An introduction to lithium 18650 batteries

May 3, 2017

The development of electrical transport, from skateboards and scooters to buses and trucks, has been an evolution of battery power. We’ve been able to store electrical energy for several centuries, but it has not been very efficient until recently. In China, where the ebike has been wildly successful, lead acid batteries were dominant over the early years. So even ebikes with an old battery tech worked pretty well.  The American media picked up on the rise of ebikes in China around 2009, so this was not very long ago. Lithium batteries were  expensive until very recently. Right now, lithium chemistries are driving massive changes in transportation, and taking ebikes way beyond the limits of lead acid batteries.

When you compare  how well lead acid (the battery in your gas car) and lithium batteries work in an ebike, you can see why lithium is a great cell for an electric bike. To start, lead batteries are heavy, which is not what you want on a light vehicle. Lead batteries suffer enormously because they lose capacity at high discharge rates. In other words, if you draw a lot of current from a lead battery, you will get a lot less total amp hours or watt hours. Lithium batteries are much lighter and lithium batteries can be made that allow high current draws without reducing the overall amp hour capacity very much.

Every battery has a chemistry and every chemistry has a set of characteristics. The basic feature of any chemistry is the voltage a cell will produce. For lead acid, it is around 2 Volts (so, a 12V lead-acid battery has six cells). For lithium, it is around 3.6 Volts. Your standard alkaline battery, a double A or AAA cell, is 1.5 Volts. So the cells voltages are all over the place and you must match voltages to the product you are using. Any battery has a range of voltages, from the fully charged voltage to the discharged voltage. Most batteries that you install, a D cell or AA cell, is installed in series. You need ‘a set of batteries’ for a toy. Often there is an upside down, right side up pattern to the cells in the battery holder. That’s how you produce a bigger voltage for more lighter or a more powerful motor.



The now-common and popular 18650 format, on the left.


A good battery has to be more than just a light battery with great capacity. For any heavy load use, it also has to be rechargeable. It’s a lot like the arguments about electric cars, right now. A car has a small battery, a low capacity, so…the range is not good. An electric car lacks a fast charge port so there is no easy way to replenish the range, on the road. Charging is critical in cars, but less so in ebikes. Range is critical in battery vehicles, but charging often can compensate for that.

There have been other chemistries that were very popular, either in consumer products or in prototype electric vehicles. The Nickel Cadmium battery comes to mind, because it was the first consumer rechargeable battery system. The follow-up to the NiCad was the NiMH, a less toxic battery with better capacity. Neither of these batteries matter much, right now, but they are still in use. The NiMH is useable at low temperatures. Lithium batteries should not be charged at freezing temperatures.  Lead Acid continues where cost is the main consideration, but…the cost of lithium is dropping rapidly.

The next battery, the next potential battery,  is more interesting, today, than those of the past. There are a lot of ideas, a lot of ‘breakthrough’ technologies. But, there are many small tweaks that can be made to lithium batteries, and there is a lot of production capacity in place. The world is getting locked into lithium because we are seeing Gigafactories to make these cells.



A Battery Management System (BMS) adds a layer of passive protection for battery pack users.


Batteries in vehicles have to be reliable. You can’t swap out EV batteries as easily as you can in a flashlight. The battery pack will require a fairly high level of management. Most of this is automated. There are circuit boards with micro-controllers that do most of the work. There are several critical issues with lithium batteries. Lithium batteries are fairly complicated and they can do nasty things if you don’t follow the correct procedures. It’s not only the risks of lithium, but the sheer numbers of cells in an electric vehicle. Every cell should be as reliable as possible and every problem with a cell should be as isolated as possible.

Here are some basic rules for Lithium:


  1. Charge them, but don’t charge past 4.2 Volts
  2. Discharge them, but don’t go below around 3.0 Volts
  3. Limit charge current to roughly 0.5C, or as specified
  4. Limit discharge current as specified
  5. Limit charge and discharge temperature, as specified
  6. Limit charging to the specified ambient temperatures (not hot, not cold)
  7. Monitor cell voltages for balance. Balance charge as needed
  8. Consider capacity and charge cycle lifespan
  9. Consider physical protection of the pack
  10. Consider short circuit protection
  11. Consider individual cell disintegration issues


The first issue has to do with charging. For the most part, the risks of lithium cells, in a pack, are in the charging and the discharging. That is when the pack is stressed. The explosion and fire issues seem to revolve around charging, and there’s not enough data to say what all the problems are. It’s probably defective cells and the fact that cells that are close to each other can cause a chain reaction of overheated or explosive cells. It is important to understand that any pack is made up of multiple cells, cells that are in close proximity. How close is a design and safety issue, but they may be physically touching. If a cell gets hot and it is close to other cells, they may get hot, and the process of breaking down can feed across the pack.

The core idea of a battery pack seems simple. You take a single cell and build a pack by adding other cells. There are two ways to add cells because there are two connections on a cell, the positive and the negative. You can go positive on the first cell, to negative on the second cell, and this increases the voltage. You can go across the connections, positive to positive, and increase the capacity of the pack you make. There are 52 cells in a common ebike pack, and 7,000 cells in a Tesla. But the same rules apply. Add cells in series for more voltage, add cells in parallel to increase capacity. Generally, the cells are put in series for voltage and then the sets of cells in series are paralleled, to increase the watt hour capacity.

A pack is a pack in the sense that the components are electrically bound together. Each cell will respond to a charge current within the laws of electricity. When you apply 50+ Volts to a set of 13 cells in series, one of the cells doesn’t ‘hog’ the whole 50 Volts, causing it to  explode. At a basic level, the cells all absorb an amount of current that maintains an equal, or fairly equal, voltage across the cells.  You can Google the word ‘series’ and contrast it with ‘parallel’, for a full understanding. Anyway, when you build these 13 cells into a pack, in series, and apply about 54 Volts to it, you (should) get an equal charge in each of the cells. You will get  the same number of amp hours, when the cells end up at the same voltage. But, there are always possible problems or glitches. The 52 cells in a ‘standard’ 48 volt pack mean that, if any one of the cells is defective, it can be a problem. Welding cells in series, and then in parallel, makes finding one bad cell difficult.



The cell-ends with the blue insulating washer are the positive ends, the flat metal bottoms are the negatives.


Charging simply means you increase the voltage of the pack and you add capacity to the pack. The two things go together. If a 48 volt pack is almost fully depleted, it will be at around 42 Volts. If you put it on an automatic charger and wait until the charger stops charging, the battery pack will be around 54 Volts. That’s a big range, and it is something to get used to.  A high voltage and a low voltage condition, up to a point,  are not defects, they are electrical properties of lithium batteries. But it’s very useful to know what the ‘normal’ range is, and to know what the normal full charge and full discharge voltages are.

Battery capacity is based on what goes in, and what can them come out, charge and discharge. So a battery has a maximum capacity, and a capacity as it sits, right now. One of the great features of Lithium cells is that almost everything you put in a pack can be used when you start to discharge the pack. It is very efficient. If you want to know the precise capacity of your pack, you need an amp hour or watt hour meter. You can do the measurement when the battery is on the bike, or do it when the battery is on the charger. Your battery has a voltage rating and an amp hour rating, in general. Since Watts = (Volts X Amps), the watt hours are the Volts times the amp hours.




Remember, though, that the Volts drop as you discharge the battery. This makes amp hour measurements complicated. Watt hours are more precise because watts are more refined a measure than amps. Here’s why. Suppose the battery is fully charged at 54 Volts. You ride and draw 6 amps. That is 324 watts. But at the bottom of the battery, what happens? When the battery is at 42 Volts, and you draw the same 6 amps, that is 252 watts. That’s a lot less watts. So to get the watts you need, as the battery depletes, you draw more and more amps. To get 324 watts, you need about 8 amps. If a battery has a capacity of 16 amp hours, and you have drawn 8 amp hours, the battery is more than half depleted because every amp is now a lower voltage amp. If you have a 500 wh battery and you know you have discharged 250 watts, it probably is about half discharged.

Charging lithium batteries is somewhat more complicated, and more critical, than lead acid batteries. You charge a set of cells the same way you would charge one cell, multiplying for the cell count, the ‘S’ number, cells in series. If you just wanted to charge one cell, you would want to know the charge characteristics of the cell. Most 18650 lithium cells charge to 4.2 Volts maximum, and the charge current is about 0.5 times the capacity. A 3000 mah battery would be charged at 1500 mA maximum, using this rule. Cells have data sheets that specify precise numbers for charging.

It sounds like the charge would be done in two hours, but this isn’t quite correct. The protocol for charging a cell (or group of cells) explains why charging at half the capacity, for two hours, will not fully charge the battery. You can’t put 1500-mA into the battery for two hours, from start to finish. The charge will taper, especially as the cell reaches 4.2 Volts. The charge rate, called CC or constant current, will be absorbed by the battery at the beginning.

Let’s assume the battery is fully discharged, and it is a 3000-mAh capacity cell. We are trying to put 3000-mAh back into the cells. The voltage rises from 40v or so to 54.6V. That isn’t a full charge, but the current will now taper at that fixed and final voltage. The voltage rises to 54.6V, but that voltage has to be maintained for a while.

The charger is fairly easy to set up for lithium cells. You take the number of cells in series (S) and the voltage will be that number, ‘S’, times 4.2. So for 13 cells it is 13 x 4.2. That’s for the maximum safe charge. Part of the nomenclature of batteries and packs is to  distinguish between different type of series connection. If you take 13 cells and wire them, plus to minus, you end up with a 13S battery pack. If there is one set of 13, that is 13S / 1P. If there are more sets of 13 series cells, that becomes the P number. A 52 volt pack is a 13S / 4P much of the time.

The voltage is based on constants, so 13S is about 48 Volts all the time with standard cells (NOT LiFePo4). But the P number, the parallel number, won’t give you capacity, or amp hours, unless you know the mAh rating of the cell. You might think a 14S / 3P pack would have less capacity than a 14S / 4P pack, but it depends on the capacity of the individual cells. The cell  used in a battery pack  must be used for the entire pack. There is no way to add different cells and have the pack work or be safe and stable.

A pack with more rows of cells, more parallel sets, will have more capacity if you are using the same cells for all the packs. Different packs end up being charged differently, and capacity fixes the maximum charge amps while the number of cells in series determines the charger’s maximum voltage. The number of cells in parallel will determine the capacity, C, of the pack, when the ‘P’ number is multiplied by the (single) cell capacity in amp hours or milli-amp hours. The charger may be set up to charge at C/2, in general, but check the sheet. Even if you are using the same cells in several packs, the charger setup comes down to the specific capacity of all the cells, and the number of cells in series.




The Luna advanced charger has the ability to slow-charge at 3A, or fast-charge at 5A. It also allows the user to select 80% for long-term storage, or 90% for maximum battery cycle life. We do not recommend ever charging a battery to the maximum of 4.2V per cell.


Charge voltage gets very confusing when people have 48 and 52 volt batteries, and they look at a charger that outputs 54.6 Volts. Generally, when looking at a charger voltage, you should divide it by 4.2 to get the number of cells in series. To add to the confusion, the nominal voltage of a pack is often the number of cells in series multiplied by 3.6 Volts, or 3.7 Volts. The charge voltage is a maximum voltage, the pack voltage is an average voltage across a discharge cycle.

Charging a set of 13 cells in series means the charger voltage will be 13 x 4.2, or 54.6. As was mentioned earlier, the maximum current will be delivered in the early stages of the charge, but it will taper off. What happens is that the charger supplies a lot of current until the voltage reaches about 54.6. Once that voltage is reached, the current tapers, but the voltage is maintained (constant voltage). The charge to  lithium cells must end. A pack  can’t float all the time like lead acid. That charge termination point is generally specified, but a rule of thumb would be 10% of the CC (constant current) charge rate.

In concrete terms, if we have a 13s pack with a 10 amp hour capacity, we can say:


  1. The charge rate is C/2 or 10/2 = 5A  (C is capacity in AH)
  2. The charger voltage is s 54.6 or 13 x 4.2 (max voltage for 100%)
  3. The taper charge end current is 10% of 5 amps, or .5 amps
  4. The CC is set at 5 amps max
  5. The CV is set at 54.6
  6. The taper charge continues at 54.6 continues until current drops to 500-mA
  7. You don’t leave lithium cells on the charger (no float charge)


The issue with a simple charger concerns that 4.2 volt constant. If you use that number, that voltage, you get what is called a ‘full’ charge, after the taper. But a ‘full’ charge will not give long battery pack life. Managing the cycles a pack will produce is critical with expensive packs, like those in electric cars.

If you break down what a charger has to do, any commercial charger is likely to be a complete solution. That means that you will generally buy a charger that is tied to one battery pack, or one voltage of battery pack. So if you have a 48 volt pack, it might work with another 48 volt pack, but it might put out too much current, or be too slow. It may have circuits that interface to a specific battery. They may make it hard to use third party equipment on some bikes.

Since most chargers have no settings (you just plug them in and hook up the battery), you can read the basic property of the charger on the label, somewhere. For a 48 volt pack, the voltage is likely to be listed as that 54.6 we computed above. They should list the amperage rating of the charger, as well. But that’s about it. These chargers are one trick ponies. Whatever voltage the charger is listed as supplying, the ‘S’ number should be that voltage / 4.2. In general, the nominal voltage of a battery pack is the ‘S’ number * 3.6. The charger voltage is a maximum number and the pack voltage is more of an average voltage as the pack is discharged.

If you want to understand the basic design of a charger, some kind of  emergency charger, it’s going to start with the voltage limit or voltage output. The other basic requirement is to limit the current into the pack. Creating a precision voltage is quite easy these days, since most of the cheap DC converters on Amazon and Ebay are precise enough. These converters are usually DC to DC, step up, or step down. You can make 54.6 Volts from a 12 volt battery, but it should be a big battery. You also need to limit the amps the charger will deliver. There are modules, mostly out of China, that let you fix the voltage and the amp limit. This is the core of a lithium charger, but it won’t end the charge without something more.

The ‘best’ setting for a charger is worthy of some discussion. If the maximum, or 100% charge, is reached using a 4.2v setting, anything less than 4.2v will produce less (%) of a charge. The advantage to, say, an 80% charge is that it puts less stress on the battery, so the battery will go through more cycles. There isn’t too much precision in saying how many more cycles, but it’s probably more than double at 80%. Knowing this should make you want to know more about how to make an 80% charge actually work. The easy answer is to buy a bigger battery pack, and only use a part of the capacity. This may or may not work. I like small battery packs that are easy to carry, easy to hide. Charging to 90% is advantageous and may be easier to live with.

It might be advantageous to make a pack that was designed to last 10 years. That is what the automakers are doing. It means setting the pack and charger up so that the pack is not fully charged or fully discharged. If you use the pack in the middle of the capacity, it can last for thousands of cycles. But you might want to design a ‘bullet proof’ case, and arrange the cells to prevent thermal interactions where one cell gets very hot and that spills over. The problem is that many packs are proprietary, and the ebikes they mate to may disappear. With generic packs there is more room to make a ten year pack, engineered for a long life. But it may not matter much, as cell prices drop, and cell capacities improve. New may be better. Would you pay $600 for a ten year pack and charger, versus $350 for a 300 cycle battery?




Here we are showing the red “Deans” connectors, the white housing is the common “Molex” connectors, and the most popular connector for charging is the yellow XT60.


Chargers that work on more than one pack may be a good idea for some people with multiple bikes or packs with different voltages. There is a fairly expensive ‘Swiss Army Knife’ charger. Hobby chargers, mostly used with LiPo packs, have an autosense function that figures out the number of cells in series. This sort of protection is good for any charger that can be used with multiple voltages. If you have different voltage packs, it is a good idea to make a connector to each pack and charger, something unique so you have to plug in the right charger.


Written by George Sears, May 2017

  • Marc Plante

    Thank you for the great material.