looking for the simplest balancing circuit

Joined
Apr 25, 2014
Messages
2,782
Location
Sausalito
i had some pcbs made that i attached to cylindrical cells with a spot welder and nickel between and it was good. then I wanted to design it with balance traces to attach a bms. Now i want to just integrate the simplest and most reliable balance circuit on the board where i was fitting traces before in the center. it should fit if i just knew what i was doing. its for a123 iron cells from batteryhookup.com (love that sale and got 480 cells). ive got an 8mm wide strip going the full 500mm length of the pcb where i could fit it...if i knew what i was doing.

ive found a program to convert my stl file to gbr so the pcb house can make it. got a 5 day trial, so 5 days to figure it out.

in the past i figured the conductivity of 2oz copper at this thickness and with vias going to a mirroring of the top and showed to be enough. this design has even wider traces.

its 12s3p
 

Attachments

  • Capture 2.PNG
    Capture 2.PNG
    15.5 KB · Views: 1,059
Do you want to have a balance circuit SMD soldered directly onto the PCB? There are pretty simple individial SMD balancers out there that you just need a resistor for load.
Though with those it wont have any way to disconnect the load from the batteries, only balancing them when fully charged.
Also it's going to be somewhat tricky to get small SMD parts onto a PCB with thick enough traces to handle the current, since thicker gauge traces have less precision at smaller feature scales.

Is there any specific reason why you're doing all of this on a PCB though, I'd be worried about warping in the pack along with current handling.
 
I would just extend balance wires outside the pack, to then be able to do per-cell (group) testing & balancing as periodic maintenance.
 
i ended up just adding balance traces so a balancer can be plugged in. enough trouble getting my step files converted to gerber as is without adding even a single bit on the board

thanks for your thoughts. if any of you is good at designing boards and are interested in getting a battery built with a123 iron 26650 cells id make it and send for my cost n free shipping. (speaking of shipping lithium batteries i found that legally as soon as any are connected the design must be legally approved for shipping and that costs a fortune, but i know of a lot of people selling ion packs they make and sell and ship ground without a problem. these i plan to build sell and ship. theyre iron at least and feel safer doing that and this pcb design, at least a different version i had made, has been pretty tested and i feel safe with it.)
 

Attachments

  • Capture.PNG
    Capture.PNG
    14 KB · Views: 983
"The simplest" is to use batteries that are good enough to not need balancing. Then you only need access to each series level to occasionally check balance manually. I have multiple packs of automotive grade cells still in factory made modules with cells within +/- 0.01V and have never been balance charged since removal from the car. I've spent less time in total manually checking the packs than you would spend connecting a balancing circuit and a BMS, and have none of those additional 100's of points of potential failure.
 
Problem is you only know after EoL whether or not the cells are that well matched.

I personally would not use a pack without balnce leads brought out

BMS needs to be easily replaced, even if not used for balancing

at least the ability to check cell/group level voltages

which then gives you the ability to correct issues if you do see any.

Really, for each individual cell would be best, without physically atomizing the pack.

Just because a pack stays perfectly balanced while still above 80% SoH, is no guarantee that wonderful state of affairs will continue for the remainder of its useful life.

Finally, IMO pack design is best if it allows for shoddy or secondhand cells to safely be accommodated.
 
Ultimately wanted to get the balancers put on the pcb but decided that’s too risky with how brittle components are and the flexing of a pcb on a skateboard. Did add a balance plug. If anyone is interested in (36) 26650 A123 cells put on here I’ll sell them for 250$ From batteryhookup. I forget how much copper it is but remember being equivalent to 12awg and w better heat dissipation. It’s double sided w vias connecting.

This is a great resource for finding people who can design a pcb
https://www.kicad-pcb.org/
 

Attachments

  • B405CFA0-3377-4A6C-B4E9-202F869E3247.png
    B405CFA0-3377-4A6C-B4E9-202F869E3247.png
    69.7 KB · Views: 871
I would have thought the length of the "wiring" path to each cell should be equalized, to avoid creating imbalances?
 
That would be ideal electrically. The design compromises electrical equality in the p group for an ideal mechanical design where the cells being perpendicular allows unrestricted flex of the pcb and no cells touching or straining the nickel connections and safe. One day ill take the p groups apart and see how each cell compares.
 
John in CR said:
"The simplest" is to use batteries that are good enough to not need balancing.

I agree. However, I'd still prefer to use balancing charger or BMS during first few months of using new battery.
I finished recently my 1 kW charger with built-in 100A BMS for my 24S6P A123 26650 battery.

J1u37Ig.jpg
 
LSBW said:
John in CR said:
"The simplest" is to use batteries that are good enough to not need balancing.

I agree. However, I'd still prefer to use balancing charger or BMS during first few months of using new battery.
I finished recently my 1 kW charger with built-in 100A BMS for my 24S6P A123 26650 battery.

J1u37Ig.jpg

LiFePo4 needs balance charging, so the way I do it doesn't work well with your type of pack. Lifepo4 can be in balance and still reflect +/- 0.05V differences between series groups, so it's impossible to know their status without balancing regularly. The cells in the packs I use stay within +/- 0.005V at all points in their SOC, so being a human BMS is quite simple. The packs stay so well balanced and cells are so well matched that simply noticing the increase in voltage sag on the very rare occasions that I surpass 80% DOD helps me avoid needing a cell level LVC. That may change once my packs get 6 or 8 years of service, but by then good BMS's should be cheap and reliable. Plus I could always go the simple route for protection and run 2 volt meters on my packs (one on the top half and one on the bottom half). With that simple info on an even # of cells in series would be all that's required to know if a cell is going low, because the 2 voltages would start to show differences under load or at rest.

With my A123amp20's, I have some incomplete BMS's that are populated only enough to act as balancing boards and bleed off 50mA when voltage is above 3.65V . They've proven reliable as high voltage bleeders, so the dual voltmeter approach combined with the controller's LVC works for me as a simple and effective system. I completely encapsulate the bleeder boards and glue an aluminum heat sink bar across the row of circuits that each burn off less than 0.2W during balancing, and it's proven bulletproof so far with minimal attention required. A couple of times a year I do check cell level voltages during the top-of-charge portion of the charge cycle to make sure everything's working correctly.

I run in the 100's of amps on almost all of my bikes, and until BMS designers lose the idea that full system current should pass through the BMS, they will never become sufficiently reliable for a price I deem appropriate, and there's just no way I'm letting a Battery Murdering System anywhere near my precious batteries. I have no inclination to look at mountains of cell logging data that's available with the newer breeds of BMS's, which is the only way to ensure that they're doing there job instead of murdering your pack, so I'm sticking with my simple and effective manual approach that requires an average of just a few minutes a year of my attention.
 
John in CR said:
That may change once my packs get 6 or 8 years of service, but by then good BMS's should be cheap and reliable. Plus I could always go the simple route for protection and run 2 volt meters on my packs (one on the top half and one on the bottom half).
.................
I run in the 100's of amps on almost all of my bikes, and until BMS designers lose the idea that full system current should pass through the BMS, they will never become sufficiently reliable for a price I deem appropriate,

I wouldn't hold my breath.
Unreliable BMS's was exactly the reason why I avoided them in 2013, instead of using it, I ended making my own charger and balancer for my pack 7 years ago.
Well, come 2020, not much has changed since then.
Granted, BMS I got is def step in right direction. Programmable, BT and PC interface, but 100 Amp only on the discharge side?
Meh.
Chargery has basically same setup as mine, only 25A instead of my 12, with fancy displays and whatnot, but its U$700 delivered!


LruJ9Zt.jpg
 
LSBW said:
John in CR said:
That may change once my packs get 6 or 8 years of service, but by then good BMS's should be cheap and reliable. Plus I could always go the simple route for protection and run 2 volt meters on my packs (one on the top half and one on the bottom half).
.................
I run in the 100's of amps on almost all of my bikes, and until BMS designers lose the idea that full system current should pass through the BMS, they will never become sufficiently reliable for a price I deem appropriate,

I wouldn't hold my breath.
Unreliable BMS's was exactly the reason why I avoided them in 2013, instead of using it, I ended making my own charger and balancer for my pack 7 years ago.
Well, come 2020, not much has changed since then.
Granted, BMS I got is def step in right direction. Programmable, BT and PC interface, but 100 Amp only on the discharge side?
Meh.
Chargery has basically same setup as mine, only 25A instead of my 12, with fancy displays and whatnot, but its U$700 delivered!

LruJ9Zt.jpg

The day I have a charger and bms that costs $620 will be the day I buy an electric car that I didn't build myself, and I just don't see that happening unless when I meet with the transit police chief he rules that 3 wheelers will be treated like cars and not electric 2 wheelers. The quite generous law we have where a 5kw motor is the cutoff above which they're electric motorcycles specifically says 2 wheels, so I want to get a letter from him saying 3 wheel electrics will be treated the same as those with 2. 3 wheel human powered vehicle are fairly common for small cargo and vendors, and 3 wheel gassers are treated like motorcycles not cars, so I am somewhat confident that he'll agree that 3 should be treated like 2.
 
I just got a bunch of lifepo4 cells and am wondering what you mean when you say this:
“ Lifepo4 can be in balance and still reflect +/- 0.05V differences between series groups, so it's impossible to know their status without balancing regularly.“
Are you saying lifepo4 cells are different than others in some way?
Maybe u got the same cells as me from batteryhookup


u can get a bms that sends cell voltages to ur phone for like 50$. And not discharge through it
 
Good quality LFP in good condition (ideally new) may not need **any** balancing, for a decade or more.

Depends on the use case of course.

New from the factory they all arrive **perfectly** balanced at the mid-point.

By moving to top- or bottom- balancing

and harmfully (or at least pointlessly) seeking to maximize cap utilization

you then **cause** the need for more frequent, or at least regular balancing.

By all means check regularly, even monitor at the cell level while in use and while charging

but only balance when actually necessary.
 
I very generally, agree completely with you John, please take the below nit picking as educational only, not at all arguing with you specifically.

John in CR said:
LiFePo4 needs balance charging
Only true for those specific packs that do.

Certainly not true as written. Shall we say, an overgeneralisation.

> impossible to know their status without balancing regularly

Wut?

What do you mean by status there?

The cells in the packs I use stay within +/- 0.005V at all points in their SOC

The very definition of no balance needed.

Even true if that is only true at your selected "target voltage" for checking balance.

I 100% agree about cell-level LVC properly cared for, a decent pack does not need that

only when abusively low DoD is required, as a last-ditch failsafe.


> until BMS designers lose the idea that full system current should pass through the BMS, they will never become sufficiently reliable for a price I deem appropriate

+1000

to that and the rest


 
Theres a new style of balancer I've been seeing on youtube, and I think they are called active balancers. When the bulk charge is complete to 4.1V per cell (give or take), the circuit then feeds the highest cell voltage to the lowest cell voltage.

If all the cells are within 0.02V (or whatever the parameter is), the circuit does nothing.

I dont know if they are reliable, or even a good idea, but they look very simple. After seeing many conventional BMS failures, I am wondering if their supposed simplicity might also make them more reliable.
 
There are literally dozens of different types all over eBay and Ali etc, seen also on Amazon priced higher.

called "non protective BMS" or equalizers but I prefer calling them dedicated balncers.

Better ones go for hundred$ at high amp balance rates, high voltage S-counts like 24S, ability to balance at any voltage, and especially if boxed with leads all ready to use.

Others are modular, about $20 per cell, just buy what you need.

Some are still bleed-resistance based, rather than shuttling energy transfer.

Many of the latter only transfer to adjacent cells, not direct to the lowest.

The term "active" to me, means the balance rate is not dependent on the voltage delta, if it is a 5A balancer it does so at that rate right up to the balanced point and the stops.

Passive ones start out fast when the difference is greater, get slower as the gaps close.

That Prowse YouTuber published a video complaining they don't do what he expected, which had nothing to do with the purpose of balancing in the first place, but that is par for the course for YT.
 
john61ct said:
I very generally, agree completely with you John, please take the below nit picking as educational only, not at all arguing with you specifically.

John in CR said:
LiFePo4 needs balance charging
Only true for those specific packs that do.

The cells in the packs I use stay within +/- 0.005V at all points in their SOC

What brand LiFePo4 cells do you use? My experience with A123 and other is that due to the surface charge at the top that has only a tiny bit of capacity, it causes voltage differences so I can't tell if the pack is in balance or not. If they're advanced enough now that packs stay in perfect balance for long periods like the Ford and Chevy cells I used, then it would be good info to know.
 
I don't take LFP CV past 3.45Vpc, any higher is no point as you say no useful energy stored there.

Resting 3.33-3.35Vpc is Full.

Also as stated the balancing point can be at midpoint, or bottom, no reason to balance at the top, many reasons not to.

Nothing to do with any advances true for nearly a dozen years now

nor super levels of quality, even applies to a set of cells with mismatched capacity.

Just realize and accept that the weakest cell / group dictates the pack's usable capacity, and ensure your usage stays away from the voltage "shoulders" at both ends.

Makers I think well proven now, besides A123, include Winston (Thundersky), Sinopoly, CALB and GBS.

Wish there were more, especially in those tiny cylindricals. . .

But I guess if one has lower ambitions for maximizing longevity, being happy with 3-5000 cycles, say 10-12 years, then there's a whole mid-tier of lesser players, and that's where I'm trying to learn more.


 
john61ct said:
John in CR said:
john61ct said:
I very generally, agree completely with you John, please take the below nit picking as educational only, not at all arguing with you specifically.

John in CR said:
LiFePo4 needs balance charging
Only true for those specific packs that do.

The cells in the packs I use stay within +/- 0.005V at all points in their SOC

What brand LiFePo4 cells do you use? My experience with A123 and other is that due to the surface charge at the top that has only a tiny bit of capacity, it causes voltage differences so I can't tell if the pack is in balance or not. If they're advanced enough now that packs stay in perfect balance for long periods like the Ford and Chevy cells I used, then it would be good info to know.

Resting 3.33-3.35Vpc is Full.

Also as stated the balancing point can be at midpoint, or bottom, no reason to balance at the top, many reasons not to.

...even applies to a set of cells with mismatched capacity.

Okay, I see now the other post was hyperbole, since in one post 0.005 doubled to 0.010, and throwing "mismatched capacity" into the mix means you're dropping "all points in their SOC". While I understand you're able to run without a BMS, LiFePo4 does require somewhat more monitoring and periodic balancing.
 
john61ct said:
I don't take LFP CV past 3.45Vpc, any higher is no point as you say no useful energy stored there.

Resting 3.33-3.35Vpc is Full.

Also as stated the balancing point can be at midpoint, or bottom, no reason to balance at the top, many reasons not to.

Nothing to do with any advances true for nearly a dozen years now

nor super levels of quality, even applies to a set of cells with mismatched capacity.

Just realize and accept that the weakest cell / group dictates the pack's usable capacity, and ensure your usage stays away from the voltage "shoulders" at both ends.

Makers I think well proven now, besides A123, include Winston (Thundersky), Sinopoly, CALB and GBS.

Wish there were more, especially in those tiny cylindricals. . .

But I guess if one has lower ambitions for maximizing longevity, being happy with 3-5000 cycles, say 10-12 years, then there's a whole mid-tier of lesser players, and that's where I'm trying to learn more.

Having a midpoint balance is a "problem" with LiFe, since it's their advantage - the %SOC is extremely hard to detect in the midpoint since the voltage is at around 3.3V, and with the accuracy of voltmeters and wire resistances, you may find that even 0.05V difference in midpoint %SOC can result in noticeable difference when you are fully charged or fully discharged.
This is why I think a top balance is the most accurate, and it gives you the ability to charge the pack without having to monitor the individual cells on every charge since this is the chosen balance point. (And charging to no more than 3.45V as you say)
Actually, charging to 3.4V is already full, just takes longer - but then if you start to have balance issues, then if several cells "pushes" themselves to 3.45V then it means some other group would be at 3.35V, and then that second group is quite far in %SOC from the first, shortening your range significantly. (This is an extreme example but you get my drift - even 3.38V after charge is still not full)
 
spinningmagnets said:
Theres a new style of balancer I've been seeing on youtube, and I think they are called active balancers. When the bulk charge is complete to 4.1V per cell (give or take), the circuit then feeds the highest cell voltage to the lowest cell voltage.

If all the cells are within 0.02V (or whatever the parameter is), the circuit does nothing.

I dont know if they are reliable, or even a good idea, but they look very simple. After seeing many conventional BMS failures, I am wondering if their supposed simplicity might also make them more reliable.

I considered those, and they waste almost no power since they just shuttle the energy via capacitors.
But due to the accuracy of measurement, and if the voltage on the BMS wires sags due to significant resistance on the connection terminals, then you risk putting the pack away from being balanced...
This is only good, in my opinion, while letting the battery idle, or, during a very very slow charge. Never while in use.
 
thunderstorm80 said:
Having a midpoint balance is a "problem" with LiFe, since it's their advantage - the %SOC is extremely hard to detect in the midpoint since the voltage is at around 3.3V, and with the accuracy of voltmeters and wire resistances, you may find that even 0.05V difference in midpoint %SOC can result in noticeable difference when you are fully charged or fully discharged.

Sorry but it is most definitely **not** a problem IRL.

From a practical POV just use cell/group voltage as an accurate proxy for **relative** percent SoC.

Actual Ah remaining will **always** vary, to some extent.

The **only** point of balancing is to keep the inevitable imbalances from reducing pack capacity to lower than the weakest cell / group.

It is impossible to remove that upper limit, tryi.g to do so will just lead to problems, or drive you crazy.

Of course you can choose to top balance because you are most concerned about pushing cell level V too high and do not want to measure that while charging.

But 100% false to say that is "more accurate".

And then you **better** ensure you do not ever let cell level V go to low when discharging, because you have created much greater differences at that end.

IMO woulfd be foolish not to have BMS based LVC, unless you left a **lot** of range in there every trip.

If you want both maximum range and good safety without a BMS cutoff, then balance at the bottom, and use cell level to to HVC when charging, accepting imbalance at the top.


> Actually, charging to 3.4V is already full, just takes longer - but then if you start to have balance issues, then if several cells "pushes" themselves to 3.45V

Charging to 3.40 means it is impossible for any cell to go higher.

> even 3.38V after charge is still not full

False, if you mean after resting 24+ hours.

If there is any **voltage** higher than even 3.35V after that, it is surface charge only, getting there only damages cells (reduces lifespan, even if only a few percentages)

and does not actually result in any significant Ah actually stored.

 
How safe are the a123 cylindrical 26650 green B cells? I do know that if u drill through them they’ll ignite but at least won’t light the rest of the pack on fire..wasn’t me but my neighbor). But would it be possible to get them to ignite from purely electrical damage such as over charging or discharging? Tempted to see if I could cycle one a couple times from maybe 2 to 6 volts and then short it without fire. I’m guessing it won’t catch fire
 
Back
Top