DIY Battery for E-Bike


1 µW
Sep 28, 2023
Hello Everyone,

I originally made this post on the diyelectriccar forum, but the guys there suggested me that this forum has more knowledgeble users for what relates to my question, so im pasting the same here.

I'd like to start by thanking anyone with enough patience to read it all, i have a degree in electronics but my daily job is in a completely different field so,
with my knowledge aging and rotting away, i want to double check with you experts to make sure i don't do any dumb mistake.

I recently had the luck to acquire a batch of still sealed laptop batteries from a closing office.
These were in a spare-parts stock and have 0 cycles on them as they were never used.
I opened all the batteries and harvested the cells, then used my D2 Digicharger to charge them all, plus estimate the capacity by measuring how much time they took to charge (the D2 uses constant 500mA for charging).

I have in total 232 Cells, but they are of different capacities.
I split them in 2 batches:
130 Cells (labeled Old Batch) are those i harvested almost exactly a year ago and after fully charging them i put them aside waiting to get more (measured most of them today, after a year of storage they dropped to 4.10-4.16V so they look healthy).
102 Cells (labeled New Batch) i harvested and charged in the last 2 weeks.

My objective is to convert my bike to kinda-hybrid, it has quick-release front tire, so i wish to buy a front wheel hub motor that i can connect and disconnect quickly in order to "choose in the morning" if for the day i need it electric or pedal-only.
The motor controller will be connected via Ubolts and the battery will be on the package carrier secured with bolts and straps.

I was thinking of using one of these motors.
Currently i usually cycle 50-70 km on pedals, i wish to achieve a similar (or more if possible) range with electric (so with minimum pedaling), but my main doubt is whether or not 36V is enough to overcome steep-ish hills or if i should get a 48V motor instead.

I attached a PDF/Excel of all the cells i have, if they were all the same i wouldn't have any problem, but they are different capacity/C-Rate, so i'm not sure how to proceed.

If i am to combine only the NCR18650 i have (they are the cells i have most of with 99 identical ones), i can do a 13S7P using 91 Cells, that would give me a 48V 19.60Ah (917Wh cca.) pack and im not sure these numbers are sufficient for the range i wish to achieve.

If instead i am to combine different cells, what would be the best way to do it?
First thought goes torward combining them in the parallels, so to have multiple parallels of same capacity that will self balance and then series them (so for example 7P of NCR and an additional 2 UR18650FM that are similar specs so to have a 13S9P or more), or more advancedly as rePackr - 18650 pack builder shows.

Could you please advise if it is a sound thought?
Considering i would like to have as many stats as possible, could you advise on a good BMS with sensors?
I also plan to have multiple temperature sensors inside the battery in various spots and have an ESP8266 send them to my main screen so i can keep an eye on the temperatures at all times.

Please feel free to tell me i'm an idiot or such, i wish as much input as possible to understand what can be done better, starting with motor, controller, battery plans etc.

Thanks to everyone in advance.


  • Cells.jpg
    1.1 MB · Views: 5
  • CellsDB - Cells.pdf
    130.6 KB · Views: 3
  • CellsDB.xlsx
    14.9 KB · Views: 4
Last thread here I read on mixing cells equally into p-groups didn't show much success. Putting a couple strong cells in each p-group was the same as just adding the same number of weak ones. I guess you wouldn't really mind if they are similar spec anyway. Ideally I think each of your p-groups would work out to matching capacity and internal resistance, which are testable attributes of the cells and p-groups. Since your cells don't match exactly I'd personally also attach an active balancer or get a BMS with that capability instead of the usual cheap ones limited to top balancing during charging.

My 13Ah 48V battery gets me about 15 miles on a 20"x4" tire fat tire bike going 28mph using a baserunner controller and 750 watt motor, if you want a range example. Cycle Analyst shows I go way above that wattage, though. More like double.

Btw, storing lithium cells at full charge damages them. You shouldn't store them at full voltage if you want long life out of them.
thanks for your input, an active balancer is definitely something i'd want, however i was checking some Daly ones as they seem decent quality, and i'm very confused by their pricing scheme:
They have a 13S 40A BMS + Active Balancer for 90 Euro 13S 40A
They have a Active Balancer for 30 Euro Active Balancer 13S
plus a 13S 50A BMS for another 30 Euro 13S 50A BMS

is there a reason i'm missing to for the stark price difference between "bundle" and "separate" or it's just that the bundle also comes with "smart" features (that i don't need anycase)?

In the end i selected cells of the same type (Charge/Discharge currents and capacities), i will have a 13S9P, each cell has 5A max discharge, hence the 50A BMS.

Thanks again for any input.
Yeah, smart feature is usually an extra Bluetooth dongle that plugs in. Then you can connect an app and toggle settings/manually kickoff balancing, etc..
I suggest that you purchase a YR1035+ IR meter from Aliexpress, about $36-40 USD. Four point probe. Measures AC impedance which is a good predictor of which cells to mix. It's not the same as the DC series resistance, but correlates well enough, Good cells for ebiking are below 20 milli-ohms. At 40 milli-ohms, there's voltage sag unless you have a big P-group. The measurement takes seconds, unlike capacity tests which can take a half day. It's not an expensive tool and works as a voltmeter accurate to 1 mv, unlike most home mutimeters.

I found that P-groups using cells with a big difference in IR have unpredictable results. They should be within 10% of each other.

I have a capacity tester, but it only does 4 cells at a time, at a discharge rate of .5 to 1.0A. Ittakes about 8-10 hours to charge/discharge a 3.5AH cell. Capacity testing is slow. For new cells, I usually get 90-95% of the rated capacity. I believe your discharge test is a bit conservative on its findings.

My ebikes, 500W motors, based on years of measurements use about 6 WH/km (10WH/mile) at an average speed of 20 km/hour with healthy pedaling. Your motor candidates looke like nomal 500W motors. Go with 48V,