Sid Battery Balancers.

Forgive my ignorance....
Do balancers really get exercised all that much after the first few cycles?

Once the first few charges are done, i'd think there'd be little more than a "teacup" worth of work for balancers to do for ongoing cycles - basically as the cells age and degrade at different rates, the balancers have only the variation of degradation to deal with.

Would Steve's idea not work better in situations where full-charge is avoided for lifespan preservation? For example, my system may not reach 100% for dozens of cycles in a stretch.
Top balancers require a cell to reach 100%, which some of us specifically avoid to increase the lifespan of the bank.

So like I keep reiterating, an "ideal" BMS on an "ideal" pack would have no work to do--in which case basically any design would work.

 But for where the rubber meets the road, with older (read: cheaper) battery packs that may or may not be matched up very well, etc., etc., the smaller/lighter systems fail miserably.  Sean's been down that road--for whatever reason, his Li-Ion battery banks constantly go out of balance on a day-to-day basis.  The only thing that's been able to prevent them from blowing themselves up due to overvoltage has been a set of 5A shunt balancers.  A Chargery BMS simply couldn't do the job.

In my case, with a homebrew LiFePo4 bank, the balancers worked hard the first few charge cycles--and now they never even get the slightest warm.

...but for another LiFePo4 bank that a stupid mouse ruined (overdischarge), it NEVER stays in balance.

So YMMV...it's more of "fitting the right tool to the right usecase."  Also keep in mind that cells are not guaranteed to degrade at the same rate--and the cell that degrades the fastest very quickly takes the brunt of heavy cycling.  If the balancing can't keep up, this cell will fail very quickly due to overvoltage.

Depends more on battery chemistry than anything else.  Li-Ion are relatively easy to balance, but LiFePo4 are very difficult to balance with a gentle method--mostly because there is basically no cell voltage change between 20-80% SOC.  (Yes there's going to be a tiny amount--but when you put that up against cell resistance and varying loads of a house application, the SOC voltage change is basically impossible to measure!)

I guess my responses are more along the lines of responding to someone who's trying to say that there's a better way to do a particular task--but without considering the situations/cases where the "better" method actually falls far short.

But I'm not saying that the "better way" is useless (at all!)  If that's all your batteries need--by all means use it!  I do agree that shifting power around is far better than just burning it up in heat.  It was only after I took into account the technical challenges (as well as the limitations) of the power-shifting methods, that I ended up settling on the shunt "total loss" method.

At least for Sean, though, he doesn't have an option!  Another customer couldn't get his huge battery bank balanced with the Chinese BMS units--likely because the leakage current of his huge bank exceeded the balance current of the BMS units.  It just took a few days for the Genetry balancers to heave his huge bank into balance, though...

In my case I'm running 12v replacement lifepo4 batteries and the balancer I got works well powermr 4 battery 10A max (don't know the method)

I also connected them all in parallel for some time before building the bank

Without the balancer my pack stayed balanced for weeks within 10mV per battery

So i have a system that needed very little  balancing 

I also don't generally reach 100% so balance needed to happen at running voltage

Although the loads effect voltage more than charge state the loads are effecting all the batteries at once effectively making it null

If you can provide a photo of the board, it should be quite easy to determine the balancing method.

So the important keyword here is that you have a LiFePo4 battery bank.

Unless your cells are >3.4vpc when resting (i.e. 3.4 * 4 = 13.6v, so above 13.6v), all of the above statements will ALWAYS be true providing all of the cells are somewhere between ~20-80% SOC.  In other words, even with potentially up to 60% SOC imbalance, the cells will "appear" balanced at all times, with a voltage of ~3.2-3.3v each (LiFePo4 nominal voltage is 3.2v.)

It's only when the cells reach near-full that their voltage suddenly (i.e. over a few minutes, when compared to a 5-6 hour charge period) will start to rise past 3.4v.  If the battery pack is under heavy charge at that point, the voltage will just keep steadily going up and up well past 5-6v (I've heard of 12v per cell or more) if a balancer does not bypass/shunt the excess power around the cell.

Fortunately, LiFePo4 is a generally very safe chemistry--and serious overvoltage generally doesn't result in any fire/explosion or other hazards.  (Unlike Li-Ion or Li-Po.)

If you take your LiFePo4 bank to even 3.5vpc (3.5 * 4 = 14v), you'll be able to tell if they're balanced.  "Low" cells will still be at ~3.4v.

That sounds right

Each battery is 12.8 nom

14.4 +-.2 charging max

Each has internal BMS so I only need to be concerned with the balance between batteries not cells

When charging each will run around 13.5 and slowly climb to 14 than suddenly jump

The charge state can be read off the voltage but within a very narrow band so voltage equalization is a valid way to balance them

My balancer is a sealed box with 8 wires on it, 2 per battery

I can hear it running and it keeps them under 10mV (0.00 difference by my meter)

7 hours ago, Sid Genetry Solar said:

So like I keep reiterating, an "ideal" BMS on an "ideal" pack would have no work to do--in which case basically any design would work.

 But for where the rubber meets the road, with older (read: cheaper) battery packs that may or may not be matched up very well, etc., etc., the smaller/lighter systems fail miserably.  Sean's been down that road--for whatever reason, his Li-Ion battery banks constantly go out of balance on a day-to-day basis.  The only thing that's been able to prevent them from blowing themselves up due to overvoltage has been a set of 5A shunt balancers.  A Chargery BMS simply couldn't do the job.

In my case, with a homebrew LiFePo4 bank, the balancers worked hard the first few charge cycles--and now they never even get the slightest warm.

...but for another LiFePo4 bank that a stupid mouse ruined (overdischarge), it NEVER stays in balance.

 

So YMMV...it's more of "fitting the right tool to the right usecase."  Also keep in mind that cells are not guaranteed to degrade at the same rate--and the cell that degrades the fastest very quickly takes the brunt of heavy cycling.  If the balancing can't keep up, this cell will fail very quickly due to overvoltage.

 

Depends more on battery chemistry than anything else.  Li-Ion are relatively easy to balance, but LiFePo4 are very difficult to balance with a gentle method--mostly because there is basically no cell voltage change between 20-80% SOC.  (Yes there's going to be a tiny amount--but when you put that up against cell resistance and varying loads of a house application, the SOC voltage change is basically impossible to measure!)

 

I guess my responses are more along the lines of responding to someone who's trying to say that there's a better way to do a particular task--but without considering the situations/cases where the "better" method actually falls far short.

But I'm not saying that the "better way" is useless (at all!)  If that's all your batteries need--by all means use it!  I do agree that shifting power around is far better than just burning it up in heat.  It was only after I took into account the technical challenges (as well as the limitations) of the power-shifting methods, that I ended up settling on the shunt "total loss" method.

At least for Sean, though, he doesn't have an option!  Another customer couldn't get his huge battery bank balanced with the Chinese BMS units--likely because the leakage current of his huge bank exceeded the balance current of the BMS units.  It just took a few days for the Genetry balancers to heave his huge bank into balance, though...

Absolutely. If it works, do it.

I was asking purely from a place of ignorance. I don't make my own battery banks. (Though i might try it if i can understand all of these balancing quirks...) I just buy the inexpensive 12-volt LiFePO4 banks and leave it at that. To date this has worked very well... i'm continuously getting what the packs are rated at, and that's what matters to me.

That's basically what I'm doing except I'm running 4 12V lifpo4 in series and they do need some balancing

Without a balancer on it after a few weeks I had a battery undervolt shutdown before hitting the inverter cutoff

https://images.app.goo.gl/AivV6KU2X4CkroEF9

Not my image but solves most of your concerns

Not that the shunt type doesn't have it's place but this can keep it from being needed

OK, so the switched cap method that can only move power between 2 adjacent cells.

True, it likely will work fine for most light scenarios--but it has a severe Achilles' heel:
...if you have (like my mouse-damaged LFP bank) 10 cells in a row all fully charged (3.55v and climbing), and 1-2 low cells.

What happens to those 10 cells that are already full?  The switched capacitor methodology does absolutely nothing for 9 of the cells in a row--as they're already the same voltage.  Voltage keeps climbing past the 3.6v max, now they're all perfectly balanced at 4.7v--and the voltage keeps climbing.  The problem is that there is no way to "bleed off" or otherwise "get rid" of the excess power.

Yes, the cap switching from the end "full" cell into the "low" cell will be doing a bit--and it could be argued that the balancer will be "pulling" from the other full cells.  Problem is that this is at an exponential decay rate--by the time you get 3-4 cells away from the "low" one, the balancer might as well be doing nothing at all.

In most cases, the BMS overhead will disconnect charge if a single cell voltage gets too high.  And disconnect "discharge" if a single cell gets too low.

Couple that with a balancing mechanism that can't pull the cells into range, and you will pretty quickly end up with a battery pack of which you can only use a percentage of the total capacity.

To be fair, though, with matched-up cells, this is unlikely to happen for quite some time.  For the rest of us with hodgepodge cells, though, it's a daily reality!

Which is why I said that the bleeder still has it's place

If both methods are employed at once the bleeders will only be active under severe imbalance and would generally indicate a cell internally bleeding charge that should be replaced (as in your case)

The advantage of my idea is its actively balancing at all times allowing capacity variance to be averaged

The advantage of yours is overcharge protection

In a pack that stays balanced on its own mine would get it there without requiring a full charge and maintain it without full cycling 

Yes, the switched cap design would work much better than total loss shunting if the battery SOC charge was a perfect linear line.  And that is entirely dependent on the battery chemistry.

The problem is that if you have LiFePo4, the battery voltage between ~20-80% SOC is basically the same--making it impossible to balance out actual cell levels "at all times" without coulomb counters on every cell.  As a result, balancing inevitably only happens either between 0-20% (if bottom-balancing) or between 80-100% (top balancing, much more common).  And if on LiFePo4, you'd better either have a really strong balancer to hold the cells in line when they suddenly crest the hill...or a dedicated charger that automatically moduates the charge current as necessary to stay within the limits of the balancing solution (i.e. not an MPPT solar charger with 4,000W of panels on it!)

With LiFePo4 you kinda have to reach a full charge in order to be able to balance the batteries in the first place.  I keep saying "20-80%" as a safe ballpark, but I haven't actually measured the SOC remaining when the voltage suddenly starts to rise.  It could be as close as 10-90% or even 5-95%--what I do know that there is very little capacity in a LiFePo4 bank at 3.5-3.6v before it falls down to the nominal 3.2-3.4v range (where it stays until nearly exhausted, at which point it will start to fall down to 3.0v).

As I keep reiterating...every battery bank is different.  And so different balancing methodologies each have their place. 

For those of us that sourced our batteries for ~$100/kwh, our balancing needs will very likely be different from those who bought brand-new ready-packed batteries at $800/kwh.

I think it really is the top 10% when it really jumps

The documentation that came with mine shows voltage is pretty linear between 20 and 80% SOC

It's just a very small amount

Fair point on Cost of capacity as mine are about $300/KWh and have internal BMS that shuts them down in case of OV, UV, short and balances the individual cells

Where my external balance needs are for 12.8v blocks not 3.2v so my SOC effect on voltage is 4x per battery and 16x full bank

Thought I would jump in here to say that I'm very pleased with the performance of Sid's balancers on my huge BYD Lifepo4 bank (sixteen 220ah 24v modules connected for a 48v bank, with all cells paralleled between modules).  The balancers are mounted on a very large heatsink (overkill, I know!) and there is a temperature controlled fan on the back.  The balancers have been keeping things balanced for months now without issues (that's the idea, right?) 

Anyone familiar with the 220ah BYD modules knows they are several years old.  (Nobody seems able to say just how old they are.  My guess is about 2012)  Older lifepo4 banks need more robust balancing ability, which this system provides.  I'll add a pic showing the readings of all 16 with the bank at approximately 56.3v at float.  Note only about .04v difference between lowest and highest, which I consider great!  Hoping to get many more years of use out of these batteries.

<a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/6-dochubert/?do=hovercard" data-mentionid="6" href="/profile/6-dochubert/" rel="">@dochubert Was pretty sure you'd ordered a set, but I couldn't recall.  Didn't want to mention you and be wrong 😉 .

Seeing your cells >3.5v each means that they're definitely balanced...that's well past the "knee" of LiFePo4.  The ~0.04v difference could be explainable by resistor tolerances on the balancers--not to mention MCU ADC (analog to digital converter) tolerance.  But it's not even the slightest bit of a concern...as none of the cells are >3.6v

Not sure how "integrated" or "automated" you care to get, but if you wanted to, you could pull an RS-232 (TTL level) data string off the balancer modules for an external readout / status monitor...

I'm very happy with the setup since installing it a few months ago.

My feeling exactly!

I'm happy with it as is.  Works great without any attention from me at all (I do look at it once in awhile...), and that's how it should be. 

Thanks again for a great product!