Amount of cells? many or few

[MK2R]

Then would it make sense to spread the pack a bit apart, even thou it will consume some more nicklestrip to connect it to allow for more air to get between the battery cells, or is this dense solution still alright in general?

at 2kw and 13s33p it is 1,26A per cell, so an average of about 0,6C (cells from 2000-2600mah 0,63-0,48c, counting on an average of 2250, 275wh per paralell unit)

I still want a like button so i can push it on your responses Fatty (yours and more, like amberwolf, very helpful user imo, and many more) :wink:

hope i am getting my other lii-500 soon, at this rate i am only testing about 12-ish cells per day, with 2 of them i'll be able to test at least 20-24 cells per day. so instead of 40 days of testing i'll be able to cut it down to about 20 and since i already got 200 good cells, i will just pick the 229 better / best cells of what i am testing right now, so hopefully i'll be able to start welding the pack before i am done with testing them all.

anyways thanks for your help fatty! =)

 

[99t4]

I still want a like button so i can push it on your responses Fatty (yours and more, like amberwolf, very helpful user imo, and many more) :wink:

Closest thing I can find to a like button for Amberwolf is this at the signature of every Amberwolf post:
"If you found this advice helpful, supporting contributions are accepted here."
(Links a paypal)

 

[MK2R]

personal like button then This forum really got me interested in ev related projects. mostly DIY stuffs. going to bid on some autions tomorrow (today) and see if i can get some LiFePO cells, 56 cells at 10ah, but we'll see how it goes, if they get to pricy i'll let them go. was thinking if i manage to get them at a fair price i wanna try some low-power solar with some mppt unit and see how they behave. from what i read up on LiFe is safer to use for that than Li-ion.

oh well, let's see how it goes =)

 

[fatty]

Then would it make sense to spread the pack a bit apart, even thou it will consume some more nicklestrip to connect it to allow for more air to get between the battery cells, or is this dense solution still alright in general?

Sure, you can space them out more if you're printing your own holders. Either way, I'd expect that allowing cell- and pack-level convection will be more important that just density. That is, make sure your holders don't seal -- provide air gaps and pathway for natural convection.

 

[MK2R]

Then would it make sense to spread the pack a bit apart, even thou it will consume some more nicklestrip to connect it to allow for more air to get between the battery cells, or is this dense solution still alright in general?

Sure, you can space them out more if you're printing your own holders. Either way, I'd expect that allowing cell- and pack-level convection will be more important that just density. That is, make sure your holders don't seal -- provide air gaps and pathway for natural convection.

unfortunately i can't print them, i am locked to the cell holders in picture (50cells) many holders of 10 cells (arrangement 3-4-3), don't have a 3d printer. i will see what i can do, gonna try some setups in the program later

 

[flippy]

1mm spacing is enough. if you need forced air cooling you are doing something wrong. you need to change your cell choice with a lower IR or increase the battery size.

 

[MK2R]

wow, just 1mm is enough? i measured it and it is 1.8mm in between each cells. i thought it was a very thight cell holder (a lot tigther than the square version).

quick question, this might sound very stupid okay, but hear me out. if i got 2 cells at 2250mah, they make 4500mah together, would 3 cells at 1500mah be compareable? same amount of mah, but 3 instead of 2, same-ish C rating as new etc

 

[flippy]

if i got 2 cells at 2250mah, they make 4500mah together, would 3 cells at 1500mah be compareable? same amount of mah, but 3 instead of 2, same-ish C rating as new etc

comparable to what?

what metric are you interested in improving?

 

[fatty]

unfortunately i can't print them, i am locked to the cell holders in picture (50cells) many holders of 10 cells (arrangement 3-4-3), don't have a 3d printer. i will see what i can do, gonna try some setups in the program later

Sounds fine. Just ensure there is a path through the holders for natural convection -- might have to drill some holes.

 

[fatty]

quick question, this might sound very stupid okay, but hear me out. if i got 2 cells at 2250mah, they make 4500mah together, would 3 cells at 1500mah be compareable? same amount of mah, but 3 instead of 2, same-ish C rating as new etc

If they share the same discharge C-rate (for example, 2C continuous and 3C non-continuous), then yes, they would be equivalent. But if they are entirely different (say, a high energy cell versus a high power cell), then no, they would not be equivalent.
Compare the datasheets and calculate to see.

 

[MK2R]

forgive me if this sounds stupid, but what if i build a 13S pack with some units with 33 cells, some with 31 and some with 35P, as long as they are the same cell type and if i balance each unit in regards to mah, it would be rather okay then?

 

[flippy]

forgive me if this sounds stupid, but what if i build a 13S pack with some units with 33 cells, some with 31 and some with 35P, as long as they are the same cell type and if i balance each unit in regards to mah, it would be rather okay then?

again, read my question and fatty's post before going further. you are basically asking "what car should i buy" or "what computer should i buy" without explaining ANYTHING relevant to make an informed response.

 

[MK2R]

hi flippy!

See it more as a general question, if u read the whole thread it's basically about battery packs and how to alter the parameters of the pack itself, what will change if you do this and that.

For example you got 5 cells of samsung 2200mah, 2 are brand new, and 3 are older and cyceled a lot. set output in the example will be 1A. if u add the 2 brand new cells that might actually test out at 2250mah, they'll get to 4500mah, rated max output is 5A nominal 2.2A. but the old cells are actually down to 1500mah, so together they'll (if u do it in math at least) get the same energy storage at 4500mah, but what happens to the output? you will only draw 0.33A per cell instead of 0.5A with the new cells.

how will they perform? the question is when you're mixing cells in each paralell group what matters the most? the total energy storage per unit or is there a correlation between numbers of cells aswell?

basically why i am asking is that i am testing a lot of used cells rn, 700. and i want to decide which cells to use and how i can balance each. the C rating will be rather low, so i will use more cells in general to get a high energy storage, i am thinking about 2.5kwh or more, i am planing on using anywhere between a few 100w up to 2kw.

So basicaly if i use the same type of cells, let's say a fixed value of 60Ah per unit, does it matter if one pack is 33cells and the other one is 32 (slightly better tested cells) and some pack might be 34 (slightly worse average mah per cell)

ps: flippy, your name is so similar to fatty that i actually thought it was him that answered, my apologizes!

quick question, this might sound very stupid okay, but hear me out. if i got 2 cells at 2250mah, they make 4500mah together, would 3 cells at 1500mah be compareable? same amount of mah, but 3 instead of 2, same-ish C rating as new etc

If they share the same discharge C-rate (for example, 2C continuous and 3C non-continuous), then yes, they would be equivalent. But if they are entirely different (say, a high energy cell versus a high power cell), then no, they would not be equivalent.
Compare the datasheets and calculate to see.

I understand, i am just not sure how cells are compared to the datasheets when they are getting cyceled a lot, i could imagine there is a somewhat impact on the nominal and max A output. or should it follow the C rating over aging as well?
Example a 2.5mah cell 2C is 5A, when it gets older and goes down to 2.3mah, 2C will be 4.6A and so on

 

[flippy]

Hi flippy!
See it more as a general question, if u read the whole thread it's basically about battery packs and how to alter the parameters of the pack itself, what will change if you do this and that.

such a random way of looking at battery packs is not useful, its beyond guesswork without more data.

For example you got 5 cells of samsung 2200mah, 2 are brand new, and 3 are older and cyceled a lot. set output in the example will be 1A. if u add the 2 brand new cells that might actually test out at 2250mah, they'll get to 4500mah, rated max output is 5A nominal 2.2A. but the old cells are actually down to 1500mah, so together they'll (if u do it in math at least) get the same energy storage at 4500mah, but what happens to the output? you will only draw 0.33A per cell instead of 0.5A with the new cells.
how will they perform? the question is when you're mixing cells in each paralell group what matters the most? the total energy storage per unit or is there a correlation between numbers of cells aswell?

that is not how this stuff works. your reasoning is roughly this:

if you have a brand new car engine that produces 300HP, will 3 completly worn out engines that only produce 100 out of 300HP still do the same work if coupled together?

would you put an engine that only produces a third of its rating in your car or would you agree that the risk of blowing up the engine is a considerable risk when an engine is that far gone?

basically why i am asking is that i am testing a lot of used cells rn, 700. and i want to decide which cells to use and how i can balance each. the C rating will be rather low, so i will use more cells in general to get a high energy storage, i am thinking about 2.5kwh or more, i am planing on using anywhere between a few 100w up to 2kw.
So basicaly if i use the same type of cells, let's say a fixed value of 60Ah per unit, does it matter if one pack is 33cells and the other one is 32 (slightly better tested cells) and some pack might be 34 (slightly worse average mah per cell)

you have to test each cell for capacity and IR in order to make any reasonable guesstimation.

but cells that are below 75% of original capacity should be discarded no not used. especially not in mixed packs.

 

[MK2R]

you have to test each cell for capacity and IR in order to make any reasonable guesstimation.

but cells that are below 75% of original capacity should be discarded no not used. especially not in mixed packs.

Thank u flippy! i am checking IR on all cells, they are about the same, the only that differs are mah. The cells i am testing are rated at 2000mah (at 1C), most cells have tested at 1850-2100mah, but a few have tested at around 1600mah, so less or more 80%. but if i get this correctly, i should try and stick to the 1900-2100 cells and do something else with the lower capacity ones?

Most cells are in the IR range 25-35


Speaking of that car engine with 300hp, i don't see the batteries as the engine, i see them as the fuel tank and fuel lines in one, you can either have a big tank to get far but slow, or short but fast. the maximal output of the battery is like what the fuel pump can deliver at a steady state (fuel pressure and flow rate) and the fuel lines are like the connections betweens the batteries and they determin how much of the fuel pumps capacity is going to be allowed through them (thickness of the nickle strips etc determins safe amp, same as a clogged fuel line will limit the delivery of fuel, but in the case of batteries a bad connection or broken strip will have more serious consequenses)

 

[fatty]

Thank u flippy! i am checking IR on all cells, they are about the same, the only that differs are mah. The cells i am testing are rated at 2000mah (at 1C), most cells have tested at 1850-2100mah, but a few have tested at around 1600mah, so less or more 80%. but if i get this correctly, i should try and stick to the 1900-2100 cells and do something else with the lower capacity ones?

Flippy has given good advice and I would defer to his judgement on discarding cells under 75%.

Don't build a pack with unequal numbers of parallel cells. Just distribute the cells such that each parallel string gets an equal number of better, worse, and total cells. To use your example, each parallel string gets 32 cells total, of which 13 would be better and 19 would be worse.

You can do this by "binning" the cells after testing -- literally put the cells into different boxes of "better" and "worse", or 1600-1700, 1700-1800, 1800-1900, and 1900-2000mAh. Then divide the number of cells in each box by 13s to find how many of each to put into each parallel string. You can widen the bins to make it easier, or narrow them to reduce string variance. You could even enter all the capacities into Excel to numerically minimize string variance.

 

[MK2R]

Flippy has given good advice and I would defer to his judgement on discarding cells under 75%.

Don't build a pack with unequal numbers of parallel cells. Just distribute the cells such that each parallel string gets an equal number of better, worse, and total cells. To use your example, each parallel string gets 32 cells total, of which 13 would be better and 19 would be worse.

You can do this by "binning" the cells after testing -- literally put the cells into different boxes of "better" and "worse", or 1600-1700, 1700-1800, 1800-1900, and 1900-2000mAh. Then divide the number of cells in each box by 13s to find how many of each to put into each parallel string. You can widen the bins to make it easier, or narrow them to reduce string variance. You could even enter all the capacities into Excel to numerically minimize string variance.

I am actually entering them all into excel Since i am going to test 700 cells i will probably have a lot of good and worse ones. so far i've tested a bit, i have already discarded the ones that are under 1800 into a separate holder. but maybe they are not that bad after all. it would be such a waste to just discard them, i want to make something useful out of them too. i am not thinking too many of them will be 2100mah, the ones in that end have tested at anywhere between 2105-2132. but the rest of the cells i am ordering in brackers of 2040-49, 2050-59 etc. i am sure that ordering them in a range of just 10mah isn't the best way, since a cell can show different results from test to test (but still in the same range, what i tested at 2045 today might be 2052 tomorrow etc).

However, one of the cells i tested had an IR of 10, i tested it in all 4 cell holders in the lii-500 and it showed similar IR, i am not sure what to do with that cell, it tested as the rest at about 2ah, but i am not sure of what to do with that cell. and i really need to look into more efficient way of testing them, i am even tempting to buy capacity testing units and build a big board with 10 or so of them, and then maybe get a cc/cv and put it at 42V and 1A and use one of the old 10S bms units and make a 10 battery charger with it. oh well, so far when i looked up things the liitokala is still the cheaper option unless i buy stuffs from china, but then i would be done with the testing when they arrive here. oh well!



A serious question now to both of you, i like you guys!

Your opinion. I will be operating the pack at a max 2kw, i don't want it to be heavier than the 13s33p. if i average the pack at 2ah per cell, it will be a little less than 3.2kwh. 0.63C discharge at full load.
1: I have the option to use 200 samsung/panasonic in the pack, (100 samsung is 2400-2500mah similar discharge C rating as the ones i am testing, the 100 panasonic is 2500-2600mah, datasheets say 10A discharge). In this case i will need to use 229 of the cells i am testing, and if i use the 229 best cells it might be a pack that have a cell average of 2200-2250mah, a 3.5kwh pack or more, discharge at 0.57C.
2: Just use the cells i am testing and don't mix in the samsung/panasonic cells.

When we are talking in terms of this low c-ratings, how much does it matter if you have a high output cell vs a lower output cell? From what i see in datasheets at discharge testing at higher A with a cell with lower IR you can draw a higher current but still have similar amount of mah, but with a lower rated cell as output A increases the mah.

I know the slightly higher rated pack will be cycled less times to deliver the same amount of energy, but apart from that at lower discharge rates, i hope both packs would be useful? not perfect in any way or form, but functional.

sorry for those long replies. i really appreciate your guidance and help, i really do. i am trying to gather as many ideas and opinions before i start, and as it looks right now i am testing 12-16 cells a day, it is fun, but the time between putting in each new cell in there is not fun.

thanks for all the ideas so far, i really do appreciate it! <3

 

[fatty]

I am actually entering them all into excel Since i am going to test 700 cells i will probably have a lot of good and worse ones. so far i've tested a bit, i have already discarded the ones that are under 1800 into a separate holder. but maybe they are not that bad after all. it would be such a waste to just discard them, i want to make something useful out of them too. i am not thinking too many of them will be 2100mah, the ones in that end have tested at anywhere between 2105-2132. but the rest of the cells i am ordering in brackers of 2040-49, 2050-59 etc. i am sure that ordering them in a range of just 10mah isn't the best way, since a cell can show different results from test to test (but still in the same range, what i tested at 2045 today might be 2052 tomorrow etc).

However, one of the cells i tested had an IR of 10, i tested it in all 4 cell holders in the lii-500 and it showed similar IR, i am not sure what to do with that cell, it tested as the rest at about 2ah, but i am not sure of what to do with that cell. and i really need to look into more efficient way of testing them, i am even tempting to buy capacity testing units and build a big board with 10 or so of them, and then maybe get a cc/cv and put it at 42V and 1A and use one of the old 10S bms units and make a 10 battery charger with it. oh well, so far when i looked up things the liitokala is still the cheaper option unless i buy stuffs from china, but then i would be done with the testing when they arrive here. oh well!

[EDIT] Turns out these are salvaged generic Chinese cells. They all do need to be tested.

Then make as big a pack as possible with all the remaining cells. With more cells in parallel, variance will tend to average out. With more aggregate capacity, you won't have to worry about running the pack down low enough to where any remaining variance is relevant.

 

[flippy]

Speaking of that car engine with 300hp, i don't see the batteries as the engine, i see them as the fuel tank and fuel lines in one, you can either have a big tank to get far but slow, or short but fast. the maximal output of the battery is like what the fuel pump can deliver at a steady state (fuel pressure and flow rate) and the fuel lines are like the connections betweens the batteries and they determin how much of the fuel pumps capacity is going to be allowed through them (thickness of the nickle strips etc determins safe amp, same as a clogged fuel line will limit the delivery of fuel, but in the case of batteries a bad connection or broken strip will have more serious consequenses)

its an analogy, not to be taken litteraly.

if you want to keep the fuel tank analogy: if a jerrycan that supposed to hold 5 liters has been dented and beat up so much that it only holds just 2 liters and leaks like rudi gulliani in a courtroom: would you consider that jerrycan useable of would you toss it?

remember that this jerrycan can light itself on fire and burn your house down and if 1 goes all his 700 friends also light themselfs on fire.....

 

[MK2R]

Thanks for the response flippy. i am aware of the risk, that's why i am aiming for lower current per cells.
the battery packs comes from (scooters?) whatever u call them in english


A pictures sometimes say more than a 1000 words. anyways u rent them in the city and just park them anywhere, and they usually get treated bad and it's too expensive to fix them so they just charge whoever rented them and buy a new one. and the scrapped ones are sold cheap to companies that buy bulks of them and just strip them apart and auctioneer out batteries packs etc (at a 10S5P config). so most batteries are pretty good (the cell cover say nominal 2000mah, i don't know at what ampere discharge, but when they test as high as 2130 at the lii-500 i am assuming that the 2000 rated is tested at a higher amp discharge)

 

[flippy]

do you actually know what cells you have? do you have model numbers and datasheets?

ps: i dont care in the slightest about their capacity. that is completly irrelevant in this stage of information discovery.

 

[MK2R]

do you actually know what cells you have? do you have model numbers and datasheets?

ps: i dont care in the slightest about their capacity. that is completly irrelevant in this stage of information discovery.

FST 18650-2000mAh
CAP: 2000mAh VOL 3.6V/4.2V
NB21017 084448

This is what the cell say
did look up the datasheet some time ago, don't remember it by heart. what i find on the internet are blue cells, mine are teal.

should be good for 2kw at 429 cells. or should i increase it to get down the discharge rate even more

 

[flippy]

FST 18650-2000mAh
CAP: 2000mAh VOL 3.6V/4.2V
NB21017 084448
This is what the cell say
did look up the datasheet some time ago, don't remember it by heart. what i find on the internet are blue cells, mine are teal.
should be good for 2kw at 429 cells. or should i increase it to get down the discharge rate even more

so in short: you have fake cells from already cheap cells.

those scooters are stupidly cheap and the batteries are never good.

without a capacity test and IR test you simply dont know what you have. i seriously doubt those cells would even manage their rated capacity when brand new.

test every cell and do an IR test and see what mess you actually have. my guess is that they are already used cells before they went into those scooters....

 

[MK2R]

FST 18650-2000mAh
CAP: 2000mAh VOL 3.6V/4.2V
NB21017 084448
This is what the cell say
did look up the datasheet some time ago, don't remember it by heart. what i find on the internet are blue cells, mine are teal.
should be good for 2kw at 429 cells. or should i increase it to get down the discharge rate even more

so in short: you have fake cells from already cheap cells.

those scooters are stupidly cheap and the batteries are never good.

without a capacity test and IR test you simply dont know what you have. i seriously doubt those cells would even manage their rated capacity when brand new.

test every cell and do an IR test and see what mess you actually have. my guess is that they are already used cells before they went into those scooters....

Flippy, have you not read the thread at all? I am testing them all before i am using them... So far i've tested 75 cells, they average 2016mah. the lii-500 say they have the same IR as my samsung inr 26j which are rated at 5.2A nominal output.
If they are used before they went into the scooter packs or not, i don't know. but they seem to be testing out with numbers similar to what the datasheet say that they should as new. there is no marks of previous spot welds to them apart from the ones that have been in them when they where designed for their 10s5p packs. so if they've been reused it might have been the whole battery pack as a whole maybe?

Well i do understand u don't want to waste your time on reading it all so i am informing you kindly like this instead, i am testing them all, putting data in a excel spreadsheet and so far most cells are testing out just fine. however i still have 625 to go.

 

[fatty]

Flippy, have you not read the thread at all? I am testing them all before i am using them...

It's a little hard to follow because the info is spread throughout your replies.

So you have:
100x Samsung 26J rated for 2600mAh 5.2A that test at 2400-2500mAh?
100x Panasonic ??? rated for ??? that test at 2500-2600mAh?
500x Chinese FST "rated" for 2000mAh that average at 2016mAh?