Thanks a lot. There are so many equalizer boards, but nearly no videos like yours to describe the strategy behind them. Will be great if you continue. For example, describe and compare other equalizer types like inductive to capacitive and etc. Great job done 👍
You present so excellently and so clearly even a grader will understand how active balancers work! Keep it up and I hope you come up with more similar vids.
I personally prefer the serial balancers (I use to call them shuffle balancers) which balance a certain cell with its neigboring cells. They usually work as fast as the serial / parallel balancers because typically there are just a few cells in the chain out of balance. And these balancers require just half of the switches and s-1 capacitor banks to work. They require just a simple clock generator and 2 high current MOSFET drivers to work. They have furthermore less inner resistance. I build them by myself and I am happy with their performance. The clockfrequency I use is relatively high (50 kHz) in order to take advantage from a nearly constant balancer current which is proportional to the voltage difference between the neighboring cells and the cell to sink or source the current. Due to the lower inner resistance of the circuit I can reach 15 to 20 A balancer current at a voltage difference of 0.4 to 0.5 V with commercially availlable MOSFETs and low-ESR electrolytics (the last time I have used 2200 µF, 4 V electrolytics with 7 mOhm ESR 4 in parallel to achieve 1.75 mOhm ESR and 8800 µF and n-channel MOSFETs with an Ron of 1.1 mOhm and a total circuit inner resistance of 5 to 6 mOhm to balance LiFePO-cells with an inner resistance of 7 mOhm).
Thanks for the video. It's very clear and helpful. The active balancer will only balance cells. For those who use LiFePO4/LTO cells on car battery, it will not help reducing the over voltage by car generator. You need passive balancer.
Passive balancer only bleeds a small amount of current which means if your charge current is above the bleed current... you battery will still overvolt and that will be bad. Plus you're bleeding power for no real reason. Even a cheap BMS with a low passive balancer would be able to stop the charge if they overvolt so that would be needed in addition to the active balancer which will balance the cells much faster and at all voltages. A better option that works more efficiently would be a buck converter that takes the variable voltage of an alternator (around 14 volts but it's not very stable if loads or RPMs change rapidly) and converts it to a stable voltage within the range of your pack.
@@asificam1 For LTO battery, over voltage is ok for a short period of time. Passive balancer is the only way to lower the voltage, and you can increase the current if more heat dissipation is allowed. Some BMS do OVP by cutting the input from alternator. This is not advised on vehicles as the voltage will suddenly jump up and could be harmful to some devices.
@@phuang3 Oh, for the sole battery in the vehicle... yeah cutting power would not be advised then... alternators need a buffer of some type as they're not overly stable. How much current is going into the cells once charged? I think that might depend on max pack voltage and the typical (~14.1 V) voltage of an alternator. At a small enough voltage difference and a sufficiently large passive balancer with enough heat dissipation I agree it would work... It just sounds so wasteful to just dump power to maintain the battery.
Wow! Thanks for making it so clear. Please do build a bigger one. You will have to make sure though, bigger doesn't always mean better because bigger capacitors will take more time to charge and discharge which means the frequency of switching will drop accordingly. if it is in kHz now then it could be in 0.1kHz. I guess we will never know until we simulate or build a circuit. Please do. Thanks
Nice Explanation. The power loss for an active balancer like this is roughly the (switching frequency)*(single capacitance value)*(variance of cell voltages). The switching frequency should be roughly 1/(10*Rds,on*C) to ensure the capacitor voltages reach the cell voltages and the cap voltages have time to equalize when they are all connected in parallel. Higher switching frequencies in general, allow you to drive down the voltage differences between the cells (decrease the variance) while the cells are charging or discharging, improving efficiency. Minimizing the capacitor values is also important for good efficiency. The Rds,on of the mosfets will determine the "rate of charge imbalance" the system can correct for. Lower Rds,ON allows for faster charging/discharging of the string while still keeping it near perfect balance. There might be a power advantage of running a balancer like this at a low duty cycle (10%) to minimize the power draw of the balancer itself. You might need to only turn it on for a small number of short bursts (20?) through the entire charge/discharge cycle, especially if the cells are well matched. Just remember, you really only need a balancer when one (or more) of the cells in the string is approaching the min (2.5V) or max (4.2V) cell voltage (that includes under high current conditions). So, you can use a balancer to get more charge into and out-of your battery pack, but you probably shouldn't be pushing your battery pack this hard anyway if you are trying to maximize the life of your pack. You can effectively prevent over-charging and over-discharging the old-fashoned way, by just monitoring your battery pack's voltage and not pushing it to the limits. Just curious, what are the capacitance values and mosfet Rds,ON (or part number) for this circuit?
Great video, thank you! These devices have gotten so cheap... I picked up a 21s one for only 36 bucks shipped, took 10 days to arrive from china. These boards, in combination with a cheapo regular dissipative BMS (usually I find one with a UART) are great... the active balancer keeps the pack balanced, the cheapo bms reports cell voltages to whatver microcontroller you use to monitor the pack...
Nice video, thank you! Thinking about the simpliest way to implement balancer with same principle lead me to idea that probably this should be combined with cell monitoring function. It should work on the principle - cell monitoring connects the highest voltage cell in the pack to capacitor. Then it diconects from higest voltage cell and connects the capacitor to the lowest voltage cell. Think most ot smart BMSs and balancers work on this principle. Some one was wasking for soic 8 which is probably gate driver chip and most probably is XJNG2103 (chinese brand with inverted outputs).
The switch marked series puts the caps in series, yet each cap is paralleled with it's adjacent cell. I should have made it more clear, it's 2p8s at this point. The caps are in series, but they are paralleled to the cells. Thanks
@@BradCagle yes, but it seems closer to 2P4S. But not even that, because every cell is parallel to it's adjacent cap. So there are extra connections which don't exist in a 2P4S. i think it can't be described with xPxS notation.
@@johnaweiss Sorry yes the diagram is 2p4s, the actual balancer I had in my had was 8s. I'm commenting without re-watching the video and making mistakes LOL
Right, did I misspeak? The caps have a series mode, and a parallel mode. When the caps are in series mode, each cap is paralleled to it's adjacent cell, in this case its like a 2p8s config. Then the caps go to parallel mode to avg the voltage across them all, at which time they are no longer connected to the battery cells.
These work great but they are always on. You want them to turn on at above 3.4 volts per cell and off below that. Better to have it off at float (timer?) There is a provision for a switch (remove solder jumper). A charge control relay can be employed but most are inaccurate in that range. I got a box full. Check the specs, you need a resolution of .01 volt, .10 is no good. Balancing below 3.4 volts can give you all kinds of balancing issues.
Yeah, they are on all the time. You probably don't want to put a battery in storage with one of these connected. That being said, I've used these in several batteries and have had 0 problems with them running all the time. Just really nice and balanced packs, and helps to keep those cells that like to hit 3.65 before the others more tame.
@@BradCagle I use 16S version Heltec on 32kwh system and it works well as long as you don’t try to do any balancing below 3.4 volts. Let the cells go where they want below and up to that point. If you try to balance at say 1/2 charge you’ll be forever playing cell voltage bop-a-mole until you disassemble and do a good top balance again. They will balance at full charge but the closer they get to balance the less balancing power they have. It could take several hours. If you’re throwing a lot of charge current and you are approaching full charge even a good bank can go a bit wonky. To reduce that, just don’t charge over 3.525 per cell - bulk with little to no absorption time. Little capacity to be gained and a lot less headaches. I’m going with a new type that has up to 4 amps of balancing power even at low cell differences. The latest version has an app for Apple instead of Android. I think Neey (2~24S 4 amp) makes it or it’s a knock off. EBay for around 160$. Should have it the first of the year.
@@SkypowerwithKarl Interesting, I never experienced the bop-a-mole in the middle of state of charge. I have a 4s 100ah pack with a 4s balancer, it gets fully cycled everyday so I might check to see if its gone out of balance. Yeah, my sweet spot for lifp04 is about 3.45v per cell this keeps the cells out of that steep curve where one or two tend to rocket up, and still gets the pack fully charged. I'm still kicking the idea around of building my own balancer with higher power MOSFETS, and maybe super caps. Toying around with the series parallel switching frequency would be fun to see how much power I can get it to transfer. I would use an Arduino to control it so would have some programmability smarts.
@@BradCagle Sound like you know what you are doing. I know just enough NOT be dangerous lol. There are a ton of balancers out there now, four times as many as a year ago. The Neey brand seems to be the one to beat. (4 amps at near full state of charge) It uses super caps and uses whole battery to assist. If you are doing it for your own entertainment, great but for market, I think you got way too much competition. The fine tuning through the apps is phenomenal. Where there’s a huge need is in affordable high current top balancing. To parallel top balance four 280ah cells from the factory voltage (no series pre/charge) will take a week with a quality 10amp power supply, no banana plug, # 8 wire and lugs. To top balance sixteen 310ah… not gonna happen with 10 amps. You need a minimum of 60 amps, preferably 80. (Only limited by terminal resistance/connections). Build a CC switching power supply that first charges to 3.4 then absorb, then to 3.5, absorbs then to 3.65 volts and cuts off. It shouldn’t fold back too early before reaching set voltage light cheap power supplies. Do that and the world will be beating a path to you. Look at my cluj “Fast 80 amp top balancing for your LiFePO4 cells” on RUclips. I’m sure you can do better.
Ok, got it. On the other hand this means that if you connect two of those to your battery, and the high frequency cycles are not in sync, they might just push current back and forth from balancer to balancer, right? Or does connecting those in parallel to one battery work?
I've thought about this, and it may be that the sync would be a problem. It might also be possible to find the control signal on one board, and route it to others.
The way I understand it is: Any scheme that uses switching to reconfigure the connection of the capacitor (or capacitors) as in this case from series to parallel, and back is called flying capacitor. The number of capacitors used is not a factor. The scheme is called "Flying" since the caps do not have a fixed electrical position in the circuit. Thanks
Thanks a lot. There are so many equalizer boards, but nearly no videos like yours to describe the strategy behind them.
Will be great if you continue. For example, describe and compare other equalizer types like inductive to capacitive and etc.
Great job done 👍
The best explanation I found on RUclips.
Thanks! Glad to be helpful :)
Extremely good explanation, i have very limited electrical knowledge and this was super easy to understand.
You present so excellently and so clearly even a grader will understand how active balancers work! Keep it up and I hope you come up with more similar vids.
Thanks!
This video has help me get over the hump of confusion, and helped me understand how these things work! Your diagram is especially helpful.
Glad it's useful. Thanks for watching!
I personally prefer the serial balancers (I use to call them shuffle balancers) which balance a certain cell with its neigboring cells. They usually work as fast as the serial / parallel balancers because typically there are just a few cells in the chain out of balance. And these balancers require just half of the switches and s-1 capacitor banks to work. They require just a simple clock generator and 2 high current MOSFET drivers to work. They have furthermore less inner resistance. I build them by myself and I am happy with their performance.
The clockfrequency I use is relatively high (50 kHz) in order to take advantage from a nearly constant balancer current which is proportional to the voltage difference between the neighboring cells and the cell to sink or source the current. Due to the lower inner resistance of the circuit I can reach 15 to 20 A balancer current at a voltage difference of 0.4 to 0.5 V with commercially availlable MOSFETs and low-ESR electrolytics (the last time I have used 2200 µF, 4 V electrolytics with 7 mOhm ESR 4 in parallel to achieve 1.75 mOhm ESR and 8800 µF and n-channel MOSFETs with an Ron of 1.1 mOhm and a total circuit inner resistance of 5 to 6 mOhm to balance LiFePO-cells with an inner resistance of 7 mOhm).
That's really awesome! So you've got it pretty tuned. That's my favorite, I love to optimize :) Thanks for sharing.
Bloody top job of describing how they work, well done!
Thank you. I'm a fan of your work!
Best video on the subject.
No views.
Typical.
Thanks for the video. It's very clear and helpful. The active balancer will only balance cells. For those who use LiFePO4/LTO cells on car battery, it will not help reducing the over voltage by car generator. You need passive balancer.
Yes it will only balance cells. You need a BMS still to protect from over/under voltage on each cell
Passive balancer only bleeds a small amount of current which means if your charge current is above the bleed current... you battery will still overvolt and that will be bad. Plus you're bleeding power for no real reason. Even a cheap BMS with a low passive balancer would be able to stop the charge if they overvolt so that would be needed in addition to the active balancer which will balance the cells much faster and at all voltages. A better option that works more efficiently would be a buck converter that takes the variable voltage of an alternator (around 14 volts but it's not very stable if loads or RPMs change rapidly) and converts it to a stable voltage within the range of your pack.
@@asificam1 For LTO battery, over voltage is ok for a short period of time. Passive balancer is the only way to lower the voltage, and you can increase the current if more heat dissipation is allowed. Some BMS do OVP by cutting the input from alternator. This is not advised on vehicles as the voltage will suddenly jump up and could be harmful to some devices.
@@phuang3 Oh, for the sole battery in the vehicle... yeah cutting power would not be advised then... alternators need a buffer of some type as they're not overly stable. How much current is going into the cells once charged? I think that might depend on max pack voltage and the typical (~14.1 V) voltage of an alternator. At a small enough voltage difference and a sufficiently large passive balancer with enough heat dissipation I agree it would work... It just sounds so wasteful to just dump power to maintain the battery.
If only I could could double-like your video!!! Thank you.
You're welcome
I feel the same.
Wow! Thanks for making it so clear. Please do build a bigger one. You will have to make sure though, bigger doesn't always mean better because bigger capacitors will take more time to charge and discharge which means the frequency of switching will drop accordingly. if it is in kHz now then it could be in 0.1kHz. I guess we will never know until we simulate or build a circuit. Please do. Thanks
You're welcome. Right, I'm not entirely sure bigger will be any better, would be interesting to experiment though!
Nice Explanation. The power loss for an active balancer like this is roughly the (switching frequency)*(single capacitance value)*(variance of cell voltages). The switching frequency should be roughly 1/(10*Rds,on*C) to ensure the capacitor voltages reach the cell voltages and the cap voltages have time to equalize when they are all connected in parallel. Higher switching frequencies in general, allow you to drive down the voltage differences between the cells (decrease the variance) while the cells are charging or discharging, improving efficiency. Minimizing the capacitor values is also important for good efficiency. The Rds,on of the mosfets will determine the "rate of charge imbalance" the system can correct for. Lower Rds,ON allows for faster charging/discharging of the string while still keeping it near perfect balance.
There might be a power advantage of running a balancer like this at a low duty cycle (10%) to minimize the power draw of the balancer itself. You might need to only turn it on for a small number of short bursts (20?) through the entire charge/discharge cycle, especially if the cells are well matched.
Just remember, you really only need a balancer when one (or more) of the cells in the string is approaching the min (2.5V) or max (4.2V) cell voltage (that includes under high current conditions). So, you can use a balancer to get more charge into and out-of your battery pack, but you probably shouldn't be pushing your battery pack this hard anyway if you are trying to maximize the life of your pack. You can effectively prevent over-charging and over-discharging the old-fashoned way, by just monitoring your battery pack's voltage and not pushing it to the limits.
Just curious, what are the capacitance values and mosfet Rds,ON (or part number) for this circuit?
Great video, thank you! These devices have gotten so cheap... I picked up a 21s one for only 36 bucks shipped, took 10 days to arrive from china. These boards, in combination with a cheapo regular dissipative BMS (usually I find one with a UART) are great... the active balancer keeps the pack balanced, the cheapo bms reports cell voltages to whatver microcontroller you use to monitor the pack...
I love these. Great for packs with dodgy cells. Thanks!
Thanks for making this video, great explanation. It is much appreciated.
You're Welcome!
Thanks alot for understanding. Love from pakistan
Thanks!
Nice video, thank you! Thinking about the simpliest way to implement balancer with same principle lead me to idea that probably this should be combined with cell monitoring function. It should work on the principle - cell monitoring connects the highest voltage cell in the pack to capacitor. Then it diconects from higest voltage cell and connects the capacitor to the lowest voltage cell. Think most ot smart BMSs and balancers work on this principle. Some one was wasking for soic 8 which is probably gate driver chip and most probably is XJNG2103 (chinese brand with inverted outputs).
Thanks. This video and these comments helped a lot.
Glad it helped!
Yes! Explanation was amazing! Thank you!
Anytime!
It's a very helpful video. Thank you.
I learned a lot from your video.
Glad it was helpful!
Great explanation 👌 👏👏👏
Glad you liked it
Thanks..Nice explanation..
Do we connect it with batteries the same way as we connect balancing connectors?
Great explanation thanks
You are a genius bro.
3:45 It appears when the switches marked "series" are closed, that puts each cap in parallel with it's adjacent cell, not series.
The switch marked series puts the caps in series, yet each cap is paralleled with it's adjacent cell. I should have made it more clear, it's 2p8s at this point. The caps are in series, but they are paralleled to the cells. Thanks
@@BradCagle yes, but it seems closer to 2P4S. But not even that, because every cell is parallel to it's adjacent cap. So there are extra connections which don't exist in a 2P4S. i think it can't be described with xPxS notation.
@@johnaweiss Sorry yes the diagram is 2p4s, the actual balancer I had in my had was 8s. I'm commenting without re-watching the video and making mistakes LOL
@@BradCagle Your diagram seems correct, but i think the connections aren't 2p4s.
7:28 Yes, except this balancer never puts the cells in parallel with each other, only with the caps.
Right, did I misspeak? The caps have a series mode, and a parallel mode. When the caps are in series mode, each cap is paralleled to it's adjacent cell, in this case its like a 2p8s config. Then the caps go to parallel mode to avg the voltage across them all, at which time they are no longer connected to the battery cells.
Thanks
You're welcome
mantab
These work great but they are always on. You want them to turn on at above 3.4 volts per cell and off below that. Better to have it off at float (timer?) There is a provision for a switch (remove solder jumper). A charge control relay can be employed but most are inaccurate in that range. I got a box full. Check the specs, you need a resolution of .01 volt, .10 is no good. Balancing below 3.4 volts can give you all kinds of balancing issues.
Yeah, they are on all the time. You probably don't want to put a battery in storage with one of these connected. That being said, I've used these in several batteries and have had 0 problems with them running all the time. Just really nice and balanced packs, and helps to keep those cells that like to hit 3.65 before the others more tame.
@@BradCagle
I use 16S version Heltec on 32kwh system and it works well as long as you don’t try to do any balancing below 3.4 volts. Let the cells go where they want below and up to that point. If you try to balance at say 1/2 charge you’ll be forever playing cell voltage bop-a-mole until you disassemble and do a good top balance again. They will balance at full charge but the closer they get to balance the less balancing power they have. It could take several hours. If you’re throwing a lot of charge current and you are approaching full charge even a good bank can go a bit wonky. To reduce that, just don’t charge over 3.525 per cell - bulk with little to no absorption time. Little capacity to be gained and a lot less headaches. I’m going with a new type that has up to 4 amps of balancing power even at low cell differences. The latest version has an app for Apple instead of Android. I think Neey (2~24S 4 amp) makes it or it’s a knock off. EBay for around 160$. Should have it the first of the year.
@@SkypowerwithKarl Interesting, I never experienced the bop-a-mole in the middle of state of charge. I have a 4s 100ah pack with a 4s balancer, it gets fully cycled everyday so I might check to see if its gone out of balance. Yeah, my sweet spot for lifp04 is about 3.45v per cell this keeps the cells out of that steep curve where one or two tend to rocket up, and still gets the pack fully charged. I'm still kicking the idea around of building my own balancer with higher power MOSFETS, and maybe super caps. Toying around with the series parallel switching frequency would be fun to see how much power I can get it to transfer. I would use an Arduino to control it so would have some programmability smarts.
@@BradCagle
Sound like you know what you are doing. I know just enough NOT be dangerous lol. There are a ton of balancers out there now, four times as many as a year ago. The Neey brand seems to be the one to beat. (4 amps at near full state of charge) It uses super caps and uses whole battery to assist. If you are doing it for your own entertainment, great but for market, I think you got way too much competition. The fine tuning through the apps is phenomenal. Where there’s a huge need is in affordable high current top balancing. To parallel top balance four 280ah cells from the factory voltage (no series pre/charge) will take a week with a quality 10amp power supply, no banana plug, # 8 wire and lugs. To top balance sixteen 310ah… not gonna happen with 10 amps. You need a minimum of 60 amps, preferably 80. (Only limited by terminal resistance/connections). Build a CC switching power supply that first charges to 3.4 then absorb, then to 3.5, absorbs then to 3.65 volts and cuts off. It shouldn’t fold back too early before reaching set voltage light cheap power supplies. Do that and the world will be beating a path to you. Look at my cluj “Fast 80 amp top balancing for your LiFePO4 cells” on RUclips. I’m sure you can do better.
@@SkypowerwithKarl oh, anything i do is not for profit, it would be released to the community for DIY, and learning.
Ok, got it. On the other hand this means that if you connect two of those to your battery, and the high frequency cycles are not in sync, they might just push current back and forth from balancer to balancer, right?
Or does connecting those in parallel to one battery work?
I've thought about this, and it may be that the sync would be a problem. It might also be possible to find the control signal on one board, and route it to others.
@@BradCagle Good point! Hoping that an electronics-savvy reads this ..
8 pins dual mosfet code ? have any idea?
I don't, I figured they are mosfet drivers maybe.
Hi, I have a curiosity, could you tell me the printed code of the 8-pin component? (soic-8)
Thanks!
I'll take a look, and see if there is anything printed on it.
@@BradCagle Thanks a lot !
@@IL_Praticone No problem. I'll check it when I get home from work.
Sorry took so long. There are no markings on the connectors. I measured the pin pitch (spacing) at 4mm centers, and the pins are about 1mm thick.
@@BradCagle sorry I didn't explain myself well, I meant the code above the black electronic component with 4 pins
good scheme, but it is not flying capacitor scheme, in flying capacitor scheme there is only one capacitor
The way I understand it is: Any scheme that uses switching to reconfigure the connection of the capacitor (or capacitors) as in this case from series to parallel, and back is called flying capacitor. The number of capacitors used is not a factor. The scheme is called "Flying" since the caps do not have a fixed electrical position in the circuit. Thanks
Sad