Fighting Back Against Dodgy Dyson Batteries

If you’ve ever worked with multi-cell rechargeable battery packs, you know that the individual cells will eventually become imbalanced. To keep the pack working optimally, each cell needs to be analyzed and charged individually — which is why RC style battery packs have a dedicated balance connector. So if you know it, and we know it, why doesn’t Dyson know it?

It’s that question which inspired [tinfever] to start work on the FU-Dyson-BMS project. As you might have surmised from the name, [tinfever] believes that Dyson has intentionally engineered their V6 and V7 batteries to fail by not using the cell balancing function of the onboard ISL94208 battery management IC. What’s worse, once the cells get as little as 300 mV out of balance, the controller considers the entire pack to be shot and will no longer allow it to be charged.

These missing resistors deserve justice.

Or at least, that’s what used to happen. With the replacement firmware [tinfever] has developed, the pack’s battery management system (BMS) will ignore imbalanced cells so you can continue to use the pack (albeit at a reduced capacity). Of course the ideal solution would have been to enable cell balancing on the ISL94208, but unfortunately Dyson didn’t include the necessary resistors on the PCB. Though it’s worth noting that earlier versions of the board did have unpopulated spots for them, lending some credence to the idea that their omission was intentional on Dyson’s part.

But not everyone is onboard with the conspiracy theory. Over on the EEVBlog forums, some users pointed out that a poorly implemented cell balancing routine can be more problematic than not having one at all. It’s possible that Dyson had some bad experiences with the technology in earlier packs, and decided to move away from it and try to compensate by using higher-quality cells. That said, at least one person in the thread was able to revive their own “dead” battery pack by installing this unofficial firmware, so whether intentional or not, it seems there’s little debate that usable batteries are indeed being prematurely marked as defective.

Proper cell balancing is key even in DIY projects, so we do have to agree that it seems more than a little unusual that Dyson would intentionally turn off this important feature in their packs. But the jury is still out on whether or not Sir James is trying to pull a fast one on his customers — as Hanlon’s Razor states, “never attribute to malice that which is adequately explained by stupidity”.

62 thoughts on “Fighting Back Against Dodgy Dyson Batteries

    1. I had a few Dysons and almost exactly after the warranty expired the battery stopped working. It has to be intentionally designed like this. For the prices they charge it should last longer. I got the extended warranty to be clever but found out that the battery is not covered! I will never buy a dyson Hoover again.

  1. Definitely a case of malice over stupidity – Dyson have been building in obsolesence to their crappy vacuums for years. I fell for the hype and bought a DC04 – never again. I have a Henry now, much better in every respect.

    1. Dysons are for people who want to show off, Henry are for people who want to actually clean stuff.

      Also, Henry deserve a lot more love for not only still being manufactured in the UK but for lasting for decades and being built to be repaired indefinitely.

      Henry master race!

      1. “to show off” – That’s a bit of a silly remark. Don’t think I’ve ever “showed it off”… I have a DC10 which was purchased for portability – It’s used up and downstairs, in the garage and also the car – It’s downside is barttery capacity as we cannot do the full house on one charge. We have had several wired units in the past and they are just too much of an inconvenience for moving around and plugging for each room.

        Henrys are ok but like most units there’s EU regulations that meant their power was reduced – We have had a Henry in the past, doesn’t clean much better than the DC10 to be honest.

        We also have a VAX carpet cleaner – now that’s an awesome unit and it’s built like a german tank. solid, powerful and reliable.

        I’ll be replacing the cells with my own rather than replace for another model so this is good to know – thanks!

        1. Henrys are amazingly repair-friendly – we had one fail after 7 years, opened it up for a look, you could tell it was designed to be serviced – all the same screw heads and all easily accessible.
          Turned out to be a pair of motor brushes that had worn away completely, replacements were under £10
          Maybe get an old one and do some basic repairs?

    2. cant speak to the battery stuff but i just replaced the belts on my 2005 DC14 Animal. it has survived a house of 3 kids 2 dogs and 2 cats that killed other well “liked” brand vacuums in as little as 2 years. Repair cost for belts and filters 36 usd…

  2. James Dyson is a mechanical engineer. I get the feeling he much admires the mechanical enginuity of his products but doesn’t get the software and electronics parts quite as much.

    1. Dyson went to art school, initially studying interior and furniture design before moving into industrial design. He is a designer / marketing guy, not an engineer.

      1. He is generally credited with inventing the cyclone vacuum cleaner. I would say that he is a real “inventor” rather than just an industrial designer or salesman. Though he is clearly a salesman too, as he has managed to successfully sell his inventions.

        1. Dyson did not invent the cyclonic vac, that was already patented in the late 1920s. Dyson designed a miniaturized variant for household use that is exceedingly loud and expensive. And he actually managed to sell a large number of these strange devices and set up an ersatz religion like Jobs did.

        2. Well he took an idea he had seen in industrial dust extractors and had the idea of applying it to domestic cleaners, after that I think his real achievement was step by step more than double the price customers were willing to pay for a humble floor cleaner by clever marketing and wacky design despite the very questionable real world performance of his early cleaners, particularly as their expensive filters became blocked.

        3. He was the first person to market a portable vacuum cleaner with cyclonic separation although they had been long used in central domestic systems and industrial dust separators. I bought a couple of Dyson vacuum cleaners and they were nasty, noisy and unreliable. I use a Henry now. He got lucky with his vacuum cleaner and made a fortune.

          1. A few years ago I was looking for a vacuum cleaner and I checked out reviews of the Dysons. All the reviews I could find mentioned the very high noise level which is what drove me to replace the crappy, cheap, noisy vac I had been using for years. I ended up buying a Meile vacuum that is so quiet that I have to push my cat out of the way when I am cleaning and it sucks so hard it could take the chrome off a trailer hitch. It wasn’t cheap, but neither are Dysons. I’m not sure why anyone would buy a Dyson.

    2. Dyson is not an Engineer. He went to art school. He is a marketeer and a showman.

      Remember all the hype about his invention of a ‘digital motor? Or, to use its othet name, a motor.

      Overhyoed, plastic, fiddly, shiny things for people who know no better.

  3. He mentions V6/V7 as compatible with his fix, and V10/V11 as “probably not compatible”. I have a V8…?

    I would love to know if ones newer than his (>V7) actually have the correct balancing implemented.

    1. A V8 should be compatible but the LEDs will be non-sensical because the V6/V7 use a single RGB LED but the V8 uses three blue LEDS + one red LED, which the programming isn’t designed for. If you look in the EEVblog thread, someone tried it on a V8 and confirmed this I believe.

      Send a photo of the PCB in a GitHub issue if you want and I can try to tell you more thoroughly! :)

      1. Great project! Thank you. I tried your good .hex firmware on V8, it works with no correct LED indication as expected. After start charging as a rule error 8 appears, though there is no any overcurrent. I gives 26.1 VDC 500mA limited power charge. Even 25 VDC was tried, a bit better, but was not effective. So .hex with 5ms option is not effective also. What can we do with faulty error?

        1. Are you using a bench power supply for charging? If so, the response time for it to go in to current limit mode might be too long. If this is the case, I’d suggest setting the output voltage to zero, setting the current limit to correct value, enabling the output, then slowly dialing up the voltage until you hit 26V. This would prevent a brief current spike from occuring when the PSU outputs 26V, current goes too high, PSU finally realizes and reduces voltage to meet current limit. All just a theory though. Hard to say what’s going on without getting a scope current probe on the power input.

          Also, the GitHub issues section would be the best place for this discussion, even though unfortunately I can’t help everyone.

  4. This does not bode well for the future of Humanity. How, pray thee tell, are we meant to envelope the Sun itself in a Dyson sphere, if, at the earliest solar flare, an energy imbalance is detected and the entire sphere is declared to be a failure?

    1. FAQ:
      Q: My Dyson sphere only achieved partial envelopment and then after a solar storm, it no longer turns on. How do I fix this?

      A: Solar storms are considered acts of god and are not covered under warranty. If rebooting your Dyson sphere does not fix the problem, please contact sales regarding our latest Dyson Sphere V2, now with Storm Stopper™ technology!

  5. So, he didn’t actually solve the problem of the non-functional balancing circuit. I wonder if it would work as the designer originally intended if those resistors are placed?

    I have a slightly different Dyson who’s battery has degraded to be almost useless, so I was looking to replace those cells sometime. Knowing how to open them is very useful!

    1. Dyson removed the spots for the resistors on the newer boards. Suppose he could have tried to install them on the older board, but the firmware mod is compatible with a wider array of boards.

      1. I’d have to either find a way for the PIC to determine if balancing will work, or add a software flag to enable the balancing code, if I added it. Also, my vacuum is a V7 :)

    2. You are correct that, unfortunately, I didn’t solve the issue with the non-functional balancing. On the V6 batteries, you could theoretically install the missing resistors and handle the rest in software, but the balancing would be very slow since it would also have to go through a 1K resistor Dyson included on the other side of the internal balancing MOSFETs. On the V7 and up models, they actually shorted the internal balancing MOSFETs of the ISL94208 on the PCB so you’d have to cut the very fine traces here and then solder on balancing resistors. I think you could actually just place a short in place of the balancing resistors because of the 1K resistor already in circuit.

      All said, I figured that of the limited number of people willing to take apart their battery, buy a programming tool, and install my firmware, there would be even few willing to make those very delicate modifications. And anyone willing to do all that probably already owns a bench power supply they could use to rebalance their battery pack manually.

      I’d initially planned on adding cell balancing until I saw that the ISL94208 balancing connections where shorted together. That’s something that just can’t be fixed in software. Maybe one day I’ll add balancing to one battery, just to show it can be done.

      1. Maybe another option is to go in, solder on one of the RC balance leads leave that hanging out a small hole and toss on one of the little handheld RC balancers when needed or charge it with the balancing RC charger.

  6. Didn’t buy a Dyson because of prices, did not expect this. I feel bad if I don’t give lithium batteries all the considerations they deserve. Wouldn’t feel qualified to make lithium-based cordless devices professionally if I can’t get balanced charging right, which shouldn’t be hard nowadays.
    It would be less problematic if we do a better job recycling old lithium but currently I try to make full use of anything rechargeable.

  7. I have the same with a Vax battery, got a new one free of charge as it failed in its first 3 years. Next task rip it open the old one, found one of the five 21700 cells was slightly below the others so guessing the same as above. Not got any further in investigating but has a NXP LPC824 chip controlling it.

  8. Well they’ve also replaced their separate carbon/HEPA filters in their purifiers – where you could simply vacuum the HEPA a few times (outside so you don’t put all that crap back into your home) and reuse it with a fresh carbon filter each time – with new one piece filters so when the carbon side is shot, the whole thing is garbage. So I wouldn’t put it past them at all.

  9. Are bad batteries the reason I see so many Dyson refurbs for sale all over the internet? I wouldn’t know where to buy a Dyson product that wasn’t a refurb!

    I’ve been trying to figure out if there are so many refurbs because they have a serious quality control problem or if it’s a marketing gimmick that lets them sell their stuff with a shorter warranty at a lower price point to the hoi polloi while the full warranty, full price units sell to the ruling class.

    Is it possible that the refurbs get the crappy batteries that don’t balance the charge and the full priced units get a properly designed/built battery pack?

    1. There is another route to getting a Dyson; their employees can build their own ones to take home, we ended up with a cordless one because my housemate dated a guy who worked there ;)

    2. Interestingly, I bought this my V7 vacuum as an official refurb, where in the battery went in to lock out after maybe six months, starting this whole project. Of course the warranty is much shorter and had expired. It’d be interesting if the cell quality in the refurbs wasn’t as good somehow, but I highly doubt they are using different BMS boards for refurbs vs new models. I have a total of 8 battery packs, only one of which was my original battery on the refurb, and none have balancing resistors.

      1. The battery pack was a 3s1p 11.1v pack. The total voltage was a little more than 2 volts, indicating that something was definitely wrong. I measured each cell’s voltage individually, and two of them were around 3.7 volts, but one was at -5.2v. Adding up the voltages 3.7+3.7-5.2 gave me around the same voltage (about 2 volts) that I measured on the whole battery pack.

        1. This happens when the discharge current is not cut off in time and you keep on discharging an “empty” battery (Which should have about 20% capacity left). At some point the weakest cell is 100% empty and when you keep discharging after that, the other cells push the weakest over the polarity swap threshold. When that happens the cell is pretty much kaput.

          1. This pack was inside a water tech pool vacuum. The battery was connected directly to a motor through a toggle switch (which is hard to tell the state of) with no uvlo (or so far as I could tell, it might have been on the bms board), so the switch was probably left on for too long while the low motor resistance discharged the battery without the motor turning, causing that polarity swap.

  10. Hm…if you ignore the imbalance with a different firmware, I expect you have a nice fire starter some month later when you load a deeply discharged cell some month later.


    1. Aan over discharged li-on cell can actually go negative voltage. If this has happened the cell is junk.
      You can try and charge it again. Very slowly (limit the charge current to no more than 20mA). However the cell will never be much use again after going this far below the discharge threshold.

    2. While the firmware ignores the degree of cell imbalance, it will cutout the discharge if any cell is below 3V, and charging will cutout if any cell goes above 4.2V. The PIC and ISL94208 are able to measure each cell in the pack. I believe this should be as safe as anything since all cells stay within their safe operating range, even if one cell is 3.1V and another other is at 4.2V. That latter scenario is probably not going to have any usable capacity, but I think it should be safe.

  11. FU-Dyson-BMS – amazing story. The evidence is so obvious. If I was an EE at Dyson, having witnessed this I would (at-a-minimum) resign. Better-yet I’d become a whistle-blower and out Dyson for sabotaging their own products in the name of profit.

    Or is this simply a case of bad engineering? Nah, you don’t put the pads for the balance resistors on the PCB then just “forget” to populate them – unless you know exactly what you are doing. Read the ISL94208 datasheet, the built-in cell-balance is obvious.[1] Not that it matters today, due to the chip-shortage the ISL94208 BMS chip is in trouble.[2] Dyson will have to find another BMS solution, hopefully this time (for their bottom line) one WITHOUT built-in cell balance!

    Check out the Renesas/Intersil ISL94208EVZ ISL94208 Eval Board schematic, the cell-balance circuitry can’t be missed.[3][4]

    * References:

    1. ISL94208 4 to 6-Cell Li-ion BMS with Cell Balance at Renesas/Intersil

    2. ISL94208 @ Mouser (yeah still available but limited quantity and long lead-times)

    3. ISL94208EVZ ISL94208 EVAL BOARD at Renesas/Intersil

    4. ISL94208EVZ ISL94208 EVAL BOARD $181.31 non-stocked 36-weeks lead @ Mouser

  12. I hate Dyson company. They have such awesome innovation with their products but the execution is just utter garbage. The cheapest plastic you can find, and build everything as cheaply as possible. Once finished, sell for an astronomical price. A winner formula?

    I have two Dyson vacuum cleaners (one mains powered, one battery powered) and their sucking power is unparalleled but otherwise they are the lowest quality pieces of equipment in their category. Even a 100 euro Electrolux manages much better.

    1. Yeah, are Dysons marked “Made in China”? If not, I’ll bet the components are all made there and assembled elsewhere. Considering that likely fact, they are hugely overpriced, especially if they’re as badly made as you say.

      IF a brand has a reputation for much greater quality (durability, usability, etc.), I don’t care where it was made and will pay a higher price than for your average, zero quality control, zero customer service Chinese garbage. When it comes from China, a higher quality product typically involves a Western world company conducting major quality control at their plant in China. I’ve read guides from entrepreneurs who say that that is a must if one wants a high quality product made in China other than something that’s made entirely by automation, like a populated PC board. As a general rule “quality control” and “intellectual property” isn’t their thing and its cultural.

      JOURNAL OF CHINESE ECONOMICS, 2014 Vol. 2. No. 2, pp 73-78
      Call for Copy – The Culture of Counterfeit in China
      by Ling Jiang

      To Steal a Book Is an Elegant Offense: Intellectual Property Law in Chinese Civilization (Studies in East Asian Law, Harvard University) – 1997

  13. Get a Rainbow vacuum and be done with it. It is expensive, but it’s well-built, reliable, and has no consumables since it uses water to filter the air. There’s no ‘vacuum’ smell either. We’ve had ours 20 years now, nothing has ever gone wrong with it, as it should be with a vacuum.

    1. I’m guessing you have a D4C or similar model? The lack of consumables Is only mostly true nowadays. They’ve since added a post motor hepa filter to the machines, at least in the USA anyways.

      I’ve seen a couple rainbows come across my repair bench with the motors burnt up because they forgot to change the filter, or the existence of the filter, and it over heated from the reduced airflow too many times and die.

      I heartily agree about them being well built, if expensive.

  14. This seems like a problem caused by ‘excessive integration’, ‘too tight coupling’ etc. Why is the BMS built in to the rest of the custom vaccum electronics, when it could be a commodity item?

    Of course, because it’s very slightly cheaper to manufacture fully-integrated tightly-coupled electronics rather than separate modules for each function.

    But look at the actual cost, to the consumer, to the environment. Much much more than the pennies saved in manufacture.

    We need to move back to more modular appliances.

    1. This (kinda)!

      I just started the process off reworking 2-6 old dead 10s 12V NiCd Makita batteries to 3s(2p) or 4s Li-Ion and am appalled that no normal current Li-Ion battery pack (Makita, Bosch, Ryobi and so on) has all balance points connected to the outside (afaik).

      Put the BMS in the charger and the undervolt+overtemp protection/detection in the tool (or battery pack)


      I choose 3s to hopefully get an undervoltage “feeling” (3*3V=9V should be noticeable enough) and the original packs have 4 connectors to the outside. So I’ll repurpose those for +, – & balance and maybe add a connector for a temp sensor separately.
      Next step is rip out the original charger electronics and use the charging socket with an imax-b6 charger or something along those lines.

      Thought about 4-5s + one internal BMS per battery pack but…. dunno. Wasn’t sure enough the ones on eBay can actually deliver enough (peak) power.

  15. You guys are missing the silver lining here. Dyson uses AWESOME 21700 cells, and since the packs ‘die’ with hardly any use, they are a gold mine for batteries.

  16. My Dyson stopped charging a couple years ago, so I bought a third party battery pack from Amazon that has more capacity than the original. It’s been working well so far. And it hasn’t burst into flame, which is also nice.

    However, the day the new battery stops working I’m going to print one of these: It’s a bodge that marries an easily swappable Ryobi One+ 18V battery pack to the vacuum. The Dyson packs are 21V, so the Ryobi delivers less suction power. But I have a lot of Ryobi packs and chargers floating around the house, and I plan to never run out of charge again.

  17. Interesting – this explains something.

    Several years ago my Mum had a Dyson that died – it just wouldn’t charge at all. Given the price tag she took it back to the shop to see what they could do about it.

    The salesperson offered to replace it with a new one on the spot – no hesitation whatsoever.

    I guess they had seen enough of them by then.

    (and that’s why it pays to buy some things from your traditional ‘bricks and mortar’ store :-)

  18. Some interesting takes on Dyson here. I’ve never worked for them, but I have worked for a competitor on cordless product.

    It’s been years since I dug into the cell balance questions. Warranties with cell manufacturers generally specify requirements around cell imbalance, I believe there were safety concerns with significant cell imbalances. My recollection is that including a cell balancing mechanism either adds a significant number of components or drives you to more expensive charge control chips. Not a problem for home projects, but cordless packs capable of running vacuums seriously stress the costs people are willing to pay for a vacuum. We never saw pack failures in the expected life of the products due to cell imbalance, and the expectations were that if a cell did that there were bigger issues with that cell. At the end of the day, not setting a product on fire with customers was the key requriement. No idea on Dyson’s decision making, but assuming it’s malice ignores a lot that happens in corporate product development.

    Details on some of the other comments made:
    -Product from China can run the gammut of high quality to low. Having staff closely tied to the factories clearly helps. But companies went to China for cheap products and are surprised when they get cheap products.
    -Remanufactured products are generally (although not exclusively) are generally store returned products cleaned up by a third party with varying levels of interaction with the OEM. My experience is that they didn’t actually fix internal product, and probably didn’t buy new batteries, but it is possible they do for high margin products.
    -Dyson does not use the cheapest materials, they in fact use some of the most expensive. They do run really, really thin walls because consumers first complaint is weight.
    -Dyson builds product in Malaysia, not China. Last I heard motors still came out of the UK, not sure about batteries.
    -Dyson doesn’t even claim to have invented the cyclone for vacuums, it was an industrial thing long before it came to consumers. He did figure out how to make it a consumer product.

  19. I’ve had 10 dysons and currently have an upright.
    I’ve bought none of them.
    All were thrown away by people as not working and needed minor repairs like unclogging or changing a fuse or the switch.
    Which is more indicative of the throw away society where a £400 vacuum can be retired for not being arsed to clean it – who knew a vacuum cleaner might itself need cleaning?

    But that said I totally buy into the obsolesce by design. Having both worked in manufacturing and product design and witnessed many things made out of materials that they knew wouldn’t last but just happen to sell replacements for – cars being a prime example, VAG group being the worst.

  20. I own Dyson V8 and it’s battery had limited capacity from the start, Dyson is not using “high quality cells” either, just average, probably cheapest they can get. The best solution for me was to buy a Makita power tool battery adapter from Amazon, it comes in many different styles, for different battery brands. Now I can use my 10+ power tool battery stock with the Dyson. They have higher capacity, they last longer, and I can swap them when they run out.

  21. I have what appears to be the same issue with a VAX ONEPWR Battery. I took it apart and found that two of the five Cells were 3.32V the other three cells were 3.35V. I trickle charged all the cells individually to 3.65V. The VAX BMS still says that the Battery is faulty and refuses to Charge it. There was an article about it at;

    I wonder if anyone has started a FU-VAX-BMS project?

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.