I'm building a multi purpose battery box for the car. It'll have at least a USB and a 12v cigarette lighter output, and retractable clamps for connecting straight to the car battery in the event that the car won't start, which has been happening lately. It'll charge from the car's cigarette lighter when the car is running.

My question is, what would the recommendation be. I'm using 18650 lithium cells, since I want at least a 12v output, should I put 3 or 4 of them in series with several series sets in parallel, or should I put them all in parallel and step up the voltage to whatever is needed? 4 fully charged cells in series would be 16.8 volts, too high for the car battery, but approaching dead it would be around 12v. If I step up the voltage I can control the output precisely, and I know the cells can handle the current output. Any advice?

  • 1
    \$\begingroup\$ I would NOT be connecting LION batteries to a car like that. The require some special treatment for charging. \$\endgroup\$
    – Trevor_G
    Commented Mar 11, 2017 at 21:58
  • 3
    \$\begingroup\$ Starting a car requires several hundred amps. Designing a step up converter for that is going to be fun. Please make a video! \$\endgroup\$
    – Christian
    Commented Mar 11, 2017 at 22:00
  • 1
    \$\begingroup\$ A step up converter will introduce current limitation and losses. It is best to put them in series series and add sime high current diodes in series before the + output cable to reduce this voltage to less than 15V. But products like this already exist on the shelf. Do a search on ebay for "Jump Starter lithium" \$\endgroup\$
    – cyberponk
    Commented Mar 11, 2017 at 22:01
  • 4
    \$\begingroup\$ I recently faced the same problem as you with my car. Let me tell you, forget about jump starter. Check your alternator first. Start your engine and turn on everything electric on the car and check battery voltage. If anything less than 13V is measured, replace your alternator with a new one and your problems will end. If you have more than 13V, then your battery is the problem. Replace your battery and that´s it. Making a jump starter for a faulty car is a total overkill, will consume much time and effort and might not work in the end. \$\endgroup\$
    – cyberponk
    Commented Mar 11, 2017 at 22:05
  • 2
    \$\begingroup\$ Let me clarify a bit, I wrote the post in a hurry before going into work. I don't plan on connecting the batteries straight to the car battery. I also don't plan on using it as a jump starter, I know that takes hundreds of cranking amps. I do know, however, that in a pinch you can trickle charge a car battery with 10-12 AA batteries in series over about 10 minutes. This is what I plan to do. I'll have a protection circuit, but I'll be able to trickle the car battery with 2-3 amps over some time so it'll start. Handy if I'm in a situation where no one can give me a jump. \$\endgroup\$ Commented Mar 12, 2017 at 5:58

2 Answers 2


I don't have enough reputation to add a comment, but I think it is important to note here that the form factor alone doesn't tell you either the nominal voltage, capacity or the safe charging voltage of a cell - these metrics vary between chemistries.

A lithium iron phosphate (LiFePO4) cell has a nominal voltage of 3.3V and must not be overcharged above 3.65V - at 4.5V it is in danger of violent rupture, disassembly and fire - despite being the relatively "safe" li-ion chemistry.

Lithium Cobalt Oxide (LiCO2) cells can be charged up to 4.7V, though this is frequently limited to 4.2V as there are diminishing returns but increased hazards above this voltage.

The low voltage cutoff presents other dangers - LiFePO4 cells can safely be discharged down to 2 or 2.5V, while for LiCO2 it is 3V. Ignoring these limits causes permanent damage which can result in dangerous failures at a future time in the life of the cell.

I've given two examples, but there are a number of different chemistries within the li-ion family, and it is the chemistry itself which is of utmost importance when deciding on safe operating conditions.

Unfortunately the chemistry is not always obviously labelled, which is why it is imperative that you use the datasheet to identify the Safe Operating Areas of the specific cell you have - whatever the formfactor (18650 is just any cell 18mm wide by 650mm long).

From these key characteristics (nominal, max, min voltages, temperature range, capacity) the chemistry can usually be deduced where it is not overtly stated, but either way a Material Safety Data Sheet ought to be available for any cells worth using in your design.

  • \$\begingroup\$ I'm using 18650 cells salvaged from old laptop batteries. Thus far I've used a lithium charger circuit to charge them, either single-cell or parallel, without issue. This exact project kinda died a while ago, but I've still been doing other things with the batteries. From my research and testing (aka poking with a multimeter while it's charging or under load), they seem like bog-standard 3.7v nominal regular old Li-ion cells, charge to 4.2v cutoff at 3v. \$\endgroup\$ Commented Mar 20, 2018 at 19:27
  • 2
    \$\begingroup\$ 650mm long? 65cm? Over half a metre? \$\endgroup\$ Commented Sep 12, 2018 at 5:07

You are duplicating exactly what dozens of commercial Car Battery jump start products do.

It's just fine to use Li-Ion batteries for this application and you should connect 4 cells in series, and then parallel multiple strings to get your desired cranking current. There is absolutely no problem with paralleling Li-Ion batteries with Lead Acid car batteries, providing the Li-Ions are not fully charged (they'd be up around 17 V).

The commercial products you see out there do NOT fully charge the Li-Ion cells, they charge them to about 14 V, whitch is the voltage of a fully charged Lead acid battery (13.8 V actually). For most of the commercial product you actually charge them from the cigarette lighter in the car so they only get charged to the same voltage as the Lead Acid battery ...13.8 - 14 V.

Here is the typical 18650 battery discharge curve:

enter image description here

If we use this chart to surmise what may be available for your batteries you could assume that to get cranking current of say 400 Amps you'd need 13 strings of 4 cells (total 52 cells) each providing 30 Amps. That's a lot of cells!!!

But the commercial products seem to do it with a much lower volume of batteries, so what discharge rate are they using??

The discharge rate for any battery is limited by the internal resistance and the current capacity of the leads. Here's some data that's very relevant.
So from a single cell you can expect short circuit currents approaching 200 Amps! Not advisable for any length of time because the battery is heating up very rapidly.

Being more conservative, you could perhaps expect that you might use a limit of say 90 Amps ....so would need 6 strings (total cranking current of approximately 540 Amps).
24 cells seems much more reasonable, and an aiming fully charged voltage of no more than 14 V. This does reduce the capacity somewhat, but not by more than a few percent. If the cells are about 2500 mAH then your battery pack is about 15000 mAH at 14 V.

How close are we to a commercial solution? This unit from Amazon quotes a capacity of 18000 mAH and a cranking capability of 700 Amps, so perhaps they use 7 - 8 strings of batteries. You could compare a whole lot of examples and get a good feel for the number of strings of cells they are using.

Warning - Remember that these battery packs are meant for intermittent use and you are pushing the envelope on maximum current draw. You need to decide on a really conservative solution with 52 cells or much more aggressive 24 cell solution. If you are buying Ebay cells of unknown characteristics I would certainly consider being more conservative.

  • \$\begingroup\$ I don't plan to use the thing to actually jump the car, just to trickle the car's battery over 10 minutes or so. I know that you can use 10-12 AA batteries in series over about 10 minutes to trickle enough charge into the battery to get it to turn over, at which point the alternator will take over. I'll be limiting the output current to the battery clamps to around 2-4 amps. I know I could buy something, but I like building stuff myself, gives me a chance to learn, plus I'll be able to build it exactly how I want. \$\endgroup\$ Commented Mar 12, 2017 at 6:08
  • \$\begingroup\$ I'm thinking 4 series sets of 10 parallel cells, which will fit nicely in the enclosure I have with plenty of room for the charging and output circuitry. The output voltage will have to be regulated, but I can do that, and I'll have separate circuitry for the 5V usb output and "12V" clamp outputs, but it will all be connected to the charge/discharge regulator. \$\endgroup\$ Commented Mar 12, 2017 at 6:12
  • \$\begingroup\$ Most car batteries are in the 25000 - 35000 mAH range so you building a battery with only 18000 mAH means you are unlikely to be able to even quarter charge the car battery in a reasonable time. ...I'd just jump start the car. \$\endgroup\$ Commented Mar 12, 2017 at 6:36
  • \$\begingroup\$ My girlfriend and I go hiking, so often we're in a place where finding someone to jumpstart the car isn't really an option. I know for a fact that the 12 AA battery trick works, I've done it before. A car battery may have a large capacity, but it doesn't have to be at 100% capacity to start a car, so trickling it over some time should suffice. The cells I plan on using are panasonic, advertised with a 2600mAh capacity. Here's the video where I learned the AA battery trick. \$\endgroup\$ Commented Mar 12, 2017 at 6:53

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.