looking for the simplest balancing circuit

[john61ct]

Avoiding wasting those tiny levels of energy is a silly goal when mains power is available.

I agree that having dedicated balancers hooked up fulltime is not a good idea, but not because these balancers don't do a much better job than those built into normal BMS, they do that.

Based on KISS IMO all electronic devices should be easily replaced, removed for whatever reason, without pack disassembly.

 

[thunderstorm80]

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

From my experience with my 26650's they experienced short-term short circuits, impressive C rate abuses, and after 9 years they lost only 20% of their capacity.
My A123 20Ah prismatic cells have had longer short circuit by accident - They let such a high amount of current that the insulation catched an actual fire, but the cells stayed intact. This mistake degraded their health of course but no fire damage, smoke, etc...
As far as I understand, part of the chemistry involves CO2 which is fire retardant.

Yet, A123 went bankrupt while members on ES still buy LiPo and literally play with fire. What a strange world...

 

[john61ct]

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

I've seen LFP puncture test that resulted in zero flames, only a little smoke.

I believe banks built from them to be in general, as safe as lead, 5000x safer than the other LI chemistries usually used for propulsion these days.

Of course a runaway charger left connected for long, all bets are off, can definitely ignite any LI.

 

[thunderstorm80]

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

I was quoting the word "problem", as a joke. Of course this is an advantage. It makes measuring %SOC by voltage only a harder task within the 30%-90% spectrum, that is all.

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".

True. It's more simple to me, as a user. It's very hard not to fully charge A123's unless you want to spend many many hours, by charging to less than 3.4V@cell.

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.

I anyway keep myself far away from going below 20%-30% SOC, as part of this top-balance strategy.
This gives me longer time intervals between manual balancing which I do every month or so.

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.

This I don't understand. Let's assume one cell has 0.4Ah less than the others.
If I balance everything at the bottom, then even if I use 3.4V charge @ cell, then this single cell will reach 3.6V first, while the others are still at around 3.37V or so (the voltage during final charge but still not topped-out). Since I have 24S battery, then this voltage spike of that single cell won't be as noticeable to the charger as if I had, for example, 12 cells in series.
In my example, close to finishing the charge, the total pack voltage can be 3.37*23 + 3.9V ~= 81.4V, when I charge to to 3.4V. (finish at 81.6V).
Of course that suffering cell will gradually lose that surface voltage and so the others will slowly rise until the are full at rest, but this creates an accumulative damage to that cell.
This is why I don't think it's a good idea to bottom-balance as it requires me to balance-charge every time...
Or - you mean that if I had this situation at the bottom - then the differences in the voltages are much greater - 2V-3V?
I still find the opposite extreme state from the balanced start a place I should never reach. With discharging it's easy. I just stop according to the Ah. With charging, it's hard NOT to charge the cells fully... Unless you want a very very very very slow charge.

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

But due to imbalance issues - some cells can be well above it, some below.

> 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.

Hmm... so if I perform such slow charge, aiming to 3.38V per cell and if my strategy is top-balanced, that kind of charge can actually renew the top-balance condition since any surface voltage will be waning in the time scale of the charge ?

Anyhow, thanks for the bright ideas. I learn more from this!

 

[thunderstorm80]

And speaking of surface voltages, a zener diode could be perfect to bleed out any excess voltage above a certain threshold, but due to their "knee" point behaviour it's not practical.
Are there more expensive zener diodes which are have a sharper knee point? 24 of these (for my 24S battery) and I never have to worry about HVC as long as the charging current is not too high.

 

[john61ct]

Of course this is an advantage. It makes measuring %SOC by voltage only a harder task within the 30%-90% spectrum, that is all.

No, it does not.

Cannot be done that way at the pack level with any accuracy anyway.

Nothing to do with which style of balancing - top middle bottom - you do, the two topic areas are not related, at all, from a practical IRL POV.


> True. It's more simple to me, as a user. It's very hard not to fully charge A123's unless you want to spend many many hours, by charging to less than 3.4V@cell

You can **charge** at 3.65 voltage no problem. For maximum longevity, do so with a profile that results in a **resting** isolated voltage between 3.33-3.35V

There are hundreds of different combinations of

initial (CC stage) current rate

finishing (CV stage) current rate

Absorb (CV) Hold Time, including zero seconds (CC only)

profiles that will get you there, and that's without varying the charge voltage setting.


> I anyway keep myself far away from going below 20%-30% SOC, as part of this top-balance strategy.

very wise, but begs the question, how do you measure that SoC while discharging? We talking about a propulsion use case here?

 

[john61ct]

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.

This I don't understand. Let's assume one cell has 0.4Ah less than the others.

If I balance everything at the bottom, then even if I use 3.4V charge @ cell, then this single cell will reach 3.6V first, while the others are still at around 3.37V or so (the voltage during final charge but still not topped-out).

To repeat, charging "3.4V charge @ cell" means stopping when the first cell reaches 3.4V. Should be completely impossible for any cell to go higher.

If you are talking about using the "average" cell voltage, based on the pack voltage divided by S-count

that would be stupid, and I have nothing more to say about that.

> In my example, close to finishing the charge, the total pack voltage can be 3.37*23 + 3.9V ~= 81.4V, when I charge to to 3.4V. (finish at 81.6V)

Where the heck is that added 3.9V coming from?

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

> But due to imbalance issues - some cells can be well above it, some below.

No, see above again if you still think that way.

And please FFR, be more precise in your writing.

The setpoint setting on the charger

must be distinguished from what you measure during charging at the pack (or cell) terminals.

And neither has anything to do with the "true cell voltage" after resting isolated.

_______
> Since I have 24S battery, then this voltage spike of that single cell won't be as noticeable to the charger as if I had, for example, 12 cells in series.

You should not be allowing any such spike. If not basing whole-pack charge termination directly off the weakest (highest) cell level voltage

then **you** must calibrate manually the whole-pack setpoints in order to get the same result.



> Of course that suffering cell will gradually lose that surface voltage and so the others will slowly rise until the are full at rest, but this creates an accumulative damage to that cell.

Nothing will rise after charger is off, why would you think that?


> This is why I don't think it's a good idea to bottom-balance as it requires me to balance-charge every time...

Wut??

"Balance charge" means "trying to balance while charging".

Above I was not talking about that at all.

Terminating charge based on the state of the first / weakest / highest cell is just called "properly regulated charging" in my book

nothing to do with the act of balancing.


By balancing at the bottom, the delta between cell/groups at the top will be maximum.

By balancing at the top, the delta between cell/groups at the bottom will be maximum.

Middle balancing minimizes the delta at both ends.

> I still find the opposite extreme state from the balanced start a place I should never reach. With discharging it's easy. I just stop according to the Ah.

Wow, so that answers the above question begged earlier.

I do not know of **any** coulomb counting SoC meter I would trust to protect a pack at the bottom.

I sure hope you are using per-cell voltage at least as a failsafe backup


> With charging, it's hard NOT to charge the cells fully... Unless you want a very very very very slow charge.

Yes overcharging is the risk factor needing avoiding, and very low current rates make that **much** harder.

Undercharging is very healthy, not something to be avoided unless in practice depriving you of significant range.

 

[john61ct]

> 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.

Hmm... so if I perform such slow charge, aiming to 3.38V per cell

So there you mean, with a charge profile that ensure no cell is allowed to go higher than 3.38V ?


> and if my strategy is top-balanced, that kind of charge can actually renew the top-balance condition

I do not understand that last phrase

> since any surface voltage will be waning in the time scale of the charge

Nor that last

please try to rephrase both, maybe using simpler basic wording?

 

[john61ct]

And speaking of surface voltages, a zener diode could be perfect to bleed out any excess voltage above a certain threshold, but due to their "knee" point behaviour it's not practical.
Are there more expensive zener diodes which are have a sharper knee point? 24 of these (for my 24S battery) and I never have to worry about HVC as long as the charging current is not too high.

The surface charge issue is not a problem that needs to solved, just a description of phenomena that occurs when overcharging.

My point bringing it up is, you should be setting up a charging regime that ensures **there is no significant surface charge effect in the first place**.

If you care about the longevity of your bank that is.

There is no countervailing practical benefit for not doing that though, only downsides.

IOW there is no reason for any cell to ever reach higher than 3.45-3.47V

Perhaps for benchmarking, capacity testing, other specified maintenance protocols done periodically

your monthly balancing regime for example, can go a little higher **occasionally** if desired

but not during normal cycling charging.

 

[john61ct]

LiFePo4 does require somewhat more monitoring and periodic balancing.

Again, I disagree.

All LI chemistries need proper monitoring at the per-cell level.

LFP needs much **less** frequent rebalancing than the other LI chemistries commonly used for propulsion

so long as the usage is not abusive and their SoH is high.

I have no experience with long-term abusive usage, nor old / worn cells, so cannot speak to those conditions.

 

[john61ct]

Okay, I see now the other post was hyperbole, since in one post 0.005 doubled to 0.010
..
"all points in their SOC"

You falsely quoted that as having been written by me?

> throwing "mismatched capacity" into the mix

Even a fresh bank from the best manufacturer, where each cell costs over $500, will have **some** level of Ah capacity difference

and that usually grows over time as cells wear.

Therefore even if it is not a problem, it behooves the owner to be set up to spot when it becomes one, if nothing else to inform her judgment oof EoL having arrived.

 

[john61ct]

Sorry Hummina for hijacking the thread

hope at least some of our back and forth is useful to its purpose

 

[John in CR]

Okay, I see now the other post was hyperbole, since in one post 0.005 doubled to 0.010
..
"all points in their SOC"

You falsely quoted that as having been written by me?

here's your post https://endless-sphere.com/forums/viewtopic.php?f=14&t=106004#p1559415

 

[John in CR]

LiFePo4 does require somewhat more monitoring and periodic balancing.

Again, I disagree.

All LI chemistries need proper monitoring at the per-cell level.

LFP needs much **less** frequent rebalancing than the other LI chemistries commonly used for propulsion

You agree and then disagree. Experienced users with good familiarity with their packs don't need cell level monitoring if using good quality automotive grade cells. I have 3 packs of used auto cells in primary service with over 6 years of daily use between them, and after ensuring they were in perfect balance when I put them in service, none have gone past +/-0.01V, so none have required balancing. At first I checked more often, but now I check cell level voltages once every 5 or 6 months just to be sure, but I'm confident that I would note a change just from monitoring pack voltage as my fuel gauge. I doubt anyone would consider that "proper monitoring at the per-cell level", especially since as it turns out that since before 2017 I haven't actually needed to bother checking cell level voltages other than to give myself peace of mind.

OTOH, every LFP pack that I have used has required periodic balance charging, because the cell level voltages showed enough variance that I couldn't trust whether the packs were balanced. Even just sitting at a storage SOC of 50%, the self discharge differences were enough to create out of balance conditions. My non-LFP automotive packs show almost no self discharge at all, not counting my 7 year old rack of V1 Nissan Leaf modules.

At first you seemed to indicate that my minimalist human BMS approach that's worked well with V1 and V2 Chevy Volt modules and 26ah bolt together Panasonic cells out of a Ford, could be used with LFP packs, but I just don't see that's the case.

 

[Hummina Shadeeba]

And speaking of surface voltages, a zener diode could be perfect to bleed out any excess voltage above a certain threshold, but due to their "knee" point behaviour it's not practical.
Are there more expensive zener diodes which are have a sharper knee point? 24 of these (for my 24S battery) and I never have to worry about HVC as long as the charging current is not too high.

What are u saying? Could u get a zener diode to burn off current at when beyond maybe 3.4v?

 

[thunderstorm80]

And speaking of surface voltages, a zener diode could be perfect to bleed out any excess voltage above a certain threshold, but due to their "knee" point behaviour it's not practical.
Are there more expensive zener diodes which are have a sharper knee point? 24 of these (for my 24S battery) and I never have to worry about HVC as long as the charging current is not too high.

What are u saying? Could u get a zener diode to burn off current at when beyond maybe 3.4v?

Yes, with a resistor in series of course.
But it won't work smoothly as the zener diode start to "open" a bit before the designed voltage, so it will always bleed out some current. Plus, just with resistors, if you have zener diodes across all your cells they won't be exactly equal so they will drain the cells differently.

 

[john61ct]

Okay, I see now the other post was hyperbole, since in one post 0.005 doubled to 0.010
..
"all points in their SOC"

You falsely quoted that as having been written by me?

here's your post https://endless-sphere.com/forums/viewtopic.php?f=14&t=106004#p1559415

Aha, mea culpa that was my quoting your #11 but failing to insert the quoting indicator.

I apologize!

 

[john61ct]

Well becoming familiar over time with a given pack does indeed lead to "proper monitoring", in the sense that you just know what the pack level voltages are, in order to avoid getting too far into the shoulders at the cell level.

My main experience is with LFP, not the EV stuff, and let's just say it differs, I do not ever see any self-discharge or much imbalancing if stored isolated in cool conditions.

I'll just leave it there