18650 cell get hot when i light up LED

Hi i am trying to charge 18650 with solar and the turning on the 25W LED. The simple diagram is attached with this post

There is CN6009 to step up 8V to 21.5V so that LED get bright. The LED did light up and i keep it up for some time

the problem is after some time one of the 18650 get very hot. LED continue to stay light as was but one 18650 got very hot. i am pretty much sure there was no short circuit

So what is the problem with this circuit do i need any resistance to control current to LED ?

Electronics become my hobby for some time now and playing with Arduino


Your diagram shows no solar charging. A COMPLETE diagram might be useful.

A datasheet for the LED would also help. Generally high-power LEDs should be driven by a constant current source NOT by a simple voltage output.


25 Watt LED!

The 18650 battery would get hot!

For 25W power draw, the step up converter will draw around 4 Amperes. Batteries won't last long.

25W at 2x3.7V makes for 3.4A - not taking losses in the boost converter into account, which may increase current draw to close to 4A.

A quick Google search gives typical internal resistance of such batteries of around 40 mΩ, so that'd make for some 600 mW dissipation. That's not enough to make such batteries very hot.

So maybe you have a cheap battery - higher internal resistance for more heat. Or you're over-driving the LEDs, it sounds at least like you're trying to do so, and you're drawing way more current than this. Especially a combination of the two will make for really hot batteries.

Thank a lot to all replies here are some explanations

First i'm a software professional and analog electronics is not my domain, but for the last few months gain good expertise with setting up Arduino based projects and circuit around it. My eventual aim is that build 18650 based battery banks and power some of my home load through solar. then the second Aim is to build up hands on expertise with IOT with some serious projects connected with web.

here are replies of above

jremington you are right it will may not last long but this is just experimental circuit. eventually i am planning to build a battery of 4s5p(or more) 18650 battery pack connected with solars and monitored by arduino with 2 or 3 LEDs for 8 hour backup. Later i will expand this system and run few LEDs and fans completely on solar and system will be intelligent to charge with solar or utility depending on availability and then try to commercialize the product. Here in Pakistan we have a lot of power shortages and such system has a lot of potentials

wvmarle yes these are new cheap china batteries (i think less than a $) as i explain above this is experimental learner project

Paul__B yes they may get hot i need some assistance how to make it work

slipstick that is a good indication i will check that suggestion. as you can see i simply connected LED with batteries


For home solar lighting, lead acid batteries are cheaper, easier to charge, more tolerant of abuse, less failure prone and much less dangerous than lithium batteries. All kinds of 12V appliances are available to work with them.

Finally, high power charge controllers and solar panels designed for lead acid batteries are cheaper and more readily available as well.

The OP says that only one of the 18650s was getting hot. I’m wondering if one of the cells became depleted, forcing the other cell to carry the full load, reverse charging the depleted one.

If that’s the case, I think maybe that cell is now a dangerous fire hazard.

do i need any resistance to control current to LED ?

Yes, you should have something to control the current. As slipstick says, with high-power LEDs (1W or more) a special constant-current (controlled-current) power supply/driver is normally used. (That's not an easy thing to build yourself.)

With regular little LEDs we use a series resistor and drop about half the voltage across the resistor and half across the LED. We use Ohm's Law to calculate the required resistance based on the (approximate) voltage across the resistor and the required current (which is the same through both series components). Then with current limited/controlled, the voltage across the LED "magically" falls into place.

A series resistor is inefficient because with the same voltage & current as the LED it has to dissipate the same power as the LED. That's not a problem with a regular little LED, but the same method with a high-power LED requires a high-power resistor and it wastes power so it's not often done. (You don't actually have to drop equal voltages across both components but it "works better" with more voltage across the resistor.)

LEDs (like all diodes) are non-linear and they are "current operated". The current spec is "exact". The voltage rating is approximate and varies from part-to-part and with temperature. Their resistance changes inversely and drastically as the voltage changes. A slight over-voltage can result in over-current and a fried LED or power supply. A slight under-voltage will disproportionally dim the LED because the resistance drops, multiplying the effect of the lower voltage.

LEDs are the opposite of how almost everything else works... With most "things" you apply a constant voltage and the current "falls into place". With LEDs you apply constant current and the voltage "falls into place". Regular diodes are similarly non-linear but the (forward) voltage drop is a fraction of a volt and there is normally something else controlling/limiting the current.

shahzad73: wvmarle yes these are new cheap china batteries (i think less than a $) . . .

My (limited) experience with "cheap chinese lipo batteries" from eBay is that the capacity is sometimes much less than advertised. I have some batteries that are labeled 2200 mAh that I've tested to be only about 450 mAh.

Such batteries probably aren't going to be happy with a 4 A load, they'll discharge under 10 minutes, and without protection circuitry to shut down on discharge you're likely to get a reverse charge current in one of them as suggested in post #7 which is potentially very bad.

For low current single cell experiments they're good enough, but I would be very careful about using them for high current series connected applications.

While the various warnings on how to control a 25W LED are apt, the OP's diagram uses an image of an actual luminaire, but also cites a random figure of "21.5V" (not to be confused with 1.21 jiggawatts!).

He needs to cite exactly what this beastie is before we can make useful suggestions. As to using series-parallel 18650 arrays, the safe charging arrangements become quite complex - though we must assume the car makers have this well in hand. :grinning:

MrMark: My (limited) experience with "cheap chinese lipo batteries" from eBay is that the capacity is sometimes much less than advertised.

That's not "sometimes". That's pretty much "always". I have had 18650 LiPo batteries in my hand that are marked 4,700 mAh - technically impossible - while weighing less than quality 2,200 mAh batteries (which came at 5-6 times the price).