How Does An Active Balancer Work?

Tear down of a common capacitor based active balancer.
AliExpress - Heltecbms Official Store
s.click.aliexpress.com/e/_DeC...
(referral link)

[@gilrand6224]

That is an excellent explanation of how this balancer works. Thank you very much for taking the time to explain it. I was hoping that the board had some means by which it would shut itself off when the balance was within specification, but that may be wishful thinking. I would like to leave the balancer connected with the BMS, but I was hoping it go to sleep when not needed.

[@ScottElliott888]

A lot of mis-information on the net about active balancers & this puts it all to rest. Great program you have there on your PC too👍 Thanks heaps.

[@conradfuller6697]

An excellent investigation and description, thank you for taking the time.

[@StuartPittaway]

Thank you

[@artypete]

Very interesting! Didn’t know active balancers worked quite like that. Well done 😊

[@StuartPittaway]

There are several types, this one is capacitive balancer.

[@jbuszkie]

Thanks for this! I always wondered how they worked but was too lazy to try and figure it out!!👏

[@StuartPittaway]

Happy to help!

[@tigiewig]

got here from off grid garage . thanks for excellent explanation of how these work . 👍

[@StuartPittaway]

Welcome!

[@jozefmitosinka7611]

very good work, you put a lot of effort into explaining the active balancer, thank you

[@StuartPittaway]

Glad it was helpful!

[@anilnaar]

Hello Stuart. Thanks, you explain complex circuits in a very easy way. May I know which PCB design simulation software you are using in this video?

[@wayne8113]

Thanks Stuart

[@akaigx9]

The current consumption of 4S of such a balancer is 5 mA, that is, 2.7 W per day - this is so as not to think about turning it off.

[@Krmpfpks]

That’s a great video! Thank you so much, I am currently learning about batteries to refurbish a used powerwall and I bought such an inverter. How fast are these pulses switching between charging the capacitors and charging the batteries?

[@GarethJones-dk9yp]

Great video. I use this active balancer in conjunction with the Diybms on my 7s lithium ion pack and does a great job keeping them in balance. However, I do only recommend using these for top balancing. I switch the enable contact on at 3.9V and only charge to 4V/cell. You must admit a diybms V5 with active balance BMS would be awesome 😉

[@StuartPittaway]

Great tip!

[@ConorFenlon]

Is your battery bank 7s1p? Or more than 1p?

[@GarethJones-dk9yp]

7S40P. LG MH1. Around 120Ah.

[@ConorFenlon]

@@GarethJones-dk9yp Oh sweet!!

[@AnupamVipul]

Great work I just have one question what is the reason for small caps ? why cant big cap with small value of resistor 4 inrush current limiting be used for quicker balancing ? Is it cost cutting measure ?

[@StuartPittaway]

Good question, I'm not really sure. They are "solid polymer capacitors" - which have better capability with ripple, so should last longer. They are expensive though, so expect this is a cost cutting limitation. You would potentially also get much larger current spikes across the PCB which again would need to be factored into a larger board or higher quality parts etc.

[@wmatozinhos1]

shared for us this circuit archive, very good your apresetention

[@petrovich5188]

Thanks. Very helpful.

[@namenotshown9277]

I occasionally have to balance some 7s/1p packs which I do manually. I'm wondering why balancers dont use this method which is fairly straight forward. My method is as follows: 1. measure all cell voltages 2. charge the low cell voltages, and discharge the higher cell voltages I use icharger in 1s mode to charge a cell, and to discharge i use a 12v automotive blinker globe ( 5watts) globe. It doesn't take all that long to get the cells to pretty close voltages, so my brain is doing the work judging when to stop charging/discharging, which I guess might be the tricky bit to design into a circuit to do same. There are things like voltage sag to consider, but it becomes quite intuitive quickly. Its top balancing also when cells are close to full. I've often wondered how those balancers worked..........thanks so much , its a clever design indeed.

[@zupciaaa2288]

How he set the values ​​of the Schmitt flip-flop in the usual ones

[@Cptnbond]

The green unknown ("U") components are as you said very likely thermal fuses. I have a similar, albeit larger bough from LCSC. Cheers.

[@StuartPittaway]

Thanks for the info!

[@ktcustom778]

These are PTC resettable fuses. From my testing, they look to be +/- 6A rating. @@StuartPittaway

[@loucinci3922]

Andy from off-grid garage sent me over to check out this video. Hope it will be a good one - LOL. Thanks for sharing

[@StuartPittaway]

Thanks for coming

[@somebody1869]

I’ve got a 16S to use with my 16S pack. It’s mostly only to be turned on during charging, using the “run” solder pads through a switch.

[@OffGridGarageAustralia]

And I think this is actually more efficient and effective than having it running all the time.

[@Krmpfpks]

I heard its even better to only have it on right after charging when the battery is full. But that‘s hard to do without a micro controller tapping into the charger state.

[@kiwiscanwifi]

would be great to see a purpose built unit plug in to work with the new all in one controller..

[@StuartPittaway]

Nice idea

[@diylithiumguy]

Great video!

[@StuartPittaway]

Glad you enjoyed it

[@CliveTrezona]

Great simulation Stuart. There must be a way to adjust the balance turn on voltage by changing a few component values?

[@StuartPittaway]

Yes, think that 3 pin chip swappee out will do it

[@turkcanatilir3943]

Stuart, thanks for this post. Have you ever thought about keeping some extra cells for using while decharging the battery pack and switching some of the cells on or off (Using MCU and some Mosfets) according to their SOCs ? Your answer is important for me. Thx again. Good work.

[@StuartPittaway]

No, I've not considered that. The SOC across all the cells is generally the same, unless you have very poor quality cells to begin with. The battery is only as good as its weakest cell, so swapping others in/out wouldn't be an improvement.

[@forrest2851]

thank you!!!

[@StuartPittaway]

You're welcome!

[@metinozturk19]

Hello, can you give the link to the falstad you made? I need it very much, thank you in advance.

[@jamest.5001]

What if a boost converter or jewl thief is used to charge the capacitor or the low cell? If the capacitor is charged to 4.5-5v or even a capacitor doubler, charge two capacitors then put them in series to charge a cell. Its going to cost space and components, maybe it is worth it if it can work Soo much better, having a boost converter or jewel thief making noise may not be so acceptable. However a fet switching caps in series should be much more welcome.. possibly use a small micro controller to discharge the cap before connecting the call to charge it , so the high cell charges it faster. Then going in series as it connects to the low cell. I wish i had more knowledge about electronics, i could have started learning 30-35 years ago and know a ton more! Thank you soo much for all the effort , and for sharing this awesome project!

[@ScottElliott888]

These active balancers are dirt cheap hence kept to an absolute minimum, yes, the cct could be improved 10x fold. 😁 Mine cost 62peso (Philippines) online from China. ($1US) 😉

[@lmwlmw4468]

Great video.

[@StuartPittaway]

Thanks!

[@444guns]

thanks for the video

[@StuartPittaway]

You're welcome

[@jamesward5721]

A multimeter, bulb as discharge load & any old power supply/charger kinda does the job too, if you're not into complexity. Test each cell, discharge the over-spec ones using the bulb till they matchish, charge the lazy sod ones till they match the median. Or not, each to their own, there's always more than one way to skin a cat.

[@tobimai4843]

Actually a pretty smart and cost-efficient design. Most Active balancing ICs from Linear etc. usually use boost converters to boost the cell voltage to the battery voltage

[@ricolauersdorf687]

Yep, but boost converters are much more efficient.

[@Ozzy3333333]

Got a part number? I am not buying your boost statement.

[@teardowndan5364]

@@Ozzy3333333 What part of it? Switched capacitors aren't particularly efficient: if you want to equalize voltage between two identical capacitors where one is charged to 10V and the other at 0V by directly paralleling them, half of the coulombs get transferred and you end up with two capacitors charged to 5V. Q=CV^2/2, which means half of the energy was lost to I2R and EMI in the process. Do the same experiment by putting an inductor between the two capacitors to store the potential energy associated with the voltage difference instead of letting it turn into heat and RF radiation, now you get two capacitors charged to about 7V. If you want to efficiently transfer energy between cells, you definitely need something between capacitors. With a buck-boost/SEPIC/CUK converter, you can transfer energy between battery cells more effectively and efficiently than using switched capacitors. You could even use an isolated DC-DC converter to either divert current from the strongest cell back to pack voltage or directly charge the weakest cell from pack voltage.

[@Ozzy3333333]

@@teardowndan5364 as I said Got a part number? I am not buying your boost statement.

[@teardowndan5364]

@@Ozzy3333333 The part number is: conservation of energy. You need some sort of DC-DC conversion between things at different voltages if you want to avoid losing the energy embodied in the voltage difference, no ifs or buts about it. You can simulate it in LTSpice or any circuit simulator of your liking using any parts you want, the result will be the same.

[@user-yh9qf5gd4l]

Great explanation! What simulator did you use?

[@StuartPittaway]

www.falstad.com/circuit/

[@Andrea-ss7fy]

Hello Stuart, really interesting!! Just a small question: how is the current supposed to flow back from capacitors to batteries if there’s a diode? Thank you :) Edit: I auto-reply myself: that on the schema is not a diode but a current probe :)

[@StuartPittaway]

Yeah - its a probe on the circuit simulator.

[@CollinBaillie]

Hey Stuart. Your simulation has resistors on each capacitor. Is this just to enable a slower charge/discharge in the simulator or does the actual balancer have these also?

[@StuartPittaway]

The balanceer doesn't have those resistors, these simulate the internal resistance of the capacitor (ESR). It's a very low resistance.

[@user-qd1yu8ig9b]

Hello, Thank you for good video. I want to know the name of circuit simulation program.

[@robertcollins4663]

It is the "Falstad Circuit Simulator" which is online only but someone made a standalone called "CircuitsJS1"

[@johnnymac1976]

Question: Would it be possible to swap out the LTO voltage detector chip to one that switches on at say 1.7V?

[@StuartPittaway]

Yes it would

[@martinjaure1310]

Which kind of JST connector does it have?

[@merieginvandyk768]

what program is used for the simulation? I am building a 48V battery pack that has 14 battery modules in series.

[@CrAzYDr1veR]

would't it be fun to have a balancer like this as a addon for your new 16s module? Be able to remove all the extra wires...

[@StuartPittaway]

Yes, maybe one day!

[@phuang3]

Big thanks to you. This is exactly what I am looking for. BTW, what if the charging voltage is too high? how does it lower the volatge? Let's say I got 5S1P SCiB battery. It's max input is 13.5V, but I'm charging it at 14.0V. All cells are over voltage of 0.1V. How does it get rid of 0.1V? The passive balancer uses resistor though.

[@StuartPittaway]

If all the cells are over voltage, then it can't get rid of the energy - there is some waste, but ideally you don't allow any cell to exceed its maximum safe voltage (which is where the BMS should stop charging)

[@phuang3]

I use it on my motorcycle. The generator's output is higher than 13.5V when RPM is high. I already got the passive balancer, and it works perfectly under my situation.

[@Wo7f_Creek]

name of this simulation program

[@lequocminh2566]

I have a pack of super - capacitor 6s (16.2V, but in my car max charging voltage only 14.4v). I wonder that can I use LFP battery balancer 6s for my stuff?

[@StuartPittaway]

These boards seem to be compatible with most types of cells so I don't see a reason why is wouldn't work with a capacitor

[@nelsoncruz7278]

Can I use 7s active balancer to balance 3s battery without damaging the balancer ?

[@StuartPittaway]

Sorry I don't know! So many variables

[@chapulino1000]

excellent reverse engineering research, thank you very much for sharing it

[@StuartPittaway]

No problem

[@nadeemsahto-zc2nn]

I have 8s lethium battery how i conect 4s two blancer in series?

[@StuartPittaway]

You wouldn't - you need to buy an 8S balancer.

[@Harani66]

these seem to come in different current ratings.. Should I be choosing one based on the overall size of battery or individual cell capacity or the max differential it could be shifting between cells (C rating ) ? or

[@StuartPittaway]

Generally, for any lithium battery 1amp balance is perfectly okay, unless they are really bad cells.

[@asificam1]

What's the practical difference between the flying capacitor type (show in video) and the inductive type (the tiny type with inductors)? Is there any real reason to use one over then other? I know sort of the difference of using inductance vs capacitance, one uses capacitors, one uses the back EMF from inductors and changing currents like a buck or boost converter. But I don't know why you'd use one system over the other for an active balancer.

[@bartklump]

Can you shed bit more light on how the schmit trigger is able to send a signal to each of the mosfet gate driver ics ? Is the 10V going to each of the mosfet gate drivers ? or what does the 10v regulator do exactly besided powering the scmit trigger? are the mosfet drivers powered from the cells directly ? (1 or 2 cells) ? is there a decoupling capacitor inbetween the schmit trigger output and the gate driver input ?

[@ming6842]

I have exact same question about this

[@StuartPittaway]

I've not gone into the detail, but I think its a schmitt trigger oscillator which fires pulses into the MOSFET driver chips.

[@bartklump]

So its sending the signals through a capacitor, which apparently can do magical things when connected to a mosfet. I don't really understand the principle, but its sort of like a charge pump, the way I understand it.@@StuartPittaway

[@StuartPittaway]

@@bartklump I believe so, ultimately it's a pulse generator to toggle the nosfets

[@bartklump]

yeah but thanks to the capacitors, its able to switch the higher voltage differential mosfets aswell. Since the driver sits at ground level.@@StuartPittaway

[@tomxx81]

Hi Stuart, any hint how to force balancing of LFP sooner? In my case, cells 6 & 12 are running away on top, but balancer doesn't balance because of cell 1 still at 3.36V while 3 & 6 at 3.5V. I'm controlling RUN solderpad with a relay (from HA, triggered by max_cell_voltage), so thinking about switching it to LTO - it should start balancing immediately. Any comments? Thanks!

[@ktcustom778]

Trigger a balance start at 3.45V for LiFeP04. with a 3.5V-3.55V 30min - 1 hr absorb, depending on overall pack health. ***Attempting any balance below the 3.45v upper knee serves only to imbalance the pack. My annual top balance usually takes over an hour at 3.5V for my well matched 280Ahr packs. My worst pack takes over 3 hrs twice per year. In your situation, switching to LTO should serve your intentions well so long as balancing on/off is relay triggered and voltage biased..

[@tomxx81]

@@ktcustom778Thanks a lot, but there's an issue (by design) that balancer starts only when Cell 1 reaches mentioned voltage.

[@michalctakanych]

Дорогой друг.Всем известно,что на переменной частоте ,емкость и индуктивность имеют разное результирующее сопротивление для протекающего через них тока.Неграмотные китайцы притулили на эту схемотехнику длинные провода мелкого сечения.Поэтому это работает в часть силы.Для обеспечения большого тока отдачи,выводы на самом аккумуляторе необходимо зашунтировать несколькими смд керамическими конденсаторами по 1 мкф.Обрати внимание,что аккумулятор может скушать большой ток,только при большой разнице напряжений.Гораздо правильнее ставить вместо аккумулятора резистор и смотреть на нем напряжение (ток) при тесте.

[@swastikdas4251]

Hi like to connect with you sir , I am also trying to make same circuit my question is that if you turns the gate on how can it charge the battery ? every mosfets gate is turned on at same time how it is possible ?

[@CannonballCircuit]

Is it fair to say this active balancer works best at low and high SOC ranges, and there's a range in the middle where the cells can't be effectively balanced with this method?

[@StuartPittaway]

With LFP cells yes - they only balance when over 3.4V

[@lezbriddon]

Thats pretty complicated so I hate to think how mine works as it only has 2 caps in parallel to make one big one, and only takes power from the highest cell to push into the lowest, the switching to achieve that on a 24s setup must be a lot of mosfets...

[@jameshancock]

Wish it was easy to change the turn on to 3.45V

[@StuartPittaway]

It might be possible, just swapping the 3 pin chip

[@jameshancock]

@@StuartPittaway once you get all of the conformal coating off and find a chip with a higher value… was thinking maybe put a resistor in series?

[@IggyJackson]

Which balancer are you using on your pack?

[@StuartPittaway]

Same as the one in this video, just 16S

[@michimichi7534]

Balancing LFP Cells below 3,4V cell voltage is not a good idea in general. As the voltage to SoC curve is that flat, t here is the high chance to destroy a former top balancing.

[@StuartPittaway]

Agreed, don't know why they didn't set the on voltage to a higher level

[@chuxxsss]

Good work, Stuart. Will reverse engineered.

[@StuartPittaway]

Thank you

[@sriramulamahesh8305]

Please design 24s lifepo4 active balancinsar.rar file

[@StuartPittaway]

I probably won't have the time or use for such a board sorry!

[@absolute___zero]

I saw this video of yours and I was happy because I thought all my battery balancing problems will be solved with this circuit. Then I found another video of channel Off Grid Garage, video hash is yPmwrPOwC3g , titled "Automatic Active Balancer - The Evolution in Balancing. But..." and I became very unhappy. He is saying that using this equalizer circuit that you are showing will actually disbalance all the cells int he battery. How can he be right if the circuit actually is made for balancing??? It is like saying 2 + 2 = 5 in mathematics.

[@StuartPittaway]

The issue is if the balancer is running all the time - ideally you only want it to run when the cell voltages are high - particularly for LFP chemistry cells.

[@absolute___zero]

@@StuartPittaway hmm, then it looks like a broken "solution" why would one be using a "solution" that creates another problem? In that case we need a circuit that includes a "maximum voltage" parameter, to deactivate/activate itself. Or something like that.

[@absolute___zero]

@@StuartPittaway wouldn't it be better to connect each cell to a battery charger circuit for 50 cents, like the ones that are included in lithium batteries for cell phones? CC-CV charger it is called. If the battery isn't accepting charge, the circuit won't draw current, so you wire every cell with such circuit and connect it to MPPT device. Active balancer circuit costs about 12 bucks, but lithium ion chargers cost like 50 cents and you don't have a problem with discharging. It will be cheaper and you don't need to buy active balancer circuit and your hedaches are over.

[@StuartPittaway]

@@absolute___zeroThese balancers work well with "normal" lithium cells but for LFP, they need top balancing due to their charge curves. I agree, the balancers should have some voltage sensor on them - but they are built for low cost.

[@StuartPittaway]

Electrical isolation is the problem then - the circuit you describe wires all the cells together in one big parallel battery - there isn't any isolation between the cell chargers.

[@CaptainProton1]

Really dangerous, if your voltage delta is too big it will chuck a lot of amps to shunt and can blow the capacitors causing a fire.

[@StuartPittaway]

I guess that's why the fuse is on the input

[@nicholashowell]

My experience of these is that they don't really work well and are not worth the bother.

[@StuartPittaway]

Was the battery very unbalanced to start with?

[@nicholashowell]

@@StuartPittaway Not as such no. I had built 3, 48v battery packs using old 18650 cells and because all the cell blocks behaved differently they would get out of balance as they discharged. I tried to use active balancers to keep the cell blocks closer in voltage but it didn't work. Often the cell blocks could be different from the next one by 0.1V without the balancer doing anything. When you have a 14s the cell at negative end could be over a volt different from the one at the positive end and the balancer would do nothing. Perhaps it was just the balancers that I was using. I got them from Ali-Express but they were not the same as the one you are looking at.

[@nicholashowell]

@@StuartPittaway Perhaps I should send you one of the ones I was using for you to examine? I'm not using them anymore due to them being rubbish.

[@StuartPittaway]

@@nicholashowell do you have a link to the model you have?

[@sjdtmv]

Thanks Stuart