Bike Battery

 When I went out yesterday, I used the bike intensively, and I completely discharged two of the batteries, and I'd guess I half-discharged the third. So I do have enough batteries to keep me going all day - but one day, I'll need to replace them, as they only last a couple of years.

It has occurred to me that, since I'm just putting the bike battery in the panniers, it can be any shape or size I want. And that means, maybe I could make my own. Here's a useful link

First, some nomenclature. A cell is one (usually) cylindrical thing that pushes out electricity; a battery is a whole bunch of cells, wired up to give the voltage and current that you need. So, for example, with two 3.7v cells, you can put them in series to get a 7.4 v battery. Or put them in parallel to get a 3.7v battery with double the power. And inside my 24v battery, I see seven separate objects labelled "XH259-10J", which are in an aluminium enclosure, and which I will call "cans".

I did a bit of research. I can buy a li-ion cell size 18650 for around $3, That give me 4.2 volts, 2.2 amp-hours. Seven of those in series would be 29.4 volts (25.2 when discharged), 2.2 amp-hours, and with five of those sevens in parallel, I have 11 amp-hours, which is the capacity of the batteries I'm using. so that's about $105. Add $16 for cell holders, that's $121. The cells weigh about 50 grams, so that's 1.75 kg. My existing 24v (actually 29.2v when freshly charged) 10AH bike batteries are 3.12kg; that includes the casing and the battery management system (BMS).

What's a BMS? I think it stops you from overcharging the battery, or from letting it get too discharged (Li-ion batteries stop working altogether if you discharge them too much). And it stops you from pulling too much current. It can also worry about the temperature of the cells.

A BMS can also balance the cells. Without balancing, the cell of smallest capacity is a “weak point”, it can be easily overcharged or over-discharged while cells with higher capacity undergo only partial cycle.

If I use protected cells, I might be able to work without a BMS, although I would put in a fuse to protect against short circuits. Protected cells have a small circuit board that stops the charging or discharge of the cell in certain circumstances. Some of the circumstances in which the circuit include: Over-Charge, Over-Discharge, Short-Circuit, and in some cases, overheating. The protection PCB resolves most of any safety issue for LiCoO2.

LiCoO2 (lithium-cobalt, commonly called Li-ion) or LiFePO4? LiFePO4 has a smaller energy density and are, it's said, slightly safer. But LiCoO2 is lighter. Voltage is around 3.7-4.2 for LiCoO2 and 3.2-3.6 for LiFeP04. So, 7 LiCoO2 cells are 29.4V, whereas 8 LiFeP04 cells are 28.8v.

LiFePO4 can be charged 2000 times; LiCoO2 can be charged 1500 times. Not a difference that sounds important to me.

LiCoO2 can be charged at 2 amps, discharged at 5 amps. So four in parallel would discharge at 20 amps. LiFePO4 can be charged at 2-6 amps, discharged at 30 amps. So four in parallel would discharge at 120 amps. I'm not bothered by how long it takes to charge, as long as it will charge overnight. So I'd be happy to charge a 10Ah battery at 1 amp, since that would take 10 hours. For discharge, a 24v 250w motor wants 10 amps, a 500w motor would want 20.

LiCoO2 lasts for 2-3 years; the electrolyte crystallises, and you lose capacity.

LiCoO2 gives 4.2 to 3.6 volts. and stores (in 18650 size) about 2200 mAh (2.2 amp-hours)

4.2V Full 100%
4.1V About 90%
4.0V About 80%
3.9V About 60%
3.8V About 40%
3.7V About 20%
3.6V Empty

LiFePO4 stores (in 18650 size) about 1100 mAh

So, all things considered, it looks like I want  LiCoO2 (Li-ion) protected cells, and expect to repelllace them every three years.

Looking at my existing 24V battery, inside there are seven cells each in an aluminium can (or so it appears, I think that actually each of the things I see is a battery of seven cells) and a pcb (the BMS, I guess). Each canl is 127x65x16 mm, which is about the size for seven 18650 cells (they are 18 mm by 65 mm long, hence the name 18650). The cans are 4.1v, so they must be seven cells in parallel inside the thin metal can. So that would mean 49 cells altogether. Seven freshly-charged  LiCoO2 cells gives 29.4 volts, which is what I get on a newly charged battery. To give 10 amp-hours, each cell would be 1430 mAh, which sounds right for this sort of cell - I've seen it quoted as between 1300 and 2400. 18650 cells weigh about 45 grams, so 49 of them would be 2.2 kg (when I weighed them, they were 3.12 kg, but that includes the casing and the BMS).

So, suppose I made up my own battery. I put seven 18650 cells in parallel, and that makes one block.  I run seven of these blocks in series, that's 2.2 kg and 126x126x65 mm. I should be able to get 5 amps per cell, but it's better (you  get more out of the battery) if you only take 3 amps. But with seven in parallel, that's 21 amps, which is plenty.

Here's a very comprehensive battery test for 18650 cells. One of the first things that you notice, is that when the manufacturer says a cell is 3000 mAh, you'll probably only get 2500, and it could be a lot less.