The care and feeding of a123-based packs...

Hey guys, I'm making my first A123 packs, just need some advice.

I'm making 3 packs, 6S3P each. 54 cells total, 18 cells per pack. I spent 5 days and cycled each cell, with my Bantam charger, 5amp discharge and 5 amps charge and wrote the mAh down on each cell. The smallest cell from the bunch I will be using is 2092 mAh and the largest is 2219 mAh. Most of the cells are 2140 to 2160. Now, I was told the cells will produce more mAh if charged and dischared at a higher rate. I think I was getting 50 more mAH on each cell when i was charging them at 8 amps, before I decided to charge them all the same at 5amps. Also it's been 60 to 65F degrees where I have been cycling them. I read you guys were checking the voltages after the charge, but It's to late for me to do that. All the cells are 3.35-3.43v after resting for a few days though.

Right now i have the cells layed out:
pack 1: 2092 to 2143
pack 2: 2145 to 2160
pack 3: 2161 to 2219

Is this how I need to make the packs? I wish they were all like pack #2, but I can't afford to be buying more cells. lol. The bad thing is I will be putting these in series to make 59.4V nominal and the first pack will run out before the last pack. I should stay at 2.5v discharge per cell, (45V) to be safe? I won't be using a BMS. I will only have a watt/volt meter display on the handle bar in the circuit.

I was thinking, maybe I should just mix the cells so they are all equal packs? 2092 to 2219 (example)

I just realised that I need to make sure the 3P cells are as equal as possible, no matter what the Series cells are because I will be using a 6S balance tap with the Bantam charger. The charger only can balance the 3 cells as 1.

Thanks
 
TMaster said:
Hey guys, I'm making my first A123 packs, just need some advice.

I was thinking, maybe I should just mix the cells so they are all equal packs? 2092 to 2219 (example)

I just realised that I need to make sure the 3P cells are as equal as possible, no matter what the Series cells are because I will be using a 6S balance tap with the Bantam charger. The charger only can balance the 3 cells as 1.

Thanks
Click to expand...

Yes, you'd connect every 3 cells in parallel, then series the groups of 3 cells. After paralleling the cells, every 3 paralleled cell would act as a single cell with c1+c2+c3 capacity and r1*r2*r3/(r2*r3+r1*r3+r1*r2) internal resistance(inverse of the sum of inverse resistances). You want these sums to be as close as possible within every series. Yes this means, that you need to put a lower capacity cell in parallel with a higher capacity one, so you get one with a mean value. The same stands for internal resistance, it's just the relationship is not quite linear. Member Doctorbass can explain this more thoroughly, he has software for matching cells. Capacity and internal resistance matching is the most important when building a lithium pack, but for running with pack level LVC only, yes the capacity match is most important. So make sure the C1+C2+C3 sums are closest.

Good luck with your build. I'm going to do the same with a 18s2p pack configured as 6s6p for charging with a RC balance charger.
Zsolt
 
Ok so I will be using the RC balance charger. For series I should but the low capacity cells with the higher capacity cells, but make sure the paralleled cells are close to equal in capacity as possible. This route I will have all the packs end discharge the same when running, but the voltages of the entire pack will be less, because the higher capacity cells in the pack won't be charged up full because of the lower capacity cells in the same pack? I just want to confirm all of this before I start soldering everything. I don't really have a good way of testing the internal Resistance.

example for one pack:
S1- paralleled:(2096,2105,2106)
S2 -paralleled:(2125,2135,2119)
S3 -paralleled:(2145,2154,2156)
S4- paralleled:(2156,2160,2160)
S5 -paralleled:(2175,2188,2166)
S6 -paralleled:(2213,2196,2203)
 
In the parallel group of 3 cell, every cell out of 3 will see the same voltage, charge to same voltage and sit at the same voltage every moment of discharge, because of parallel connection. If you balance charge the 6s3p blocks, every cell end at the same HVC voltage set in the charger, fully charged.
You'd want the sum of 3p cell capacity to be close to each other.

Example for one pack:
S1 - paralleled: (2092, 2156, 2219) sum = 2092 + 2156 + 2219 = 6467
S2 - paralleled: (2125, 2154, 2188) sum = 2125 + 2154 + 2188 = 6467
S3 - paralleled: (2105, 2160, 2203) sum = 2105 + 2160 + 2203 = 6468
so that S1 capacity of 6467 is close to S2 cap. of 6467, and ~cap S3 of 6468 and approx the cap of S4, S5, S6...S18
Delta capacity in this case is 1mAh(ideal)
This is the beauty of series-parallel connection, gives additional balance if properly assembled.

Not like
S1- paralleled:(2096,2105,2106) total = 6307
...

S6 -paralleled:(2213,2196,2203) total = 6612
Delta capacity from S1 to S6 is 305mAh, definitely not good.

Working out the numbers optimally is a complex problem though, with 2s is easy, you sort them and pair the weakest with the gratest and go on with the remainder. With more in parallel (3 in this case) you can try something similar, but you might not find the best combination for every 18 serie. Doctorbass had a computer program I think written in MatLab that would simulate cell configurations and make changes in placement till the sums were optimal.
You can search for pack cell matching in the forums. I think I got the basics, but there are a lot more info on this.
 
So it won't damage the cells by paralleling a lower capacity with a higher capacity cell? I guess I just don't understand how the charger charges them up. So your saying the voltage is what matters? Like, The voltage all peaks each parralleled cell at the same time with the same votlage but the capacities will be different?
 
TMaster said:
So it won't damage the cells by paralleling a lower capacity with a higher capacity cell?
Click to expand...

No, paralleling a lower capacity with a higher capacity cell does no harm at all. State of charge (SOC) correlates with resting voltage (in the case of a123 far from linear), but full capacity is considered at 3.6v or 3.65 for a123 LiFePo4.


Like, The voltage all peaks each parralleled cell at the same time with the same votlage but the capacities will be different?
Click to expand...
Yes


I guess I just don't understand how the charger charges them up.
Click to expand...
You might want to read The secrets of battery runtime and other related topics from BatteryUniversity for better understanding.
Great info under Charging lithium-ion batteries, but it is about LiIon, Not LiFePO4 a123.


More info on cell matching here and here.
 
Thank you for all the information! I will sit down with the calculator and get everything matched up! :D

Picture of the pack formation. I CNC'd out some custom battery bars, .032 thick 110 copper. The bar end are .25" wide, so this is equvilant to 10AWG wire. Good for 55 amps, but I'm only going to be pulling 22-30 with the controller.
 

Attachments

  • IMG_5826.JPG
    83.1 KB · Views: 1,102
I spent about 2.5 hours but I finally got all the cells matched up. Ugh what a task. I feel for the guys paralleling large numbers of these!!!
 

Attachments

  • matched_cells.jpg
    40.7 KB · Views: 1,080
Your CNC work looks great.
And the calculations too. Only that I expected the average capacity to be around past 2200, works out to be only 2148mAh. Are you sure charging at higher rate gets better capacity? What about discharge, higher discharge rate also increases capacity? I think I read about BMI cells having this property but didn't knew the same applies to a123. This is unusual with conventional cells.
 
I was told by someone very experienced with A123's that the higher discharge and charge rates do produce more capacity. The guy said he charges his A123 cells at 20 amps to start, then as the capacity fills up the amps drop down, (use the RC chargers) I tried a single cell at 10amps and it seemed to produce more capacity. The cell started out at 10 and within a few min the amps were starting to drop. I have no idea how the BMS chargers work though. According to the data sheet the reccomended fast charge is 10 amps.

When I cycled all these cells I only used 5 amps DC and 5 amps C. So the capacity may be a little less. Also It was cold. The air temp was 60 to 65F. I will take a few of the left over cells I have today and charge them at 10 amps. My charger will only discharge up to 5amps though. I will report back the results compared to the previous.
 
I suspect the higher capacity strictly comes from an increase in the cell temperature. If so, you could get the same result by just heating the cells rather than heating them via high charge rates.
 
Ok I can confirm I charged the same cell at 10amps instead of 5 and It produced 52 more mAh. So that makes my cell average in the pack 2200 instead of 2148. :D It probably would produce even more if I would discharge them at a higher rate.
 
webfootguy said:
I suspect the higher capacity strictly comes from an increase in the cell temperature. If so, you could get the same result by just heating the cells rather than heating them via high charge rates.
Click to expand...

I may have to try that!
 
I took the heat gun and warmed up the cells some and Yes it does work. I got another 30-40 mah out of it just from the heat. So I turned a 2178 cell into a 2260. ;)
 
lifepo4ever said:
Do I need a bms for one 4sp1 pack of A123 ?

my idea is to use it for my atv and charge it with a car charger thank you for all your help :D
Click to expand...

You shouldn't....as long as YOU don't overdischarge it! I've had a 4S1P pack in my sons ATV for almost 2 years now. Charged only by its alternator......

Just balance the pack first by making it a 1S4P pack...let it sit for a day to equilize. Originally the ATV had a 12v 4AH sla, the A123 pack turns the motor over much better. Last I checked the voltages were within .05 of each other. Make sure you use a charger that is CC/CV/Flt and does not do desulfacation.
 
i have a lifepo4 charger here 2 amp hour and 3.65 volt , can I charge each A123 cell individually with that and after make the assembly of the pack ?
 
lifepo4ever said:
i have a lifepo4 charger here 2 amp hour and 3.65 volt , can I charge each A123 cell individually with that and after make the assembly of the pack ?
Click to expand...

Yes. For an sla charger use a "smart" charger.
 
lifepo4ever said:
thank you but what that mean ? : CC/CV/Flt and does not do desulfacation i am not a engineer lol!
Click to expand...
cc: Constant current. This is the bulk charge step which brings the cells from their discharged voltage to the charged voltage.
cv: Constant Voltage. The charger will hold the cells at the max charge voltage a while to top off the charge, and let a BMS do its work balancing the cells.
flt: Presumably float. The charger will drop to a lower float voltage. Not sure if this is useful for Lithium, but it avoids fluid loss in lead acid cells.

Desulfatation: the charger will cycle the cells to a higher than peak charge voltage. In my case for a 12V smart charger, peak charge is 14V and the desulfatation is a 10s pulse at 15-15.5V every 30s. This is again only useful to lead and could be harmful to lithium if the BMS does not cut off.
 
hey thank you for all your reply ! my car charger have a 13.50 volts peak my other one is a smart charger lifepo4 2 amps and 3.65 volt peak

so I will use the 3.65 volt charger before assembling the 4sp1 and when I need to charge the pack I will take the car charger but will not be able to balance the cell . doesn't matter anyway because it for starting my atv and the circuit will charge the battery
 
TMaster said:
I spent about 2.5 hours but I finally got all the cells matched up. Ugh what a task. I feel for the guys paralleling large numbers of these!!!
Click to expand...


Tmaster.. I have a program on a Mathlab platform that calculate every paralle serie pack you want to match them perfectly!.. not just the capacity but can ALSO match the internal resistance!.. all that in.. less than 10 sec!! not 2.5h! :wink:


I posted that two years ago in my 432 cell project.

This is some exemple of my cell matching using mathlab and excel

file.php


file.php


file.php



(30s4p.. each cell is labeled from 1 to 130)

file.php




Also.. I would suggest you to not cut the original nickel strip of the A123 cells too short like you did... but instead you should keep them twice long and ply them on themself.. I mean if you solder too close to the spotweld the heat is too much transfered to the cell inside... and it is also more difficult to solder with normal solder iron... instead just solder on the next plied half of the strip..

http://endless-sphere.com/forums/viewtopic.php?f=14&t=15165&p=225369&hilit=a123+nickel#p225369

file.php


Doc
 
You must log in or register to reply here.

Similar threads

M
A123 ANR26650M1B Spot Welding
Replies
24
Views
1,044
gromike
G
Skyler
LFP reverse polarity.
2 3 4 5
Replies
105
Views
4,612
DaLanMan
DaLanMan
J
Li-Ion Battery Pack Building Techniques
Replies
5
Views
869
amberwolf
amberwolf
O
Lifepo4 cylinder 32650 cells simplified stationary design idea?
Replies
7
Views
181
jonescg
jonescg
B
A123 pouch cells LIFePO4 balancing issue
Replies
8
Views
742
amberwolf
amberwolf
Back
Top