how fast to charge 8 lithium ion battery in parallel

Thread Starter

christos christou

Joined Aug 10, 2018
9
hello everybody i'm working on a small vacation project to explore things, i have 8, 18650 battery connected in parrallel 3.2V-4.2V and each cell has 3500mAh total of 28Ah how fast can i charge it ?
 

oz93666

Joined Sep 7, 2010
737
With respect to the expert above ,you most definitely can charge in parallel ... this is standard procedure in all powerwalls...

Each cell will always have the same voltage as the others , that means higher capacity cells will receive slightly more current than lower capacity cells .... this is no problem as long as you do not exceed the upper limit ...

The recommended charge current for most 18650's is around 1/2C to C to get a good life ....

If you charge your 8p pack at no more than 20 A this will be fine ...

The important thing about these cells is the voltage you charge to ...this has a great effect on the life of the cell . you can safely charge standard cells to 4.35 V they hold 15% more power than the normal 4.2V, but will fade after 100 cycles if you charge only to 4V you will get 1000 cycles
 

digsys

Joined Jun 27, 2018
27
Pretty much in total agreement with oz :)
All our solar EV race packs, and most EVs are made with cells in parallel, often many. It's standard practice in many other systems as well.
When / if a single cell fails, it nearly always goes high-impedance, which gives great redundancy and can be detected if needed.
IF you mix cells, age and capacity, you will see inefficiency and ageing build up over time. It can become an issue if you charge at near max,
as some cells may go over. If you're not after MAX life or performance, it may not be an issue for you.
Agreed with the 4.35V (we only got ~10cycles), and have settled on 4.1V (there's bugger all energy to gain after that).
 

Hymie

Joined Mar 30, 2018
924
The problem is if seven of the eight cells fail (going high impedance), then the charging circuit will be charging a single cell at 20A. Admittedly the charging circuit will stop once the battery voltage reaches the cut-off voltage (4.2V), - but by then thermal runaway might have occurred, resulting a catastrophic failure of the battery pack.

Perhaps this is the reason these packs fail under charge.
 

ebp

Joined Feb 8, 2018
2,332
I would be very leary of charging paralleled lithium ion cells purchased from anywhere other than an authorized distributor for a major brand at any current higher than a single cell could handle. Unless it has changed recently, the consensus has been that cells from "random" ebay vendors and their like will have capacity far lower than claimed. If the vendor also sells underpants or makeup cases, I would not trust their claims about lithium cells. A specialty seller might have decent product. Since actual capacity is often well below the claimed capacity, the assumption of allowable charge current of C or C/2 based on claimed capacity fails. I would have doubts that even two out of 8 would be well enough matched to reasonably share charging current.
 

MrAl

Joined Jun 17, 2014
7,882
hello everybody i'm working on a small vacation project to explore things, i have 8, 18650 battery connected in parrallel 3.2V-4.2V and each cell has 3500mAh total of 28Ah how fast can i charge it ?
Hi,

This turned out longer than i thought because i wanted to fully explain my line of reasoning.

Charging each cell individually is always the best bet, but when they are permanently wired in parallel you dont have any choice. I would not want to do it this way, but i realize people do it all the time. Worse though is charging in series.

The question of how to charge one cell by itself comes first.
The answer is the max charge current is determined by the manufacturers data sheet and what they show as maximum charge current. It's that simple.

The question of how to charge TWO cells in parallel comes next.
Given that we have two healthy cells that can take a max charge current given on the data sheet as 2 amps and they are connected in parallel and that current splits in a parallel circuit, since each cell can take 2 amps and that makes the total current 4 amps, it would seem reasonable to charge the two in parallel at 4 amps.
There's only one problem with that line of reasoning.
One of the assumptions about the charge currents in a parallel circuit is that the current "splits" into two branches, but what we dont know is how evenly it splits between the two branches. If we charge at 4 amps total are we getting 2 amps into each cell or are we getting 1 amp in one cell and 3 amps in the other cell?
If we get 2 and 2 that's fine, but if we get 1 and 3 that means we went over the max charge current of one of the cells. The max was stated as 2 amps on the data sheet but we are putting 3 amps through the cell which is not recommended. So what is the true answer to this question then of two in parallel when the max for any single cell is 2 amps?
The only answer can be the max total current for BOTH cells in parallel must be 2 amps. That is because one cell might not draw any current at all, or some very low insignificant current which means the other cell gets the full 2 amps. If we did allow 4 amps total that one cell would be getting the full 4 amps and that is 100 percent over the max current spec in the data sheet.

Now we can move on to 3 or more cells in parallel. What is the total max current allowed.
Unfortunately, the best answer to this question would be the same as with two cells. That the max current is the max for ONE single cell alone, and for the previous example that would be 2 amps.
It's unfortunate because it's a rather low current WHEN ALL CELLS ARE FUNCTIONING PROPERLY but when only one cell is functioning properly that one cell gets the ENTIRE current which is 2 amps and that is at it's max current already.

The reason this can happen this way is because each cell has a characteristic voltage and an ESR and both of these determine what current the cell will draw when placed in parallel to another cell with a different characteristic voltage and different ESR.

Now the obvious question is, just how likely is it that we will see a very large difference in one or more of these characteristics between cells in parallel.
The answer to that involves a probability, but unfortunately there is little to base the solution to this on. For example if we have 5 in parallel, what is the likelihood that one cell will fail and so if we charged at 8 amps (which would max out 4 cells and now that's ok if 4 cells are working correctly) we'd still be ok. There's almost no way to predict this so there could actually be 2 cells that have higher characteristic voltage and so 8 amps would be too high.

So the best answer is to charge at the max current of ONE cell even though there are N cells in the pack in parallel. Anything higher might work for a while but the risk goes up that a very bad problem such as fire might occur.

The other solution is to have the pack wired in parallel for normal USE, but when charging disconnect the cells and have each cell charge individually and match the terminal voltages before they are connected back together again. THAT is the fastest way to charge safely.
 
Last edited:

Thread Starter

christos christou

Joined Aug 10, 2018
9
thanks you all for your answers so i took take to really read all the asnwer one by one they are all useffull, all 8 battery are identical max charge voltage is 4.2 +- 0.5 volt
max current charge is 1.0C (3400mA) standart charge is 0.5C and 1700mA end current cutoff 50mA
Nominal Voltage Average 3.635V
energy (3500mA/3400mA) : charge/discharge
after reading i was conviced that i could charge my power bank at 4V 28A for around 1h charge and the 1000 cycle sounds pretty good
and when i read the last post of MrAI i was chocked by truth that one cell could draw 27A and the rest draw just 1A that could be catastrophic !
and i can't disconect them for each time of charge !
charging 3500mA*8=28A at 3.5A would take around 9h:8' that is huge
is there any logic circuit solution to disconnect the battery automatically while charging like a gate ? (i know this sounds a litle stupid)
 

oz93666

Joined Sep 7, 2010
737
i was chocked by truth that one cell could draw 27A and the rest draw just 1A that could be catastrophic !
This in my opinion is ridiculous scare mongering by people who do not work with these cells ... myself and digsye have practical experience ....... I am an active member of this forum https://secondlifestorage.com which deals exclusively with the use of these cells ... the members there all operate in parallel banks ....millions of people with ebikes also use these cells in parallel ...look on youtube , they make their own batteries by joining cells in parallel ... also e bike batteries sold by companies have these cells in parallel ...so do nearly all laptop batteries operate 2p3s (cells in parallel )
 

digsys

Joined Jun 27, 2018
27
Plus dozens of solar car teams over 25+ yrs, chalking up millions of Km .. plus thousands of Teslas and other EVs. some with 22-28 hard
wired parallel blocks, chalking up billions of Km. I seriously doubt everyone has got it wrong. A 5 sec google search found this -
www.dataweek.co.za/59316n
 

MrAl

Joined Jun 17, 2014
7,882
thanks you all for your answers so i took take to really read all the asnwer one by one they are all useffull, all 8 battery are identical max charge voltage is 4.2 +- 0.5 volt
max current charge is 1.0C (3400mA) standart charge is 0.5C and 1700mA end current cutoff 50mA
Nominal Voltage Average 3.635V
energy (3500mA/3400mA) : charge/discharge
after reading i was conviced that i could charge my power bank at 4V 28A for around 1h charge and the 1000 cycle sounds pretty good
and when i read the last post of MrAI i was chocked by truth that one cell could draw 27A and the rest draw just 1A that could be catastrophic !
and i can't disconect them for each time of charge !
charging 3500mA*8=28A at 3.5A would take around 9h:8' that is huge
is there any logic circuit solution to disconnect the battery automatically while charging like a gate ? (i know this sounds a litle stupid)
Hi,

That is 4.2v plus or minus 0.05v.
Good luck with it.

One thing you did not mention was if you are using protected cells or non protected cells. The protected cells have a built in protection circuit which avoids things like over voltage, under voltage, and over current.
Non protected cells have to be dealt with very carefully.
 

Hymie

Joined Mar 30, 2018
924
If potential disaster were to happen and most of the charge current passed through one cell, its temperature would rise (rapidly if all 27A through one cell).

Depending on the physical configuration of the cells, you could wire up two of these, monitoring the temperature close to all the cells, and switch off the charging circuit if either exceeds say 80C.

https://www.ebay.co.uk/itm/W1209-DC12V-50-110-C-Digital-Temperature-Control-Switch-Sensor-ATF/192424469588?ssPageName=STRK:MEBIDX:IT&var=492422927289&_trksid=p2060353.m1438.l2649

Or you could go for a single shot device and tape one of these to each cell, with them wired in series to cut off the charging circuit should any go open circuit.

https://www.ebay.co.uk/itm/5pcs-SF76E-Sefuse-Cutoffs-NEC-Thermal-Fuse-77-C-77-Celsius-Degree-10A-250VHGUK/202262797582?hash=item2f17cd5d0e:g:qT4AAOSw2cNak9pT&_sop=15&_pgn=1&_sacat=0&_nkw=thermal+fuse&_from=R40&rt=nc
 

MrAl

Joined Jun 17, 2014
7,882
Hi,

Yes temperature and time are often the fallback safety mechanisms put in place in case any of the normal methods dont catch the end of charge for some reason. Some maximum temperature and some maximum time.
 

Thread Starter

christos christou

Joined Aug 10, 2018
9
thanks all again !
just wannted to know so to be shure cause the batteries are the same and they are not protected cell, so if i put protection circuit between every positive and negative of two battery to limit the current that pass trought it to 3400mAh that would make second battery takes less than 3400mAh and and until the last could take 0Amp right ??
 

Hymie

Joined Mar 30, 2018
924
You could monitor the current through each cell by placing a low value resistor (say 0.1Ω) in the negative line of each cell. With 2.5A flowing through each cell, the resultant voltage would be 0.25V.

You could then build a complex circuit that monitors each voltage and makes appropriate circuit control should the voltage be too high or low.

A simpler circuit arrangement would be to feed all of these voltages to a single comparator circuit that switched off the charge if one of the voltages exceeded a preset level (say 4A).

You could even consider placing a fuse (say 3.15A) in each battery path, but remember this would limit both the maximum charge and discharge currents.
 

MrAl

Joined Jun 17, 2014
7,882
Hello again,

I did not know you had the ability to separate the cells and remake the pack. That changes everything because then you could add a monitoring circuit of which there are several kinds.
 

digsys

Joined Jun 27, 2018
27
... so if i put protection circuit between every positive and negative of two battery to limit the current that pass trought it to 3400mAh that would make second battery takes less than 3400mAh and and until the last could take 0Amp right ??
Actually nope - IF you simply set a 3400mAh charge, then all parallel cells will divide the current equally (or close enough). Should 1 cell go O/C
for any reason, the rest will re-divide equally, but slightly higher. You are correct in that you can never exceed the maximum.
No-one in multi-parallel designs puts any type of circuit monitoring on every cell. 99.9% of the time it isn't needed .. and worst of all,
it greatly reduces MTBF ! Adding active elements increases danger levels dramatically. Plus, in larger packs, determining / fixing what
actually failed, and getting to it is a nightmare.
Nearly all pack makers use a polyfuse on each cell (these stay off and reset automatically after the fault is fixed).
If you want to prove some of this yourself, the idea of adding a 0.01ohm resistor (0.1ohm wastes too much voltage), and monitoring the cells
on +/- voltmeters (analogue ones are perfect and require no DC supply) is a good learning exercise.
When you get too much conflicting information, proving it yourself is the best option.
 
Top