21700 battery advice needed

You do not have much choice, Samsung 40T3 is your best bet which you can get at $10 each from 18650batterystore. Molicel P42A is cheaper but out of stock everywhere. Theoretically Samsung 50S can also deliver 20A CDR but there is no way to purchase it. And forget about protected cells if you need 20A CDR. Seems like you are using it with your light, why do you really need 20A CDR? 3x cells means that your light need 60A CDR to drive, how can it keep temperature down for the board and LED with continuous 60A current?

s., also saw you mentioned 18000 lumen, but I assume you light cannot deliver sustained 18000 lumen, right? Also I would rather change the spring than adding a magnetic block

In fact I just pulled the trigger for the molicel P42A- yes they are out of stock but not everywhere. I also saw new acebeam 5000mah batteries available in a month, $20 each and 20A discharge current - protected and already 75 mm long- was tempted to go for it, when I noticed it said 20A PEAK discharge current so no telling what continuous.

I haven’t fully tested how long the light can sustain 18000 lumens but according to manual, it was at least for a minute or minute and a half if I remember correctly. Visually it sustains about the same level , possibly the whole minute but I really should test it in my light box.

Changing the spring may be a better idea but somehow it feels more invasive and less simple to me, but it may come to that. Hopefully I can get the molicel plus side stick out a bit so I don’t have to also solder a little blob on that end too. Thanks!

Cool, P42A is a great battery but its initial voltage sag is obviously worse than 40T which could affect the output. If you are really pursuing the brightness 30T will work the best followed by 40T/VTC6A and then P42A.

60A current drawing with 3x21700s is very significant and I would put safety first. In this case changing the spring might be a better (also a permanent) option. Just my opinion.

I didn’t know about the voltage sag. I guess I will see how it works. The spring may be a good option but the springs installed should already be designed to withhold the current that is required, so I am not too worried… maybe I should be….lol

You can easily see the difference using HKJ’s battery comparator. And no, my safety concern is for the magnetic block you are going to use, not the spring.

Molicell P42a.
Cheaper than a 40T
Out performs a 40T overall
Slightly higher capacity than 40T.

To get to 75mm, you’re going to HAVE to either use a neodymium magnet OR a solder blob.

The magnet will have higher resistance than a solder blob (Not sure how much more as i haven’t tested it), so you have to judge if you’ll be willing to live with that trade off.

There are hardly any good button tops that can provide high current cheaper than $10. Button tops in general add a fair amount of resistance because they are usually spot welded and connected via a tiny nickel strip.

Personally I would go with a P42a with a solder blob as I believe that it will be worth the trade off. I am aware that soldering directly to the top is dangerous and “could” shorten the life of the cell, but I think it will out perform one equipped with a magnet throughout its cycle life.

Have same length problem with unprotected 26650s in Sofirn SP70 and have used magnet between Shockli cells since purchase a couple years ago and this simple solution works quite well and cells seem fine and consistent in discharge.

That’s perfectly fine as the current drain is much much lower. (2s1p I suspect it to be less than 6-7 amps)
over 20amps though and you’ll see the resistance add up

I used to solder blobs on top, but sometimes it was messy and sometimes the blobs didn’t want to stick to the metal or fall off over time. I might either solder the blob 5 mm high which will already take care of the length and then I don’t need a magnet on the other side, or glue a small ( button too size) magnet on top and solder it on the bottom edge all around. We shall see which works better

What would be the safety concern with the magnetic block?

Been pondering this and I don't think I have great answers, but I do think this is worthy of doing it the right way and thinking it through carefully because we're not talking about a trivial amount of current. Are you modding/creating in a host, or altering an existing light, or just trying to save pennies on the batteries it was designed for?

Ultimately, it makes so much more sense to change the host parts in a way that they will take ordinary-length cells, rather than having to change each cell each time and attempt to get their lengths consistent.

If the tail spring board is easy to remove, totally try a longer spring there that can handle the current and retain all or most of its height while doing so over time. If that can/should be done on both ends, assuming the driver board is compatible and accessible, then do that...or perhaps a button/slug can replace a spring.

I really really do not recommend soldering cells. Spot welding is the way to go, when done correctly, but soldering adds a lot of risk especially on the negative end, and you need to add a considerable amount here, not just a little dab. If soldering is the way you want, then do it on the positive end/cap and be careful about it so you don't short, and quick about it so you don't damage seals or cause it to vent. Overheating the negative end can spell bad news for the cell either in performance or safety and that's true of these high drain cells that can take a lickin' as well as it is for ordinary lower drain cells that are somewhat more sensitive. And of course if you were to use protected cells then you wouldn't be able to solder on the negative end anyway. I've soldered buttons and blobs and tabs on dozens of NiMH and NiCd batteries over the years, but I won't solder lithium...did once, but all things considered it's not worth it to me unless it's just a blob on the top to make a small button but I haven't needed that in any of my lights in a very long time. I just wonder about all of the heat and current in this application. Plus soldering to nickel can be unreliable even with scuffing. We've seen springs literally melt from high current that was done wrong, and cells that overheated to a dangerous level through current or accidental shorts, so I wouldn't put a lot of faith in solder here even though their melting points are above what we'd expect to see.

I searched the forum for a bit but couldn't find posts that I think I'm remembering from the past about magnets and resistance/current limit. It may have been from Dale (DB Customs) in some thread, really can't remember. But I was thinking that there was a medium-level of current that was a practical limit for magnets but of course most who use them are using tiny ones as buttons/spacers, not thicker ones to fill up space and cover a larger cell. I would try some kind of slug to use as a spacer on the tail end if you don't want to modify the light itself. If nothing else, a copper pipe cap could be trimmed and filled with solder - a 1/2" is around 5/8" o.d. and I think a 3/4" will be too wide without a fair amount of sanding. I looked on Amazon to see if there were any convenient discs/slugs and there are lots of jewelry stamping blanks but most are way too thin - thickest I saw was 1/8" and they were pretty expensive for a smaller quantity of 10 or so. Washers and bushings present the problem of the spring going in the hole, so you don't get the effect you're after. If you have a very well stocked hardware store there are copper bolts that exist and you could cut the head off of one and grind it to suit....a little pricey but you could do the same with an electrical split grounding nut thingy. Heck, really for this purpose I think aluminum would be fine and if you can score a short length of 3/4" bar you could just trim pucks from it.

So the cost of parts and consumables, repeatedly for a trio of cells and more cells in the future, plus risks that are not zero...sounds cheaper and better to just spend a couple extra bucks on the ridiculously priced cells that fit the design of the light...and don't buy lights like that again. :)

Please don’t use a magnet on the positive end that comes into contact with the spring, if the magnet slides off to the side it could cause a short circuit

The positive end doesn’t have springs , it’s a ring that the top of the battery comes in contact with. This is still a valid concern though, because if I solder a blob on the positive end , it could detach over time and cause a short touching the other, bigger diameter ring there that seems to intend to connect to the negative end of battery

These are all good thoughts. I’m not trying to mod the light. I just don’t want to spend $75 on a spare pack of batteries and also it pisses me off that the manufacturer makes their money trying to get the customer shell out $75 on spare parts instead of making the battery size compatible with all the others , or springs that can stretch longer .

I’m not crazy about welding magnets, the spacers are a better idea, it’s just that magnets seem to be an easier and more available solution . If I can find/ make a spacer, that would be a much better option. Cutting off a bolt head seems like a simple enough idea, or some kind of aluminum puck cut to size. This being said, I still would need to solder a blob on the positive end, because it needs to stick out -there are so springs on the positive end, it’s a ring touching the end of battery.
And yeah I’d never buy a light like that if it wasn’t the only light of having this power and throw and the beam profile that I like. I don’t like flood lights, I want a wife powerful center that can throw. No other light I know does that.

One correction I just noticed it’s written on the flashlight they are using 35A batteries , so 20A probably wouldn’t work anyway

35A CDR is the upper limit for 40T and P42a (none of them get 35A CDR rating from Mooch) and I think only 30T can safely deliver 35A CDR (not too hot). You mentioned that the original batteries are protected button tops? To me $20-25 price tag is very reasonable if they can safely deliver 35A CDR. I guess they used technology from the future.

I couldn’t find the discharge rate specs for the actual battery they use, they just don’t disclose it I guess. But the light itself says it uses 35A batteries. Whether that’s peak or continuous, or a recommendation what to use or a lie, I don’t know. I think I will take my chances with the molicel. Since they are connected in parallel, I don’t imagine the light requires 105A discharge rate to produce 18000 lumens, I hope not….for comparison I use 3- 20A discharge 18650 batteries in my M18 Emisar and get the 14,000 lumens OTF with no apparent issues so far.

You may get a YR1030/1035 to measure the internal resistance of these batteries at AC 1kHz. The impedance of 30T should be between 6 and 7mOhm while 40T and P42A should be between 9 and 10mOhm. The impedance at AC 1kHz will definitely reveal the truth of the battery.

I be been wanting to invest in a decent tester as mine is falling apart . Will this YR1030 have pretty much all the functions that a regular tester does? I just looked at the picture of it, it doesn’t seem to have too many buttons/ knobs etc. lol. Thanks )

I would suggest the latest version YR1035. It uses a 18650 inside while YR1030 uses a lii-polymer battery. The battery in YR1035 is a crap but you can easily change it.

I don’t think it is a comprehensive battery tester but it can measure impedance at AC 1kHz with decent accuracy. You may also check HKJ’s testing on YR1030 and YR1035.