How to Parallel Balancing. (YEP 99% of us is doing it wrong)(PART#1)

Will Prowse said:

Wait guys. What if you have a active balancer on your system... And your BMS cuts off charging, but as you discharge, and your pack goes out of balance at low SOC, and triggers active balancer, and starts charging a individual cell! At an unsafe temperature. Haha I never thought of that before! Well, if it's discharging quickly, the internal cell temperature would probably rise enough to enable charging.

Click to expand...

Maybe my understanding of 'active' balancers is wrong, but i thought that they 'actively' balanced all the time on both charging and discharging, ensuring that the cells stayed balanced all the time. I didnt think it was something that was 'triggered' at set points etc. So i'm thinking, the pack never goes out of balance in the first place.... but i'm only just trying to get to grips with this stuff(no experience).
With passive balancers, that gets done at the top to passively bring them into line, but not with active....(i think)

I know i would like a fuller explanation Will, if i'm totally wrong on this.

SolarRat said:

And that's the way we should all treat these unknowns. Better safe than sorry until there are definitive answers.

Click to expand...

So what is unknown? If the spec calls something out why should you not believe it? What other specs should I not believe? Max voltage? Min Voltage? Or do we get to selectively choose which ones to believe or not believe?
If I saw one spec that was out of line with a everyone else, I think there is room for concern. However, I have seen a LOT of cells speced at -5. Just because that does not fit someones belief system does not mean it is wrong.

Would I charge at -5C, No. I would probably set my BMS to cut off before that, just like being conservative with any other spec.

FilterGuy said:

So what is unknown? If the spec calls something out why should you not believe it? What other specs should I not believe? Max voltage? Min Voltage? Or do we get to selectively choose which ones to believe or not believe?
If I saw one spec that was out of line with a everyone else, I think there is room for concern. However, I have seen a LOT of cells speced at -5. Just because that does not fit someones belief system does not mean it is wrong.

Would I charge at -5C, No. I would probably set my BMS to cut off before that, just like being conservative with any other spec.

Click to expand...

Yeah, but they dont state implicitly that you can charge down to that temp at 1C, just that you can charge down to that temp. They are not stating any falsehoods, maybe just economical with all the facts perhaps???

ghostwriter66 said:

Another interesting note he said was that really only the American's seem obsessed with cell balancing. He said that he only uses BMS for High Current cutoff, Low Current Cutoff, and Low/High Temperature cutoff -- he said that if you get 4 batteries from the same batch that them getting enough out of "whack" (my word) to be concerned about would take a long time ...

Click to expand...

This is something I have come to believe as well... If you you have reasonably matched cells, why do you need to top or bottom balance at all? As far as I can tell, balancing really only allows you to squeeze those last few electrons in at the top of the charge and get those last few electrons out at the bottom of the discharge. If your voltages are set reasonably conservative, What difference does it make? Also, if you are using a system with a balancing BMS, it is going to redo it all over time anyway.

If you are powering an RC plane, I can see you would want to balance.... this is where people are going to try to squeeze as much as they can out of the battery. That is not what most of us Willites are doing. Our priorities are almost the opposite. Rather than squeeze every bit out, we want to make the batteries last, so we are going to be setting the voltages conservative.

Excelent Thread and good discussion.

Just to toss in a bit of randomness and/or monkey wrench, just for kicks & giggles.
What if you have a BMS which is capable of balancing cells During Charging / Discharging and while in Storage (no amps being drawn) States ? Example BMS is the Chargery which does so if enabled. It isn't a big active balancer, being only 1.2A but seemingly capable for most use cases I would think. Seems to me, if you need more powerful balancing than that, there are likely some issues with the cells.

Link for the Progressive Dynamic products for those who may be interested:
https://www.progressivedyn.com/specialty/pd9100l-series-lithium-converters/

SolarRat said:

Since we are already way off topic, here is some good research data on charging near or way below freezing:

Fast charging of lithium-ion batteries at all temperatures

Range anxiety is a key reason that consumers are reluctant to embrace electric vehicles (EVs). To be truly competitive with gasoline vehicles, EVs should allow drivers to recharge quickly anywhere in any weather, like refueling gasoline cars. However, none of today’s EVs allow fast charging in...

www.pnas.org

" Schimpe et al. (11) cycled identical graphite/LiFePO4 cells at different temperatures. Cells at 25 °C lost 8% capacity in 2,800 equivalent full cycles (EFCs). At the same capacity loss, cell life drops to 1,800 EFCs at 15 °C, 1,400 EFCs at 10 °C, and 350 EFCs at 0 °C. "

Still want to try -5c?


There are literally thousands of these research papers out there for us to gather info from, a bit here and a bit there....

Click to expand...

The PNAS citation is very interesting. I had previously thought that 0 C charging would result in an immediate "catastrophic" capacity loss. This suggests 8 percent loss after 350 cycles at freezing. I will need to read the paper and see how evenly the accelerated low temperature loss is spread across those 350 charge cycles. If it is not front loaded too much, the consequences of one or two cold days might not be as grim as I had thought (although obviously to be avoided if possible).

What are Y’all using for a power supply?

Steve_S said:

Excelent Thread and good discussion.

Just to toss in a bit of randomness and/or monkey wrench, just for kicks & giggles.
What if you have a BMS which is capable of balancing cells During Charging / Discharging and while in Storage (no amps being drawn) States ? Example BMS is the Chargery which does so if enabled. It isn't a big active balancer, being only 1.2A but seemingly capable for most use cases I would think. Seems to me, if you need more powerful balancing than that, there are likely some issues with the cells.

Link for the Progressive Dynamic products for those who may be interested:
https://www.progressivedyn.com/specialty/pd9100l-series-lithium-converters/

Click to expand...

SolarRat said:

Since we are already way off topic, here is some good research data on charging near or way below freezing:

Fast charging of lithium-ion batteries at all temperatures

Range anxiety is a key reason that consumers are reluctant to embrace electric vehicles (EVs). To be truly competitive with gasoline vehicles, EVs should allow drivers to recharge quickly anywhere in any weather, like refueling gasoline cars. However, none of today’s EVs allow fast charging in...

www.pnas.org

" Schimpe et al. (11) cycled identical graphite/LiFePO4 cells at different temperatures. Cells at 25 °C lost 8% capacity in 2,800 equivalent full cycles (EFCs). At the same capacity loss, cell life drops to 1,800 EFCs at 15 °C, 1,400 EFCs at 10 °C, and 350 EFCs at 0 °C. "

Still want to try -5c?


There are literally thousands of these research papers out there for us to gather info from, a bit here and a bit there....

Click to expand...

Here is an interesting graph from the Schimpe paper showing 5% loss of capacity when storing LFP 100% SOC at 35 C for ~200 days, and a second showing how progressively lower SOC storage affects this.

KG6S said:

What are Y’all using for a power supply?

Click to expand...

I used a Mean Well HRP-300-3.3 for top balancing 3.2 V cells in parallel. Worked perfectly for me.
[Edit] One issue I had using the Mean Well was the lack of metering. A nice benchtop power supply offers the advantage of readouts, but I couldn't justify the $$$ for this use. I also realized my AH meter wouldn't operate on less than 5V and didn't support using an external power supply. Fortunately I was able to measure amps using my shunt, a volt meter, and good old Ohm's law. This will do the job for my once a year top balance. If I were doing it more often, I'd certainly pick up a cheap watt meter that can either run on 3V or off an external power source.

Ceberha said:

The PNAS citation is very interesting. I had previously thought that 0 C charging would result in an immediate "catastrophic" capacity loss. This suggests 8 percent loss after 350 cycles at freezing. I will need to read the paper and see how evenly the accelerated low temperature loss is spread across those 350 charge cycles. If it is not front loaded too much, the consequences of one or two cold days might not be as grim as I had thought (although obviously to be avoided if possible).

Click to expand...

Here is the relevant figure from the Schimpe paper - it seems that the loss of capacity when charging at 0 degrees C is quite rapid, but one ore two charges at freezing would not "kill" the battery. Sadly no data in this paper on charging below freezing. Of note, this test was done at a high 1C charge rate.

Solarfun4jim said:

Maybe my understanding of 'active' balancers is wrong, but i thought that they 'actively' balanced all the time on both charging and discharging, ensuring that the cells stayed balanced all the time. I didnt think it was something that was 'triggered' at set points etc. So i'm thinking, the pack never goes out of balance in the first place.... but i'm only just trying to get to grips with this stuff(no experience).
With passive balancers, that gets done at the top to passively bring them into line, but not with active....(i think)

I know i would like a fuller explanation Will, if i'm totally wrong on this.

Click to expand...

Assuming you stay above 10% SOC, I agree the cell voltage deltas will be below the balancer trigger voltage. Below 10% SOC, some cell voltages may fall off steeply, which would trigger balancing.
Several cold sunless days in succession with no low voltage cutoff protection could put your pack in this scenario. Wouldn’t low voltage cutoff set around 20% SOC help prevent this situation?

cass3825 said:

Assuming you stay above 10% SOC, I agree the cell voltage deltas will be below the balancer trigger voltage. Below 10% SOC, some cell voltages may fall off steeply, which would trigger balancing.
Several cold sunless days in succession with no low voltage cutoff protection could put your pack in this scenario. Wouldn’t low voltage cutoff set around 20% SOC help prevent this situation?

Click to expand...

Cass, i dont have enough understanding to venture an opinion. I am still trying to figure exactly how each system functions. in 'active balancers', I had thought, that at any point in the discharge(or charge for that matter) cycle, when one cell got out of whack by a certain percentage or mv, then the balancers would immediately shift the amps to that cell to bring the voltage back into step with the rest. I had thought that is what the 'active' part meant(ie a constantly ongoing process), where as, with passive, when the first cell out of whack reached the threshold limit voltage, amps were bled off as heat, thus the other cells could get up to the same balance as the 'whacky' cell. Unfortunately, i simply dont understand enough about each system functionality, so i'm just guessing at this point.

Your understanding is more than you are giving yourself credit for.
The key point is charging vs. discharging in subfreezing temperatures. I believe it’s perfectly ok to discharge LFP when cold, but charging is detrimental. Passive balancers discharge the cell with the highest voltage. Perfectly acceptable in cold conditions. Active balancers, on the other hand, store the excess energy from the cell with the highest voltage in a capacitor or inductor, then direct the stored energy into the cell with the lowest voltage, potentially damaging the cell.

Solarfun4jim said:

Cass, i dont have enough understanding to venture an opinion. I am still trying to figure exactly how each system functions. in 'active balancers', I had thought, that at any point in the discharge(or charge for that matter) cycle, when one cell got out of whack by a certain percentage or mv, then the balancers would immediately shift the amps to that cell to bring the voltage back into step with the rest. I had thought that is what the 'active' part meant(ie a constantly ongoing process), where as, with passive, when the first cell out of whack reached the threshold limit voltage, amps were bled off as heat, thus the other cells could get up to the same balance as the 'whacky' cell. Unfortunately, i simply dont understand enough about each system functionality, so i'm just guessing at this point.

Click to expand...

Yes but because the discharge/charge curve is relatively flat with LiFePO4, active balancers usually don't kick on till high or low SOC. The cells can be out of balance at 50% SOC, and the voltage difference between the cells can be so minimal that balancer won't be triggered. Also, some active balancers need a higher SOC voltage to push current from one cell to another. Depends on the circuit design. Some work horribly slow at low soc vs high soc for this reason. Also, cells like lto with lower nominal voltage will pass very little current to the cell next to it with most balancers. It again, depends on design of balancer. Some will boost voltage with tiny converter circuit and that helps a lot.

Steve_S said:

Excelent Thread and good discussion.

Just to toss in a bit of randomness and/or monkey wrench, just for kicks & giggles.
What if you have a BMS which is capable of balancing cells During Charging / Discharging and while in Storage (no amps being drawn) States ? Example BMS is the Chargery which does so if enabled. It isn't a big active balancer, being only 1.2A but seemingly capable for most use cases I would think. Seems to me, if you need more powerful balancing than that, there are likely some issues with the cells.

Link for the Progressive Dynamic products for those who may be interested:
https://www.progressivedyn.com/specialty/pd9100l-series-lithium-converters/

Click to expand...

By the way we have been playing for the last week with the Chargery 16T ... wow .. I'm not going to say anything yet about why this "may" be my favourite BMS becuase I have not played with it enough .. but -- WOW

Will Prowse said:

Yes but because the discharge/charge curve is relatively flat with LiFePO4, active balancers usually don't kick on till high or low SOC. The cells can be out of balance at 50% SOC, and the voltage difference between the cells can be so minimal that balancer won't be triggered. Also, some active balancers need a higher SOC voltage to push current from one cell to another. Depends on the circuit design. Some work horribly slow at low soc vs high soc for this reason. Also, cells like lto with lower nominal voltage will pass very little current to the cell next to it with most balancers. It again, depends on design of balancer. Some will boost voltage with tiny converter circuit and that helps a lot.

Click to expand...

AND one of the things he showed us yesterday with some of the "Chinese" BMS's he had brought -- was that the leveling of each of the battery cells did not kick in and begin "leveling" until the voltage of each cell had individually hit the 100% SOC mark per cell .. so for some of us on some of these major BMS's - actual did not benefit from the cell leveling since we kept our charges at 90% or somewhere BELOW 100% SOC ...

PS - there were others though that were active leveling the entire time - I think Chargery is one of them - but you have to turn that function on -- SOOO finding the right BMS for how YOU are going to use your battery is just as important -- there is NOT a one size fits all ... but some are damn close .. LOL

ghostwriter66 said:

By the way we have been playing for the last week with the Chargery 16T ... wow .. I'm not going to say anything yet about why this "may" be my favourite BMS becuase I have not played with it enough .. but -- WOW

Click to expand...

I look forward to your report. I have not played with one but I have read the users manual. The manual is (in my oppinion) hard to read but once you get past that it looks like a pretty nice BMS with a lot of capability.

ghostwriter66 said:

By the way we have been playing for the last week with the Chargery 16T ... wow .. I'm not going to say anything yet about why this "may" be my favourite BMS becuase I have not played with it enough .. but -- WOW

Click to expand...

I just flashed / updated the firmware which was easy enough, note, you have to have the USB cable plugged in with the BMS on "before" starting the application, select the com port and the right file and voila. Settings need to be reset to whatever you tweaked, by default the BMS restarts in LIPO mode. Sure is nice to have one that can be updated and tweaked. I know I could never revert to using a non-interactive BMS again which I could not monitor or tweak.

SolarRat said:

If the cells aren't yet balanced (and the obviously aren't if we are now balancing them), one cell may hang at 3.38v while another zooms up quickly to 3.5v. That higher cell may stay in the upper knee for hours waiting for the lower cell to catch up. Equalizing them at 3.4v first insures this doesn't happen. That's the way I read between the lines?

Click to expand...

This is physically impossible when all cells are connected in parallel, which is a premise for this procedure.
As cells become fuller and fuller, their individual SOC gets equalized naturally because current splits naturally between cells based on their demand, which is controlled naturally by change in IR as the cell gets charged.

The idea not to leave cells above the knee for "too long" is a sound idea in principle and has merit, but what people get hung up on due to their ignorance is how to quantify "too long". The effect of "too long" is measured in weeks/months, not hours or minutes.
There is absolutely no harm in staying above the knee for many hours, just don't do it for weeks or months.

The rest of my post is a general rant, not pointing to @SolarRat or anyone in particular, not even @ghostwriter66 who's info is generally true and helpful, but not very well understood, no offense...

Making something sound like a Bible because it came from some "Design Engineer at largest LFP factory", please. If you can't post the name of the factory and educational background of this "Design Engineer" which can be independently verified, then it's no different than any other piece of FUD on the Internet.

I am a Design Engineer at one of the largest LFP battery makers in the US, with UL1973 certifications and many 1000s of systems deployed all over the world for over 10 years now. Does it make my statement credible unless I post the name of the company and my name, so it can be verified? Absolutely not. As far as this forum is concerned, I'm just an anonymous dude sitting at my keyboard making stuff up.
My point is, please don't believe anything unless you can verify it with credible sources.
There is way way too much FUD around Lithium batteries. At least 50% of info on this forum and most others is complete nonsense based on ignorant repeating of unverified and misunderstood sources.

electric said:

The effect of "too long" is measured in weeks/months, not hours or minutes.
There is absolutely no harm in staying above the knee for many hours, just don't do it for weeks or months.

Click to expand...

electric said:

My point is, please don't believe anything unless you can verify it with credible sources.

Click to expand...

I agree with your second point, so...you got any data to back up the first? Not that I don't believe you may be right, but I like to see some sort of research that confirms this. Otherwise, I'll stick to my better safe than sorry approach.


As far as the cell drift part, yeah, I was drinking and had a brain fart. I should delete that so it doesn't confuse others...

My data is my personal experience, which means squat unless I post my name and company, which I choose not to at this point, maybe later when my company has a competitive product to offer for DIY market. There are scientific papers you can find easily via Google, where names of scientists and their institutions give them plenty of credibility. Effects of temperature, SOC and calendar aging on cycle life of Lithium batteries.
Everything @ghostwriter66 posted has merit, but you really need to understand the relative scale of effects, the scale of time factor, taking everything into account together and looking at it practically in a real world application, where batteries need to be used, not ridiculously pampered based on a single parameter, while ignoring other, much more pronounced parameters, for which you may not even have control.
Plan to get 3-5 years of continuous day to day use ( depending on environmental factors ) and anything above that will come as a bonus.
I have LFP cells still in use after 10 years and some didn't even last 3 years, too many variables to consider unless we go into specifics of each application.
Sorry for ranting and not giving specific answers. I don't have time to write long posts with links and data, I got a family to feed. I usually post just to stop the FUD and give people a chance to learn on their own. Don't blindly follow anyone, not even @Will Prowse

FilterGuy said:

So what is unknown? If the spec calls something out why should you not believe it? What other specs should I not believe? Max voltage? Min Voltage? Or do we get to selectively choose which ones to believe or not believe?
If I saw one spec that was out of line with a everyone else, I think there is room for concern. However, I have seen a LOT of cells speced at -5. Just because that does not fit someones belief system does not mean it is wrong.

Would I charge at -5C, No. I would probably set my BMS to cut off before that, just like being conservative with any other spec.

Click to expand...

I would like to apologize to the forum about my rant. The nice thing about this forum is that we can all share what we know or beleive and then we can build a better overall understanding together. I should not have let go with my rant because it does not help us build a better understanding...it just throws a wet blanket on the conversations. It would have been better for me to ask questions to explore the statement so we can learn together about the subject.

FilterGuy said:

I would like to apologize to the forum about my rant. The nice thing about this forum is that we can all share what we know or beleive and then we can build a better overall understanding together. I should not have let go with my rant because it does not help us build a better understanding...it just throws a wet blanket on the conversations. It would have been better for me to ask questions to explore the statement so we can learn together about the subject.

Click to expand...

Consider that people writing specs are not Gods and have their own agendas. Specs get changed over time as people learn new things. Initial LFP specs from 2008/9 had 4.25V charge voltage and 3C rate, then over time got updated with more realistic numbers.
Specs are made based on lab tests with controlled environment and time factor is often simulated or assumed. Over the years real effect of time is becoming more clear as more data is collected.
Each spec has a range of values where all other specs were perfectly controlled. Not possible in real life.
For example, cycle life is at room temperature, or temperature range is at specific SOC and C rate. You can't pick extremes from each specs and combine them, it will never work.
Another example, can you charge below 0C ? Of course you can, but at what rate and how much life you sacrifice? You don't know because specific parameters have not been tested, because testing is expensive. If you need it for your application, you test yourself, or pay someone to test for you.

Lab tests are time constrained because cost of testing is a an expense and cell maker needs to make money to survive another day. Product must go out the door today, while lab is still testing long term parameters.
All this assumes an honest cell maker or seller, but so many are just plain liars. It's unreal how much stuff is just made up, but this is where our world is right now, can't trust anyone. Always look for reputation of the source and how they value their name.

While I'm ranting, might as well throw another one in , something that people argue about to no end.
What is a definition of a full charge or charge voltage? 3.4V to 0A , or 3.5V to 0.1C , or 3.65V to 0.2C, or 3.8V to 1C
Answer - all of the above can be true or false, depending on bunch of other things.
What does it even mean to be "full"? Based on what metric or spec? Every metric has a range, there is no single answer.
Answer is picked by an engineer based on multiple data points and application specifics. Resulted battery will perform and last based on how well the engineer understood the parameters and made appropriate trade offs. THERE ARE ALWAYS TRADE OFFS !!!!
So, since in the battery world 2+2 can be 4 or a range from 3 to 5, this opens the door for endless arguments and FUD.
I have seen it since I joined this field 12 years ago and just gets worse and worse.
End of rant... everyone have a great weekend!

electric said:

There are scientific papers you can find easily via Google, where names of scientists and their institutions give them plenty of credibility. Effects of temperature, SOC and calendar aging on cycle life of Lithium batteries.
Everything @ghostwriter66 posted has merit...

Click to expand...

I've read hundreds of scientific papers the past few months, and don't remember ever seeing one that tested how long in the knee is "safe", just to avoid them, which is why I was questioning it. Do I think a few hours in the knee is a problem...probably not. But when balancing 16 1000ah cells in parallel, with a 5 amp power supply say, that can be days or weeks. It may still be days with a 50a supply, so it just got me wondering.

I value what you say because you are one of the knowledgeable ones here...but like you, I take all info without test data to back it up with a grain of salt. Heck, even some of the things in research papers I take with a grain of salt! Trust but verify, and it's tough to verify without test data. So I'm back to "better safe than sorry" until more data is put out there.