Selecting cells for assembly (parallel groups)

Hello all,

after getting great piece of advice from the forum I selected the battery technology suited for my ebike (Li-ion), and bought cells (Sanyo NCR 18650GA) to build my battery pack (14s5p).
The pack will be built for my ebike (phaserunner + MacMotor), and designed for a max output current of 30A (6A/cell ~= 1.71C). I will use a 35A BMS from Bestech.

I will open a new topic concerning the physical build of the battery pack, but today I am concerned with the selection of cells to put together in parallel groups.
In this matter, I built a small board able to fit 10 cells in series (10s) that I connect to my balance charger (iCharger 1010b+).



The plan is to test each cell characteristics, eventually connect the charger to the computer to plot charge/discharge curves, and balance them properly before connecting them together (it is better to avoid brutal self-balance right ?).
I would now need some advice on how to select the best cells to put together.



After connection, I can use the charger to monitor individual cell voltage and internal resistance.

View attachment 1

However I wonder whether this information is enough to select the group of cells, or whether I should get more information (plots) by doing a charge/discharge cycle ?
I will try to figure out the connection with the computer to extract plots but what parameters would you recommend for such "benchmark cycle" on these cells ?


On a side note, this charger has different lithium configurations to start from, but nothing ready to use for Li-ion cells (or does it ?).
Looking at the table of pre-configured characteristics, I plan to select Lilo (I like to stop before 4.2v to save number of cycles, but all these values can actually be customized).



Do they actually mean Li-ion with "Lilo" ? :?



Does the whole strategy sound correct ?

For those of you familiar with this charger, do you know whether the balance charge program actually balances around existing voltage or whether it goes up to full (or nominal) charge ?


Ok that's already a lot of questions for a first post, I keep other ones for later

Thank you very much for your help !

Assuming your balance charger is connected to each cell in the string, then it will charge till one or more cells is full, stop, discharge the full cells, then continue charging till all are at 4.2v using the lipo setting. I just can't really see if you have that balance plug connected in the picture, to each cell.


Lilo setting will charge and balance them to 4.1v

When using the bulk charge settings, then it charges till the pack is at 4.2v average. Don't use that setting for charging until you know the cells are balanced.

Personally, I'd sort a bit simpler, but still start with the balance charge. I think any with low capacity will show it in internal resistance number.

Then individually check internal resistance. Chuck any with way off numbers. Spread the ones with slightly higher resistance through the pack, so they are not all in one parallel group. Then after full charge, set aside a week or so, and recheck for any that self discharge. Any that go from 4.2 to 4.1v when new are suspect. 4.2 to 4.18 no big deal. or whatever the number is for those cells. If all discharge to 4.12v then it is what it is. I just mean suspect any that don't fit the curve of the whole group.

Chances are, the quality cells you bought, you won't be finding too many that fail.

Thank you for your answer !

I just can't really see if you have that balance plug connected in the picture, to each cell.

Click to expand...

Yes the balance plug is connected to each individual cell. That allows to get the characteristics (volt & resistance) of each cell shown in third picture.
Ok I understand that the balance charge will balance after full charge (and not around the existing voltage).
I will indeed balance cells before assembling the battery and attempt any bulk charging (and a BMS will be used too).

Reading you post, I am tempted to plot a curve of charge/discharge to identify the 'outlier' cells. Still need to figure how to do that.
I guess that your strategy (charge, wait, then check self-discharge) will point to the same ones... I might do both and report results actually.

Is there any advantage of doing a full charge/discharge cycle before using the cells in the bike (they have been stored a 3.7v for several months) ?

BTW, do you know whether measuring the internal resistance should be done at full or nominal charge ?

phase runner has low voltage cutoff therefore you don't need chinese crap bms

I use a cycle analyst for lvc and a bag of other reasons with bms. I would only rely on the BMS for emergency battery saving cut off. Not the rule.

cheapcookie said:

phase runner has low voltage cutoff therefore you don't need chinese crap bms

Click to expand...

I will NOT build a pack without BMS.
Low voltage cutoff is nice but HVC, balancing, and short circuit protection are mandatory !

Anyway I don't want to start a "BMS or NOT" topic, there is a ton of them already.

Just get all cells to same voltage and spot weld them in parallel group of 5 then series. Put bms on with an extra set of sense wires to monitor each parallel group and or to balance if ever needed. Now get on the bike and go ride before summers over.

999zip999 said:

Just get all cells to same voltage and spot weld them in parallel group of 5 then series. Put bms on with an extra set of sense wires to monitor each parallel group and or to balance if ever needed. Now get on the bike and go ride before summers over.

Click to expand...

Ahah thank you for that
I am maybe overthinking it, but that's what I typically do

Well FWIW, my house burned because a battery with a nice bms lit up. nuf of that discussion though.

I don't know what your I chargers can do, discharge graphs of each cell would be sweet!!!

Mostly I just wanted to say don't put the weak cells all in one group, unless you intend to replace all those soon. IR should point out any really crap cells. I seriously doubt you will find any that self discharge, but that test is easy. sit a week, look for a drop.

Should be able to really quickly sort for IR. You don't need numbers, just look for cells that drop more under load than the norm. Light bulb, voltmeter, 2 sec discharge, next. You'll see a bad cell by the bigger voltage drop. And they did this at the factory too I bet.

dogman dan said:

Well FWIW, my house burned because a battery with a nice bms lit up. nuf of that discussion though.

Click to expand...

Thank you for the info, since my charges will happen inside an appartment with no garden or anyplace to keep a fire 'safe' I will think about building an insulated charge/storage box (maybe a big bucket with sand or something like that)... Any advice ?

dogman dan said:

I don't know what your I chargers can do, discharge graphs of each cell would be sweet!!!

Click to expand...

Spent my yesterday evening trying to extract information from usb connection... Quite a hassle (needs a windows computer etc...) but I think I figured it out.
So far I extracted raw data during the balance charging and I will try to plot nice graphs with that. If I make that work I'll try with charge/discharge cycles.

dogman dan said:

Should be able to really quickly sort for IR. You don't need numbers, just look for cells that drop more under load than the norm. Light bulb, voltmeter, 2 sec discharge, next. You'll see a bad cell by the bigger voltage drop. And they did this at the factory too I bet.

Click to expand...

Ok, I didn't think about the voltage drop, I'll see whether this correlates with IR measured by the charger.

I'll keep in mind to spread the cells in different parallel groups, but anyway I'll keep you informed with all info I can get from this 10 cells batch.
Probably no news before next Monday though because I'll be away this weekend.

Another question, what is considered the reasonable threshold (in volts) for charge/discharge in order to maximize the number of cycles. I know everyone has his own recipe but I heard a lot about discharging no more than 20% and charging no more than 80% (except during balancing or occasionally). Can you tell me what volt/cell does 20% and 80% capacity relate to (approximately) ?

Apparently the default discharge cutoff on the charger is 3.0v, is there a risk of cell damage by going so low few times (I was planning full charge discharge to estimate total capacity) ?

Thanks again for your help !

Bouteille51 said:

Another question, what is considered the reasonable threshold (in volts) for charge/discharge in order to maximize the number of cycles. I know everyone has his own recipe but I heard a lot about discharging no more than 20% and charging no more than 80% (except during balancing or occasionally). Can you tell me what volt/cell does 20% and 80% capacity relate to (approximately) ?

Apparently the default discharge cutoff on the charger is 3.0v, is there a risk of cell damage by going so low few times (I was planning full charge discharge to estimate total capacity) ?

Click to expand...

Ok after more reading I realize it depends on the current drawn, temperature, etc... I'll plan my high and low voltage cutoffs using that if that makes sense.

with 18650s i only go down to 3.3v ( you can normally see the obvious point on a discharge graph) and up to 4.1v to get the highest cycle life vs capacity from them keep the charge rate low when you can ( 1/4c when you have 4h and no more than 1c when you need a faster charge )

And whats up with A and B at 15A ? all other ones are a close match and looking at your discharge chart 10A is the maximum for those battery's

Also have you accounted for the v drop on the wires ?

with 18650s i only go down to 3.3v ( you can normally see the obvious point on a discharge graph) and up to 4.1v to get the highest cycle life vs capacity from them keep the charge rate low when you can ( 1/4c when you have 4h and no more than 1c when you need a faster charge )

Click to expand...

Thank you for your advice. For the 14s5p pack, in normal conditions I will bulk charge @ 5A so ~1A for each cell (~0.3C).

My apologies I forgot to cite the source for this figure.
Here is the full webpage containing all characteristics: http://lygte-info.dk/review/batteries2012/Sanyo NCR18650GA 3500mAh (Red) UK.html

Bouteille51 said:

with 18650s i only go down to 3.3v ( you can normally see the obvious point on a discharge graph) and up to 4.1v to get the highest cycle life vs capacity from them keep the charge rate low when you can ( 1/4c when you have 4h and no more than 1c when you need a faster charge )

Click to expand...

Thank you for your advice. For the 14s5p pack, in normal conditions I will bulk charge @ 5A so ~1A for each cell (~0.3C).

My apologies I forgot to cite the source for this figure.
Here is the full webpage containing all characteristics: http://lygte-info.dk/review/batteries2012/Sanyo NCR18650GA 3500mAh (Red) UK.html

Click to expand...

i would do your own testing as none of those tests were done on with welds just a plastic cell holder :/

Disclaimer I do not use a BMS and have it for 1,020 cycles so far. Yes danny dog I ride bareback because I use 20ah quality cells. Your cells should be good out of the box if real Panasonic GA cells .

Most of my questions were solved when my charger usb port fried, together with my PC motherboard.
Basically same story happened to this guy: http://www.helifreak.com/showthread.php?t=725276

I now need to replace my computer and I won't be able to connect the charger to any other computer.
So I'll just use it to balance charge the cells, and check the IR before assembling the battery pack.
No more questions about capacity, charge/discharge plots, and subtle optimizations...

I'll probably get my bike running sooner than expected after all...