Using laptop BMS board to charge 18650 cells

[WallBender]

Just starting out on my project to build a battery for my house (long term) and a small jump pack for jump starting cars. Long story short I ordered a ton of stuff from china to start my project, but... I already have 195 cells from 23lbs of scrap laptop batteries from the local metal recycler center.

Currently I'm using a couple of head lamp/lights to charge a total of 4 cells at a time (ones that came out 2.5v or more) and it's a very slow charge method (250ma per cell max). I'm looking to cheat the system a little and actually use those BMS boards from the laptop batteries. Sadly I didn't log which ones came from what models and such and have already harvested the cells out, so I'll have to reverse engineer a little and probably make some guesses/fry some boards, but most boards are setup for 3 series, so it should handle 3,6, or 9 cells at once. More cells would be less current each, so using less would be a quick charge on cells that are most likely good, and using 9 or maybe even 12 would be for a slow charge for questionable ones after they are pre-charged to the min voltage of the BMS.

I figured I'd document the progress some here and I should put out a disclaimer that clearly I'm no expert and info on these boards is very slim of the few I've tried to look up. Some do require a way to "activate" the battery to allow charging/discharging.

If I can pull this off, it puts value back into the BMS boards instead of just tossing them to the metal recycler again and I could have effectively not bought any charger boards. Sadly I don't currently have any holders, so plans are to just solder wires on since I'll be soldering the packs together in the long run anyway. It will just make me pre-tin a bunch of the cells . Could also use magnets from hard drives and such that I have laying around.

For my first project I'm thinking of a charging station with 2-3 BMS boards with min 3 cells per board and max 6-9 and power it with a pc power supply. Possible issue might be that I'll need 12.6v to fully charge them, but on my quick test setup 12v is working from a 1a wall transformer.

Are there any foreseeable issues wiring the power in to the BMS boards in series? For the BMS boards with the activate mode, I could add a switch to turn them on/off or switch the power in connection.

Anyone else looking to do similar, post and let me know so atleast I know this info will help someone. Comments, Ideas, Suggestions etc very much welcomed. I have a ton of wire(15k+ feet) and fair soldering skills and an alright understanding of basic electronics

For those with OCD you might want to look away as I'm basically the complete opposite, stuff for my self is about function over looks, but I do go OCD for anything I sell.... talking about that is there interest in a 3-9 cell charger? It would be completely pre-made, just a power supply would likely be needed.

I also have a ton of charger boards coming from china... way more than I really needed but the price was worth just buying the 100. I could build single cell chargers off those as well which was my original plan for myself. I know the cell holders are probably pretty crap for the wiring, so I'll probably be re-wiring them with 18awg wire.

Anyway I'll get some pics to show the progress so far.

---

Here's my quick test setup. This BMS board required the pink and white wires hooked up.

Here is some info on the BMS board (some from google)

Board: LI-IMT60R1
Balances Cells?: Nope
Brand: Lenovo
Battery Model: LI-IMT60
Cell Layout: 3s2p



Might be a little hard to see in the photos, but the RadioShack multi meter is setup for amps and is wired in line from the wall charger (12v 1.2a) to the BMS board. Oddly it over volts the cells a little while charging (4.26v across a single cell) and started off drawing a tiny bit over 1 amp when I first started it and after 10 minutes or so it's dropped to 426ma. I didn't check the cell voltage before charging but with the power off they read 4.07-4.08v so it should be in the CV (constant voltage) charge state since the cells are pretty full.

I'm not sure what the best way to describe the wiring on these boards, an image is probably best but takes time to build assuming I run though most of the boards I have for this project.





Safety note, I do have a 5amp fuse on both connections on the BMS board, first because it connects well and second it fuses my multi meter and wall charger. It would be ideal to also run a fuse on the batteries, but with so few risks should be quite low for cells that already have good voltage.

Photos:

Power supply on (charging)

Cell Voltage




Amp Draw from power supply




Power supply off







Board overview

 

[WallBender]

Small update on this, I saw voltages as high as 4.37v at the cell, but charging finished by the time I made the last post. Battery is reading 4.18v and right before it turned off I think I saw the voltage swing up/down like the BMS was checking the cell voltages. Seems like first test run is a success and the cells charged fairly fast from ~4v the first time I checked them (after charging a little at 1amp) to 4.18v (all cells the same within 0.01v, started out with the other two 0.05v less).

Interesting on the kind of high voltage while charging, I wonder if the batteries that came from the pack were spec'ed for the higher voltage, or Lenovo's way to cheat the end of the charge and over voltage a little to charger quicker with the expense of wearing the cells more. Maybe this is why laptop batteries fail fairly often?!


Update: Soldered some wire onto magnets and soldered to the board to make hook up a little easier. Checking how dead the BMS will allow to charge the cells. First test was average around 1.8v and it started off drawing nearly 2 amps. Probably need to up my power supply since it's rated at 1.2 amp lol. The cells measured 1.6v, 1.7v, and a tiny bit over 2v, and after charging for a short time they came up pretty close to each other ~3.48v +- 0.2v so I suspect the BMS is balancing on the charge cycle instead of the "wasteful" discharge cycle.

I'm going to throw a few 0.5-1v cells in and see if it attempts to charge them =). Well looks like it didn't like the super low voltage cells which isn't surprizing. I had to put all cells in with close to 2v each to get it to start up again. 1.5v was too low to kick it on and ~1.7v it kicked on. Looks like the power supply I'm using does have a temp sensor since it kicked off after a while running around 1250ma. Looks like all cells are within 0.1v.

 

[DarkRaven]

I've found lots of LiHV cells in laptop batteries. These are the ones designed for a voltage of 4.35V. Maybe this BMS belonged to such a battery.

 

[WallBender]

Would the LiHV cells be the same size as the 18650's? I did get a couple batteries with the flat style cells (pouch) but those parts are separated. Also the temp sensor for the BMS is on the board instead of directly on the battery (poor design?).

Based on photos of the battery pack, it had the 18650 style cells. Probably should have kept track of what cells came from which battery and such, but I wasn't going to use the BMS boards.

I have a bunch more boards so I'll have to guess the sensor wires on a couple more and hopefully find more that don't require the "active" state to bring it online since there's little info on the pin outs.

 

[DarkRaven]

Yes, they come in standard 18650 format as well. Samsung ICR18650-30B is one example, they are rated for 4.35V +/- 0.05V. Found lots of them in Dell batteries.
Temperature sensor on the board instead of the cells is fine. They are tightly packed and when the cells get hot then the adjacent board will also get hot very soon. There will be a delay but this isn't a huge problem since you can always compensate by setting the threshold a bit lower if needed.

 

[egam]

Cool Idea. Thank you for sharing. I am saving all of the electronic debris that comes from the packs. I was thinking about salvaging parts off the boards, but now I need to rethink that. The fact that this balance board equalizes on charging is pretty cool. I don't like balance boards that bleed off power as heat at the top of charge.

 

[WallBender]

@DarkRaven

Makes sense, it is probably the only BMS board in the lot I have that's on the board, most were glued between the two cells in the center of the packs, or had two temp sensors. I'll have to keep a better eye on voltage when looking up cells, I thought they'd be pretty much all the same when the cells are the same size. I was going to toss the pouch batteries to the recycler, but might try to find a use for them since if I remember right they support a higher draw and would be more ideal for a jump pack (don't have enough for that though atm).

@egam

Yea that was kind of the idea of why I wanted to share my findings, I'm sure most people harvesting 18650's are basically just tossing the boards. Almost every board I have is different, so I'm not sure how common balance charging is or if the cells I charged just happen to get closer in voltage just because of how they charged (they are all untested atm). I'll have to probably put 3 tested cells in with one a lot lower than the other two and see how many cycles it takes to balance if at all. I guess running a amp reading on the two between wires would show the current differences or if they are purely sensing wires and all the charging current is via the two end wires (overall neg/pos).

Anyway, for those that would like to use this idea for a charger, it's ideal if you note the wiring down before you remove the cells. I've only tried one board, but I probed a fair few things around trying to get it to charge and didn't blow anything. This board might just have protections so your results might be different. I've reversed the two sensor wires and shorted each of the other pins to ground that were not the +/-. FYI first two and last pins are both the same connection on the BMS, I'm only using the outer two for simplicity.

What would be really neat is to figure out if it's possible to "reprogram" the BMS for things like temp settings, current/layout (3s1p vs 3s2p vs 3s3p), and such. I've seen some packs have the temp sensor also wired on the connector for the laptop to read and there's a communication pin.

Probably should also mention the 2p and 3p setups I'd suggest only using cells paired up by similar voltages so the balancing current isn't too high. Each of the series isn't as critical but you don't want fully dead batteries mixed with fully charged. Not the ideal charger I know, but should still be useful since most of us won't need 20+ chargers after the initial run though except for retesting / diag work on packs.

Anyway, fun project so far, I'll try to run though a couple more boards today and hopefully have a better idea how they balance. Need more multi meters! lol.

It might be a little overkill but I'm thinking about grabbing 3 or so super cheap multi meters to monitor the list below. Anything else I should check?

Amp inline between 12v source and BMS
Voltage from 12v source (to watch for dips or if the power supply kicks off for some reason)
2x amp readings for the two center sensor wires as I've been calling them
At least one voltage reading across a cell, 3 would be ideal for real time watching the process better

That comes up to 7 meters, just have to see if I can find some super cheap ones local or not. If not more china orders! lol

 

[jdeadman]

Wow that's interesting. Of the packs I pulled apart I don't look closely to the boards and thought them to be just connection boards not balancers. I'll have to look at the next one.

 

[WallBender]

If there's an IC to watch out for to be able to tell if a board has balancing that would be good info to log.

I wired up a much better power supply from a dell desktop (18VA on 12v rail) and the last cells I had in it wouldn't start charging which was really weird. I threw in some pretty low voltage ones ~2.4v and it's jumping 0 to 3 amp draw on and off, I can actually hear it clicking per cycle probably from the less than ideal connection on the magnets. I'll give it a little to see if it will start charging. These are my "maybe" cells, so fair chances I'm hitting bad ones. Not sure if a BMS would do this, but a possible theory is that once it hits a target voltage, it tests the internal resistance and if it's too high for what it's expecting it deems the pack bad and won't charge. Just a theory though.

This cycling on/off is strange though, math calculates at 500ma per cell in it's original layout which seems a little high for low voltage cells (10.23v across all 3 after a few minutes). I did put similar cells in and they are within 0.01v now (3.4v, 3.4v, 3.39v). The way I have it wired atm it will be a little hard to see the current from the balance leads.

Going to test the balance a little different. Since the board is acting weird with low voltage cells that start off around 2v or so I'm going to throw in some 3v+ ones and mix a low one in with 2 higher ones and log the before/after volts after a while charging and once it finishes (if all goes well) I'll give an ending voltage.

Start volts:
3.6v
3.67v
3.22v

Update... charger wouldn't start switched the 2nd and 3rd cell and it started drawing 0.8a and stopped again. Maybe it balances between the cells first then charges? I'll keep an eye on it. I suspect if my guess is right, I wouldn't need the power supply on till they balance anyway. If that's how it works, that's a pretty neat design/idea and is efficent power wise even though it takes the battery offline (charge and output voltage) of what I'm seeing so far.

New Start volts:
3.6v
3.22v
3.67v

Ok it's been almost 10 minutes, lets check the voltages.
3.61v
3.35v
3.68v

Interesting the 0.8a draw pushed them up a fair bit, but the balancing (if it's doing that) is quite slow but kind of appears it is doing something. I used my Fluke meter and hooked up between the cell ends and each wire and the only one I could get a reading on was the low voltage cell was bumping 0.01a on and off, I couldn't see a reading on the other 2 cells. I suspect it's designed to balance when assembled over days/weeks instead of "on the fly" balancing.

I messed around with the meter a little and switched to the 300ma amp spot (was using 10a to prevent blowing fuses) and got a reading 0.06 which I suspect is 6ma or maybe 60ma. Turning off the power supply actually made it balance more, same setting pushed to 0.22. Maybe it's logic is balance while not in use, super slow balance while in use, but it does seem to be actually balancing.

Since it doesn't need the power supply to get the same readings I think I'll poke around another board.

 

[WallBender]

Second BMS board is interesting, it didn't need the sensor wires hooked up. Using 3 cells balanced already at 3.45v for testing so good voltage too and it started drawing around 2500ma. Might be more of a dumb board which would work well for charging cells. Holding wires is a little hard to test balance current though so I'll leave it at that. Board does look much older so that might be partial why it will charge w\o monitoring cell voltages.

Board: CL740 REV4 (also has reference to CL741
Balances Cells?: Untested
Brand: IBM?
Battery Model:
Cell Layout: 3s2p

 

[jdeadman]

Cool looks like a cool project.

 

[WallBender]

Currently trying to figure out this active pin thing. Finding some info but seems like nothing is powering up the board to accept the charge . Seems like dell's have a sensor wire in their chargers to tell the system how big of a charger it is. Not sure how that's communicated to the battery, but all dell batteries I can't get to accept a charge.

Also update on the first BMS board, after an hour or so of sitting, the voltages didn't change at all. I did rip a trace up before so that might be related.

I also tested a dell laptop I have laying around, and it puts out about 18.5v to the battery pins, but the power supply is around 19.5v which I suspect is too high. My 12v power supply is probably too low for most of these boards and maybe I killed the first board from under voltage.

Still poking around, finding a spec sheet on one of these boards would make things easier lol.

EDIT....

Alright I guess I overlooked a small detail... I burned up one board and since then no other boards worked including ones tested working before! Derp on me one of my fuses blew =). New fuse and the Toshiba board works now w\o active pin. I like the layout and such so I might extend the wires for easier handling and test that one a bit more and try not to knock any traces off this time. Based on what I've read/understand on electronics I'm thinking it has cell balancing so hoping to test that. Seems the balancing would be during charging based on one source of what I've read. Several older batteries didn't have balancing.

 

[WallBender]

Update on the 3 cells I noted before, I used the toshiba board to charge them 5+ hours and it's currently drawing around 21ma so basically it's done charging. Turning off the power supply and reading the voltages right away clearly shows it didn't balance while charging infact only one cell is truely within range for fully charged (4v+ in my eyes). Here's the resulting voltages noting that the 2nd one is a higher rated cell from the factory 2600mah vs estimated 2200mah (blue sony but not in cell db).

3.91v
3.76v
4.06v

Not bad for a "reclaimed" charger board (just to get the charge atleast close to 4v), but clearly it must be designed for per-balanced cells, or somehow it balances on the discharge cycle or while idle. I suspect it cut off the charge because of the highst cell voltage while power was still on was 4.11v which is a good sign it doesn't overcharge cells and I suspect it does the same thing in reverse on the discharge cycle with a low voltage limit before it cuts off. This learning process is interesting but that board wouldn't be one I'd want to rely on long term for charging/testing, just for bringing up low cells to an acceptable voltage. I'll have to figure out what the lower limits are and hopefully it has a genital "pre-charge" cycle before it starts dumping 500ma+ into each cell.

I don't have photos of the board but here's some info off it for reference (myself and others)

Toshiba 3465
DD201004200/6 R3
Wiring format is name - color (board label):
Neg - Black (0V)
VL - Blue (4V)
VH - Yellow (8V)
Pos - Red (12v)
Power into the board is labeled on the back side, the pin next to 8v is the pos and the opposate side is neg like most I've found so far.

Pretty interesting info I've discovered so far, seems to make sense since I've had a fair bit of issues with low capacity batteries from around the era most of these batteries came from. My china knockoff Toshiba battery in my current laptop has been spot on, but I didn't buy the cheapest one either and I think it's a 12 cell pack. Really this kind of shows just how critical it is to balance batteries if you want their full power storage.

I'll have to come up with a list of board labels and what they mean some time so others could give this a shot too. The layout is logical on most, but some have things reversed from what you'd think like the main pos/neg being the center wires on the board. Anyway going to get another board fired up and see if I can find one to balance these cells. I'll document the good ones better since they would be the ones I'd be using long term for balancing (3p at a time) assuming the speed isn't too crazy slow.

 

[daromer]

Dell and HP if im not mistaken uses a voltage ref for setting the charge part. I have had to add a resistor to get those powersupplies running when using them for other than charging the computer in question.

 

[WallBender]

Yea seems it's the common non-working boards are from dells and hp packs which was most of the lot I got. I read a bit on the dell power supplies and they are 3 wire, pos, neg, and a signal to say what size of charger it is, just not sure how that translates to what the battery receives. Good news is I have a dell laptop sitting around, but not sure which pin would be the signal. I think it is the 8 pin version of the battery pack.

Side note, 2nd 5amp fuse blew, not sure which board popped it time to replace it again lol.

 

[WallBender]

Haven't really gotten around to doing much more with this project yet, kind of really need my battery holders to show up with might be a month or two from now .

Anyway for some more numbers for those interested. At the 12v input most seem to draw around 2 amps of current on low batteries and if my math is correct, that should be around 6.5 amps across the 3 batteries or around 2162ma per cell in a 3s1p setup. Might be a little high for 3 cells, but I guess it kind of works as a quick charger. Once the batteries near the top voltage the current drops pretty quick to 1amp, then slowly drops from there of what i've observed. One was running 28ma draw @ 12v so it seems it doesn't fully turn off on it's own and the cells were around 4.2v (going off memory).

Seems like the laptop board based chargers would be good for larger cell count of semi-balanced cells, or to top up 3 larger packs at a higher amp rating while still charging correctly for li-ion. I guess ideal would be a 3s2p setup since that's what they are designed for.

Using tp4056 boards might be better overall since I've read it's possible to combine them together in a way to get more than 1 amp out of them. I'll have 100 showing up and only 40 cell holders, so I'll have to play around with higher than 1 amp output options using them to test it further and have plenty encase I hook up backwards. Half thinking about wiring up the temp sensors from the laptop packs on the tp4056 boards so I don't have to monitor temps manually. Not sure what the board does exactly, I just know the chip has a leg dedicated for temp sensor and is factory grounded. I assume from that the temp sensor goes between the leg and ground and if I remember right, higher the temp, higher the resistance. I'll have to test the temp sensors and see if they are similar resistances at room temp, but should have enough for the charger I'm planning to build.

 

[cmg_george]

Hi @WallBende, thank you for sharing your experience.I had the same idea and i'm playing with a BMS from a laptop battery (i don't know the laptop model) but i'm disappointed that looks like is not making any balance. The BMS i have is 4S and i like it becouse is very small but without balance is useless...any idea how to add balance?

 

[Korishan]

It would be hard to find out how to add balance as you can't get the specs on the TI IC (which most use). Unless, of course, you have access to commercial level programmers/designers and can sweet talk them into getting the info for you.
Also, if you want to get some data from the boards, you need to use the EV2300/2400 communications controller. The EV2300 is about $100, and the EV2400 is about $200

 

[cmg_george]

Well for sure i will not go for EV2300/2400
There is no wonder why laptop batteries is the first thing to fail in laptops... i was not expecting no balance. I also saw that it lets cell go up to 4.3...something that i also don't want so i think my small and nice BMS is going to trash.

 

[WallBender]

Sadly it seems most/all laptop bms boards don't balance, atleast the ones I was messing with. You could charge a group then connect them in parallel to let them self balance (voltages must be close though). Another option but much slower is to charge them in parallel with a higher amp charger. Like the Chinese boards a lot of people buy on ebay can handle 1amp charging if I remember right. If you put 4 cells in parallel you'd have kind of a slow charge at 250ma per cell, but they would self balance too.

If you mixed cells with high capacity and low, the high capacity cells could end up being under charged and the low capacity ones have a higher charge, if they charge quicker than they balance.

Also to point out, those china boards are pretty easy to setup if you have a 5v power supply that can put out the amps required for x chargers. Commonly used is a PC power supply. If you hook a cell backwards to the china charger it will burn out though, so that's something to keep in mind.