What happens to different capacity packs in parallel??

[SimpleCJ]

I'm going to be using used A123 cells for my eBike and I'm wondering how this is going to work out...

I tested 77 A123 M1 cells all individually and recorded their capacity, then I grouped them into as closely matched sets as possible.

I ended up with 9 - 8s packs (26v). The ave capacities of each pack are 2.2Ah, 2.15Ah, 2.1Ah (x2), 2.0Ah (x3), 1.95Ah, and 1.8Ah

Each pack will have it's own balancing and power connector.

My question is, does anyone know what will happen, with all these packs in parallel, when the lower packs go flat? Will the flat packs continue to be pulled down further or will the increased resistance of those packs offset the load to the packs with remaining capacity? What about charging them all in parallel? When the lower packs top out, will more of the charge be pushed into the better packs or will they all continue charging while the lower packs go over their charging limit?

I understand quite a bit with these cells. They are very tolerant of overcharging/discharging, but without individual balancing, the difference between cells can become critical, that's why I'm doing individual packs so I can periodically balance each pack, but what about when I'm using them or charging them all at once??

 

[Lapwing]

Have you measured how many volts total for each series string of cells in the match group you have selected?

 

[SimpleCJ]

Well, the nominal voltage for 8s is 26.4v with a 28.8 peak charge voltage.

After grouping into 8s packs, I cycled the cells to make sure they were all dropping off and charging up close to each other. So, each pack is 8s1p and I'll have 8 or 9 of them wired in parallel. I'm thinking that the 1.8Ah pack is not really a good one to use, the rest of them are fairly close in capacity, but I don't know if they're close enough...

I can balance charge each matched pack in under 25 minutes at the 10A setting (which means about 4 hours total charge time for 8 packs whether I do it individually or as a 8p pack. When I first grouped them and did the first cycle it took over an hour to balance charge them, two cycles later and the charge time is less than half.

Here's a discharge and charge graph. I know the numbers are hard to read, if I leave them bigger it squashes the graph. I need to figure out how to hide the scales for all but one cell.

 

[Doctorbass]

I see you are using the Megapower or the Robbe power peak with rs-232 interface.. with Logview right :wink:

This is very usefull ! I love it

Capacity matching is very simple.. cap is just adding as you parallel cell.. nothing to borther with...

Look what i've done with my makita cells and many test for building battery packs.. that may be usefull for you :wink: :

A123 Capacity et Energy vs differents charge currents graphs
http://endless-sphere.com/forums/viewtopic.php?f=14&t=2717&hilit=logview

My mathlab softwatre to match cells with incredible precision!
http://endless-sphere.com/forums/download/file.php?id=8817&mode=view
http://endless-sphere.com/forums/viewtopic.php?f=14&t=2764&st=0&sk=t&sd=a&hilit=432+konion&start=15

Ideally you would need to match with capacity AND INTERNAL RESISTANCE... that's a perfect cell age and quality indicator! My megapower do that measurement.

Doc

 

[SimpleCJ]

Actually I'm using the iCharger 1010B+ with LogView, but ya, it's nice to be able to watch and record the data like that... when the USB link isn't failing...

I don't have enough cells to attempt precision matching, I have what I have. Also, I know IR is an indicator, but it's not constant, not even for an individual cell. State of charge effects it as well as age. I did check a bunch of them (my charger does that too) and most seemed close enough ~12-18 mOhms...

I understand that adding them in parallel just adds capacity, I'm not asking about capacity calculations. I'm asking, will continuing past the capacity of the lower cap packs draw them below cut-off voltage? What about when they are all charged together in parallel? Will the lower packs get overcharged?

I know some of these answers will be defined once I get everything put together and actually drive it, but that's a ways away...

Would you say it'd be better to leave out the lower 1.8Ah pack so the rest are closer to each other, maybe just for charging??

 

[Doctorbass]

Actually I'm using the iCharger 1010B+ with LogView, but ya, it's nice to be able to watch and record the data like that... when the USB link isn't failing...

I don't have enough cells to attempt precision matching, I have what I have. Also, I know IR is an indicator, but it's not constant, not even for an individual cell. State of charge effects it as well as age. I did check a bunch of them (my charger does that too) and most seemed close enough ~12-18 mOhms...

I understand that adding them in parallel just adds capacity, I'm not asking about capacity calculations. I'm asking, will continuing past the capacity of the lower cap packs draw them below cut-off voltage? What about when they are all charged together in parallel? Will the lower packs get overcharged?

I know some of these answers will be defined once I get everything put together and actually drive it, but that's a ways away...

Would you say it'd be better to leave out the lower 1.8Ah pack so the rest are closer to each other, maybe just for charging??

this is a question of current sharing. in parallel, none of the packs (i prefer using cell) will draw the others ubtil it is deffective or shorted.

The lower Ah cell will just stop giving current when 100% dod reached while the cell with higher ah continue to give current when paralleled together... their voltage will always be the same. if you could place an ammeter in serie with every cell then paralleled, you sould see the amp of the lower cell droping faster than the better one.

Doc

 

[SimpleCJ]

Ok, thanks Doc!

So that means that I should be able to squeeze out all 18.3Ah before the whole set goes flat, cool. I still think I might run into issues trying to charge all 9 packs together as a 8s9p config with no balancing. I guess I'll just have to wait and see how it works out, I just hate the thought of having to charge every pack individually, but I guess that's what I get for using used cells huh?

Have you seen the new cartrige shaped lithium cells that are coming out soon?

24v 4.2Ah!!! Would love to get like four or five of these for my bike, that'd be sweet!!

BTW... I say pack because it's an individual set of cells with it's own power and balancing lead. Cells make up packs... but I guess in my config, you could consider each pack as a "cell" with all the "cells" in parallel... whatever floats your boat...

 

[Doctorbass]

Ok, thanks Doc!

So that means that I should be able to squeeze out all 18.3Ah before the whole set goes flat, cool. I still think I might run into issues trying to charge all 9 packs together as a 8s9p config with no balancing. I guess I'll just have to wait and see how it works out, I just hate the thought of having to charge every pack individually, but I guess that's what I get for using used cells huh?

Have you seen the new cartrige shaped lithium cells that are coming out soon?

24v 4.2Ah!!! Would love to get like four or five of these for my bike, that'd be sweet!!

BTW... I say pack because it's an individual set of cells with it's own power and balancing lead. Cells make up packs... but I guess in my config, you could consider each pack as a "cell" with all the "cells" in parallel... whatever floats your boat...

Yeah.. i know these toshiba lithium titanate cells...

they charge in 5 minutes, have around 6000 cycles... but they only have half the energy density of ironphosphate (a123 etc) they have around 55Wh/kg.. so you need twice weight of those for the same Wh..

I know that if you use A123 and that you charge them oly at 100% soc, you may get around 1000cycles.. BUT at the end of these cycles, you will still have 90% of capacity! that is stillmore than toshiba cell

AND GUESS WHAT.. that will take 10000 cycles to drop to 50% capacity ! according to the A123 ingeneer...

That's Bill dube that wrote that and i know he is probably one of the best person on the earth to have used them!

http://www.rdrop.com/pipermail/oeva-list/2007-May/001278.html

You need to dust off your calculator and actually do the math.
>
>At 50% capacity loss (10 times 5%) you will get 10,000 cycles (10
>times 1,000.) This is direct quote from the engineers a A123
>Systems. I don't make this stuff up.
>
>By the way, can you find another brand of battery that lists the
>cycle life for 10 C discharge?
>
>At 08:33 PM 5/17/2007, you wrote:
>>What that chart shows:
>>
>> 1C the loss is 5% at 25C after 1000 cycles
>> 2C the loss is 10% at 45C
>> 2C the loss is 20% at 60C
>>
>>All I can really say is you are now Bill the Battery Salesman
>>to say expect 10,000 cycles for an EV.
>>
>>Jack Murray

Doc

 

[SimpleCJ]

Ahh... didn't think about the weight factor, you're right!

I love the A123 Cells, been using them for over a year now in my two Traxxas RC Rustlers (one brushed, one brushless). I also built a HUGE pack for Kent Riches, friend of Bill Dube, last year.

Check out the the project... http://www.rcuniverse.com/forum/m_7541729/tm.htm

The bike went 125mph last November...

I'm actually using the leftover cells from that project for this eBike project and I'm going to supply Kent with some of my findings. For his packs we did solid parallels, meaning there is no way to balance charge individual cells. We didn't think much about that at the time. No one on our team really knew much about batteries or BMS and I was the only one who really did a lot of research as well as hands-on stuff with these batteries. We had all the cells tested, but they were tested in groups of up to 16p and not on the individual level. Now that I see how varied the capacities of my sample of 72 cells are, I'm wondering if Kent is going to run into problems with low capacity cells being in parallel with good cells. Seems like when he charges the packs back up, those low cells could, after a few cycles, become critically over-charged.

They say these cells don't explode, but I saw some pictures of a small scooter where the A123 cells suffered a catastrophic failure. I'm sure it wasn't as bad as what would happen with LiPo's but it was still catastrophic.

Kent is sitting on a 128s8p brick of these things and traveling at high speeds, I would hate to see something go wrong. Main problem is that we were given used and apparently abused cells to work with and had almost no previous experience with that type of project. Had they all been good new cells, I wouldn't be so concerned, but with my continued research I believe there may be a problem.

What do you think Doc??

 

[slayer]

interesting simpleCj
I can see in your pics that your teammates seems to be at least checking the voltage of each cells before did you sort them out and group them in state of charge before assembling ...did you recharged them individually before also, or you just assembled them as you received them


Also where did you get them ?

 

[SimpleCJ]

I did charge low voltage cells and discharged overcharged cells that came back from testing in California (only a few dozen total), but we had over 1200 of them and I'm sorry, but to do them all individually with my hobby charger would have taken FOREVER!!!

We were matching them into similar voltages, down to the hundredth of a volt. Problem is that we had assumed that they were all of equivalent condition. The cells were matched into sets of 32 by voltage, but as I have found, that is certainly no indicator as to cell condition or capacity. After the first discharge and charging cycle I'm sure that the voltages across the cells in parallel will be off because of the varied capacities of the cells.

I explain the source of the cells and our initial processing and testing of the cells in that thread....

 

[slayer]

awesome , and i can see you made a spot welder that is great

 

[SimpleCJ]

Works great for the NEG terminal, but not so great for the POS terminal...

 

[Doctorbass]

Wow Amasing work you did !!

about cell measuring, yes.. i've been a guiny pig for testing as high as 432 cells for my 100V 23Ah battery pack!

the cells i used was paired so i just had to measure 216 cells (pair total of 3.0A). I first charged every of them to 4.200V at 5A (1.7C)and stoped charge at exactly C/20 wich is 150mA.

Then i measured every of them by discharging them at the max rate my megapower can do wich is 5A. it take around 45minutes for every cells pair..

45min x 216 = the equivalent of an entire week non stop!!...

BUT.. i got the exact mAh and exact RI of every pair..

then i matched them by doing a compromise between good RI and capacity... cells that had a good capacity and bad RI where matched with cell with low capacity and good RI ..

every parallel pack of 1s 18p was as accurate as 0.09% from each others!!! like perfect!!

And today they are still perfectly matched... no bms.. I just monitor every parallel group once a month to ensure..

Could you ask to Bill if he matched the cells of the killacycle with capacity and RI?

Doc

 

[Doctorbass]

You can stop bothering with lower capacity cells that are in parallel with the rest of good cells... they dont negatively affect your pack.. it's just a question of less total capacity on the parallel pack than if all cells would have good capacity..

just ensure that every paralell packs have the same RI and and TOTAL capacity.

like i said.. it's just a question of current sharing between paralleled cells while dishcarging or charging.. and their voltage is always the same cause they are paralleled...

Doc

 

[SimpleCJ]

Wow Amasing work you did !!

Could you ask to Bill if he matched the cells of the killacycle with capacity and RI?

Doc

Bill is sponsored by A123, he gets all brand new cells so matching won't really be that big of a deal. Brand new cells should all have nearly the same capacity as well as IR right?

 

[Doctorbass]

Wow Amasing work you did !!

Could you ask to Bill if he matched the cells of the killacycle with capacity and RI?

Doc

Bill is sponsored by A123, he gets all brand new cells so matching won't really be that big of a deal. Brand new cells should all have nearly the same capacity as well as IR right?

Yes they should.

I admit i am more familiar with used cells.... ALOT OF THEM :wink:

Doc

 

[swbluto]

... but not so great for the POS terminal...[/quote]

Double entendre?

 

[SimpleCJ]

... but not so great for the POS terminal...[/quote]

Double entendre?[/quote]

What are you, the grammar police?? Give me a break! Don't bother posting if you don't have any valuable input.

 

[Stan]

I've got a Bionx PL350 kit. It's running a brushless 36 volt motor in the rear hub powered by a 36 volt 10 Ah Bionx Lithium-ion Ibattery. Could I parallel a 36 volt 20 Ah LiFeP04 Ping battery through the Bionx Li-ion battery's XLR charger port which reads 41.5 volts at full charge at pins 2 and 4 of it's female XLR port?

 

[swbluto]

... but not so great for the POS terminal...[/quote]

Double entendre?[/quote]

What are you, the grammar police?? Give me a break! Don't bother posting if you don't have any valuable input.[/quote]

I take it you're not easily humored?

Ok, Mr.I'm-too-serious-to-not-appreciate-the-humor-in-unintentionally-creating-double-meanings.

(I actually suspected it was intentional and pretty clever. But, maybe not. Just in case you didn't get your own unintentional humor...

It could mean that a positive terminal that doesn't work all that well and, rightfully, it's a [b]P[/b]iece-[b]O[/b]f-[b]S[/b]hiitakeMushrooms. :wink: )

 

[Toorbough ULL-Zeveigh]

it went over my head originally as well & that's despite the fact I keep my eye open for d.e.'s & t.e.'s.
So it's appreciated from this corner at least ur pointing it out. :lol:

 

[SimpleCJ]

I get it now, went over my head at first too, so no it wasn't intentional. Sorry, I was in a bad mood the other day...

 

[Doctorbass]

I get it now, went over my head at first too, so no it wasn't intentional. Sorry, I was in a bad mood the other day...

Hey SimpleCj.. are you still with us? :wink:

Doc

 

[swbluto]

I should be imprisoned for derailing a thread. I believe derailment is a federal offense in the US.