[010] Why do M18 batteries charge to 3 bars?

[@WaynesWorld999]

I've always questioned weather they are balancing properly, or at all. This really could be called industry malpractice, or at the very least, anti consumer.

[@toolscientist]

Yeah, it seems Dewalt and now Milwaukee are happy as long as the battery makes it past the warranty period. I wonder how many more brands I'll find that don't balance.

[@discojesus00]

@@toolscientistare you going to be testing other brands or maybe m12 as well?

[@toolscientist]

Yes, I want to test them all (within reason - I'm poor, and a lot of the cheaper US and EU brands aren't available in AU). I have 5 or 6 M12 tools and 2 batteries from my own collection, so I'll do them soon.

[@PowerScissor]

Balancing old M18 packs is such a pain. I don't even bother putting all 4 screws back in my battery cases I have to open them so often. So frustrating that Milwaukee refuses to make a charger /pack combo that actually balances the cells properly. It literally takes me an extra 15-30 minutes to do it manually...and it would be so simple & convenient to just leave the battery on the charger for an extra 15 minutes if it would just do it.

[@slinkyfoxx]

How do you do it? I got a 8amp/h on redemption about 8 months ago and it only ever charged to 3 bars from new.

[@PowerScissor]

​@@slinkyfoxxJust take case off, measure cells with DMM, then alligator clip the low cells to a bench power supply until they are all even.

[@toolscientist]

Doesn't even need a new charger. All it needs is a software update in the batteries. The 1 kOhm resistors makes me think they had problems. The Texas Instruments reference design uses 47 Ohm resistors.

[@SirSpence99]

@@PowerScissor What settings do you use for your supply and for how long? Also, do you use anything to stop the charging or do you just check voltage occasionally?

[@gf2e]

If you’re doing this often, I’d suggest getting a balance charger, perhaps like one from iSDT. You can add a 6 pin balance cable to all the cells. Then you can balance without disassembling it. I can’t find the charger for my black and decker battery, so I just use a balance charger.

[@Кто-тоТам-ъ5н]

I am watching you from Russia and I think that you are doing a decent job for all owners of the M18 tool (even though I am not one of them) but I am still grateful to you for this, thank you!

[@arva1kes]

Someone was very horn..I mean excited , then video was made.

[@emmettturner9452]

It would be amazing if someone designed an open-source balancer for 20V Max (18V Max) and M18. I’ve been doing it manually for 20V Max batteries for years now using alligator clips to the balance leads or back-probing pins to the terminals. With a dozen DCB107 chargers around I’ve always been tempted to take one apart just to build a jig using the terminals. Taking one apart to replace the main PCB with an open-source balancing board would be even better. :)

[@toolscientist]

Lot of effort to make an open-source PCB. I'd rather just mock Milwaukee until they fix their damn firmware themselves :-) Balancing jig wouldn't be too hard. I've seen people make 3D printed jigs for rebalancing Flexvolt packs, and I've seen someone add JST connectors to the M18 packs so they could balance without taking it apart.

[@saiiiiiii1]

​@@toolscientistI applaud you for this, I just wish you platform was 1000x bigger😢

[@donalexey]

@@toolscientist I had to create a JIG for balancing Bosch batteries. But you need to disassemble it every time.

[@gf2e]

I drilled a small hole in my black and decker battery to run a 6 pin balance cable. Now I can use my iSDT balance charger with it. Some packs such as M12 have a connector with every cell already connected. But, sadly, that doesn’t work with balance chargers because the battery pack usually has a resistor in-line. You will probably always need to add your own cable.

[@toolscientist]

@@gf2e yeah, same problem with dewalt - resistors on the pins. I wonder if there's a charger out there that's really smart and can be told to ignore resistance, or really dumb and just ignores resistance anyway. Prob has to be smart, the resistors create voltage drops that would need to be accounted for when measuring the cell voltages during balance. Or you've got to turn off balancing, let the cell settle, measure, then resume balancing.

[@DarkFox6211818]

Wow. As a Milwaukee tool owner that's incredibly disappointing.

[@riba2233]

Haha, there will be some disappointed folks on Reddit 😅 Great stuff and thanks as usual!

[@toolscientist]

Looks like I'll have to do Makita now. That's twice I've been 95% sure that a battery balance charges and twice that I've been wrong!

[@riba2233]

@@toolscientist yeah lol, you def should!

[@saiiiiiii1]

Pretty sure Makita does. I've taken apart an 8 year old battery this week and cells were within 0.02v. 😮

[@KeijonAutoVuokra]

​​@@toolscientistif makita doesn't balance charge then their manuals (and marketing ig) are lying. Iirc they specifically mention that batteries get balanced when left on chargers after they are fully chsrged

[@toolscientist]

@@KeijonAutoVuokra thanks, I'll have to look that up. I like when they make specific claims (rather than vague marketing claims), easier to test them that way. My feeling is that Makita got burnt by the poor design of their first LXT packs and went overboard after that.

[@ThriftyToolShed]

Excellent video! Great information. It seems the slow chargers may benefit in this regard as well? Longer time balancing it seems. Thanks for sharing!

[@toolscientist]

There's no balancing at all, so speed/length of charge makes no difference. I guess it's possible that the slow charger sends a command to the battery to tell it to balance, but that seems unlikely that the slow charger would do that, but the rapid wouldn't. Ultimately it's the battery that decides to balance. The easiest time to balance would be after charging has finished.

[@toolscientist]

One of the giveaways that balancing never occurs is the structure of the i2c conversations. If a balancing transistor was turned on, you couldn't measure the cell voltage due to the voltage drop across the resistors. So you'd think they would turn off balancing before measuring cell voltages, just to be sure. But the only time they touch the BAL_CTL register is after they've measured the cells.

[@ThriftyToolShed]

​@toolscientist Wow! I was thinking it didat least during the constant current charge portion. Good grief...

[@no-damn-alias]

They probably have their extra service charger where they put in warrantied batteries that actually do only charge to three bars inside the warranty frame, which balances the cells and then can return it to the customer as fixed. You should sell your solution to unhappy customers as a an add on that you put between the charger and the battery for charging which activates the balancers. When the information spreads you should be able to sell a lot.

[@tonytango6676]

I suspect that they put a first year university coder to design this code and they did not understand the big picture. That was a number of years back and no one has touched the code since. And who knows, maybe they lost the source code.

[@toolscientist]

That has crossed my mind. It was developed back in ~2008, then the Redlink Plus update (digital comms with charger) was in ~2013. They definitely made an update when the 6/9Ah came out as they changed from 2.0V to 2.5V low voltage cutoff. Then there was at least another update around 2021/2022 to add the LED diagnostic codes back in. So they seem to be able to make updates.

[@gf2e]

That seems plausible. I’m also wondering if there’s some manager who says “it’s working fine, don’t change a thing, you might break stuff!!!”

[@richardkaz2336]

A bit like what Boeing did for the 737 Max software.

[@smeegle]

As pessimistic as i sound, i would not be surprised if the code is present but disabled, as death by imbalance tends to take quite a while, its kinda like a stealth planned obsolescence to sell more batts, dyson batteries and i also suspect ryobi 36v batteries do the same

[@toolscientist]

I rarely buy the planned obsolescence argument as it only takes 1 disgruntled employee to leak it and then your reputation is trashed. Consumer electronics can get away with it, but power tool brands would get savaged for doing such a thing. But yeah, I'm struggling to think of what other reason there is. Only thing I can think of is that they couldn't get it to work reliably. But then why stick with the same balancing chip for 15 years?

[@KeijonAutoVuokra]

The whole communication sequence seems like a project that was given to an intern and never reviewed. This is almost maliciously bad lol

[@toolscientist]

I'm guessing a lot of small embedded projects are just hacked at until they work sufficiently. Whoever wrote this probably left the company years ago, and no one bothers to change it for fear of breaking it. It can't be that complicated - I think an experienced embedded engineer could re-write the whole thing in a week or two (prob only a few days, followed by lots of testing)

[@Nitrxgen]

@@toolscientist The amount of code programmed into those things can't even be that much though. Considering there's no balancing, the only useful thing any of the software seems to do is flash codes and half of that is useless. The foundation for a lot of good features is there but it's not being used properly. I reckon if they give me access to their code, I bet I can cut out all the bs, all the lazy copy paste, all the errors, and actually make it do what it's supposed to do and probably in a fraction of the size. Maybe it's worth sacrificing a battery or two and overwrite the memory on those chips. I'm a heavy Milwaukee user and I thought to myself "heh, dewalt don't even balance their batteries?" but this one got me sad.

[@toolscientist]

@Nitrxgen they've also got the charger comms protocol, and I've heard there's a hidden diagnostic protocol that can pull more data. It's not going to be massive, but I'm guessing ~3k-6k lines of code. The FU-Dyson-BMS project looks around 2-3k lines (main is ~900, i2c ~750, then several 100 line files). But yeah, if someone leaked the code we could add balancing in a day. In a month we could clean it up a fair bit, add dynamic current-based low voltage cutoff, and get rid of that low-voltage bounce as well

[@antman3351]

there's probably a comment "Todo uncomment for prod" in the code above the commented out logic for balancing 😂

[@toolscientist]

Or maybe a '#define BAL_ENABLE 0' hidden in a header file that no one has looked at in 15 years 😄

[@TranTek]

Amazing how much time and dedication you put in doing this, Thank You i read many of these different year version of M18 12Ah due to imbalance, i found they rely on Samsung 40T poorly made, same cells used in DeWalt Flexvolt 12Ah, Bosch uses version 2, same issue. Milwaukee for short period of time switched to using Murata VTC6a but most of them have bad board which won’t charge. then they switch back to Samsung 40T version 3 which i have yet to see any bad one. Again no balancing, just rely on the cells itself.

[@toolscientist]

Lets hope these new Ampace tabless cells are more reliable

[@TranTek]

@@toolscientist i have them but sold them very quick and based on feedback from customer that they love it on the other hand, some of 8Ah and 12Ah board would kill and drain some bank to 0v, i caught one 8Ah that quite warm to around 35C in the area of the lights meter and drain the 2nd bank from the terminal to 1.4v so i quickly remove both ends of the main terminal i am kind of shock these multi million dollars tool companies don’t really test out what they design ?

[@queazocotal]

The BQ battery balance chip, for those wondering, supports 50mA current through the FETs. The BQ chip does not control the FETs automatically in any mode, it relies on the MSP430 (which is quite capable of running balancing code). 50mA would deal with 25mAh of capacity imbalance over one charge of half an hour period. Idling for a day could correct one cell bank by 1200mAh. (adjacent banks can't be balanced simultaneously). The microcontroller has 32K flash, and 1K RAM. Which is comedically overspecced for what it's doing here. There is no question that they ran out of space. In principle, the battery firmware may be 100% capable of balancing if instructed or configured to by the charger, and the charger just doesn't.

[@toolscientist]

They might have large code size. The 12Ah is actually using an RL78 (R5F100BDA) with 48k ROM, when the RL78 comes in 32, 24, 16, et cetera variants. So either they've written the world's most overcomplicated BMS (that doesn't balance) or for some reason the 48k chip is easier to source. But I agree that there's no way balancing was left out due to space constraints. My 31 lines and several variables can be cut down to about 8 lines and 3 variables (lowest cell voltage, highest cell voltage, highest cell index) if you only balance the highest cell. And in any case it's a trivial amount in a 32/48k system.

[@toolscientist]

Another commenter just posted their new Forge 8Ah battery and it's using the same Renesas chip, but "BEA", which has 64k flash. They must be mining crypto in there! Now I see the scam: 1. Distribute batteries that mine crypto using other people's power. 2. Don't balance so the batteries fail early 3. Harvest crypto from returned batteries 4. Profit!

[@gf2e]

Hilti battery chargers have cellular radios built in. They could totally be mining crypto.

[@der_pinguin44]

This video was ho░░ This video was great, thank you.

[@TCPUDPATM]

What a fantastic video! I’m sure the battery thoroughly enjoyed herself. Wait. *static sound*

[@toolscientist]

Who doesn't like a good probing...

[@tamask001]

Lidl made a big deal out of their new batteries a couple of years back, as they added balancing and bluetooth. At the time I was a bit shocked, because I assumed all 18V batteries already had balancing. What a fool I was.

[@jay3412]

I’d love to know which tool brands actually balance their batteries so I can switch to them. Ryobi, Ego, Stihl, Makita, Metabo? I hope you’re able to keep test other brands. This is good stuff to know.

[@toolscientist]

I've only tested Dewalt and Milwaukee so far. Unfortunately both of them lack balancing. I'll have to go back and look at Makita

[@no-damn-alias]

The only battery took I own is a leafblower. Rest is corded. Cheap cheap Amazon Teppco leafblower. Went for the 28V max super bla bla battery as I felt like it. Paid 75€ for the whole kit in 2018 and it still works fine. Balances the cells perfectly has 4Ah 28V and has over 1000 cycles on it. During fall more obviously but also in summer. We run an inspection bay and when oldtimer cars come in it helps to blow out the smell and once a day to clean the parking lot so fair to say it at least does half a cycle a day although I often did 2-4 cycles a day in fall. Only issue is you can feel the voltage drop now when you push the indicator and turn on the leafblower. Also the battery gets handwarm after a full cycle which means the cells are way warmer but they heat up so uniformely that I have a constant velocity for a long time. When cheap chinesium can do it, why can't reputable brand? Then they wonder why people buy knock-offs

[@dkhuber1]

For what’s it’s worth my 8.0ah battery has this issue. I took battery apart and tried charging cells individually. When I did this I was able to witness the battery balancing itself. But that is impractical. Next I plugged a Venom charger into the positive and negative ports on battery at 0.5 amps. The charger monitors battery voltage and adjusts charge accordingly. Since battery was receiving charge and not able to communicate with charger does that mean it was forcing battery to balance?

[@Miginc2]

Well that explains why i have to manually balance my 12AH every few months

[@Dani-js2ge]

Would be very interesting to see what Bosch Professional batteries do when charging or in use. Bosch representatives claim that their tools communicate with the battery.

[@toolscientist]

Bosch is high on my todo list. I'll need to hunt for a 2nd hand battery, tool, and charger, but they seem common enough. I'm sure they "communicate" in some way, but that could be similar to LXT/M18 where it's just a simple stop/go signal.

[@donalexey]

@@toolscientist They definitely not actively balanced (at least the core ones) although claimed by Bosch representatives that they do. i had to create a jig and balance them with external charger.

[@toolscientist]

@@donalexey damn. The Hall Of Shame might end up being quite large

[@x2thel]

Anything we can solder to the board to override this?

[@toolscientist]

You could try to build a mini board that taps into the i2c pins and VCOUT. It could then listen to the comms and stealthily inject balancing commands inbetween the normal communication. You could only balance for around 400ms at a time as you need to turn off balancing before the main chip tries to read the voltages. A better solution would be to write open source firmware. Someone did this for Dyson batteries, but I don't have the time for such a feat.

[@stevebabiak6997]

@@toolscientist - if the sample circuit from the data sheet shows 47 ohm resistors, why wouldn’t putting a resistor in parallel with the 1Kohm resistors give similar behavior as in the data sheet sample circuit?

[@toolscientist]

@@stevebabiak6997 it would, but the problem is software. I realise now I didn't explain it very well in the video, but the AFE can't balance on its own, it needs the MCU to tell it which cell to balance. As the MCU never tells it to balance, it will never balance, no matter what resistors are there.

[@stevebabiak6997]

@@toolscientist - have you tried to use a shunt resistor, just in case it might have an effect? I don’t own any M18 products so it’s not bothering me, but just curious because when things seem this close …

[@wrefk]

I believe Dyson batteries are the same way, have all the circuitry to balance but dont. Infact, supposedly if the cells in the dyson get unbalanced beyond 0.1v the bms bricks completely, and cant be recovered even if manually rebalanced

[@toolscientist]

Yeah, and some madlad made their own open source firmware to fix it. Don't think I have the time for such an endeavour, though

[@karlosss1868]

@@toolscientist By fix you mean remove the "Brick" feature then yes. But they still dont balance.

[@jamay1878]

Great video, I've replaced a couple of the BMS' on the milwaukee packs with ones from Aliexpress and I wonder if they do any balancing as they look like they have some chips for it.

[@toolscientist]

That would be hilarious if the aliexpress boards balanced and the genuine ones didn't. I might have to buy one.

[@jamay1878]

@@toolscientist They seem to have a few variants, some a lot more packed than others. The problem I had with one of them is that the nickel strips that come installed on the bms are too small for the current required and ran hot.

[@ExodusC]

This is immediately what I wondered. Short of someone coding a custom firmware for the factory BMS, if there was a Chinesium BMS that actually did cell balancing, I'd consider buying a bunch to swap in as soon as my batteries are out of the warranty period. Maybe a good idea for a video? And as you and others mentioned, I would be curious about the new Forge batteries, if there's a new BMS there that actually balances or not.

[@toolscientist]

@@ExodusC I do have a battery where I damaged the PCB during testing. I could do an aliexpress swap and test.

[@ExodusC]

@@toolscientist I really hope the tool outlets pick up on this video. The greater Milwaukee community understands the 6Ah don't die like the 8Ah and 12Ah tend to, and this is presumably one of the big reasons why. I'd love to see Milwaukee get shamed by the entire tool community into fixing this. It's 2024 and at the prices they charge for these batteries, the dang things should have cell balancing and a Bluetooth module on the BMS that can report battery health to a Milwaukee app or something. It's ridiculous they're using what's probably ancient code written by an intern for these BMS.

[@owenimholte3835]

Soooo how close are we (you) to writing new firmware and flashing the MSP430 to have proper firmware?

[@toolscientist]

It would be a big project (unless the source code leaked!), so I don't think I'll try. It would also only be useful to a small number of enthusiasts with the technical knowledge and equipment to flash ROMs. It would be a great middle finger to Milwaukee, just like the guy who did the FU-Dyson-BMS project. But my todo list gets longer every week, whilst I'm still only finishing a project every ~2 months. I've got a few other projects that I've been working on that I think will be more useful for the average person, and also fairly achievable for anyone with a soldering iron.

[@KaldekBoch]

Very curious what AEG (Ridgid) does now since my garage is all AEG.

[@toolscientist]

AEG should be interesting. Pretty sure they have standalone protection with cutoff in the battery. I've seen a few of their tools with 2 pins and some with more, so they've got something extra going on. Are you in Sydney, AU?

[@dimammx]

​@@toolscientistI'm very interested to know if it balances or not. I can confirm that the battery does have a low-voltage cut-off. The battery itself looks to cut off around 13.5-14 volts. Newer tools have redundant cut-offs in the tools themselves. I observed that they don't start with voltages below 12 ish volts. Also, if I recall correctly, there are different thresholds for stopping a working tool vs. starting the tool.

[@gf2e]

@@dimammxNice! I suspected they did have a cutoff because they switched from NiCd to LiIon with the same battery connection. But my only Ridgid battery and charger are NiCd so I haven’t been able to test :)

[@ExodusC]

I have been thinking about this video every other day since it's been posted. I am going to hazard a guess that Australia doesn't have Harbor Freight (historically notable for selling abysmal cheap tools, have been improving quality dramatically over the past many years)- but I thought it would be hilarious if their cheap power tool batteries were doing proper cell balancing. If the absolute bottom of the barrel Harbor Freight Warrior drill kit for $13 USD does proper cell balancing on its packs, I think I'll lose my mind.

[@toolscientist]

Or Aldi... I could definitely test Aldi Ferrex as it's super cheap. I was planning to work through all the pro brands first, though.

[@NoPegs]

Team yellow superiority. 😂 (Oh, wait, we don't balance at all.)

[@toolscientist]

😄 I wonder which brand is going to disappoint me next

[@NoPegs]

@@toolscientist Makita. For sure.

[@donalexey]

@@NoPegs I have been told that some Makita batteries do balance.

[@russcole5685]

Would it be advisable to exchange those resistors ourselves? Also, the extra code you wrote, is that able to be added to code in the program chip. Or is that a factory only thing. Those updates would extend the life of these otherwise great battery's

[@toolscientist]

No point changing the resistors as they're never commanded to balance. There could be a reason for such high resistors, they may have had problems in development You can't add parts of code to a microcontroller, only the whole code. So we'd need to reverse engineer (or reimplement) Milwaukee's whole code. The chips may also be locked after they program them at the factory. Overall I think easiest option is to just build a jig for easy cell connection and manually balance periodically.

[@russcole5685]

@@toolscientist yup, that is my usual go to. I have a smart charger with a Ballance lead made up just for the purpose. Every 20 or chargers I make sure to do a Ballance/maintenance charge. Milwaukee really dropped the ball on that one. Otherwise a great product. I use the same principle as when charging my LiPos for my Model air craft

[@tonytango6676]

I agree that a separate jig or whatever that you would drop the battery in to do the rebalancing. Maybe an LED cycling through the various batteries showing the voltages. Maybe a button switch to cycle between the batteries instead of an automatic switching every few seconds. If you want to get fancy, maybe a Bluetooth connection to an app on your smart phone that displays the various cell voltages. And sell this on Amazon or eBay.

[@toolscientist]

@russcole5685 there must be a good reason for it as it costs them nothing to add a bit of code. It seems their older batteries held up well without balancing, but then the 8 and 12Ah really struggled. Lets hope the new Forge batteries have balancing or at least more reliable cells

[@russcole5685]

@@toolscientist even back thirty years ago, I didn't like the idea of pushing 8.4 volts through a 1.2 volt NiCd battery. I always tried to separate the cells and charge separately. This is the older days of 6 or 7 cell batteries for our RC cars. I'm no way a electrical minded guy, but to me squeezing say 12V through a series of 1.2V cells seems a bit hard on the ones first in line. My thoughts, run in series, charge in parallel. Correct my thinking if I'm being over cautious

[@darthtater]

Good video. Infuriating that Milwaukee doesn’t enable the balance. I’ve got a new Forge 8.0, can I offer any data or pics to the cause?

[@toolscientist]

If you can read the MCU and AFE part numbers, that would be very helpful. You'll need a magnifying glass and a strong light that you can move around to get the right angle. There's sometimes too much coating, though. I've scratched mine off on a few packs, but they were out of warranty anyway. I don't want you to void your warranty.

[@toolscientist]

Another useful thing would be measuring the resistances of R7, R9, R11, R12, and R13 with a multimeter. You need decent probes and to push fairly hard to get through the coating. If you're not confident in doing this, then please don't, I don't want you to damage your new battery.

[@darthtater]

@@toolscientist the conformal coating is super tough on these. I’ve got good probes too. I tried but I felt like I was on the verge of damage.

[@toolscientist]

@@darthtater ok, thanks for trying. I've had good luck with sewing needles, but you'll need to secure them to your probes somehow. Helping hands can be risky. I let the smoke out of a battery using needles and helping hands, that's why I made these spring loaded probes that I briefly showed in the vid. Anyway, don't worry about it too much. I'll prob buy one in the next few months. It looks different enough to justify testing.

[@darthtater]

@@toolscientist bummer looks like KZbin took down my Imgur link. MCU is a R5F100BEA. Analog IC Renesas 489250.

[@kahlid-ataya]

I'm just wondering if parkside batteries and chargers do the same by the way they are way cheaper than Milwaukee and Dewalt

[@toolscientist]

They're not available in Australia, unfortunately, so I can't test them.

[@kahlid-ataya]

@toolscientist in Lebanon they are almost in every hardware store as a technician I do repairs for both the battery and the charger as well as the device itself such as impact drivers drills grinders and so on

[@toolscientist]

@kahlid-ataya yes, I've seen quite a few videos about Parkside from ErCan Everything, who I think is Bulgarian.

[@kahlid-ataya]

@toolscientist i think he is I did some videos about it but in arabic

[@mman454]

I’d be curious to see how Ego outdoor equipment batteries balance. They have a built in “maintenance mode” which actively drains the battery to 30% after the battery has been unused for a month. So they clearly have the capability built into the battery, and use it for the aforementioned purpose. Wonder if they balance or just let the cells do whatever.

[@toolscientist]

I do want to do Ego, but it'll be some time before I get to them. Thrifty Tool Shed has done a lot on them, so you might be able to find an answer on his channel.

[@dedamarsovac]

But can someone code it for lower price?

[@RRrrRRrrlandfin]

Does Makita do any better?

[@toolscientist]

I thought so, but I was also pretty sure Dewalt and Milwaukee balanced, too. Looks like I'll have to check Makita, too.

[@MattsAwesomeStuff]

I know when I buy Dewalt I'm buying cheap. What's Milwaukee's excuse?

[@toolscientist]

Only things I can think of are that they had problems getting it to work reliably and then after a while they found that cell matching was good enough. Otherwise it's just planned obsolescence, which is risky for a pro tool brand.

[@MattsAwesomeStuff]

@toolscientist marketing-wise, how about how balance charging takes longer, and they lose more in sales from the perception that they're less efficient and slow charging, than they lose by the long term reputation of having poor balancing or only charging to 3 bars (something that takes years to notice)? From the looks of the code, it's like they didn't know who to hire (how many EE's or Programmers on staff there to even help choose a candidate?), may never have known it never worked, because they hired an idiot. I'm really curious what Dewalt and Milwaukee's corporate response would be if approached. Once upon a time I wrote to Campbell-Hausfield about my compressor motor burning up, and I was put in charge with the literal chief of technology at the company in response. Some of these technical people just never seem to be approached and are fascinating to talk to.

[@toolscientist]

True. But they could just leave it to stealthily balance at the end of charge and no one would know or care. Marketing wise you could add an "optimising" LED to the charger and convince people that it's putting extra special electrons in the batteries.

[@queazocotal]

@@MattsAwesomeStuff Except balance charging really doesn't take longer. It's 2mA drain, only on imbalanced cells. If one bank only is high, for example due to low capacity, with charging taking half an hour you get back an effective 10mA due to better balance leading to whole pack rising in capacity by 5mAh over a half hour charge. The charge only takes 2 or so seconds longer in half an hour, and only as you're raising the capacity. Obviously, you could trivially also continue balancing when in a charger idle. Even if you don't want to run the balance circuitry when out of the charger.

[@MattsAwesomeStuff]

@toolscientist well yes obviously this would be the right way to go. It's such a tiny thing to keep them from falling farther out of balance. When you and I run Milwaukee things will be different, but do you think the guy who wrote that code is A-tier material? Or was his IQ below room temperature?

[@97jettatrek]

Allright. Now, what mad scientist is going to figure out how to pull data from the IC and flash the update code to it?

[@toolscientist]

I vote for the guy that did the Dyson open source firmware. Anyone but me, really 😓

[@97jettatrek]

Oh yeah! The FU Dyson BMS. Maybe I'll reach out. Can't hurt.

[@karlosss1868]

I'm just starting to add balance leads to my Dyson. I tested a bricked pack & all cells were fine but the cell causing the fault has about 5% more capacity than the others & the circuit cripples itself when the voltage balance goes out to far. There must be money to be made in making a better BMS. Maybe one that charged each cell seperately to the correct voltage and one with a button to select storage charge for those that only use a device once every two months.

[@toolscientist]

Someone made an open source BMS software for Dyson. FU-Dyson-BMS, or similar

[@karlosss1868]

@@toolscientist Thanks mate!... I saw that but the hacked circuit doesn't balance the cells. Balancing is critical as although the circuit protects from over & under voltage without balancing, the battery pack never gets to full pack voltage as it stops charging when the first cell reaches full charge and never really gets close to discharged as the BMS disconnects when one cell reaches the low voltage cut out limit. Interestingly Ive discovered internal resistance seems to have nothing to do with the cells ability to self discharge. On one of my packs, the highest capacity cell (with a low internal resistance equal to the other cells) had the lowest voltage (before balancing) and was the culprit for the BMS tripping the circuit. I would have thought it would have been the lowest capacity cell but the combination of it having the highest tendancy to self discharge (and to be honest it is only slighgtly more than the others) and the circuit not seeming to balance the cells, the cell with the highest tendancy to self discharge (not the weakest cell) was the curprit in this pack. I've hunted high & low for a circuit to have the high & low cell cut out AND a balancing feature AND an on'off switch but to no avail. I now have balance leads in all 3 of my Dyson packs and expect I'll be set for years.

[@lucidmoses]

This is a perfect example of planned obsolescence.

[@donalexey]

Maybe balancing work only on new batteries? Or only in batteries with less than (threshold here) charge cycles? Then they just out of warranty and designed to die.

[@toolscientist]

The real conspiracy theory would be if it worked on the really old packs, but they disabled it because their batteries were lasting too long

[@JordyP89]

I have soldered flying leads to an external plug where ill plug an active balancer in every 6 months or so.

[@toolscientist]

Are you CarbonCrew? They also added a JST socket to their 12Ah www.reddit.com/r/MilwaukeeTool/comments/139wpzc/added_balance_leads_to_m18_120ah_battery/

[@JordyP89]

@@toolscientist wow, almost a direct copy! Thanks for your efforts on this, we spent quite a bit of time trying to work out what you did but gave up.

[@toolscientist]

I'm here to answer the questions that nobody cares about. So welcome to the 'nobody' club 😄

[@fluteplayerify]

Perhaps it's a bug where when they check the cell voltages for imbalance, the accidentally read one of the cells twice. Assuming both batteries use the same firmware, this would cause the same cell in the pack to be unbalanced.

[@toolscientist]

There could definitely be a bug, but not from checking the same cell twice. If they were balancing the wrong cell, I'd see it. Instead they're just not balancing anything. In my 12Ah pretty much all the cells are at different voltages, so even if they skipped one or even 2 cells, there would always be a cell that's at a higher voltage than the others.

[@r0000g]

Do the M12s balance?

[@toolscientist]

2 months ago I would have said definitely yes, but now I'm 50:50. I have 2x M12s so I can test them.

[@r0000g]

@@toolscientist Well, I don't know if you do requests, but...

[@toolscientist]

@@r0000g it's coming. I have Makita 40V, M12, and Ryobi 36V as my personal tools, and I'll do those before buying others. Also have to do Dewalt's flexvolt and powerstack as I have those from my Dewalt vid but decided they were different enough to warrant their own vid.

[@stevebabiak6997]

@@toolscientist - I assume that the 36V Ryobi is the same device that is called 40V by Ryobi in the USA.

[@toolscientist]

@@stevebabiak6997 yes, just another one of those annoying things where they use different labelling and product numbers between regions. It's mostly so it's harder for regions to tell if they're getting ripped off and reduce grey-market imports, but it just makes it harder to research products and find answers for problems.

[@chadkrause6574]

It’s so you have to buy more batteries. Planned obsolescence

[@toolscientist]

Maybe, but doing it in a way that's demonstratable is pretty risky for the reputation of premium tool brand. It might just be that they couldn't get it to work reliably. Maybe even someone disabled it years ago and forgot to re-enable it.

[@chadkrause6574]

@@toolscientist I could see that, but a 1kohm resistor is what you would use if you never intended on balancing. That’s just a RC filter at that point

[@toolscientist]

@@chadkrause6574 one of my theories is that they had too much noise in their system so they had to jack up he R value to filter it out. The reference design is using 47 Ohm resistors, so they've gone 20x higher. The other option is that 20x higher R means 20x smaller C on the input cacpacitors (there's a cap between each cell pin). So maybe it's because they ran out of space on the board? But then there's no components on the underside, so there was space available but maybe they wanted to save money?

[@chadkrause6574]

@@toolscientist I guess that’s possible but they could use 0201 components elsewhere to make space, plus they seem to have room. And how much noise is there really in a battery system when charging or balancing with no load on the tool?

[@chadkrause6574]

@@toolscientisthowever, the fact they keep rereading cell voltages may mean there is a bad programmer in the mix. 4mA balancing current isn’t completely unuseable if these batteries sit on the charger a long time and have matched cells. Once the system is balanced, the balancing system is just maintaining it

[@slim56]

Why balance cells when you can just sell another ridiculously overpriced battery?

[@toolscientist]

I guess for customer loyalty. Main reason I'm shifting to Makita 40V is the Milwaukee 12Ah packs failure rate. I had 2 of 3 of my packs do the "3 bars of doom" just outside of warranty

[@timothybayliss6680]

Huh...

[@toolscientist]

As in: "Huh, I didn't follow any of that", or "Huh, why aren't you balancing, Milwaukee?"

[@alexmoloney1560]

@@toolscientist Yes

[@emmettturner9452]

The other one.

[@aaronrunkle]

This is so disappointing. C'mon Milwaukee! This is basic stuff

[@LabiaLicker]

Fascinating stuff. Do we know any tool manufacturer that does actually balance their battery? I'd love to learn more about how you do these reverse engineers.

[@toolscientist]

Well, so far I'm 0/2 brands that balance their cells. Looks like I'm going to have to check all of them. It'd be funny if some cheapo brand like Aldi was the only one doing it

[@gf2e]

@@toolscientistI can try my Kobalt 6 cell batteries. What’s the best way to do that - a good voltage difference on a cell, and how many cycles should I do to see if it is actually balancing?

[@stevebabiak6997]

I am hoping Ryobi does balancing. Because that’s what I own ;)

[@toolscientist]

@@gf2e if you can see which resistors are the balancing resistors, then measure the voltage across them. Should be zero normally and ~2V when balancing. Other option is to push a cell high by around 0.1V and then leave it on the charger for a few hours. Make sure when you push the cell high that you start with all cells around 3.6V where the discharge curve is fairly linear. Also let it settle for 10min - cells tend to lose 0.05-0.2V after you remove charging current. Then you get into the problem of do they balance during or after charge. If during, then yeah, you'll need several charge cycles and measure the voltages between each, with 30min rest before measuring. Measuring the resistors will be the easier and more definitive answer.

[@gf2e]

@@toolscientist The cells just happen to be 3.6v. Clearly, a sign from above. There are six sets of NPN transistor pairs (KNM marking), with two 0402 resistors each. Not gonna be balancing with a whole lot of power on an 0402. One of the resistors is 20k, the other 100k. That is not very promising to me unless I am missing something. 180uA is kinda low. It’s only a 2Ah battery, but.