they failed to properly identify the risks or troubleshoot a process that they perceived to have been conducted in an unsafe manner.

        Basically, this is clickbait and uninformative as is. Unfortunately a lot of ltt and other media seems exactly like this, despite whatever image they are presenting. They churn out content for money, even if they think they’re doing good.”

        Respectfully, Ryan, I disagree about the value of this video or their transparent tackling of the problem. Yes I agree that the discussion of self-bricking with the red BMS and others was good but that was only a few minutes out of the video. If it was primarily that, I’d be less annoyed. Many people responded to the video with the above points (that wasn’t a failing cell, you threw away a bunch of cells with no fault and got really scared instead of critical, you may need to “jump” the BMS with a charge after replacement) and the only one ltt responded to when I viewed it was the suggestion that they try charging the BMS after cell replacement on the non-RED camera one. Maybe If they do a follow up video, great.

        1. Do not confuse a feature with wholesale endorsement. The video isn’t perfect, but there’s still some good information in there. And for those who know better, perhaps we can look back on our early projects where we threw away perfectly good hardware because we thought it was faulty, only to realize our mistakes later, and relate to Linus for tossing out what appear to be good cells. Whether he needed that safety stuff or not, he was prepared to use it and *that* is where the excellent example is. Will they follow up? Maybe. If they do, we’ll probably cover it too.

          1. An uncritical feature that does not mention any of the potential flaws sure can feel a lot like a wholesome endorsement – especially to folks who may not know better and come here to learn things.

        2. There is a vast difference between disinformation and just not being a master of something you are demonstrating with the top end prolevel tools.

          This is a great cautionary tale – that reasonably technically minded people who clearly did a reasonable amount of research (or their pretty workshop studio would likely have burned down) can’t just do a battery swap, despite spending a reasonable outlay on tools- very useful to many viewers no doubt, as I highly doubt anybody with any technical understanding of how laptop/powertool/camera battery are made hasn’t looked at the price of a replacement vs the price of a few cells and thought about trying it themselves – and I certainly won’t suggest you shouldn’t try, just know you are almost certainly going to spend more mastering the how and finding a good source of cells than buying the odd replacement battery.

          Nothing said was really incorrect, in the same way almost everything you teach the 6-16 year old isn’t really correct, but neither is it wrong, just simplified and presented for the somewhat interested novice.

      1. IKR? How many failed RAID arrays and resulting data losses does Linus have to have before people realize LTT is a bunch of YouTubers and not a group of technical professionals with the deep technical knowledge and skills that generally only come with formal training and hands on experience?
        I mean I appreciate what they do but they need to work within their limitations. Or at least acknowledge them.

        1. I would point out many of the failures they have had over the years are things even the big companies have had at times – even the supposed tech pro’s can miss a detail – perhaps assuming the default configuration is sane, or making a parse-able and silent erroring typo in theirs, letting a certificate expire, assuming person x was dealing with y as its not your job… The difference is the biggest companies usually have a massive budget they can throw at the screwup and don’t also advertise it on youtube as the cautionary tale the way LTT does, so its not as memorable.

          I’d also say they do talk about their ‘not professional’ status plenty, while showing that at least getting started down the road to mastery of something need not be hugely difficult, or cool tool x makes something otherwise very hard kinda simple…

          Not having a clue isn’t a downside, as long as you don’t pretend to expertise their shared experiences are a good introduction for their fellow novice (and its LTT, they are often kind of proud of the halfarsed janky approach, and the rest of the time throwing fancy gear at it while getting in the expert to make it work (at least on the phone)).

  4. Was it an evil BMS or an engineer deciding to save money by not installing a flash flash chip. And that flash chip is another part that is a potential point of failure that will reduce the expected mean time between failures of the battery pack.

    I’m not saying that the outcome, if replacing the batteries and allowing the BMS to power down is not evil. But it may not have been the primary motivation. I would say that saving a few cents per battery pack also reduces the final price point, but it sounds like they (Red Digital Cinema) are really gouging their customers on battery packs. So in their case the internet probably was to be evil above all else.

    1. The problem is that this is a small volume product. They have all the same costs to produce any other mass market product – engineering, injection molds, regulatory compliance, warrantee, etc – but they have to amortize those costs over a few thousand units of production rather than the few million units of most mass market products.

      I sincerely doubt there was anyone sitting around twirling their mustache and making evil plans to make the batteries hard to service.

      1. Except that they could have designed the cameras to use somebody else’s batteries. Blackmagic Design, one of Red’s competitors, did just that. Their small Pocket Cinema cameras use Canon or Sony batteries (and their third party clones), depending on model. The larger Ursa cameras have interchangeable backplates that allow the use of batteries from four different manufacturers. Blackmagic also does the same with lens mounts, supporting multiple types. As for memory, they support a number of ways to handle that; you can buy expensive CFast cards, but you can also connect an external SSD or an HDMI or SDI recorder, or use standard SD cards (though you’ll want the fastest ones available) if you don’t need to record at the camera’s maximum quality.

        Red has followed the same business model as their legacy competitors; lock people into the system and make lots of money selling the extra bits and pieces they need. Blackmagic has notably chosen not to do that, keeping the total system cost much lower. Admittedly they’re not targeting the cinematic market; their products rarely show up in the production of theatrical films or high-end TV shows, whereas Red cameras sometimes do.

        LTT went with the high end choice because they could afford to, because their photography guy wanted to, and because they’re making enough money to afford to do it. But they could have gotten at least 90% of the quality for 25% of the price if they had looked elsewhere, and people watching on YouTube probably would not be able to tell the difference.

        1. It can be “locking people in” or it can be “guaranteeing performance and compatibilty” – sure it may be a fairly safe bet to use Sony or Canon batteries, but then you and your expensive high-end product are reliant on a 3rd party who may or may not appreciate your approach, may or may not honour any warranty if their battery has been used on 3rd party gear, etc. etc.

          It could easily be as much work to properly validate and characterise 3rd party battery packs & adapters as to design your own, and at the end of the process you own the design & performance.

          In the scheme of things, people buying RED cameras are not buying them to save a few bucks, there’s plenty of other ways to photograph things for a lot less money if that’s your primary concern.

        2. Using somebody else’s battery can be nice – but for that to be workable you really need there to be a suitably mass-produced battery that meets your electronic requirements AND you can be sure will still be there for the entire lifespan you wish to support your product…

  5. Why is this video even featured? Did he hack the DRM, reverse engineer the BMS schematic, or reveal anything that a simple Google search couldn’t? Most people are already aware of DRM protections, especially those of us surfing Digikey and designing our own. Chip manufacturers pride themselves on including the feature. The only thing to be gleaned from this video is that Linus should stick to building computers and not come out of his cave again.

  6. This video is cringeworthy. He’s so damn fussy about some welds having a bad contact and then totally overreacting. It is a PAIN to watch him bin one set of cells after another. He’s sending such a bad signal because the welder is some old junk. They should have purchased a nice Kweld-unit, it’s 10000 times better for sure.

  7. If one wants to “repair” a battery pack particularly safely one could always consider riping open the pack, discarding the lithium cells, and then inserting a custom circuit to give an output voltage which would look to the BMS like the Li cell voltage, but would in reality be some form of regulator (outputting the voltage expected from the old Li setup) fed from a bunch of NiMH cells. It would be heavy, it may be bulky, it wouldn;t be suitable for things that fly or need to be fast and agile, but NiMH won’t do anything too bad if you abuse it.

  8. Dear Linus,

    Do NOT trust a shop that sells you 3.49V lithium batteries.
    Do NOT trust a shop that sells you a mix of differently charged lithium batteries.

    When you do trust a shop and they sell you that crap, stop trusting them immediately and just take the plunge and try someone else. With a bit of due diligence you can find a reliable shop in 2 or 3 tries at most.

    Cells should be charged to 3.8V, but 3.7 is acceptable. Cells should not self-discharge when new. So either they were not charged properly at the factory or they are leaking from being old even when new…. Not good.

    I bought ten 5300mAh 18650s on ebay a long time ago. I knew at the time that this was unrealistic: I’ll accept them when they reach 50% of that rating. They weren’t. I got most of my money back and decided keep them for tests involving 18650 formfactor projects. i.e. mount the 600mAh battery while testing buy a real battery when the project goes into actual use.

    That didn’t really happen. I’m guessing I bought those about 10 years ago. Correction found the: You bought… Email: I bought them over 8 years ago (jan 5 2014). NEVER used, never recharged…. 8 years later they are still at 3.8V!

    Linus, The cells you bought are crap. Way worse than the crappy ones I bought on Ebay 8 years ago.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.