Cell Run-away description and resource review

[BradCagle]

I usually see this happen on packs built with non-matched raw cells.

A good solution is to stay on the back side of that charge knee. Somewhere between 3.4v, and 3.45v

 

[FilterGuy]

It’s always better to use real life data , hers a Winston cell (LFP with Yttrium added ) data sheet graph , note Winston specifically statenthstvtge cell will be damaged if charged above 4v

note that 100% SOC is reached at 3.65V. Greater then that leads to over charging ( > 100% SoC ), but lower C charging results in more energy being stored
View attachment 108697

A few points.

* There are a lot of charts out there with variations to the voltages and SOC, but they all tend to have an extended flat space at around 3.4V with sharp knees at either end.
* It would be interesting to see what the Winston chart looks like if they extended the curves beyond what is shown.
* Since we always advise shutting the charge and discharge off well before the extremes I show, perhaps I should cut the extremes off as the Winston charts do.

I don't think my charts are 'wrong' per se, but let me ponder how I should adjust the presentation based on your feedback.

 

[John Frum]

I am not following. If you do a search on LiFePo4 charge curve you will find many versions of this chart. Granted, we never take a cell to 4.2 volts but that is what the charge curve looks like.

I hope nobody gets the idea that its ok to charge an LFP cell to 4.2 volts.

 

[FilterGuy]

I usually see this happen on packs built with non-matched raw cells.

A good solution is to stay on the back side of that charge knee. Somewhere between 3.4v, and 3.45v

Thank you for the feedback.

I should probably add something about matched and unmatched cells. Most of the comments in the document have an unstated assumption that you are starting with reasonably matched cells.

 

[BradCagle]

Thank you for the feedback.

I should probably add something about matched and unmatched cells. Most of the comments in the document have an unstated assumption that you are starting with reasonably matched cells.

I'm curious if people think they have matched cells when they really do not. An example would be if you bought X number of cells for some vendor.
The cells are 280ah. Do people assume they are matched?

Reality is they probably are not unless the vendor tested each cell, or has the test data from the manufacture, and selected your cells based on that data.

 

[John Frum]

Do people assume they are matched?

Most of the first timers here don't know that matched is a thing.

 

[Goboatingnow]

Thanks for the feedback, but unless I missunderstand the comments, I disagree.

In a top-balanced battery, the voltage on a weak cell will start dropping before the other cells and will eventually cause the BMS to shut off while the other cells still have energy.

sorry that doesnt make sense , the cell that reach’s the top balance voltage setpoint , ie the cell with the lowest energy capacity , will trigger the balance activity , what then happens is energy is diverted away from that cell , in theory this allows the charging process to continue , ie the other cells with greater energy capacity to “ catch up “ . most Times the diverted current is less then the cell charging current , so some charging continues on the low energy cell. Usually it still reaches the cutoff voltage before the other cells , and hence remains partially out of balance , but the next balance cycle , or in a few , will typically bring the cell into full balance

the voltage of the cell being balanced may drop , only if the net cell current is negative , ie the sum of charging current less balance current. But in typically charging systems , the charge current is greater then the balancing current , so the cell voltage continues to rise , albeit slower then the other cells , this is the idea to allow catch up time.

this is why many times it takes several balancing cycles sometimes to balance cells , particularly with low current passive balancers.

Here is a thought experiment: If I build a top-balanced 12V LiFePO4 battery out of two 100Ah cells and two 200Ah cells, what will the Ah of the battery be? The answer is that it will be a 100Ah Battery. The 100Ah cells will start dropping in voltage and shut off the BMS while the 200Ah cells still have half their charge.

In a series configuration yes to a point , this is a function of BMS protection rather then the battery config , your thought experiment is often the reason why parallel “ first “ cell configurations are better then series first. By rearranging this battery as a parallel first orientation , you mitigate the problem to an extent , because paralleled cells of different capacity share the load more or less.

If there are parallel cells within a single battery, the BMS sees the parallel cells as a single mega-cell. In this case, the capacity of the mega cell is the approximate average of the parallel cells. (the capacity of the Mega cell is not defined by the weakest cell, but the capacity of the battery is defined by the weakest mega-cell)

yes I made this point above

Note: For series batteries, the weak battery does define the total energy of the system just like series cells defines the capacity of the battery
For parallel batteries, the weak battery does not define the total energy of the system just like parallel cells does not define the capacity of the battery.

that was my point , that a catch all statement isnt True , secondly typically large banks have paralleled elements
in a series battery, the theoretical energy density is the sum of the elements , just as it is for parallel configurations ( it’s not “ average “ by the way ) , the issue with series strings is the energy cannot be extracted without damaging the lower energy cells. That’s the point

Even before your comments, I had decided I needed to add clarity to this page and updated the diagram to this:

View attachment 108690
Perhaps I should also add an annotation about the BMS or Loads cutting off. I should also point out the dashed part of the red line represents the unused energy of the other cells.

what you need to make clear is this is a single cell series battery ,

Fair enough, I will define it in the doc. In the meantime, by 'run-away cell', I mean a cell that has a voltage that quickly diverges from the voltage of the other cells.

there is no such thing , why would the voltage “ suddenly diverge “. This would only happen in an extremis situation, ie damaged or life expired cells , or cells that have widely different energy densities arranged in a series string.
it’s also not a term the industry uses. Runaway is used exclusively for thermal issues , because it’s runaway means a self supporting phenomenon. Out of balance is not a runaway situation

 

[FilterGuy]

Again what’s this term “ run away “ , are you talking Thermal run away

top balancing is problematic if you can’t control charge currents as you simply run out of time to draw energy out of the highest voltage cell all the while charging continues , so unless the balancing current is exceeding the charge current the time ( which is unlikely without charge current step back ) then the out of balance cell will continue to accept net charge current and it’s terminal voltage will rise , since your top balance start voltage is so close to your cutoff voktafe you simply run out of balancing time.

I guess I need to add a bit about balancers.... but I worry it will be a confusing tangent.

Since most of us have charge currents that far exceed what even the best active balancer can do, if the cells are not top balanced, there is a high possibility one of the cells will quickly rise in voltage well before the others. This can be viewed as "running out of time" for the balancer, but the effect is still that one cell voltage gets ahead of the others.

In a properly top-ballance system with reasonably matched cells, you can usually run with no balance function at all if you are not driving high C rates and are not draining the cells extremely low. A cell just does not change much during a discharge and charge cycle and will pretty much be at the same voltage when it is charged back to full.

active balancing has the advantage that no energy is wasted , you can in theory , run an active balancer , all the time during the charge process , you can even run it in discharge ,or during battery rest , this means there’s a far longer potential balancing time. Also the nature of switched capacitor balancers is they are higher current capable then resistive balancers.

It is a bad idea to run a balancer of any kind if the cell voltage is much below 3.4V. The balancer is using cell voltage as a proxy for SOC, but in the flat part of the curve, the cell voltage is a piss-poor indicator of SOC. In fact, a cell measuring 3.25V might have an SOC that is higher than a cell measuring 3.28V! If a balancer tries to balance the voltage between these two cells it will actually screw up the top balance. This is why the default turn-on point for balancing on a BMS is almost always around 3.4V.

Note: If you are interested in this, the "Off Grid Garage" youtube channel did an excellent series of videos on balancers and when to turn them on.

 

[Goboatingnow]

I should add to your thought experiment ,in a series cell battery the limit of the the energy extracted is a function of BMS disconnect strategy , not of the battery physics , for example you could in theory have the BMS do other things ,
firstly you could active balance during discharge , this transfer energy back into the lower energy cells so that all cells reach the bottom together, this is basically what active bottom balancing does , equally you could do mid balancing so that the weaker cells are effectively charged by the stronger cells as the battery discharges

This is the point , your hypotheses only holds true for certain BMS architectures . EVs for example often bottom balance ( or do both top and bottom )
Theres a good appluication note here https://www.ti.com/download/trng/do... Cell Balancing - What to Balance and How.pdf

Noet that runway due to cell imbalance can only occur where no protection circuits exists and the weaker cell is exposed to constant over voltage and over charge for many charge cycles , the overvoltage damages the already weakened cell and lowers its capacity which exacerbates the situation. Since we are talking about BMS protected cells , self sustaining runway cant occur

 

[Goboatingnow]

Fair enough, I will define it in the doc. In the meantime, by 'run-away cell', I mean a cell that has a voltage that quickly diverges from the voltage of the other cells

this is not what Texas Instruments and others define as a runway condition , a cell imbalance runaway only occurs where a weaker cell is exposed to uncontrolled series charge cycles resulting in repeated over voltage situations , this further weakens the cell and the cycle repeats until the cell is destroyed . This cannot happen in BMS protected cell charging as the cell is not ever overcharged and hence there is no significant further divergence. I have quite a bit of collected data on series cells, used in fractional C situations , They do not tend to diverge significantly over time , yes unbalanced cells cause loss of energy capability but its wrong to suggest that there is a rapid divergence

 

[FilterGuy]

sorry that doesnt make sense , the cell that reach’s the top balance voltage setpoint , ie the cell with the lowest energy capacity , will trigger the balance activity , what then happens is energy is diverted away from that cell , in theory this allows the charging process to continue , ie the other cells with greater energy capacity to “ catch up “ . most Times the diverted current is less then the cell charging current , so some charging continues on the low energy cell. Usually it still reaches the cutoff voltage before the other cells , and hence remains partially out of balance , but the next balance cycle , or in a few , will typically bring the cell into full balance

The page we were discussing was about discharge and the end of the cycle. At that point, the balance is not involved (or at least should not be)

With a top-ballanced battery, during discharge, the 'weak' cell will track the other cells reasonably well while the SOC is high but will diverge when the SOC gets low. Yes, I guess if you kept discharging the battery you could get some/most of the energy out of the other cells, but the voltage of the weak cell will get into ranges that damage the cell. However, a properly set up BMS would stop the discharge before that happens.

So, since we are being pedantic, the more accurate statement is: The weakest series cell (or meagcell) defines the capacity of a properly built battery that has typicall BMSs found in the typical solar storage system

 

[FilterGuy]

this is not what Texas Instruments and others define as a runway condition ,

OK..... My definition is what I have always seen on this forum.

 

[Goboatingnow]

OK..... My definition is what I have always seen on this forum.

my point is that run-away is a self perpetuating circumstance, rather like a " chain -reaction ", ie thermal runaway , where the heat generated itself fuels further runaway . IN a BMS protected system , I fail to see how run-away can be applied to cell balance or cell voltage , I cant see any fault mechanism where the existence of a high cell voltage then perpetuates a higher cell voltage and so forth

 

[Goboatingnow]

The page we were discussing was about discharge and the end of the cycle. At that point, the balance is not involved (or at least should not be)

thetas precisely what bottom balancing is , and bottom balancing is a far better way to extract all the energy , which is why its used in EVs, top balancing is used cause its simpler to implement. The drawback is that bottom balancing requires the battery be drawn to low SOC , which doesn't suit some applications

With a top-ballanced battery, during discharge, the 'weak' cell will track the other cells reasonably well while the SOC is high but will diverge when the SOC gets low. Yes, I guess if you kept discharging the battery you could get some/most of the energy out of the other cells, but the voltage of the weak cell will get into ranges that damage the cell. However, a properly set up BMS would stop the discharge before that happens.

I agree that in a series string irrespective of balancing , a weak cell remains a weak cell , no amount of balancing can correct that , what balancing does is ensures the other cells get more charge time , top balancing is largely carried out because cell often have very similar energy densities but for various reasons , their terminal voltage versus SOC is different , This is all top balancing can correct

BUT, as you said in a conventional BMS protected series string the weakest cells define the battery density , hence there is actually little or no point attempting to balance cells that have significantly different energy densities ( just like your thought experiment ) balancing cannot correct fundamental differences in energy densities


balancing , can be either active or passive ( I presume you understand the differences ) . Top balancing is carried out near the top , largely to facilitate low balance current equipment , if you try and balance when the charge current is high , your balance current must equally be high otherwise the cells will not balance within the charge cycle

hence as a result top balancing has to occur where the charge current has fallen considerably , which is why that occurs near the termination of charge point.

to balance a cell , you need ( a) to slow its charge rate so that the other cells catch up , (b) stop its charge altogether until other cells catch up or (c) actively discharge it so it and the other cells which remain being charged , meet at a common point

In active balance the energy is removed from the out of balance cells and transferred to the other cells, This has several advantages , the other cell get additional charge so they " catch up quicker" , and unlike passive balancing the energy removed from the out of balance cell isn't lost
active balancing facilitates discharge balancing ( bottom and mid point balancing ) , whereas passive can only do top balancing

 

[Bob B]

This forum is primarily about DIY storage batteries used in solar applications ..... it is NOT about EV's. Most DIYers on this forum are also TOP balancing their cells .... so, introducing all the counter arguments about what an EV BMS may or may not do is counterproductive.

 

[HRTKD]

my point is that run-away is a self perpetuating circumstance, rather like a " chain -reaction ", ie thermal runaway , where the heat generated itself fuels further runaway . IN a BMS protected system , I fail to see how run-away can be applied to cell balance or cell voltage , I cant see any fault mechanism where the existence of a high cell voltage then perpetuates a higher cell voltage and so forth

The topic of this thread has nothing to do with thermal run away. Run away is being used in a different context here. It is a bit confusing, but once it's understood that it's not about thermal run away, it's OK. Other possible ways to title it could be "Cell Voltage Deviation" or "Deviant Cells" (I like this one, lots of eyebrows will be raised over that!).

I think that it's acceptable to assume that the cells in the pack are all the same Ah specification. Otherwise, the discussion could get rather complicated. We're trying to help the new guys troubleshoot. Too much into the weeds and we'll lose them.

 

[Goboatingnow]

This forum is primarily about DIY storage batteries used in solar applications ..... it is NOT about EV's. Most DIYers on this forum are also TOP balancing their cells .... so, introducing all the counter arguments about what an EV BMS may or may not do is counterproductive.

yes but active balancing is getting more and more common and hence alternative balancing strategies may appear in time . my comments about balancing method was more to illustrate that the issue demonstrated by the OP is a function of a particular BMS strategy rather then a function of the battery

 

[Goboatingnow]

The topic of this thread has nothing to do with thermal run away. Run away is being used in a different context here. It is a bit confusing, but once it's understood that it's not about thermal run away, it's OK. Other possible ways to title it could be "Cell Voltage Deviation" or "Deviant Cells" (I like this one, lots of eyebrows will be raised over that!).

I think that it's acceptable to assume that the cells in the pack are all the same Ah specification. Otherwise, the discussion could get rather complicated. We're trying to help the new guys troubleshoot. Too much into the weeds and we'll lose them.

I didn't say it was about thermal runaway , I made the comparison that using the term runaway was inappropriate . The issue being discussed is more correctly described a cell imbalance issue

 

[Bob B]

yes but active balancing is getting more and more common and hence alternative balancing strategies may appear in time . my comments about balancing method was more to illustrate that the issue demonstrated by the OP is a function of a particular BMS strategy rather then a function of the battery

I'm a firm believer in the KISS methodology ... when possible. Over time, this forum has generally concluded that the most simple method for most DIY packs is to top balance and then try to maintain balance by only balancing above 3.4V.

Active balancing is a hot subject and people have their own strong opinions. I tend to think that if cells are top balance and capacity matched active balancing is not necessary .... I also believe that balancing below 3.4V is a bad idea MOST of the time.

Some just like to argue for the sake of arguing with no real concern of whether it is productive.

 

[Goboatingnow]

I firmly agree that several BMS units is better then one unit connected to every cell , thats a key point

I'm a firm believer in the KISS methodology ... when possible. Over time, this forum has generally concluded that the most simple method for most DIY packs is to top balance and then try to maintain balance by only balancing above 3.4V.

agree its certainly common

Active balancing is a hot subject and people have their own strong opinions. I tend to think that if cells are top balance and capacity matched active balancing is not necessary .... I also believe that balancing below 3.4V is a bad idea MOST of the time.

why is balancing under 3.4V bad in itself , Active balancing is superior in all balancing methodologies and has become prevalent in any sophisticated solutions especially switched capacitor designs and may cheap BMSs now use active balancing . Passive balancing is very wasteful in many cases , especially where power generation is limited . Top balancing benefits from active techniques.

Some just like to argue for the sake of arguing with no real concern of whether it is productive.

unless you examine the shibboleths , you never build a first principles understanding of why things are done the way they are and the tradeoffs . The lack of discussion leads to blind obedience