DIY 10S+ Cell Balancer

I’ve been out of town @TarzanHBK Will read through the thread when I get a chance.

This^

It pretty much comes down to these reasons: Not everyone uses quality cells. Not everyone uses quality welders. Not everyone has good soldering/welding skills.

Any of the above plus not having a BMS will ruin your batteries. As an example @denton had a battery pack in which welds failed and he was also using a charger with way too high voltage, but since he was using a BMS to charge and discharge no damage was ever done to his batteries.

1 Like

Someone with a wrong charge voltage is kind of unfair to use though as an argument no? :wink: It is not hard to make a setup without a BMS work well, so I still feel like they are not really necessary.

His charger was not the reason his cells were out of balance, it was the bad welding tabs. Still a valid argument. Look at eboosted and all the trouble he’s having with drifting cells.

It’s not trivial… One, you have to buy or make a quality welder. If you are soldering, you have to have good soldering skills and even then by the fact that you’re soldering onto batteries, you will damage them some. Two, you have to buy quality cells. Three, you have to have very conservative settings. Have you looked at whitepony’s settings? He only uses like 70% of the available capacity in his cells and doesn’t discharge them at high currents either. Four, if vibrations work a weld/solder joint off a cell, your pack is F’d. (this is where my example came into play really)

There are plenty of reasons to use a BMS. If you prefer to baby your pack instead then fine, I do it that way just because I like to know the state of my batteries all the time.

Not sure what to say - as usual I get your point but am personally not convinced that a BMS is crucial. Put some balance leads on the pack and check from time to time and it should work fine. I used a DIY welder and have no problems with my packs. People with faulty packs will obviously benefit from a BMS but I’d rather pay attention when building my pack in the first place and make sure everything is well.

All I am saying is that when you know what you are doing they are not crucial.

When you know what you’re doing, want to be conservative and prefer to do the monitoring yourself, yes. They are not crucial. For everyone else, it kinda is.

1 Like

You can’t use without BMS because or you buy a board that is a BMS or you become the BMS :upside_down_face:

1 Like

I have made the same idea disgussed here. It consists from 10 of TP4056-modules, 10 of 1A/5V power supplies and 7-segment voltage-meters. It is able to charge at 1Ah rate. Usually it took about 5h to charge after riding. It has been good enough for me.

Then i attempted to charge simultaneously with 42V/2A charger, and about 3 of the modules died. I don’t know if the 42V charger was the cause or something else. If you think about it, what could be the thing that would kill those modules if it was the 42V charger?

Anyways, it was pain to troubleshoot and repair because of the design i made. I broke it all to pieces except the voltage meters. Nowadays i’m just charging with only the 42V charger and voltage displays attached, so i see when any of the cells reach the 4,2V. Usually all the rest are about 0,1V-0,15V lower that the best.

I stumbled upon to this thread when i was searching info about those lm2596-modules. The TP4056 is not adjustable, so they can’t really be used on separate charger if you want to auto-cut the voltage when full, because the wires make additional resistance. lm2596 can be adjusted, though it will not cut the charge, but it is still better, because it heats less.

I’m actually moving to “normal” balance-charger now. A FET per cell loses the excess current into heat while series charging. But that’s out of this topic. But i’ll try the lm2596-modules as parallel-project though :slight_smile:

2 Likes

Well… that is some in-depth info you got there…

On a side note… I also witnessed that TP4056 overcharges the cell, if left for loo long… so yeh, either manual cut-out or a circuit is needed… otherwise they dont seem very reliable when in parallel config

epic necro post, nice!

Read the whole thread, and my thoughts:

  • the ultimate charging method is to cc/cv charge 1 cell with 1 charger. so @jmasta is going toward the one-true-path, not away from it.

  • cc/cv power supplies are different from charging circuits in that they do not have termination. The lithium ion charging profile is to hold constant current until termination voltage is reached, then hold the voltage constant until the current hits some minimum threshold (C/10 or C/20 usually). So cc/cv PSU will get you close but is not following charging protocol, and you’ll terminate at different states of charge on each cell.

  • if you use cc/cv psu, you’re gonna need a voltmeter per cell to tell you when to pull the plug.

  • those $1-2 electronics boards, I believe most of them overstate their capabilities. You need to burn-in test a bunch of them, and run them way under max spec, or add active cooling.

  • many lithium ion charging ICs don’t behave correctly. I thought TP4056 was okay though…

  • BMS has very limited balancing capabilities, and termination is not adjustable (are there exceptions?) so if you want to charge to 4.1v/cell to extend cycle life, balancing never kicks in

  • hobby chargers have better balancing capabilities but is also limited. so individual 1s chargers will actually finish faster.

  • failed BMSes seem to be a fairly frequent killer of packs? I feel you need to babysit the cells whether you use BMS or not.

I think this application makes the most sense for people who are running without BMS, to periodically check on the pack. It’s not so horrible to break the 6s series connection and balance charge each subpack once a month, but I think this thread’s premise is worth considering.

1 Like

@Quezacotl I’ve got to say, well done! That thing is awesome! :clap:

It’s cool to see that the idea is indeed feasible. Sounds like there might be a few bugs to work out, but what you built is basically exactly what I had in mind, even though we came up with the idea independently

I kept brainstorming this project a bit after posting but never updated the thread. I will post the schematics when I have more time. I’d love to see this progress further



Great points, @deucesdown!

This is basically what I am doing. I have a SuPower BMS and also an adjustable 12S charger. I almost always do a partial charge, which means the BMS balancing never actually kicks in (since it starts at 4.15V/cell). Despite not balancing each charge, the pack stayed very well in balance after probably 50 or more partial charges. My board has a built-in 12S cell voltage checker with a switch, so I can easily monitor individual voltages of all 12 cells. Occasionally I break the battery down into 6S “subpacks” and give them a proper balance with a hobby charger on the slowest balance setting. Currently I am using 4x 3S Lipos but will be replacing that with 2x 6S li-ion packs in series, so that I can still use the hobby charger for balancing or discharging if needed.

Side note: Even if you don’t use the BMS for balancing all the time, I still think they are very important to include in your design. It’s the only way to get cell-level voltage protection for over-discharge and over-charge; I see so many people using a bypassed BMS, which means you are losing most of the BMS’s safety features. Also, having some form of over-current protection is a must (in other words, if you bypass your BMS or don’t have one at all, you really should have a fuse in your system!)

This is something I’m struggling with too. vesc will progressively reduce power, but bms is BAM off, and throw you to the ground. the fuse I agree with, but struggle with how many and what level.

I’m similarly paranoid about e-switches, I think @lowGuido ruined me.

You should be. I mean this is your life that you are playing with.

1 Like

Your BMS cutoff should be lower than your esc cutoff. That is what makes it a secondary safety. The fuse is the last resort.

Cutoff order: ESC - linear drop in power till halt. BMS - hard cutoff as to not damage battery Fuse - hard cutoff in case all else failed

Limited balancing capabilities: Definitely but if your cells are more than 50mV out of balance after a single cycle there is something very srong with your battery.

I build batteries and power systems for unmanned aircraft and have never seen one of my packs even with up to 13 cells in parallel go out of balance more than 5mV. As long as you use quality cells from the same production batch to build a pack, you’ll rarely have balancing issues with a BMS that can do 100mA.

The problem arises with poorly manufactured packs where you are relying on the BMS to mask a greater issue.

Termination is not adjustable: Negative sir. Quality BMS with comms have adjustable limits. If you choose to go with a cheap BMS then they will more than likely have an analog balancing scheme which can be adjusted by changing certain components. Bestech BMS come with several configurations that you can choose from. If you want a real BMS, TI has reference designs but they are usually $300-400 for evaluation modules.

Either way, limits are adjustable. My BMS starts balancinng at 4V/cell and cuts off at 2.5V per cell. I requested it this way from Bestech , I think I paid $50 for it.

My supower BMS kicks in at 4.1V and has a lower limit of 2.9V. This is meant more for a Lipo pack as they dont go below 3v/cell.


Again, the BMS is not supposed to be the main protection, but a redundancy in case the ESC fails

1 Like

There are just so many S. ideas in this thread that I keep wondering maby this one of the reasons why Li-ion has got the reputation of being dangerous.

Please people, for your own sake and the people around stick by the book. This batteries can become dangerous, if not treated properly. And remember when you see a board burning that next time it can be yours.

Let Tesla, Samsung, Panasonic do the testing of methods. When they say that you don’t need BMS, and balancing than you can start with that. Until then try to stick to the working solutions. That are proved and tested in many years.

Rules :

  1. Get a quality BMS
  2. Make good welds 3)Use same batteries, from same batch if possible, at same voltage. 4)Connect balancing leads. Don’t start balancing just yet. 5)Good connectors all the way, don’t cheap out on that 6)isolate the cells from each other,
  3. Do at least 2 charges and discharge in bench. Check voltage drift from balancing leads time after time. This will tell you if your pack is properly welded. 8)Shrink wrapping after you found out that everything is OK. 9)Connects balancing leads and do one or 2 discharge and check again.
  4. After a week of use one more check.
  5. And do some checks time after time. Remember those pack are put to real stress all the time. Vibration, high current discharge, high current charge at reg breaking. So at least try to improve there where you can.

I have a non-bypassed BMS, system fuse, charge fuse, and my batteries are permanently housed in “fire-resistant” bags. I have built-in cell voltage monitors for all 12 cells. And I have a log book of every cell voltage for almost all of my discharge cycles

This idea was never meant to replace a BMS. Cell level charging has the potential to be safer than bulk charging. So don’t rag on someone for having a new idea :kissing_heart:

1 Like

It was not directed to you. It was more general…It’s not you asking that is the problem, because people who ask and discuss rearly make mistakes. But what grind my gears is people deciding that you should do things this or that way. Because it worked for them just fine, for 2-3 months. And that is just not enough testing. So I didn’t mean to offend neither you or anyone else. Just trying to raise the awareness a bit, because I feel that not everyone has understood that Li-ion should be treated with respect.

1 Like

OMG sorry for the wall of text lol

Generally I agree with you and appreciate you taking the time to type out the recipe.

But if you’re playing with lithium ion, precisely because they’re capable of so much mayhem, really I think one would be much better off considering how things work and how things break, rather than blindly following rote advice, especially if they don’t have the background to read between the lines, or the skills/knowledge/equipment to execute.

And there’s always room to improve, like all the cell level fusing discussion. Without critical thinking and discussion we’d just stay where we are. And while we can learn from industries with established standards, many of these standards are difficult or impossible or too expensive to apply to our hobby.

There’s also tons of grey area, I’m gonna rattle some off that I’m still trying to resolve in my head (I’m not looking for responses, just raising things I’ve been trying to learn about)

Are nickel strips good enough for series connections (I think no for typical 2x6355 and up), and is it okay just to do the series connections at the end of a parallel group (not ideal).

What is adequate mechanical support – puncture resistance, water survivability, friction rub-through of insulation, embrittlement of solder joints through movement, is hot glue flammable and does that matter.

What is adequate cell isolation (kapton tape? fish paper? cell holders?) Is it worth it to leave room for ventilation/cooling between cells.

Is cell level fusing worth the trouble – I think we’re pretty close but haven’t quite figured out proper implementation for eskate (do the fuse wires drop voltage, is it significant, how to acquire proper fuse wire, how to protect from vibration, how to detect broken fuse wires).

How much current can silicone wire really carry (ampacity of copper doesn’t really change, I believe the super high amp ratings of silicone wire is solely due to the 200deg rating of the insulator, and 200deg is TOO HOT inside our enclosures and next to our batteries).

Bottom balancing and bulk charging vs top balancing via bms.

If using BMS, how to detect failures (broken balance leads, fried components). I think like @jmasta, per cell voltmeter of some sort is the only way I’d feel comfortable.

Crimped vs soldered wire splices (crimp is the winner, but need proper equipment/technique)

Corrosion of contacts – charger ports and balance wires

After seeing @Eboosted board fire (sorry dude!), quick disconnect (loop key) and quick release enclosure and battery mount seem worth thinking about.

I think the e-switch situation sucks – $50 is expensive for a switch, and the $50 ones blow up (in “ON” position) on high powered boards. Mechanical switches also suck, with DC arcing problems above like 6s. Anti-spark loop key also sucks, you have exposed contacts when turned off, the contacts wear with cycles, and you can lose the key. BMS integrated e-switch seems like a good idea except for all the BMS’s blowing up, and crashes like these (looks like e-switch failed and bms cut off power, sorry Joe!). There are some high cycle circuit breaker switches for marine, that look like they might work? But bulky and expensive.

2 Likes