I don't think (but have no experience yet) that most balance current in the range of 30 to 50 milliamps will do any good with cells of this capacity. I saw the guy on YouTube reviewing balancers, and none seemed to work until he got the 5 amp current balancer (and this was very tiny cells under 10AH). Logically, milliamps into a 280AH cells means moving the decimal point a lot of digits to the left.I looked at the total delta difference between yours and mine too. Funny how close it is.
Yup, I have found balancing to be worthless. It will be interesting to see if the deltas deviate much over time. Thanks for reporting back.
EVE matched cells, some cells drops very fast
- Thread starter rgmaninit
- Start date
Solarfun4jim
Solar seduced :-)
I think even the 5A ones don't actually apply that level of amperage, until the cell difference is pretty large, rendering it fairly useless in the very short time it takes for the 'runner' to hit the ceiling.....imo.
That depends on a couple things:
First, the cells need to both be well up in the knee (where 5A can mean something in terms of shifting voltage levels).
Second, the charger needs to shift to a low-current Constant Voltage mode with currents dropping below the 5A level (or the BMS needs to trigger High Voltage Disconnect, which still allows Active Balance to even things out).
So third, Active Balance will be most successful when charging to full for some period of time before starting to discharge...
The SOC difference between 3.35V and 3.45V is less than 0.1%, so as long as cycle-to-cycle stray/mismatch stays close to those levels, you are talking about less than 4 minutes for a 5A active balancer to even out two neighboring 280Ah cells mismatched by 100mV...
(want to come clean that this is all theoretical on my part. I concluded the passive balancers are worthless, analyzed what these 5A active balancers should be able to do, and have one on the way from Heltec/China).
I will be interested knowing how well this works. The problem I see is when a little bit of a load is applied to a fully charged cell, the voltage drops quickly. As long as a small charge is applied to keep the balancer working, the balancer might work. But then the voltage rises quickly with a small charge at the top. It's going to be another balancing act...lol. I hope it works.
Yeah, that was my third point - you need a stable low-current plateau up high in the knee for active balancing to have much of any chance of doing its thing...
But a 5A Active Balancer should be able to balance out a 100mV difference between neighboring 280Ah cells in under 5 minutes, so the stable plateau doesn’t need to last terribly long.
Will report back on how it worked out.
Be sure to check ebay, they sell them there with faster shipping from the US.
Also I've found several reviewed here: https://youtube.com/channel/UC8o60l2FEHhXGZpNtDbVyBw
Last edited:
The scientific way would be to find your actual divergence and then do the math to make sure your charging voltage allows enough breathing room. If your cells aren't deviating after many charge cycles, there's no need to charge above the knee and have the BMS balance, either. Just do a manual balance once in a blue moon.
The problem with a top balance and then checking resting voltage is surface charge and whatnot. My 280's sit at ~3.55 after a week of resting! Personally I'm happy doing a top balance and then charging to an average of 3.55 vpc or so, running some cycles, and making sure everything is within reason and isn't changing. Then maybe drop down to 3.50 or 3.45 when I put the battery into service (depending on whether I feel the BMS needs to balance every time I reach 100% SoC)
Solarfun4jim
Solar seduced :-)
Once my pack is installed in my RV, they wont be coming back out to be parallel top balanced, plus i wont be taking off the connections(i like to keep the connections tight) in order to parallel wire them. Thus, my intention to only charge to 3.4v/cell, which is well inside my limit before divergence shows up. Thus they will not have any additional balance function input whatsoever. I will monitor my cell difference values and if that delta jumps above 20mv, then i know i have drift and will need to consider an active balancer....i'm hoping not.....i like the idea of set and forget.
You are certainly right.. since there are not that many possibilities between 3.4-3.6v the easier would certainly be to try some voltages and see the behavior of those guys on the tandem...?
I agree with letting them sit, connected in parallel, for several days post top balance.
With my earlier 8s EVE 280Ah packs I immediately disassembled after the end of the top balance, and as the cells settled to 3.5 and below, they diverged by 70mV or more.
Keeping them parallel to settle I've found reduces the divergence when capacity testing the assembled pack.
Solarfun4jim
Solar seduced :-)
I know i rubbished passive balancers previously(@1.2A etc) as they appeared to do little in the time they have to operate between say 3.4v and 3.6v, however on second thoughts, every time you reach the top of a cycle, the high ones are always getting held back a little and the lower ones come up. Now if this process is repeated for say 500 cycles, the the slow passive balancing will bring them all to balance at the charger cut off point. It might well be the multiple repeated iterations of the balancing that eventually gets it there??? What do folks think of this possibility?
Contrary to a lot of statements here, I've had amazing success with passive balancing. I've had packs that were out of balance by 10+Ah corrected within a day or two (makes sense when you do the math). To the point where I can top balance in groups of 4, put them into a 16s pack, and the BMS will get everything perfectly balanced out within hours. I no longer try to get the most amazing perfect 16-in-parallel 3.650 top balance sitting in parallel for 5 days resting etc etc ... The BMS gets everything in check quickly as long as I separately get all the cells to approximately 3.60+ before I put them in series.
Hell, when I take a group of 4 empty random cells, I just put them on the Chargery in series and charge at 14v and let the BMS do it's thing. It might connect/disconnect the pack over and over again, but it eventually gets balanced out. I just wake up one day to a charged and balanced pack.
Hell, when I take a group of 4 empty random cells, I just put them on the Chargery in series and charge at 14v and let the BMS do it's thing. It might connect/disconnect the pack over and over again, but it eventually gets balanced out. I just wake up one day to a charged and balanced pack.
I
If the cells are well-matched but poorly balanced, they might or might not be able to get them balanced through enough cycles.
But if you’ve got well-matched cells and get them well-balanced from the get-go, the passive balancing will probably be good enough to compensate away small deviations in SOC / voltage at full charge as they emerge...
In my experience, the passive balancers are not going to be effective at making a usable battery out of unmatched cells.
If the cells are well-matched but poorly balanced, they might or might not be able to get them balanced through enough cycles.
But if you’ve got well-matched cells and get them well-balanced from the get-go, the passive balancing will probably be good enough to compensate away small deviations in SOC / voltage at full charge as they emerge...
Solarfun4jim
Solar seduced :-)
The is great practical info cinergi. I'm only now starting to get to grips with the balancing thing.
That would mean the cells are within 0.5% SOC. (probably even closer).
This discharge curve shows how little capacity there is above 3.4Vpc Once you hit 3.5V, the cells are almost full, and you can be assured of reasonable SOC regularity.
I think its important to note that 10mv between cells at 3.2V is drastically different than 10mv at 3.6V.
I have occasionally seen people recommend this or some variant of the "assemble and parallel and let them sit" idea. All of which are totally unnecessary. Charge all the cells to 3.55-3.65V (either together, or separate if required). Then assemble.
Last edited:
Ah, yes - important to differentiate between 1-time balancing vs. constant balancing because you've got a mismatch somewhere. Obviously if the balancer can't keep up with your charge rate, you won't have a good time
Yes, a combination of well-matched cells and 1.5A passive balancing is likely to be much, much more successful than the combination of poorly-matched cells with 76mA passive balancing (and to be honest, I’ve never actually gone to the trouble of verifying that the passive-balance function on my BMS even functions as specified...).
Exactly. That's why this whole idea of getting everything to EXACTLY 3.650 in parallel is overkill. The BMS will fix it at those voltages.
I've now done at least 10 series-then-parallel method to get cells charged and balanced. It's worked perfectly every time.
Even with cells which have pretty variable internal resistance (say 30% spread on a pack). If your current rates aren't high, and you choose a lower max charge voltage, Then the BMS has overhead and time to passive balance without worrying about bumping off the high voltage cutoff for the high cells.
BTW the 1.2A (if that's what you're referring to) passive balancing on the Chargery really isn't 1.2A because it can't operate anywhere near 100% duty. So it's effectively a LOT less than 1.2A.
I have seen 300AH packs with poorly matched cells work fine with 70mA passive balancing. The key was the pack was running 0.1C average rates, and the charge voltage was well below the cutoff. 3.45-3.5Vpc max. This was combined with a absorb timer of 1 hour to allow the BMS time to work. Once the cells got pulled out of the upper range they were only 10mv apart. The key was to get the pack into the balance range regularly, so the BMS wasn't shortchanged.
On a well balanced pack this isn't needed typically, unless you are seeing high current rates.
On a well balanced pack this isn't needed typically, unless you are seeing high current rates.
From my 280Ah cells, 3.260V represents at over 14Ah (~5%) more than 3.250V, while in the difference between 3.600V and 3.590V is less than 0.85Ah (0.3%).
All of this balancing technology only has any effect if you charge well up into the knee (especially once you factor in channel read precision of +/-5mV).
Sure. The reality is, that if you are not using the top and bottom few percent of the pack, then balance is much less critical. You will only notice the imbalance when the pack starts delivering fewer and fewer AH before it drops out. In many cases that could be years of usage. There are plenty of users out there which have monitoring and disconnect only, no balancing. They just manually balance when they have an issue.
In my view most folks will probably be fine with 3.45-3.5Vpc and a 20-30 minute timer, with passive balancing. Turn the balancing on above 3.4V and target 10mv differential.
I'm going to throw this out there and you guys can chime in on what you think.
If you top balance let's say a Bank of 16 sells in series... I'm confused at why you would need to balance on discharge all the way down to cut off voltage and balance all the way back up.
Does this not throw out the whole point of top balancing if you have a new Bank of cells?.
If you top balance let's say a Bank of 16 sells in series... I'm confused at why you would need to balance on discharge all the way down to cut off voltage and balance all the way back up.
Does this not throw out the whole point of top balancing if you have a new Bank of cells?.
Most don't balance on discharge. Balancing on charge is really pointless below 3.4v in my experience, my cells certainly stay within 5 to 6 millivolts until then. From 3.4 to 3.65 is under 5% of capacity, closer to 2% in my experience. Your cells will spend very, very little time between 3.4 and 3.65, and at least mine don't show enough divergence below 3.4 to bother until your lowest capacity cell takes a nosedive around 3.0 to 3.1. So I top balance, and have (in my admittedly limited experience) seen no need for active or passive balancing. I could be badly mistaken.
Similar threads
