diy solar

diy solar

A faster & better way to top balance?

nate_syd

New Member
Joined
Apr 14, 2021
Messages
161
This has probably been done before, and maybe there's a better way - but as a n00b i read the usual sticky post on balancing & in parallel it was going to take AAAAAGES for a 24v 300AH pack at 3.6v
So i thought about it & have a really simple way to do it much faster.

The issue is that you need the biggest voltage difference (charger vs battery pack) to push current when charging, when you put all cells in parallel - this is the lowest voltage drop you could possibly have. A full pack in series is the highest voltage drop.

So here's the process...
  • Get yourself a hobby battery charger that can do LiFe chemistry, they're soooo cheap!!! & can use a 12v computer PSU, and have awesome (1.5A) balancing, and big current capacity. Drones, remote control whatever... that type of hobby charger.
  • Connect up the battery bank in series, connect all the balancing & start the charge, lets say its a 24v 8s pack
  • 1 cell will hit high voltage, remove that cell from the pack & re-assemble the pack - its now a 21v 7s pack
  • Set the hobby charger to 7s cells, connect all balancing & charge
  • the next cell will hit high voltage, remove it from the back & re-assemble - a 18v 6s pack
  • .... repeat
All cells will be removed 1 by one when they hit peak voltage, and you've kept the highest charge current at all times, you've had balancing & safety on - so its all "safe". There is NO BMS!!! so not completely safe, you need a good charger that stops when it hits full voltage, which is easy enough for a hobby battery charger.

Feel free to point out any errors or problems, or link to a better way/information if there is any. The biggest downside is the removal of the BMS, so watch like a hawk & make sure you KNOW that the charger disconnects & stops when it should on the voltage.

Hope that helps any other n00bs & speeds up everyone's balancing.
 
There are folks who were watching like a hawk and ended up with puffed, totally ruined cells.

Patience, to avoid a total loss of a cell seems prudent here. Especially since it is very difficult to get a replacement cell in a timely fashion.
 
Let's try this another way.
The big LFP cells we deal with are "tedious" to Top Balance because they hold a LOT of AH.
Take a 200AH LFP cell. It can discharge 200A (1.0C) and take 100A to charge (.5C)
From 0%SOC (2.50v) to 100% 3.65v would take 2 hours at 100A, 4 hours at 50A and so on.
To TOP BALANCE, all cells are placed in PARALLEL ! NOT Series.
They are then charged to 3.650V with as many amps as possible but NOT exceeding the cell Limits. This continues until the amps taken falls below 1.5A (this is considered as saturated). Technical spec says cutoff can occur at 0.05C, for a 200AH cell that is 10A.
Once all the cells have reached 3.65 & saturated, then they should be left connected in parallel while they settle for the next couple of hours. Typically the cells will drop to around 2.550 within 1 hour and 3.500+ within two hours. This is NORMAL for LFP.
Once settled, then configure that battery pack into Series Config with BMS and fuse and your ready to go (casing etc of course being done too).

Shortcuts & rushing usually do NOT end well, Patience Testing happens in this process. BE PATIENT ! Very patient !! smoke'm if ya gottem LOL.

One "shortcut" that is safe:
You can Top Charge 4 cells - This is the same as Top Balancing but with 4 cells at a time it will be faster to get the cells up to 3.65. Even after they settle, getting them back to 3.65 will be Much Quicker. Do the cells in "lots" of 4 cells, once you have 8 cells Top Charged, then do a Top Balance with all 8 cells in parallel, (it will be quicker) and then assemble the battery packs per standard.

I have built several packs and suffered many of the hard lessons and costs that go with that... some pretty $$ One of the reasons I wrote the Luyuan Tech LifePo4 Assembly guide (link in my signature). I gave up playing & bought a TekPower TP1540E bench Power Supply that can output up to 15V @ 40A and charging the cells at 3.65V/40A goes fairly well. The CC ( Constant Current) get's the charge to 90% and CV (Constant Voltage) takes it the rest of the way as the amps decrease as the cells saturate.

BMS' with Passive Balancing do not do much for Large Capacity cells, that is NOT their target, they can only handle small capacity cells as it only burns off Hi Voltage from cells and is very slow. At this time, very few BMS' have Active Balancing which is suitable for Large Capacity LFP cells. Active Balancers do work and QNBBM, Heltec are Known Good devices that work as designed but they are not that cheap either.

Bulk Commodity ESS grade cells CAN deviate up to 1,v per AH of capacity when reaching above 90% SOC. That's 200mv for a 200AH cell. Top Charging & Balancing can resolve "some" of that and smooth things out.

Matched, Batched & Binned cells (cost more) will behave identically to each other through their entire voltage & load cycles (just like in EV's) and they will not generally deviate more than 30mv and will often settle in at below 10mv difference.

Example of Bulk versus "properly" Matched EVE-280 Bulk = $92, Matched = $125 (and THAT is a deal, typically it is 50% above Bulk Cost) due to the 3 full cycles of Charge/Discharge at various C-Rates with measurements logged every 10 minutes.

FYI: Many Chinese Vendors rely on the fact that the Translator Software is partly borked and inaccurate... (the excuse)
Many say they sell "Matched" cells BUT what they mean to say, is they are Voltage & IR (Internal Resistance) matched so they are identical when shipped. THIS IS NOT MATCHED ! This is Voltage & IR Matched at the storage voltage NOT through operational Test Cycles which is what Proper Matching IS.

Several links in my signature will help you, so grab the downloads etc.
Hope it helps, Good Luck.
Steve
 
A+ thanks Steve.
I found my bench supplies (usual style of bench PSU) could only supply very little current to the pack when in parallel, despite a few low cells that needed a lot. I think it was Wills vid where he connected in series first & then moved to parallel.

Top balancing... i take that to mean that all cells are evenly and fully charged to their "full" capacity states. How you get there is not important, but balanced just means that all cells in a pack are 100% charged. Is there actually a difference between doing the whole bank in parallel, 4 in parallel, or my method of removing 1 by 1?
Yes the parallel method is absolutely safer (+Joe above), but the end result is technically the same?
 
A+ thanks Steve.
I found my bench supplies (usual style of bench PSU) could only supply very little current to the pack when in parallel, despite a few low cells that needed a lot. I think it was Wills vid where he connected in series first & then moved to parallel.

Top balancing... i take that to mean that all cells are evenly and fully charged to their "full" capacity states. How you get there is not important, but balanced just means that all cells in a pack are 100% charged. Is there actually a difference between doing the whole bank in parallel, 4 in parallel, or my method of removing 1 by 1?
Yes the parallel method is absolutely safer (+Joe above), but the end result is technically the same?
if you have a BMS that will cutoff when 1 cell reaches max voltage then it will be safe.
I'm not sure that you can get a bms to work with different numbers of cells though.

otherwise the danger is that you miss a cell topping off and it gets overcharged.

FWIW I use a cheap switch mode power supply nominally 3.3V 50A, adjusted to 3.65V.
Actually I've got 4 in parallel now. They start at about 40A each if the cells are empty.
 
if you have a BMS that will cutoff when 1 cell reaches max voltage then it will be safe.
I'm not sure that you can get a bms to work with different numbers of cells though.

otherwise the danger is that you miss a cell topping off and it gets overcharged.

FWIW I use a cheap switch mode power supply nominally 3.3V 50A, adjusted to 3.65V.
Actually I've got 4 in parallel now. They start at about 40A each if the cells are empty.

perfect - thanks Steve
 
Just a caution about paralleling cells for top-balance without getting them charge to at least just poking above the knee individually. (happened to me)

If you don't charge them up individually to just get into the upper knee when you slap them together, you risk damaging a cell that was in a low state of charge to begin with when initially placed in parallel.

Ex: 4 cells arrived to make a 400ah 4S bank. All cells showed roughly 3.2v on meter. Check. BUT, due to the flat curve of LFP, that can mean anywhere from 20-80% SOC. We knew that, but were in a hurry.

Put 3 cells in parellel and watched current travel quickly and settle down amongst them as they got assembled.

Then came the 4th cell. It was actually pretty low in SOC. Attached the last parallel bus-bar connection. ZAM.

Essentially what happened here in parallel, with that last low-SOC cell, was that the 3.2v 3P / 1200ah bank made a HUGE inrush current to a 400ah cell, which was not designed to have 1200ah pushed into it. That cell was never the same.
 
The primary Voltage Curve, is from 3.000-3.4000 with nominal being 3.200
Sorry but I have been at this a little while, my directions provided are safe acceptable & works. There are a few different ways to accomplish Top Charging & Top Balancing, several mentioned here, some good and some downright dangerous... like using a car FLA Charger to hard charge LFP and ending up with puffed useless cells.

Initial Charging fresh cells in series with a BMS can work partially but it will not Top Balance or level up the cells and everytime you ave to disassemble & reassemble there is an extra risk for something to go wrong.

KISS Applied is always best !

lfp-voltage-chart-jpg.27632
 
Steve - I'm in full agreement. I'm just cautioning what can happen if one is careless and feels rushed for time to do these things. Especially with a very large parallel string on the bench.

It's the real-world factor of excitement. Despite reading yours and other great directions here, we know that some will open the box, and start to slap 'em together like lego's.

Those great charts you have show under-charge voltages, not the settled at-rest, out-of-box voltages and just slam in parallel without realizing how far apart a settled SOC voltage can be.
 
Last edited:
I am glad someone gets something out of all my writings & documentation. You are Most Welcome !!! It does feel good to have someone say Thanks once in a while, for the efforts made.
Then I too shall say "Thanks". I have learned much from your posts and resources.
 
I top charged mine in series with no trouble at all. Yes the BMS was on and all was fully protected. Used a 12volt 20amp battery charger. Used two shunts to bleed power off the top cells as it charged. All arrived at 3.6 very close together on the second try. I expect next time I will do it on the first shot.

OK so the resting voltage had some variation. Switched the balancer on full time and let it run overnight. Next day charged again and the delta was 10mv or less at the top with no intervention. Reset the balance function for during charging only and I am done.

The parallel science for top balance is a perfect method. And still my series balance is plenty good for me.
 
Steve - I'm in full agreement. I'm just cautioning what can happen if one is careless and feels rushed for time to do these things. Especially with a very large parallel string on the bench.

It's the real-world factor of excitement. Despite reading yours and other great directions here, we know that some will open the box, and start to slap 'em together like lego's.

Those great charts you have show under-charge voltages, not the settled at-rest, out-of-box voltages and just slam in parallel without realizing how far apart a settled SOC voltage can be.
LOL, I hear what you're saying and on the odd occasion someone will have a cell flipped the wrong way & they very quickly realize it, a few have actually admitted to it... People don't like to say they screwed up ! should not surprise anyone right... human nature.

LFP is actually fairly "POLITE" compared to other Lithium Chemistries. You can string a set of cells together in Parallel even if they are at differing voltages and they WILL level up. IE, take 2x @ 3.200, + 1 @ 3.300 & 1 @ 3.100 Set then in parallel and leave them alone for 2 hours, they'll level pretty close. Even if you apply charge right away, the lower cells WILL absorb more at the start till they level up. * When using a proper CC/CV (Constant Current/Constant Voltage) Charger System.

An Insider Secret (sorta)
Fresh new cells leaving the factory, leave at 3.200 on the dot. That IS default for LFP as That is the Nominal Voltage and is considered the 50% Point (which it is). This is of course the optimal Storage & Shipping voltage state for LFP cells, this is also the safest voltage to ship due to Temps & Conditions... @ 3.2 they can freeze or get heated up and it won't bother them during shipment.
The trick secret. Resellers won't charge cells, they can't be bothered with the volumes they handle.
As LFP sits, it slowly discharges (normal) , so the voltage drops slightly day by day, @ 3.200 midpoint, there is a heap of reserve, so it is slow.
The lower that voltage is when you get it in your hand, is an indicator of how long the cells have been stored.
LFP Self-Discharge being on average of 3% per month while in storage. The older the stock, the lower the voltage will be.
This becomes evident when people buy cells that were sourced from Battery Brokers and you get cells with a Low VPC, some folks here have received cells @ 2.75Vpc, so you Know they been sitting somewhere for a LONG Time....
- The Quick upshot: The closer the cells are to 3.200V, is an indicator as to how fresh the cells are.
- IF all the cells come at exactly the same voltage, it is generally OK to assume they came from the same production Batch.
- IF the cells are Random Voltages, even within 0.100v difference, they are likely from Different production batches.
* Each batch can be slightly different.

Just some added info many may not even realize or be aware of.
 
Just charge them normally in series with bms.
Once bms shuts of because one cell hit 3.65 then the rest of the cells should be close to full.
Now you can top balance in parallel.
 
Once bms shuts of because one cell hit 3.65 then the rest of the cells should be close to full.
No - thats not the case & that was the whole issue i was working around. The others werent close & would take AGES in parallel, in small batches of serial its much faster
 
If you are going to charge vastly different SOC cells in series configuration, use an active balancer and a BMS that tapers current when a cell gets close to the target upper voltage.

I wouldn’t ever recommend connecting a cell to a charger without having cell level monitoring that is capable of disconnecting the charger. I either use the active balancer method, or charge individual cells while using a celltop monitor set to disconnect the charger at the target voltage.
 
Hook up in series and use a BMS to get you close to the top (but the BMS will keep you from going over on any single cell), then when you hit the top on one cell, rewire in parallel and top the cell to 3.65V.

FYI I started with 4x 271Ah CATL cells within 0.002mV of each other (3.276, 3.277, 3.277, 3.278). I tried to step charge in parallel with a 10A bench supply at 3.4V but it went right to CV and was pushing <2A. After ~12 hours I ended up disassembling the pack and reassembling as 4S but with the BMS attached. I just set it to 8.5A @ 14.0V and after ~18 hours it finally pushed 2 of the cells to the 3.65V cut-off. I rewired in parallel and pushed them to >3.63V (3.630, 3.632, 3.634, 3.637), all within 0.007mV of each other. The parallel charge still took probably 5-6 hours since the charger was in CV mode and despite being set to 8.5A was actually pushing less than 1A most of the time. If I'd tried to parallel charge the entire pack it would've taken at least a week, maybe two.

Note that as I got up above 3.45V the individual cells diverged quite a bit - as much as 106mV at one point... so don't think just because your cells looked matched at a lower SoC that they'll all hit 3.65V anywhere near the same time. As others point out trying to use voltage is meaningless below ~99% SoC

BTW by my math my cells were probably in the 20-30% charge state when they arrived.
 
In theory: What if you connect 4 cells (or whatever BMS you have supports), and when the BMS shuts off charging, remove that one high cell and replace it with a fresh one. Repeat the process each time a cell trips the BMS. Eventually you'll have them all around the same level so you can go back to putting them all in parallel to finish a low voltage equalization.
 
There are MANY threads on THIS Subject, some lasted over a year with 100's of replies.
The sum total of the experiences and methods Which Work Well has been discussed and yet the same little Dance around the Mulberry Bush to confuse and beffudle occurs again.

First Rule that Applies to Most Everything ! K.I.S.S. Keep It Simple Silly !
The more complicated & involved the more likely that Murphy will Apply his laws and Ruin your Day !

Some suggestions provided here... well if you want to Fry a BMS, go for it. Many others did that too and shared that as well... but Never Mind. Hard Lessons are always available to those who want to learn the hard way and usually the costliest way.
 
Back
Top