The BMS isn't really enhanced, it's a basic feature that can be software-programmed to custom settings by customer request. If this is actually a good idea is debatable: cell-balancing is a neccesity on batterypacks with lithium cells and DELIBERATELY bypassing a safety-feature of the BMS-system inside the pack is something I would advise against: the cells are imbalanced, the BMS acted upon that. Though the thresholds for cell-voltages usually can be set freely by customers' request and they thus can be set very tight (which is probably not really the case here), it still is a bad idea to circumvent this safety. The cells can't be balanced and thus will become ever more imbalanced as you charge and discharge it again, with the real risk that a cell starts to heat up beyond safe limits (because seriously, if they dare to let out basic safety-features on the PCB, who says they did design the rest of the software correctly (read: 'nah... we don't use the overtemperature-protection. By the time the cells could go thát bad, the BMS has already cut off the cells due to imbalance'

Salut, Chipmax2? cela m'intéresse. Ca ne marche pas avec le Pickit? J'ai moi aussi un sv10 à flasher.

@@damienraymond5190 chipmax 2 de chez EEtools, c'est un programmeur que j'avais au travail. Je pense que cela fonctionne avec le pickit mais je n'ai pas essayé.

@@thinhlong2107 Exactly the same steps.

@@thinhlong2107 Same connection from left to right, Only ground pin is in slightly different place but I believe you can figure it out.

can i ask you how to get v10

v10 has more sockets, I don't know which part to plug in

@technomicah407 What steps & settings did you take / change? I've got the same model battery, mplab & pickit 4 but i keep getting low voltage errors. I used the low voltage setting instead of the high 5v setting due to 3.3v reading from the pins. Appreciate some clarity greatly

@perspective6734 I did change the power settings for the pickit4 to provide power to the battery. Pretty sure I just left it as default, which was 3.3v.

@@technomicah What do you mean by taking a dump of the EEPROM? I ended up resetting the settings & ran it through the 5v warning & 3v supply, that worked, cheers for that

@perspective6734 prior to flashing the firmware, there is some diagnostic data stored in the chip describing why it failed and other stuff. Not really important. Glad you flashed it successfully

The 8 red flash blink code means charge overcurrent protection. Others have run in to this issue on certain vacuums it seems like, and there are issues about it on the GitHub page. I think the workaround is to increase the overcurrent protection threshold, and I might have posted a patched version of the software in one of those issues, or maybe in the EEVblog forum thread. I'm not sure why the charger inrush current seems to trigger this on some vacuums and not others. I suspect maybe some minor difference in the PCB since it seems to occur on the V8 vacuums mostly and I don't think I had one of those to inspect closely.

@@tinfever Thanks for the answer! It will be great if you fix the firmware))) In the meantime, I changed the original charger to a non-original one and this problem has stopped.

It's possible the original firmware didn't have low temperature protection or maybe their threshold was set lower. The threshold should be roughly 7C. If you wanted to, you could remove that protection by commenting out the two lines of code in the safetyChecks function that reference MIN_TEMP_C, I think.

You’re probably correct there was nothing in the original firmware regarding low temperature protection, as it’s not really applicable to a household vacuum. I’ll leave it as it is, installing your firmware was scary enough, messing with the code is beyond my comfort zone as I don’t have any spare batteries to play with. Thanks again.

I'm not sure. I've been getting some spam comments that I've been blocking but yours wasn't one of those, unless I accidentally clicked on the wrong one. I apologize if I did that. The FU stands for Firmware Upgrade ;) Thanks for the kind words!

Sorry for the delay. The PCB doesn't need to be connected to the batteries to flash the firmware since it is solely flashing the PIC microcontroller, although you would need to configure the PICKit to provide 3.3V power to the PIC if the batteries aren't connected. Of course, the firmware will not operate without cells connected and I've honestly never tested how it would behave if you tried to use it like that.

That might work but it could be tricky to scrape the conformal coating off the holes if your only access is through some small drilled holes. I think some uses have reported just jamming the wires through the coating works. Using a Dremel to cut a rectangular slot might work better? I've never tried it though. You could start with holes and then move to Dremel if needed? Also, someone made a 3d printed case opening jig that might help (again never tried it though). it's linked in the resources section at the bottom of the GitHub page.

I can't recall which component is the limiting factor, but I think there was some reason you wouldn't want to go more than 1A. I think there might have been a diode that the charging current passes through or something?

If you don't have balancing circuits available, be sure to stay away from higher charging current at all, as imbalance will get out of hand far quicker when the BMS can't do anything about it. Most likely, 500 mA/780 mA is the maximum allowed standard charge current for these cells: higher charge rates either require more convective cooling or even forced-air cooling. To add to that: why would you want to have an even shorter charge-cycle? I expect the battery to charge fully in 3 to 6 hours (3 being the most likely). Lithium cells that spend a lot of time at their full-charge voltage of 4,2 Volt degrade quickly and will limit the overall lifespan of a pack. And exactly that is probably the reason why they fail so much: most people don't know you are better off charging a lithium pack to 3,7 Volt/cell if you want to let it sit there for long times, but given it's a vacuum cleaner, most people probably keep it on the charger forever, quickly degrading the cells in about 2 years. There are few manufacturers (DeWalt is one I know and a few names in the pro-AV-industry are doing so as well) that have a 'storage-charge' function on their chargers, to charge the battery to 3,6-3,7 Volt/cell so you can store it for months without the capacity being affected noticeably, compared to stock a pack at full state-of-charge at 4,2 Volt/cell.

Thank you for this great reply and a reminder :) @@weeardguy

Technically that should be possible since I think the PICKit is using the low voltage programming mode in this case. I've never tried it or done much research, but you'd have to see if someone has done the work making the software to have a raspberry pi work as a PIC programmer. This example came up in my very brief searching: github.com/WallaceIT/picberry Not sure if the that will support the correct programming protocol for this specific PIC model or not though. I suggest only connecting the ground, ICSPCLK, ICSPDAT, and VPP pins when attempting programming, leave VCC not connected. If you can't find software off the shelf to do it, you might be best off buying $10 PICKit clone from AliExpress. You could also maybe search for Arduino to pic programmer options but most Arduinos use 5V logic levels and this PIC is 3.3V

@@tinfever thank you. Ali express knock off might be the simpler approach. I’ll give that GitHub a read. Thanks again for sharing.

Someone on the EEVblog forum thread, or on a GitHub issue (I can't recall) reported that they were able to use a PicKit 2 with some third party program that extends the PicKit 2's part support to include the PIC16LF1847 on these BMS boards. This isn't supported using the normal Microchip provided programmer software. I've only ever used a PicKit 3 or 4 though.

@@tinfever great, thanks for the quick reply!

@@SebastianHagemeister did you find it? Trying to fix for cheap and the kit3 isnt cost effective.. You had luck with pickit 2?

I had to use PicKit Minus third party software to program using a PicKit 3. I believe this will work with the PicKit 2 also.

I'm glad it helped! I'll have to check out your tool the next time I'm reversing engineering something!

Of course it won't work. You describe how your battery cells are stuffed, and no firmware in the world can fix that problem. BTW: If you battery will charge and operate (even for a few seconds), then it is NOT a "bricked battery".

Hi. I reverse engineered the PCB with just a multimeter and a lot of probing, combined with drawing traces and connections on images of the PCB loaded into GIMP. If you look on the GitHub page, you can find what I call "spaghetti wiring diagrams" which are the GIMP images I traced the logical connections over. It is hard because it is a four layer PCB so you can't see everything. I started by loading the PCB images into GIMP and aligning/mirroring the top and bottom PCB images so I could switch layer visibility and see where tracks jumped from top to bottom layers. I'd draw in the visible traces, look at the data sheets of the ICs, and start probing where I thought the connections might go, or just probe every connection that wasn't already designated. It took a long time. It gets easier though because the more you figure out, the fewer options remain for everything else. Remember that most designs are just reimplementations of the IC reference designs, so those are a good place to start. Also, make sure you are using sharp probes and enough pressure, since the conformal coating makes probing even harder. Good luck!

I don't think that lithium ion batteries can really be recovered once worn. I'm guessing the battery cells have a higher ESR due to normal wear. The max setting pulls much more current than the normal setting, which can cause significant voltage drop on the cells when loaded, which means the low voltage cutoff is hit even sooner. If normal mode isn't sufficient, you'll probably want to replace the whole pack or just the battery cells.

Unfortunately not, if you attempt balance charging of any sort without going through the PCB’s intended charging while it is on, it will brick itself with 32 red flashes. I believe discharging to balance does work however. If you have the time, I highly recommend reading the original thread in which tinfever and more worked towards a solution towards all this, as it documents many of the pitfalls and nuances they encountered through the process: www.eevblog.com/forum/reviews/dyson-v7-trigger-cordless-vacuum-teardown-of-battery-pack/

You are right in that if the firmware detects charging from an external source or a cell imbalance of more than about 300mV, it will brick itself with an EEPROM bit which will require the PICkit3 programmer to erase (see my earlier comment about dvd4me’s unbricking video).

I haven't done a ton of testing on how the original firmware behaves, but here are a few thoughts. The BMS lockout is handled by a setting in the EEPROM. dvd4me in the thread linked by Wendell found the EEPROM bits you can change to unlock a battery pack. The PIC microcontroller also turns off unless the battery is on the charger or the output is enabled. (You can measure that the 3.3V rail goes low since the ISL94208 turns it off completely). I think if you rebalanced the cells using an external balancer, while the PIC was turned off, it would probably work fine. You may or may not have to reset the EEPROM bits mentioned since I'm not 100% sure if it will go in to lockout mode if it detects the cell voltages have changed too much while the PIC wasn't turned on. tldr; I think external balancing with a possibley necessary EEPROM bit change would probably work, but I haven't tested it and can't guarantee it.

V10 vacuums use a different microcontroller and BMS chip entirely, so this project isn't compatible I'm afraid.

If the stock firmware detects the cells are out of balance even once, it will stop working permanently even if you rebalance the cells. However, It is possible to edit a value on the microcontroller's EEPROM to unlock a pack that is locked for this specific reason. More information on this specific EEPROM edit can be found in this forum thread, you have to be very careful when programming the EEPROM since it is very easy to accidentally erase the entire microcontroller program memory: www.eevblog.com/forum/reviews/dyson-v7-trigger-cordless-vacuum-teardown-of-battery-pack/75/ Flashing my firmware also fixes that lock.

The PicKit 3 is the programmer needed for the microcontroller on board, which is a PIC device. This repair does not involve replacing the microcontroller but reprogramming the one that is already there, so will definitely need a PIC specific programmer. If you have access to eBay you can get one for about $40, but not sure it is worth it if you are only going to use it just for this repair.

Is your battery running my custom firmware? I don't understand your other questions I'm afraid.

What a joke. 😢

I think that might just have been the herald of doom... true doom is when your car needs it.

Planned obsolence m8. 😑😕😔

All domestic devices with lithium cells need firmware of some form to keep the consumers safe. It's not "planned obsolescence, it's because there is high danger and liability if devices liek Dysons start catching on fire.

My fridge talks to the supermarket

It sounds like there might be a hardware failure. I would suggest seeing if 3.3V is present on the VCC pin when you press the button or plug it in. You may have to use the schematics to find the fault.

Who? It is YOU to resolve it, only if you want to resolve it. If you cannot understand, then buy a new battery pack.

Hi. Someone on the EEVblog forum thread, or on a GitHub issue (I can't recall) reported that they were able to use a PicKit 2 with some third party program that extends the PicKit 2's part support to include the PIC16LF1847 on these BMS boards. This isn't supported using the normal Microchip provided programmer software. I've only ever used a PicKit 3 or 4 though. If using a different programming tool, it would be good to double check the config bits are set correctly. I include those near the end of this GitHub issue if you need them: github.com/tinfever/FU-Dyson-BMS/issues/9.

@@tinfever Thank you for reply!

Hi. If you are running my firmware an encounter an error like that, could you please open a GitHub issue? I'll try to help there. Please provide as much info there as you can such as the original issue with your battery pack, if you are using the original charger, how frequently that error occurs, etc.

That's unusual. Error code 16 is for the brown out detection, which usually occurs when the battery is short circuited. Could you open an issue on the GitHub page? The first thing I'd suggest doing is using a multimeter to measure the resistance between the terminals on the vacuum (not the battery). It's possible something failed short-circuit in the motor causing the battery to trip that error.

@@tinfever I opened a topic on your GitHub page, I would be glad if you reply 🙏

It's extremely unlikely this would work with anything other than Dyson V6, V7, or V8 batteries.

Unfortunately not. The newer models like the v11 use a different microcontroller and BMS chip so it'd require almost starting from scratch.

FYI: The board is marked 228499-01/03. Also, there is a red "All Protect" message under the "Configuration" line in the upper right corner in the Pickit 3 Programmer window.

OK. I got it resolved. Broke my head searching for a problem. Finally, measured all batteries again and found that one of them was low voltage - 3,6 v. Probably I missed it when I was charging batteries separatelly. After charging this battery everything works fine. Thank you again!

Yes. A Pickit 3 or newer is required. You can find clones on eBay, AliExpress, Amazon. Genuine PicKits can be bought directly from Microchip or from places like Digikey or Mouser.

​@@tinfever Thanks for infos😉, -> Genuine Pickit 3 = deprecated, Pickit 5 = ~100$, ~40$ Amazon are clone too ? In this case ~15$ Aliexpress , if it works ?... I disassembled Dyson V7 batterie to replace 18650 but I think the Genuine BMS is now bricked. I wonder the better solution ? : buy a Pickit 3 clone or a BMS clone... I'll investigate

@@Redscreen-qk5pf The Pickit 3 I used was a $15 clone so that should be fine.

Someone else built something for one of the newer models and it might have been the V10. There is a link near the end of the EEVblog thread for my version.

If I remember right, the pickit2 doesn't support the PIC16LF1847. It might support a similar microcontroller part that has the same programming protocol but I can't remember for sure.

Where did you get the bms for only $4?

@@therogman7485 aliexpress have tons of bms board. Just need a little mod to fit in.

@@mingj2637 yeah but not for $4 lol it's over S15 to Canada

You had success replacing the board? I also tought about it

Translation of original comment: "Balancing is useless, too high load currents. In the event of an overload (turning on the turbo mode), weak cells will instantly fly away again, so Dyson refused to balance. This was not done out of bad intentions, it's just a really useless option, plus the lack of cooling. I just don't understand why it was posted for free? I made a decision on these batteries two years ago (I'm sure you know this) and later added diagnostics without disassembling the battery via a computer or a standalone device: kzfaq.info/get/bejne/oJ95rZeasd-pfIU.html" That's really interesting that you already developed something almost identical! I had no idea and I thought I'd searched around. Crazy how it feels like there are separates worlds out there. That's an interesting viewpoint on balancing that I hadn't heard before. Wouldn't balancing while charging keep a weak cell from drifting lower in voltage compared to the rest, and thus prolong the usable life of the battery? I do see your point though that a very poor health cell will have higher ESR so when you draw 17 amps in turbo mode, that ESR will quickly cause the apparent cell voltage to plummet and trigger the low voltage cutoff. As an example, lets say you have a pack with 5 perfect cells and 1 poor cell, and the healthy cells are at 4.2V with 30 mOhm ESR, but the poor cell is at 3.6V with 200 mOhm ESR. As soon as you start drawing 3A (the normal mode current draw) the healthy cells might appear as 4.11V (30mOhm x 3A = 90mV drop) but the poor cell would appear as 3.0V (200mOhm x 3A = 600mV drop), and so the 3V low voltage cutoff would kick in. Without balancing, you would be screwed because the good cells are already at 4.2V so you can't charge the battery, but the poor cell instantly hits 3V so you can't discharge it either. However, if you had balancing, couldn't you slowly bring that 3.6V cell up to 4.2V? The ESR would be the same but now with a 3A current draw, you'd have an apparent cell voltage of 4.2V-600mV drop = 3.6V, which would leave you some usable capacity to actually use the battery to vacuum. Your point regarding turbo mode remains though, because no matter how well balanced, even a 4.2V 100mOhm ESR cell at 17A is going to drop 1.7V to 2.5V, assuming it can supply 17A to start with. As for why I released it for free, as much as I'd love to make money from all the work I put in to this, I didn't see a good way to commercialize it. I think there aren't going to be a ton of people willing to open their battery, buy a PICkit3, and program it themselves (even if I wish more people would consider it) so I doubt I could sell more than a few copies of the firmware if I tried to. That leaves using it to repair other people's batteries, but shipping Li-Ion batteries is complicated to do legally and I don't think I'd be able to charge very much before people just decide to buy a knock-off on Amazon. Even if I charged $20 + shipping, shipping both ways would probably easily be another $20. I doubt many people would pay $40 to ship their battery to get it repaired, when you can buy a knock-off on Amazon for the same price. If there would be a high volume, then maybe you could bring some of that repair price down but I can't see ever getting enough batteries to even take an hour a day in time.

All in the spirit of open source and right-to-repair! The more freely available information there is to fix repairable things like this, the better off we all are. Unfortunate that the language barrier does somewhat limit the diffusion of knowledge in engineering fields.

@@jajefan123456789, кому лучше, а кому нет. Хорошие решения требуют много ресурсов, времени и испытаний. Не всё может быть свободным. Почему бы тогда и не работать бесплатно? Бесплатный инструмент для мастера, бесплатная работа для клиента, почему нет? :) Но нет, за работу мы хотим получать гонорар. Я делаю хорошие вещи и я не могу позволить это делать бесплатно.. Увы, как разработчик, я не работаю бесплатно.

Dyson are protecting their customers by ensuring that battery packs cannot be used or re-used if cells have reached certain critical safety limits. They couldn't be doing anything even slightly better than what is being done here. When the pack faults, the BMS trips and you replace the pack. That's EXACTLY how all responsibly-manufactured domestic products behave.

You are correct that if the cell voltage drops under load enough to reach the cutoff voltage, the firmware will fault again. If this happens, it would indicate a failing battery cell. The cell voltage of each cell must be within the safe range for the firmware to allow the battery to turn on, the cells don't all have to be the same voltage though.

Sounds like somebody tried to replace the cells with Laptop-pulled cells.

It sounds like you might be running in to the over-discharge protection, meaning one of the cells is hitting 3V or the BMS thinks one of the cells is hitting 3V. Could you try connecting the battery to the charger for any amount of time to reset the low battery flag, and then test if the issue occurs when the battery is not installed in a vacuum? If one of the cells has a very high ESR, it's possible that the cell will measure an acceptable voltage with no load, but as soon as you put a reasonable load on the cell, the voltage will drop dramatically. If you are encountering the over-discharge protection, as soon as that is tripped, the battery will always display the 3x blue LED flashes when you attempt to pull the trigger. This is reset by connecting the charger for any amount of time. The next step after that would probably be one of two options: 1) Connect an 3 amp load across the battery pack directly (not going through the BMS) and verify that all battery cells remain in the acceptable 3.0-4.2V range. This would eliminate the possibility that one of the cells has high ESR causing the BMS to immediately trip the low battery flag. 2) You could also install MPLABX and connect the PICKit to the BMS, then compile and run the program in debug mode. You can load the "MPLab Watches List.xwatch" file located in the /firmware folder in the GitHub repo in to MPLABX as the watches list in the watches window, and then pause the program while running in debug mode so you can view the actual cell voltages being measured by the PIC. You could even go further and insert a breakpoint at the entry in to the error function code so you can see all the variables and determine why it is entering the error function. You'll probably need to solder on a programming header to the BMS or find a way to make those connections relatively stable while you are working on it. While it would be possible to add a line of code that forces any given battery cell to always read as 4V so the program thinks it is always OK, I would strongly advise against doing that, especially without knowing the root cause of the issue in the first place. The reason I believe this firmware is reasonably safe is that I've implemented pretty much every protection possible. Hardcoding a cell voltage level would break some of those protections. If you have further issues, please make a GitHub repo issue and I'll try to help as soon as I can. Thanks.

@@tinfever Thank you for your reply 1 was fine Even if the load was connected, it did not drop below 3V. Even if you press the switch with the battery pack alone, it blinks blue 3 times Now I have installed MPLAB X IDE and loaded the farm However, the debug mode does not progress easily while learning how to do it with Google search.

Can i program it usign a different programmer instead of pickit 3?

Did you know what 4 times flashing means? I cant find anything related to it on the internet. I was disassembled and checked the cells on my battery one of the cells was around 3.08v and the other 5 was around 3.5 ish i was then recharged that 3.08v cell back to the 3.5v but it still gives same 4 time flashing weirdly when i hold down the switch for 5 seconds it gives 4 red and 1 blue but no matter what i tried i did'nt managed to get it to charge.