Background: I would like to build a battery pack out of 18650 li-ion cells (3S2P). Safety is my absolute priority.


Here is the list of all safety measures that I could think of:

  • Use batteries with build in protection
  • Use BMS on whole battery pack
  • Install a fuse at the output of the battery pack
  • Connect the output of the circuit through a thermostat that will measure the temperature of the cells
  • Install a fuse to each wire in balancer
  • Keep a battery pack inside Li-Po safe bag even during usage
  1. Which of those measures actually make sense to implement and what is best combination of them? I.e.: Does it make sense to have both cells with protection and BMS?
  2. Are there any other reasonalbe safety measures that I've missed?
  3. What is the difference between chargin via balancer and charging through load cables? On difference that comes to my mind is that usually balancer cables have smaller diameter and therefore they cannot handle high current.
  4. If I have a battery pack with BMS can I still charge it using balancer port?

Update 2021-11-03

Thanks to everyone for the great advice and for building my awareness of the many risks. To actually maximize safety I decided to purchase an already manufactured LiFePo4 battery pack with BMS.

  • 14
    \$\begingroup\$ The safest way to make a LiPo pack is to not make a LiPo pack. Search for Lithium Iron Phosphate. LiFePO4. Very close performance to LiPo, no risk of explosion \$\endgroup\$
    – Kyle B
    Nov 2, 2021 at 7:36
  • 3
    \$\begingroup\$ @KyleB: I dont intend to use LiPo, mainly for safety reasons. I want to use Li-ion. \$\endgroup\$ Nov 2, 2021 at 8:14
  • 1
    \$\begingroup\$ @tlfong01: No I'm not EPAC guy, I'm just the guy don't want make set a fire in his house. For EPAC I would have to build significantly larger battery packs, mine will be at most 3S2P. \$\endgroup\$ Nov 2, 2021 at 8:14
  • 10
    \$\begingroup\$ LiFePo4 is not LiPo but something much safer. (LiPo and Li-Ion are essentially the same thing safety-wise) \$\endgroup\$
    – user16324
    Nov 2, 2021 at 13:33
  • 3
    \$\begingroup\$ Whenever some manufacturer claims "it has built-in protection" you need to ask them what that means more exactly, because it can literally mean anything. It could mean reverse polarity protection with a diode etc, or it could mean a built-in fuse or it could just mean there's a NTC thermistor for you to connect to, or it could mean some smarter current-limiting electronics, or it could mean a whole built in BMS system. And probably some other things... \$\endgroup\$
    – Lundin
    Nov 3, 2021 at 9:41

3 Answers 3


If safety really is the absolute priority, you should not build your own battery pack. If priority #1 is to build it yourself (for education, customization, entertainment, or whatever reason), and safety is priority #2, then you should probably use a safer chemistry than Li-Ion (as comments have pointed out).

So I'll assume you want safety-third.

Your measures seems mostly fine, and would generally all result in increased safety. I'd slightly question placing a thermostat on the output. What you want is a thermal cut-off, which would be a part of any reasonable BMS.

The main thing it seems you might not be thinking of is the inherent danger in adding complexity. If you have a safe, self-protected cell, protected by an appropriate fuse, that system could very well be safer than a complex, but shoddily assembled battery-pack with wires running everywhere. Consider that every component you add also adds things that might go wrong, and you have to factor this into your overall risk-assesment.

As for you questions:

1/2: See above

3: Charging with a balancing cable allows the charger to charge the cells individually, so it can fully charge them all. If you charge only through the load-cables, you'll have to stop charging once a single cell reaches maximum voltage, meaning you might have unutilized capacity in your pack.

4: This depends on how the BMS is designed and operated, but there's no reason why you couldn't design the BMS for protection only, and leave charging and balancing to an external charger (and do it through a balancing cable)

  • \$\begingroup\$ Thanks for the reply. I used to say that safety was the number 1 priority, but now that I think about it, that wasn't true. I was actually more along the lines of "I want to build a very safe battery pack myself (educational goal) at a reasonable price". I can build many other things myself that won't blow up in my face, so I can buy an already manufactured battery pack. As for chemistry: I'm not attached to Li-ion at all, if there are any safer ones I'm happy to use them. \$\endgroup\$ Nov 2, 2021 at 13:13
  • \$\begingroup\$ A note on complexity is also very helpful. Some of these measures may be unnecessary and at best will do nothing, and at worst may create new problems. Thanks for making me aware of this. \$\endgroup\$ Nov 2, 2021 at 13:19
  • \$\begingroup\$ In the question 3 I meant something different: There are chargers where you need just to plug balancing cable (i.e.: amazon.com/dp/B08L7VCBXG ) There are some where you need to plug both load cable and balancing cable (i.e. amazon.com/FCONEGY-Battery-Balance-Discharger-Batteries/dp/… ). What is the difference between those two methods? \$\endgroup\$ Nov 2, 2021 at 13:22
  • 1
    \$\begingroup\$ Without reading through the datasheets of those chargers I can't be sure, but the only thing I can think of that connecting the load cable gets you, is the ability to push more current into the battery, charging it faster. Of course, there is a maximum charge-current for any battery which you really don't want to exceed, so the charging has to be configured for the specific battery. \$\endgroup\$
    – sondre99v
    Nov 3, 2021 at 7:40

It's nice to see someone caring about safety for once. Batteries can be dangerous things.

Built in protection in the cells is usually very limited, often just an unreliable thermal fuse. A best it gives very little protection, but I find it gives a completely incorrect sense of safety. If a cell has built in protection, my response is to say that's nice, and then treat it the same as if it didn't have any.

A BMS is good idea. But BMS can be at any level. Theoretically they should:

  • Stop over charging (safety)
  • Stop over discharging (safety)
  • Control charging (from a practical point of view)
  • Protect against over (and under) temperature events (safety)
  • Balance all the cells in the pack (for safety, cell life and performance)
  • Monitor pack usage for general information and for state of charge

But I am aware of may commercially available ones which do not do all these things. I have tried to put these features in commercial packs I was working on, but was told it was too expensive, and not required as we had already covered the bare minimum.

Fuses should always be fitted. But they are the final level of protection. Make sure you fit the correct one for your system, but make sure you design the rest of the system (including the BMS) so that you never need it. A fuse is for when other things go wrong. Fuses will not save the pack in case of over or reverse charging. Thermal fuses are a nice idea, but are not to be trusted any more than normal fuses. I would fit a thermal fuse as well as a standard fuse.

All decent BMSs will monitor cell temperature. It is required of pretty much all safety regulations CE, UL etc.

Li-Po safe bags is a nice idea, but again it's a thing that should help if things go wrong, but you should have a BMS which stops this ever being required.

A fuse per cell shouldn't be required as the mechanics of pack should not allow access to individual cells.

You appear to have missed off the important box for the pack. The pack should be in a nice self contained insulating box so that there is a physicals barrier between the cells and the world. The power comes out via a couple of contacts (and maybe some method of talking to the BMS if required, but that could be only one more contact).

In summary: put in everything you can. But make sure the cell pack has some physicals protection between it and the outside world. This stops it hurting the outside world, and the outside world hurting it.

A decent BMS, which does everything I listed above, would be a good place to start. There is no such things as too much protection.

  • 1
    \$\begingroup\$ Thanks for the answer. You are right, I've missed the importance of the box, I was focused on what is inside. I wanted 3D print the box for the pack, but not for safety reasons. \$\endgroup\$ Nov 2, 2021 at 13:25
  • 2
    \$\begingroup\$ If you do 3D print a box, you will need to be careful as the 3D printed plastic is quite brittle and isn't fully water proof. As a thing to sit on your desk, it's fine, but if it ever goes outside or has an exciting life, you might want to consider an off the shelft plastic box. \$\endgroup\$
    – Puffafish
    Nov 2, 2021 at 13:47
  • 2
    \$\begingroup\$ There are two kinds of built-in protection in 18650 cells: the mechanical temperature / pressure cutoff, which is present in practically all 18650 cells, and the electronic protection circuit, which is advertised as "protected cell". I would claim that these are quite reliable, but pack-level monitoring has its place also. \$\endgroup\$
    – jpa
    Nov 2, 2021 at 17:48

Usually you would only do one of these:

  • Use batteries with built-in protection circuit
  • Use BMS on whole battery pack

The reason is that the per-cell protection circuits have a maximum voltage they can tolerate. When the per-cell protection disconnects the battery, it will see the combined voltage of all other cells over it in negative polarity:


simulate this circuit – Schematic created using CircuitLab

Many single-cell protection circuits are rated only for about 5 volts reverse voltage, as would result from connecting a single-cell charger in reverse. Some will tolerate up to 20 or 30 volts, but it is often difficult to find specifications.

A whole pack BMS circumvents the problem by having only a single point of disconnect, implemented using a MOSFET that can tolerate the high voltage involved. This also reduces the resistive losses because there is no need to have a separate transistor (with some series resistance) in each cell.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.