Thermal modding for Convoy s2+

@Henk4U2 Do you think making threads to the body so that the sleeve can be screwed would be better?Thermal compound between the threads is feasible too

@badtziscool That’s why I thought about hammering the sleeve in,so that maximum contact can be achieved.In my design the sleeve gets over the first “channel” (?) so I guess i should make it shorter because there will definitely be air trapped there.Also I don’t understand if the air/hand is a rhetorical question.I guess the air?(English is not my mother language)

@Cythras I don’t know if mine is a Fet driver,it’s completely stock (i guess it is?). My 7153x6 gets fairly hot,not scorching that’s true,but if I can make it better with an inexpensive solution,I can push the led harder,right?

@whatsthepoint I know man,I’d love a copper head for my convoy.The all copper s2+ was fairly expensive,which beats the purpose of the s2.If I end up making the sleeve,I know I’m looking at minimum gains,but I wanna see if it makes a difference at all

Have you seen this thread: How hot does an S2+ get at 2.75A? Comparing an anodised Convoy S2+ with a fluorescent yellow powdercoated Convoy S2+

I think the first thing you should get is a IR thermometer and measure what your current model is producing wrt to heat. Then it would be easy to create a “mock-up” out of a section of copper pipe - something to slip fit over the head w/thermal paste and then measure and see what the mock-up produces wrt to heat. That should get you on the path (relatively cheaply) to determining if the full project is expected to create a significant difference.

Thanks!I hadn’t seen that thread,all I want now is a fluorescent s2+ :heart_eyes:
Djozz is keeping it real,that test answers some questions I had in mind.The IR thermometer idea,is great,it would indeed be easier to find out the hottest area around the head and start from there.My only concern still is,even if I get the sleeve as tight as I can and manage to get some thermal paste in there,will the heat tranfer better than letting the body be exposed to air?
I might have to start ordering some copper bushings to find out

Make the fit as neat as possible and jb weld it.

A FET driver allows the emitter to pull as many amps as it wants (that the battery can provide), and is commonly known as Direct Drive (basically battery directly wired to an emitter). Stock S2+ use regulated output via AMC 7135 chips, each providing 350mA of output. 6 in yours means a maximum current of 2.1A, 8 means maximum of 2.8A. A FET with a low forward voltage LED (SST-40, i think would be a good example) and a high discharge cell will probably pull up to 7-8A or so assuming there isn’t high resistance in the circuit somewhere (springs, mostly). You do not need to mod the body of a S2+ to achieve this.

However, high current definitely equals higher thermal output as well. If you add a copper slab to your light like you are supposing, you will increase the thermal mass of your light, AKA it will take a little bit longer to heat up. However since there is no good way to dissipate that heat (airflow over fins, fan cooled heatsink, etc) you will still be thermally limited.

TL:DR If you build what you propose, your light will heat up slightly slower but will not significantly lower the maximum temperature it runs at before it burns you (or steps down if you have a driver with thermal regulation). You can push an LED as hard as you’d like (though decent cell/spring bypasses will be required, and maximum draw depends on your LED), a modified body like yours will only slow down the initial heating up of the light a little bit (assuming it’s done well and doesn’t end up having the inverse effect of insulating the light).

@CRX JB Weld is epoxy,right?Is’nt epoxy a bad heat conductor?

@Cythras I have a BLF A6 A17DD-L FET+1,but I’ve read that the s2+ gets really hot with that driver,that’s why I haven’t tried it yet.I thought about swapping the led because the XM-L2 is older technology with something more efficient like a Luxeon V2 (I ’ve started reading about vF,not sure I get it completely).
So I ’d be delaying the initial throttling but inevitably,it would start burning,even if I do it right.I thought of doing this mod,along with a thermal enhancement kit and a copper pill so I can run about 700-900 lumen without burning the LED or risking the battery life.Do you think it’s worth it,or should I look for something like a Convoy M2/Astrolux s41?
Also thank you for taking the time to explain some things,I don’t understand everything I ’ve read from other posts

Yeah but it's steel epoxy! :D

With a neat fit the layer would be minimal.

Obviously a press fit would be best but that could be tricky.

You would be better removing the ano too if you want the best heat transference.

I did something similar with an aluminium jacket on brass.

This wasn't fixed, just a slip fit and took the heat well enough but the problem will always be saturation and removing that heat..

Aluminium jacket.

Firstly, I assume you mean battery life in terms of degradation due to heat, not runtimes which will invariably get shorter as you increase output.

The BLF-A6 Driver was designed for the A6, which is a very similar light to the S2. It will get too hot to handle in the higher modes, but has thermal regulation you can set and will automatically step down at your desired temperature. A 8 x 7135 2.8A driver (like many stock S2 come with) is about the limit of what you can sustain wiithout thermal stepdown in the S2+. So if you use the 17DD driver, any mode that draws more than 3A is probably going to eventually throttle. 2.8A with a modern high output LED (XP-L HD/HI, SST-40) will get you over 1k LED lumens, possibly close to 1200 with an SST-40. You are unlikely to hurt either the LED or the cell running at that (assuming decent thermal contact between led>mcpcb>pill), though I wouldn’t leave it unattended on max output.

You are correct in that the thermal enhancements would only delay the throttling.

So, i guess there are a couple options for you. The simplest would be to try running a S2+ in the highest maximum factory configuration- 2.8A driver, LED of choice (though Luminus SST-40 is the best in terms of sheer output for a single die 3v LED, you do sacrifice CRI/Tint choices.) You should be able to hit well over 1k lumens without having a hotrod light.

You could also install a 17DD and give it much higher possible output, as well as thermal regulation so it will never overheat and damage itself. If you don’t need turbo mode, you can always run it on a lower mode to match the performance of your current configuration. Maximum performance with a FET requires special attention to LED and battery choice though, some LEDs will burn with high drain cells on direct drive regardless of heatsinking. Also requires springs to be bypassed with wire to reduce resistance.

Another option would be to change hosts entirely- It’s hard to recommend the S41 nowadays, the emitter choices are dated and stock factory output is rather low compared to what is possible now. The Jaxman E2L is a compelling choice for your use case, as it takes the form factor of an S2+ (though slightly shorter), and adds heatsinking fins to dissipate heat, as well has having an integrated shelf transfer heat instead of relying on the threads to transfer heat from a pill to the outside body of the light. Additionally, it uses 3 LEDs on a single board, which increases efficiency at reasonable levels (more light vs heat output at the same current vs a single LED) as well as much higher maximum output should you desire to run it at higher currents. The one thing to consider would be the optics layout, the E2L uses Carclo total internal reflection (TIR) optics instead of a traditional reflector in order to focus all 3 emitters. This will generally lead to a floodier beam than a single emitter would produce. The E2L is available from Jaxman’s aliexpress shop as either a host (Add 3x LEDs on a 20mm MCPCB and a 17mm driver) or as a pre-built light with about 900 lumens of sustainable actual output through the front. The downside is that the stock driver is quite simple, just a low-medium-high with memory. No worse than the old stock convoy 3/5 mode drivers, but there are more flexible options now with Bistro(improved A6 firmware), Biscotti, etc. firmwares available.

Isit worth it? That’s up to you. Assuming you have a 17DD right now and nothing else, the cheapest route is obviously to use what you have. The A6 firmware is nice in that you can roughly match the performance you have now (with very minor runtime loss due to a FET being less efficient than 7135s), and only step up if you need short bursts of higher output.

If you don’t mind buying another light, I think you can get a new Convoy S2+ with 2.8A drivers for around $12 USD and you can try it out that way. Make sure you get something with their new firmware though, not the old 3 mode one.

If you like the idea of a triple LED setup, the E2L is basically a premium S2+ designed specifically for 3 LED setups, and is really nice. As a bonus, it legos well with Convoy battery tubes so you could make it even shorter than stock with a 18350 tube and cell.

Otherwise, there are many other options. The M1 is a larger light, and a larger light will have more heatsinking mass, as well as larger dissipation area enabling you to run lights harder and longer without thermal limits. It really depends on what you want, there really are no wrong choices here. If you end up with something that runs hotter than you’d like, just step it back a mode and use it on a lower setting.

@CRX Wow man!!!These lights are gorgeous!!!Are those knife handle pins on the bottom side?They look great!!!
I’ll try to find Jb Weld here in Greece!I’m a woodworker and I use a little trick sometimes.When you glue two or more pieces,before the glue sets the pieces get really slippery,so I make the pieces as tight as I can and I use the glue as a lubricant.I wonder if JB Weld is similar.Does it have a long open time?(Before the adhesive starts to set)

@Cythras Ok,I’m trying to take in all of the information,I’ve read your comment 4 times till now :open_mouth:

So what I get is that I’m ok with the stock shell,maybe I’ll try a few of the existing mods (copper pill,enhancement kit) and I should maybe start reading about drivers with firmware,so I can find a setting that can get me to a level I am pleased,but without the flashlight catching on fire.I like the guppydrv and the bistro so far,from what I’ ve read,but I’m afraid that I’ll have to memorize everything.All I need is a moonlight,a low and a 50-60% of the max output (something between my med and high).I get that most of the stock firmware go for a 30% because I’ve read that from there to a 50% it’s not a discernible difference,is this true?
I will definitely buy another flashlight, I was looking at the Jaxman m8,the M1,the lumintop edc18 and the Yootoo SD2.I gave my father my old Sofirn sf36 and now I’m looking for a replacement,while having the s2+ for an edc.Maybe I’ll try swapping the XML2 to a sst-40 and later I will fit a driver with more options for the s2+ and look for a larger light in replacement of my old Sofirn(that was my “big” light)
Do you have a recommendation for an LED?I like the white XPL2 V6 that my sofirn had,maybe a little less harsh tint-wise,and I’ve come to hate my T6-4C that my Convoy has.Do you have anything closer to the first one and and less like my second one?

Thank you again for your time,I’ll go back and re-read your comment!

Those are 2mm magnets in the tail.
There is 4min jb weld kwik and 6hr standard stuff, prob better using the 6hr setting formula.
You can add some fine graphite powder or diamond dust too :smiley:

Nice job making a magnetic tail part of the design!!!

After reading a little bit about JB weld and mixing parts of alumina or similar heat conductors this seems quite feasible!I might go for over 2mm thick though,make it worth the trouble.I also found another answer searching for JB Weld mods.It appears some mixtures of epoxy and metal DO conduct heat faster than air.So, I might scrap my shell just for the sake of “science” :smiley:

Don’t bring science into this! :laughing:

The A6 driver does not have thermal regulation; there’s no temperature sensor on the attiny13

The 17DD does if you get Crescendo or Bistro but not A6

Ah shoot, that’s good to know. I just knew that Bistro was basically an A6 driver+, but thought that the thermal regulation was on both.

The Attiny25 has a built in temperature sensor that Bistro uses. Attiny25 also has more “room” for programming firmware. Attiny13 does not have those things.
But Attiny13 is more stable, is not so fiddly when you want to add a lighted tailcap or so.

@CRX I’m sciencing and you can do nothing to stop me!!! :smiley:

@ treellama @Henk4U2 The board that I have is this https://www.aliexpress.com/item/32831478212.html?spm=a2g0s.9042311.0.0.42174c4dhoNOZJ
(I don’t know if it’s allowed to post a link from a page so if not,please tell me if I need to remove it)
It’s a A6 driver from what I read,but it’s also a 17DD driver.Does it have a temperature sensor or not?Also for future reference,is it better if a driver has an onboard sensor,or is a NTC type like the LB-D4 sensor better?I’ve read somewhere that the Astrolux MF-01 mini has a problem with the onboard sensor,so it needs a cooling spacer in order to get better readings.What is your driver choice for a single cell light,with moonlight,40-60% med and accurate temp control?(If allowed please leave a link so I can read about it and order it)
Last question and I won’t bug you any further,the A6 driver is a 4A-6A current driver.Does that mean that this is the maximum it can deliver,or that it can’t go lower?

Thank you for your time guys,I’m learning a lot!!!

That one does not have a temperature sensor or thermal regulation. But, even on the ones that do that I have used (Bistro, Crescendo) it isn’t great. It will keep it from overheating, but it will not regulate it perfectly to the temperature you pick. So it will ramp up and down periodically instead of finding the sweet spot and staying there.

The LD B4 might do better, I haven’t used one. H17F might do better as well, haven’t used that either.

The A6 can deliver more than 4-6A with a high output battery / bypassed spring, and it can go much lower. There is a 350 mA constant current driver on there for the low modes, which it can PWM to get a nice moonlight. It’s a great choice if you have an LED that can be direct-driven. Just have to turn it off manually if it gets too hot.

What’s in a name. The A6 driver is a 17DD driver. Diameter=17mm DD=direct drive=FET.
So this ought to be okay. The enlargement is a bit blurry, so I can’t tell if it has a bleeder.
But if you want to add a lighted tailcap later on, a bleeder is quite simple added to the driver.

There is a potential issue that treellama has pointed out.
The Bistro driver has a temperature sensor, but it responds to the driver temperature.
That is better than nothing, but monitoring the Led-PCB for temperature would be better.
Be aware, the driver you asked about goes by many names. But Bistro is not one of them.

@treellama I see,so it’s more of a guess of what the temperature is on the actual LED,depending on what the driver and the body temperature is.I might have to check the LB-D4 because it’s easier shipping it.The H17F is only available through the US from what I’ve checked.Thank you for the information!

@Henk4U2 I’m getting to know the different names and it’s kind of hard,that’s why I got confused.So I should get one of those drivers and set the limit according to what’s comfortable within the range.That’s much clearer to me,thank you!!