Multi-cell Lights w/serial setup & parallel draw

Ok, I'm going to preface this question with a concession: I may own about 40 flashlights, but that certainly does not make me a flashlight tech expert; still a newb compared to a lot of you on this forum... here goes...

With all these multi-cell lights coming out now... one of the complaints is that they are getting too big to edc... the new TK40 is one awfully big holster light...

Wouldn't it be possible to set these lights up so that they are in series in the light but the draw is parallel... load some cells from the tail and some at the head with 2 separate battery compartments, the one near the head would require a slightly smaller interior and two layers of conductor to carry the current from the rear cells past and to the head...

This would give you a more slender light... and still allow the light to draw enough current to put out more lumens. Eg. a 4-cell cr123 light that puts out, say, 4-500 lumens with a similar led setup to the TK40. (a decently low setting required so the light still has some acceptable run-time)

Again, I'm a newb, so if this wouldn't work, tell me why...

or...

make a version of the TK-40 that uses 8 AAA instead

whatcha think?

I don't quite get what you're saying, and what the real advantage would be over existing lights. There already exists many MC-E/P7 lights which use 4xCR123 or 2x18650 and are slimmer than the TK40. EDC is not what any of these lights were made for.

If you want to actually EDC it, the M60-MC-E will run in a 6P size host for >500lm.

you are right... shouldn't have even said "edc"... but the TK40 is really even too big for a holster...

I guess all I'm trying to say is that lights like the P3D, M20Warrior and TK11 use more than one cell, but use them in series... if they were parallel draw you could get more lumen-power out of them without sacrificing much in size (maybe a slightly bigger reflector in some lights in case you need a bigger led)... or if you are adding more than 2 cells, you still wouldn't have to make something as large as the TK40 is going to be by going parallel in series...

Do we REALLY need more lumens? That's a matter of opinion I suppose...

I guess, as you say, there are several lights out there which already do this... remember, Marduke, I am a newb... can you recommend a light which uses a parallel setup that you really like?

(guess I'm really hoping fenix comes out with one as I like their acceptable quality/excellent price ration)

2benlightened said:

you are right... shouldn't have even said "edc"... but the TK40 is really even too big for a holster...

I guess all I'm trying to say is that lights like the P3D, M20Warrior and TK11 use more than one cell, but use them in series... if they were parallel draw you could get more lumen-power out of them without sacrificing much in size (maybe a slightly bigger reflector in some lights in case you need a bigger led)... or if you are adding more than 2 cells, you still wouldn't have to make something as large as the TK40 is going to be by going parallel...

Do we REALLY need more lumens? That's a matter of opinion I suppose...

I guess, as you say, there are several lights out there which already do this... remember, Marduke, I am a newb... can you recommend a light which uses a parallel setup that you really like?

(guess I'm really hoping fenix comes out with one as I like their acceptable quality/excellent price ration)

Click to expand...

Keep in mind that the TK40 is only Mag 2D size. It's big, but not HUGE.

But how do you figure a parallel arrangement magically gets you more lumens?

I suppose the law enforcement spent years with 4-D cell Maglights on their belts... you are right there...

I am "assuming" that the TK-40 (I'll keep this as an example) has the battery rack it does so that the cells can be used in parallel... that using them all in series wouldn't increase the current, but only the voltage... and I guess also assuming that it uses circuitry like the P3D where the increase in voltage doesn't increase the output of the light, only the run-time... therefore my assumption that to use circuitry like the P3D (does the M20 Warrior not use the same type?) you need a parallel setup to increase output... and maybe a different/ bigger/ or multi led setup like the quad one in the TK40...feel free to correct any of these assumptions...

Ha ha... I realize I may be making a fool of myself asking these silly questions and showing my ignorance... but gotta learn somehow...

Personally, I would buy the TK40 and the size wouldn't put me off... I'd sacrifice the size for the output and uncompromised runtime... but a few here have commented that it is a bit big... thus my question of using the same setup, but making the light longer and more slender (for those that like the idea of using their flashlight as a weapon, it would be longer and more slender... think club LOL)

Marduke, is my reasoning false because other lights use serial setup but different circuitry to increase output?

Serial/parallel does not really matter, it is easy to do some electronic adjustment of the voltage/current. What matters is how much power you can draw from the cells.

One way to make small power full lights is to use IMR cells, a single 18650 (About same size as 2xCR123) can easy support a MC-E or P7 at full power. A IMR16340 (Same size as a single CR123) can also do it, but you are at the limit of its abilities and will have a very short run time (maybe 7 minutes).
But there is another problem when doing that: heat any led running at 10W needs to get rid of lots of heat.

Manufactures do usual not base lights on IMR cells, you will have to build it yourself with parts from different manufactures. One possibility is Dereelight CL1H with MC-E pill, running from two 16340 or IMR16340.

here's another assumption to correct... the P3D is efficient partly because of the circuitry it uses to limit the voltage coming in... = longer runtime... my reasoning for using parallel may be faulty... but I'm wondering (aloud) about using parallel instead of changing this circuitry... thinking multi-modes.

appreciate all the feed-back, like I said, I have much to learn... admitting that may allow me to learn much... (I hope)

2benlightened said:

I suppose the law enforcement spent years with 4-D cell Maglights on their belts... you are right there...

I am "assuming" that the TK-40 (I'll keep this as an example) has the battery rack it does so that the cells can be used in parallel... that using them all in series wouldn't increase the current, but only the voltage... and I guess also assuming that it uses circuitry like the P3D where the increase in voltage doesn't increase the output of the light, only the run-time... therefore my assumption that to use circuitry like the P3D (does the M20 Warrior not use the same type?) you need a parallel setup to increase output... and maybe a different/ bigger/ or multi led setup like the quad one in the TK40...feel free to correct any of these assumptions...

Ha ha... I realize I may be making a fool of myself asking these silly questions and showing my ignorance... but gotta learn somehow...

Personally, I would buy the TK40 and the size wouldn't put me off... I'd sacrifice the size for the output and uncompromised runtime... but a few here have commented that it is a bit big... thus my question of using the same setup, but making the light longer and more slender (for those that like the idea of using their flashlight as a weapon, it would be longer and more slender... think club LOL)

Marduke, is my reasoning false because other lights use serial setup but different circuitry to increase output?

Click to expand...

The TK40 is 4S2P, not entirely parallel. It has two sets of 4 in parallel. This is done to get the voltage above the Vf of the LED so a more efficient buck circuit can be used. They added the extra 4 cells so the current can be increased without significant voltage droop.

Either series or parallel, the total power available remains the same. But if you are going to use multiple cells, it is more efficient to make sure the voltage is higher than the Vf instead of lower, so the light just has to buck the voltage instead of boost it.

ah, now I'm getting down to it... so lets say we want, for example, a 6AA version with a buck circuit... if we went 3S2P... would that not give us a higher current (just as high as with 4S2P) but still keep the voltage higher than the Vf?... but still install 3 cells into the tail and 3 into the head as I described so you have a long slender light...

yes, that would be VERY long... so lets think cr123A's instead. should be able to make a light about 11-12 inches long... (she says that's just about right LOL)... sometimes you have to think "inside" the box...

ha ha... if I ask enough silly questions, Marduke may pass DM51 in posts

2benlightened said:

ah, now I'm getting down to it... so lets say we want, for example, a 6AA version with a buck circuit... if we went 3S2P... would that not give us a higher current (just as high as with 4S2P) but still keep the voltage higher than the Vf?... but still install 3 cells into the tail and 3 into the head as I described so you have a long slender light...

yes, that would be VERY long... so lets think cr123A's instead. should be able to make a light about 11-12 inches long... (she says that's just about right LOL)... sometimes you have to think "inside" the box...

Click to expand...

Well, under high current draw, 3 cells would droop in voltage below the Vf of the LED. When exhausted, the pack would be providing only 3v, far too low.

By using 4S2P, a dead pack is still 4v, which is >Vf~3.5v, so you maintain near constant performance through the life of the discharge.

3S2P would gain you a negligible diameter decrease, which is why it's not really worth it.

If you want a thinner light, get an MC-E 2x18650 light. There is a range of them, both cheap and quality.

that math go for AA's... or for CR123's? The 6 cell setup could still work for the 3v cr123's wouldn't it?

anyway, that narrows down the reason they went with 8 AA...

I'm gonna check out that MC-E... thanks for the advice

6 cells in a row wouldn't make for a light much fatter than an L2D; quite a huge change in diameter.... which was my other part of the question... how hard would it be to design a tube that carried two separate currents... assuming you DID want to do such a thing... or are there lights already existing which do that...

2benlightened said:

ah, now I'm getting down to it... so lets say we want, for example, a 6AA version with a buck circuit... if we went 3S2P... would that not give us a higher current (just as high as with 4S2P)

Click to expand...

I'm not sure what you really are asking here, so let me give you some numbers. It's easier to see the implications that way. If you use AA rechargeables (if you use AA, you should, and preferably sanyo eneloop or equivalent), each cell is 1.2 v.

Putting them in series raises the voltage:
4S = 4*1.2v = 4.8v.
3S = 3*1.2v = 3.6v.

You can parallell as many 3S or 4S as you want, they're still at the same voltage. 3S2P = 3.6v, 3S3P = still 3.6v. 4S2P = 4.8v, 4S3P = still 4.8v.

However, if the current draw is constant, the runtime will be doubled when going from 4S to 4S2P, or trippled when going from 4S to 4S3P. Depending on battery characteristics, you may see some other improvements also. (Each batteries delivers half or one third of the current, and this may be more optimum for some batteries.).


There are some issues, however, with parallell battery circuits. If the cells aren't charged to the same capacity, or if one cell is damaged (not having full capacity), you will have problems because one cell will be empty before all the others. Shorter run time at best, a "reverse charged" (i.e. dead) cell at worst. I guess with a 4S2P and using good quality batteries this risk is acceptable.


As to your question, the 3S is NOT really good for this application. With fresh alikalines, 3*1.5V = 4.5, which is clearly above the vf of the diode. But the AA alikalines will not be at 1.5V very long, particular under heavy load. The voltage drop quickly to 1.2v or less even though you may only have used something like 20% (inaccurate guestimate) of the total capacity. The battery configuration should be 4S to stay above vf as the voltage drops untill batteries are totally empty.

3*AA are used in direct drive light, where a resistor makes sure the initial high voltage of fresh batteries (1.5v for alikaline, 1.3-1.4v for NiMH) does not fry the diode. Direct drive lights will start out bright on fresh batteries and be gradually dimmer as the battery is drained. This is good for some applications. A well regulated light will have almost constant light output, and then die to leave you in the dark. (Some will have a moon mode for almost empty batteries). A direct drive light can be bright when fresh and have really long "moon mode" run times, squeezing the last drop out of the battery.

that was very informative, jankj, thanks... there were tidbits of info that I didn't know... name of the game... great explanation

but what happens if 3v cells were used instead? (123)

2benlightened said:

6 cells in a row wouldn't make for a light much fatter than an L2D; quite a huge change in diameter.... which was my other part of the question... how hard would it be to design a tube that carried two separate currents... assuming you DID want to do such a thing... or are there lights already existing which do that...

Click to expand...

Why would you use 6 CR123 cells in a single row? I've seen a few 3x18650 lights (same length), but 2x18650 is far more common. Again, there is no real reason to use a parallel arrangement in that setup.

cause from my understanding the Vf is related to the LED and by using 3V cells 3S2P, or even 2S2P would stay above the Vf for a large percentage of the battery life... (you're right though, couldn't mix and match cells that's for sure)

ok, here's my reasoning... I'm sure it's faulty which is why I'm putting it out there for critique...

I'm assuming that the "buck" circuitry is a good setup...

I'm assuming that with a buck circuit the brightness of the light won't necessarily go above a certain level just by raising the input voltage

eg; the 3 and 4 cell bodies on my P3D raise the voltage but only affect runtime, not brightness (and also use up more of each cell!)

I'm assuming therefore to get more brightness with this setup, you need to raise the current, not the voltage.

I'm assuming that drawing much more current than the P3D example I'm using does through a single row of cells is bad, or inefficient, or can't be done... or something; whereas adding together the currents from 2 different rows of cells, (or in this case 2 rows in line with each other) would avoid that problem.

I'm assuming that someone wants the body on their light to be long and slender, not fat... me for one. (actually, in true CPF spirit... one of each!)

I'm assuming that if there "is no real reason to use a parallel arrangement in that setup" that I am missing a tidbit of information in the above assumptions... could you point it out? Is it because there is another type of circuitry that would be just as efficient but allows more brightness without too high a current drain... or is the secret in the type of cell... do the 18650's operate more efficiently at higher current drains?

thanks for your patience with me

Four points:

1) The lower the voltage to a driver, the higher the current required to achieve the desired power output at the appropriate output voltage. Higher current means higher losses to resistance.

2) In light designs where minimizing body diameter is desirable, designing a holder or other means of contact to achieve parallel battery arrangement is counterproductive, when series battery arrangements are very simple and require little to no design effort.

3) Variations in internal battery construction may result in some batteries experiencing higher or lower current draw than others in a parallel arrangement, whereas in a series arrangement they tend to experience the same current draw.

4) I think buck converters tend to be more efficient than boost converters... not sure on this one though.

Of course, as everybody has said, this battery stuff matters a whole lot more in direct-drive lights. I have a Cree MC-E dropin prototype "room sweeper" for Surefire turboheads which runs direct-drive (the production version will have a driver! ) on two series Li-ion cells right now... the emitter is wired 2P2S. The batteries see a drain of around 1.2A at the tailcap, so around 600mA per core. However, if I had chosen to wire the emitter 4P, I would have to use a single li-ion and that cell would have to survive a comparatively-withering 2.4A draw.

Also, looking back at your more recent post, the reason your P3D gets longer runtime with more cells is that it's pulling less current because of the higher voltage to the regulator. If you were to rig up a 2P2S body for your P3D, you'd probably see almost the same change in runtime as 4S, except that it's likely the buck converter is less efficient outside its intended voltage.

If I follow your thoughts correctly, you think that the only way of reducing the current pull from the batteries is by wiring in parallel. That is not the case.

The LED is set (by the circuit) to consume a set amount of power in watts = volts x amps. The LED voltage is fixed ~3.5v, so only the current is changed (you can control it with voltage instead of current, but that's a different beast).

For a second, let's assume input = output with no circuit losses.

So, the LED consumes the same amount of watts as what your batteries provide. As you raise your battery voltage, you degrease the required current from the batteries, while maintaining the same power to the LED. The reverse is also true. By lowering the voltage, you must increase the supplied current.

So by adding more cells, you are by either method increasing your runtime proportionally compared to how many cells you are adding. Now, depending on the light and circuit used, adding them in certain configurations are sometimes more advantageous that others.

So when you have a simple linear battery tube flashlight, it is an unnecessary complication of wiring cells up to be both series and parallel in most circumstances.