Amount of cells? many or few

MK2R

100 W
Joined
Feb 19, 2021
Messages
114
Location
Sweden
Is it "always" better to have as many cells as possible, or is there a balance between min and max?

I got 200 "good" cells. panasonic and samsung, together they can deliver the current i need, testing range from 2400-2600mah each. the panasonic have a higher potential nominal current, so the samsung should be the "weak" link and limiting factor.

But how many cells can i actually have in paralell, and is there a limit where it could be concidered stupidly much?

with 200 cells, average of 2.5ah the pack would be 2.5*3.7*200 = 1850wh. i want to use a maximal output of around 2kw, so less or more discharge will be around 1.1C at max, and in most cases a lot less.


The question is, if i add more cells, even not as high quality, (let's say 1800-2000mah, but still within a safe discharge of 1C) would it make sense to add 100, or a couple of 100eds of those cells to the pack? would it change the potential heat development of the total pack for example, or would it make little to no difference and just add a larger storage?

I realize that weight and volume will increase, but apart from that, is there any benefits?
 
Good is having a 100Ah battery and only using 20Ah which gives you more total cycles.
Bad is having a 20Ah battery and using all of it when you ride.
 
The more cells you have, the lower the per-cell charge and discharge rates. That's easier on the cells and makes them last longer and perform better, all else equal.

You don't say what kind of vehicle or other application you have for these, but adding hundreds of extra cells to most small EVs would diminish their ride quality and performance.
 
Also , while the more cells you have in parallel ,the less total resistance there will be, mixing high impedance and low impedance is a bad idea , from my limited understanding. You wouldn't want VTC5a mixed with PF in other words.
 
Sorry, i didn't specify what i was intending to use it for. Less or more a powerbank to power an inverter of 2500/5000w. however i am not intending to use it to maximal output. I want to be able to charge my plug-in hybrid car, be able to use it during winter time with a timer to heat my car when i am going home from work etc. And also be able to use as a regular power supply incase power is out at the home for some reason.

So i am thinking either if u can build several banks and connect them somehow, or live with the facts that you have to lift it around. 200 cells would only would be a lot lighter, but even at 600cells we're looking at about 30kg, a bit heavy but liftable for sure
 
Dak77 said:
Also , while the more cells you have in parallel ,the less total resistance there will be, mixing high impedance and low impedance is a bad idea , from my limited understanding. You wouldn't want VTC5a mixed with PF in other words.

Mixed cells in a parallel group will work out their differences, with the stronger cells pulling most of the weight (and the weaker cells maybe even recharging the strong cells in between bursts). The main thing is for the groups in series to be all about the same in current capacity and amp-hour capacity.
 
MK2R said:
Sorry, i didn't specify what i was intending to use it for. Less or more a powerbank to power an inverter of 2500/5000w. however i am not intending to use it to maximal output.

For a stationary application, you might as well cram in all the cells you have available, as long as they're all the same operating voltage and all the parallel groups have about the same composition of strong and weaker cells.
 
Chalo said:
MK2R said:
Sorry, i didn't specify what i was intending to use it for. Less or more a powerbank to power an inverter of 2500/5000w. however i am not intending to use it to maximal output.

For a stationary application, you might as well cram in all the cells you have available, as long as they're all the same operating voltage and all the parallel groups have about the same composition of strong and weaker cells.

Okay! but in regards to composition, i can make the composition in 2 ways, either by an equal mix in regards to mah, or an as even as possible blend of internal resistance. (or both if possible)

The only videos i find they just mix the cells in regards to mah, and no1 really talks about internal resistance.

But i guess with a lower output per cell, internal resistance matters less compared to a scenario when you want to have an output closer to the limits of the cells (like 2C or even 3C).


So. if i understand this right, for my application, the more cells i use, the less i will stress each cell, no matter if they are testing out good and high quality or a bit more poor but still performing.
The reason for why i mix the paralell units to have an as equal as possible mah rating is so when i am starting to get closer to emptying the battery pack, the bms will not prematurely cut-off the pack (for example 1 unit is at 20% and some are at 45 or 50%. instead i want a scenario where 1unit reach 20% and most of the others also have reached 20-25%).
 
Chalo said:
Dak77 said:
Also , while the more cells you have in parallel ,the less total resistance there will be, mixing high impedance and low impedance is a bad idea , from my limited understanding. You wouldn't want VTC5a mixed with PF in other words.

Mixed cells in a parallel group will work out their differences, with the stronger cells pulling most of the weight (and the weaker cells maybe even recharging the strong cells in between bursts). The main thing is for the groups in series to be all about the same in current capacity and amp-hour capacity.

Ah ok. Thanks for clarifying that and being nice about it .
 
MK2R said:
Is it "always" better to have as many cells as possible, or is there a balance between min and max?
But how many cells can i actually have in paralell, and is there a limit where it could be concidered stupidly much?
More capacity is always better. The advantage of more cells in parallel is that cell variance averages out. The disadvantage, as recently discussed, is that a single (untested, reclaimed) cell can kill a whole parallel string. It sounds like you're testing every cell, so this shouldn't be a problem.

MK2R said:
Okay! but in regards to composition, i can make the composition in 2 ways, either by an equal mix in regards to mah, or an as even as possible blend of internal resistance. (or both if possible)

But i guess with a lower output per cell, internal resistance matters less compared to a scenario when you want to have an output closer to the limits of the cells (like 2C or even 3C).
With a large parallel cell count, you don't have to do either -- just distribute equally by brand; the capacity and IR variance will average out.

MK2R said:
So. if i understand this right, for my application, the more cells i use, the less i will stress each cell, no matter if they are testing out good and high quality or a bit more poor but still performing.
Correct

MK2R said:
The reason for why i mix the paralell units to have an as equal as possible mah rating is so when i am starting to get closer to emptying the battery pack, the bms will not prematurely cut-off the pack (for example 1 unit is at 20% and some are at 45 or 50%. instead i want a scenario where 1unit reach 20% and most of the others also have reached 20-25%).
This should be insignificant, and regardless, you shouldn't emptying the pack to LVC anyway. And with more capacity, this becomes even less of a concern.

You can visualize this dispersion and the effect of adding parallel cells with the attached spreadsheet.
 

Attachments

  • parallel dispersion.xlsx
    11.5 KB · Views: 17
fatty said:
MK2R said:
Is it "always" better to have as many cells as possible, or is there a balance between min and max?
But how many cells can i actually have in paralell, and is there a limit where it could be concidered stupidly much?
More capacity is always better. The advantage of more cells in parallel is that cell variance averages out. The disadvantage, as recently discussed, is that a single (untested, reclaimed) cell can kill a whole parallel string. It sounds like you're testing every cell, so this shouldn't be a problem.

MK2R said:
Okay! but in regards to composition, i can make the composition in 2 ways, either by an equal mix in regards to mah, or an as even as possible blend of internal resistance. (or both if possible)

But i guess with a lower output per cell, internal resistance matters less compared to a scenario when you want to have an output closer to the limits of the cells (like 2C or even 3C).
With a large parallel cell count, you don't have to do either -- just distribute equally by brand; the capacity and IR variance will average out.

MK2R said:
So. if i understand this right, for my application, the more cells i use, the less i will stress each cell, no matter if they are testing out good and high quality or a bit more poor but still performing.
Correct

MK2R said:
The reason for why i mix the paralell units to have an as equal as possible mah rating is so when i am starting to get closer to emptying the battery pack, the bms will not prematurely cut-off the pack (for example 1 unit is at 20% and some are at 45 or 50%. instead i want a scenario where 1unit reach 20% and most of the others also have reached 20-25%).
This should be insignificant, and regardless, you shouldn't emptying the pack to LVC anyway. And with more capacity, this becomes even less of a concern.

You can visualize this dispersion and the effect of adding parallel cells with the attached spreadsheet.

Wow. What an answer <3

I am learning more and more each day that goes by, i am still a few weeks from actually doing this. still getting parts so i am gathering information and thought meanwhile i am waiting.

I am working on battery pack ideas in a 3-d program called blender, been working with it before. so before i am even touching my spot welder i want to create my arrangement in there and possible get some feedback. i also found out that the nickle strip i ordered needs to be put in 2 layers (0.15mm x 2) to safely draw 60A. My inverter will maximally stress the battery end with 52A per parallel string, but my applications will hopefully not stress more than 45A per string. and if i do calculate everything right the inverter will be at lowest possible working voltage when my battery cells are at 3.07V and i hope it is more than safe to stop at 3.07V rather than going down to the sub 3V region


One question that is important i do believe. before i weld or connect anything, each cells should be at the same voltage to not have a sudden energy surge while they try and get in balance with eachother. but how close do they need to be? it will actually take some time to charge all the cells and the time it takes from charging the first cell until the last cell is charged will be long, i can only charge 8 batteries at a time. So what is acceptable voltage range to connect batteries? like can i connect a 4.18V with a 4.20V battery, or should they ideally be at 4.20V and same % of SOC?
 
If you hook up two cells that are identical (or two large paralleled groups), then the one at 4.2V will equalize with the one at 4.1V, and both will end up near 4.15V.

As long as they are not getting too warm to touch, you are likely OK. I would just charge both groups to the same voltage using the charger you have, and anything less than 0.1V is unlikely to ever be a problem.

If you are still concerned, run the initial equalization connection through a resistor, like an old simple 60W filament bulb (not LED, or CFL)
 
MK2R said:
I am working on battery pack ideas in a 3-d program called blender, been working with it before. so before i am even touching my spot welder i want to create my arrangement in there and possible get some feedback. i also found out that the nickle strip i ordered needs to be put in 2 layers (0.15mm x 2) to safely draw 60A. My inverter will maximally stress the battery end with 52A per parallel string, but my applications will hopefully not stress more than 45A per string. and if i do calculate everything right the inverter will be at lowest possible working voltage when my battery cells are at 3.07V and i hope it is more than safe to stop at 3.07V rather than going down to the sub 3V region
Good plan :thumb:

MK2R said:
One question that is important i do believe. before i weld or connect anything, each cells should be at the same voltage to not have a sudden energy surge while they try and get in balance with eachother. but how close do they need to be? it will actually take some time to charge all the cells and the time it takes from charging the first cell until the last cell is charged will be long, i can only charge 8 batteries at a time. So what is acceptable voltage range to connect batteries? like can i connect a 4.18V with a 4.20V battery, or should they ideally be at 4.20V and same % of SOC?
It really doesn't matter -- they don't have to be close. The high charge cell will sag down under load, decreasing the voltage delta and thus current flow, so it's self-limiting. Just don't connect a completely empty cell to a completely full cell and you'll be fine.
 
There are so many experienced and smart people in this forum, if i wouldn't have asked my questions on here and went the way i was thinking of in the beginning i would probably have screwed a lot of things up. Recently i stumbled upon a video regarding inverters (actually it was a video about solar power and MPPT units, i am still trying to figure out how they safely charge a battery in regards to cc-cv chargers).

With high output inverters, like mine at 48V and 2.5kw/5kw some people had problems with the on and off current. "the capacitors" need to be precharged like putting a small resistor between the battery pack and the inverter before you connect the battery itself to the inverter. and from what i understand, i need to do this as well, each time i do connect the pack to the inverter etc. but does it matter what resistor i use? i have some 5w spare resistors, idk if it would work? the only thing i know is that they used a 25w resistor to precharge a 3/6kw inverter to be used with a 100ah lifepo4 battery. SO less or more, in this field of what i shall use to make it safe, i have no clue what so ever.

Also why is there not a thumb up or like button? there are so many helpful comment i would have loved to leave a like/thumb up on!
 
Alright! got my first liitokala ill-500 unit, one more is on it's way, the first 4 cells tested out at 2046,2047,2073,2064 at internal resistance ranging from 32-36. the used cells in the packs are supposed to hold about 2000mah. i am not sure if i can tell anything from the first test, but if i am lucky this "50" pack of cells might be pretty good then.

This is a 3d model of the cell holders i ordered, still waiting for them thou, and the nickle strip. all in picture is in scale, even the nickle band at 0.15mm thickness

Skärmavbild 2021-04-06 kl. 22.10.56.png


So meanwhile i am testing my cells i have tons of time to design packs! :D wish me luck, cheers!
 
So. am i building this pack right? 13s33p. red "dots" are battery + end and plain silver is - end. Red and Blue cubes represent + and - at the battery pack

Skärmavbild 2021-04-06 kl. 23.00.09.png

Skärmavbild 2021-04-06 kl. 23.39.51.png

Skärmavbild 2021-04-06 kl. 23.40.15.png
 
Another idea of a 3x3 with an empty core. in total "just 25p" so for a 13S it's 325 cells, and a 14s will be 350 cells.

in this pack there will need to be insulation between each pack stacked on top of eachother. but the series connection will be basically nickle plates "U" shaped/folded to attatch + to -

Much more compact solution. i can put it in with a filled core. at 13S it will be 442 cells (total cell holder capacity will be 450 cells, so 8 will be empty slots). at a 14S it will be 448 cells, so just 2 empty slots. however capacity will be similar.

Each cellholder measures 33cm 5.5cm + a few mm on top of the 18650cell (maybe a 18650 cell will be 70mm in total hight instead of 65 since the build of the cellholder).
so a cube the size in cm: 33*16.5*21 + some spacing/insultaion where the pack builds in hight etc. so without an outer shell it's somewhere in the ballpark of maybe 35*20*30 cm. from 325 cells to 448 cells it weights about 15kg-21kg (rounding up for bms weight).
Skärmavbild 2021-04-07 kl. 04.02.19.png

This is pretty fun, and i find blender useful to work with basic 3d models. i guess when u fold packs, you need to first weld it next to each other, and then fold it afterwards over some kind of insulating part so the batteries won't touch eachother
 
If this is your first high voltage, high current build, and you don't have packaging constraints like you would in a vehicle, I'd suggest the simpler planar build first -- there's just less to go wrong. It's inevitable that you'll iterate (rebuild) or build another pack with knowledge and experience that you gain from the first one, so there's no need to overcomplicate the first one.
 
yeah, it is my first build, i am also thinking about going the more simple pack to begin with. it felt like a pretty solid arrangement at first glance, when i have the batteries tested and all set up it's not that expensive to pull it apart and get new nickle strips and rebuild it to something else. So far i've just tested 12 cells, rated at 2000mah (at what amp, idk). but at the lii-500 my first 12 (out of 450) are testing at above 2000mah (2040-2090), so if i am lucky some of the packs have not been cycled many times at all. at a price of 150 usd for all 450 in total this testing might turn out cheap per cell. oh well, only time will tell ( A LOT of time, damn testing 4 cells at a time is time consuming, i can't wait for my second charger so i can test 8 cells at a time).

Just got my daly bms today, a 13s 60A one (10A charging, 480Watt?). I will see what i can do in the 3d program when it comes to wire the bms up, i am going to ask for advice later on when that is done. I know for sure i am overthinking a lot of things, but since it is my first build i rather ask before i do something and hopefully prevent anything obviously bad happening. and i guess the first idea is safer to build, the more compact version have a lot of potential to short-circiut while u build it, just some tiny mistakes can ruin everything if you accidentally connect + to - while building
 
fatty said:
If this is your first high voltage, high current build, and you don't have packaging constraints like you would in a vehicle, I'd suggest the simpler planar build first -- there's just less to go wrong. It's inevitable that you'll iterate (rebuild) or build another pack with knowledge and experience that you gain from the first one, so there's no need to overcomplicate the first one.

Fatty, i got a theoretical question about the first layout. they are packed very dense compared to the normal square brackets u can buy. i am thinking about heat and cooling. if i will use the pack at somewhere around 0.5-1C discharge i am expecting some heat production, each cell are made to withstand this, but when u pack 300+ of them tight like this the collective heat production might be somewhat more than they can handle. should i make some design changes to allow more air between the cell packs, maybe add a fan? if possible or just go with the design as first intended :s or am i here, as in most other cases overthinking everything?
 
depends on the cells you are going to use. crappy cells with high IR will heat up more and cause issues on compact builds like your stacked idea. but flat plates have more surface area so it can shed heat better.

1C is generally not an issue unless you are living in a high temperature climate. generally you want to keep cells below 40c.
 
Alright, i guess that sounds fair. the cells i am testing is in the range 25-35 miliohms (according to liitokala). I am not sure what that actually stands for in regards to heat build up. only reference i really have is that some highpower cells might have around 10miliohms IR, and some cheap crappy chinease might be into the field of 100 or more. I only know the batteries was used in e-scooters and e-bikes, so i am guessing that whatever their factory spec was, it was enought to be useful for that i guess. Oh well, testing will continue! can't wait until i get my second test/charger. testing just 4 cells a time, it takes at least 6 hours per run.

I live in a somewhat cold country i would say, sweden is not one of the hottest places on earth, but summers can be warm, esp if you leave it in the car a sunny day i guess? but apart from that, it is not that warm in general
 
35millioms at 10A gives you a voltage drop of 0.35V and a power generation of 3.5W.

in each cell.

for 10milliohms and 10A its 0.1V drop and 1W of heat....
 
So at 13s33p. 429 cells at 1A is 429x3.7 ~1.6kw. So if i want 1.6kw, it will be about 0.35W per cell, 0.35*429 which is about 150watt of heat. and if we stress the pack at 2kw or even the maximal of the inverter (2.5kw) we are somewhere in the range 200-250watt of heat development.

I read the numbers, i am not sure how i should relate to them. i know my heater in my room runs at a lot more than that. it is equal to 3-4 "old-school" lightbulbs with tungsten. or about the same heat as 3 human bodies together produce. I wonder if passive cooling from the air is enough? I am sorry, i don't have the experince to be able to make an good estimate of what is enough cooling and heat build up over time etc. I guess in the winter when it is sub freezing outside it would benefit from this to heat up the batteries to a good working temperature, but in the summer when u get 30C outside i guess an additional 200watt is not going to do the pack any good


btw, this is correct understanding of how to hook up the bms?(red is + and blue is - )
Skärmavbild 2021-04-09 kl. 20.34.58.png
 
No, you generally should design a pack not to require active cooling -- it's a good indication you're pushing the cells too hard.

0.5-1C will be no problem.
 
Back
Top