SupFire L5 modding adventures

Recently bought a SupFire L5 from Richard (RMM) at MtnE ( SupFire L5) and nicely impressed with the quality/value. The finish is a flat gray and feels a lot richer than the price range of this light. For me, it's almost perfect for this form factor, just really wish it was a true SS bezel, and there was more beef to the heat sinking internals. After a few days, noticed the button (e-switch) was acting up, not working consistently, and mentioned it to Richard but wanted to find the cause first before doing anything bout it - was it the switch itself or maybe the driver getting flaky. So did some poking around on the driver, and it did seem to be the switch itself - that's sort of bad news, not knowing where I could get a replacement PCB mounted switch to fit, and if I did find one, how much, and really didn't want to bother Richard with it.

So, I only had one type of spare e-switch:

100PCS Tactile Push Button Tact Switch 4mm x 4mm x H1.5mm 4-pin DIP #1292

bought for $3.75 w/free shipping off of eBay, seller: electronic_diy_honest, this is what it looks like:

It's tiny, but works pretty well and I got 100 of them...

So, it looked like if I sand down the mounted switch, I could possibly mount this thing vertically on the switch, still have it positioned properly for the rubber switch boot, and can wire it up to the driver - that's what I did. Here's the button and plastic mounts sanded down:

Perfectly flush:

Here it is mounted and wired up. It's bit blurry, but the 2 tabs on the bottom are soldered to ground, I attempted to solder tabs on the back but don't think those held, so I added some super glue from the top to get between the old switch and new one - seems to hold pretty good. I wired a top tab to the point of contact on the old switch that goes to MCU pin #2 (that's the e-switch input). Funny because we use pin #2 for the ATTiny13A's... Maybe that unmarked MCU is the same?

While I was at it, though I'd do a pseudo "Mod Option #1" (i.e. Stage 1) mod, using a warm XM-L2 (T4 5B1) mounted on a MaxToch 26mm, and tweak up the amps a little. Well, actually I was forced into upgrading the LED because I attempted to route around the resistors, like we do on the F13 clones, but with the LED +, not LED -, and it went really bright for a second or so, then poof - popped a LED fine wire, just one of two. I know that because the LED would still light up but very dimly, and I could see in the silicone a bubble of sorts that appeared around one of the gold wires.

Notice below I removed the pill shelf from the L5. It easily pushed out from the driver cavity. Again, unintentional because I was trying to push off the MCPCB that was epoxied down through a wire hole, and instead of pushing up the MCPCB, the entire pill shelf pushed up Smile. Well planned mod for sure...

The shelf is only about 1.25 mm thick, so the MCPCB's are thicker Frown. So with the shelf out, it's a lot easier to work on the LED MCPCB mount. So, sanded it down smooth, drilled and tapped holes for screws, etc...

Added a single R500 resistor. The resistor math said it should do about 3.25A. I wanted to keep the mod relatively mild, and not planning on adding any heat sinking at this time.

here it is re-assembled - pushed it down tight, added some AS5 (a little) at the small ledge the shelf sits on - maybe that helps. I used GC Extreme between the polished MCPCB and pill top shelf:

here's the top view. Nice and neat - I used the round LED alignment piece, but not the big plastic shield (these parts not shown) that was on there originally, that covered the entire MCPCB:

After assembly, the initial tests were disappointing, getting only 2.57A at the tail. However, I failed to realize the KK 26700 cell I was using was somewhat down at maybe 4.0v and I didn't have the spring wired yet. So, a quick test in the lightbox revealed a big benefit to copper wire the spring. I can easily determine that by testing it assembled, than test it with the tailcap off and using a heavy wired to jumper the battery - end to the body.

Before the tail spring wire: 660 lumens OTF (30 secs)

After wiring the tail with 22 AWG: 796 OTF -- Wow! A 21% gain!!

I don't have pics, but the tail spring sits in the tail unattached/loose, under a brass cap. So, I simply added the 22 AWG copper wire down the middle of the spring, that sat it back in the assembly the way it was. Only other thing I did was treat the top and bottom of the spring with NO-OX-ID. The 21% boost was even slightly higher than what I recorded with the heavy jumper wire - this I've seen before, and I sort of explain it by the relatively poor connection the jumper wire has to the body from simply holding it against the threads.

So I then went ahead and wired the driver spring to see if I can record a higher tailcap amps reading. I did get 2.72A now, a small bump from 2.57A - interesting...

Then finally got a fresh KK 26700 off the charger, and re-tested:

  • KK 26700 at 4.22v, 3.11A at the tail, lumens: 928 @start, 904 @30 secs, throw: 23.5 kcd (measured at 5 m)

Should be noted I had this L5 already configured with a 37 mm UCLp lens.

As a reference, pure stock #'s I measured, probably a XM-L2 T6 1A (I think with a fullly charged KK 26700):

  • 2.20A at the tail, lumens: 748 @start, 687 @30 secs, throw: 18 kcd (measured at 5 m)

I really like the way this limited mod came out. Great tint and pretty good output and throw, plus a respectable runtime from a cell that has 5400 Mah capacity (rated at 5000, but the Opus charger says between 5300 and 5450 or so). Not too crazy about the stock UI of course - 5 classic modes, always starts at hi (hi-med-lo-strobe-SOS), and you can turn the light off from any mode by holding the switch down for 2 to 2.5 seconds -- way too long... It could be a worse UI, the hi-med-lo modes are actually nicely spaced, and nice to have the hold long OFF feature.

The ultimate mod though would be to add heat sinking, run it at a bit higher amps, with around 4A or DD using a BLF17DD piggybacked, and my e-switch driver. I really like side switch lights though - much better UI's are possibly and for high amps, the switch is not taken a beating of the power cycles and is not in the circuit adding resistance.

For the replacement switch, I'm really liking it so far - it is working very reliably, but doesn't have the click feel the original switch had. Unfortunately, many of those PCB mounted switch have a button that is rectangular, and seem to be less reliable if they are pushed in slightly to either side, and with the typical slop the driver mount has in these hosts, they are susceptible to problems. This replacement switch, mechanically, seems like it will be better because the button is round and all metal, unlike the plastic stock rectangular button. Time will tell if it holds up.

For $3.75 for 100 of these things, I'm hoping to find more uses because I got 99 left Smile.

Update Oct. 8, 2014:

Wanted to boost the power a bit more, so went ahead and added another resistor. While I had it apart, though I'd add some missing pictures from above.

Here's details views on the tailcap assembly. The housing is quite chunky:

All the parts with my 22 AWG wire added to the spring:

Close-up. I left the spring just sitting in there, and seems to really work well based on the 21% bump in output:

Although the retaining ring is plastic, it's pretty hefty and should hold up well:

Driver side, wired up with 22 AWG:

The "front stack". I replaced the stock o-ring with a thinner one to make up for the extras thickness of the UCLp lens (shown):

This is the plastic MCPCB cover, which I no longer use:

Fully mounted/wired, really like'n these MaxToch MCPCB's. I still use the stock LED alignment piece:

View with the extra resistor (R620), and another view of the mounted switch. Not much of a bump - results below:

After adding the R620, again, I expected a more significant bump, but with the same make cell, KK 26700, and at the same level of 4.22v:

  • KK 26700 at 4.22v, 3.37A at the tail, lumens: 972 @start, 945 @30 secs (no throw measurement)

8.7% bump in amps resulted in a 4.5% bump in OTF lumens. Probably about right...

Some Measurements:

  • Lens: 36.8 mm x 1.42 mm
  • Reflector: 36.8 mm O.D. x 33.3 I.D. x 22.65 mm depth, 9.0 mm hole

For Reference of reflectors in roughly the same size head:

  • F13 clone Reflector: 36.8 mm O.D. x 34.5 I.D. x 26.15 mm depth, 6.9 mm hole
  • Eagle Eye X6 (BLF Edition) from J.M. review, Reflector: 29 mm I.D. x 24.57 mm depth
  • UniqueFire M6 Reflector: 35 mm O.D. x 32 I.D. x 25.3 mm depth

Note: all my lumens and kcd measurements are "my own" taken on an older LX1330B meter - kcd taken indoors and lumens in a PVC constructed lightbox, accuracy can be freely debated (CYA all the way...)

Thanks Tom E. It was a nice read. The tabs on those switches are so small. I failed in using one in the OL build but have used one in my latest build with the Oshpark board making things a lot simpler.

Thanx - forgot who/where the link on these switch's were posted, but guess a few of us bought these same exact ones. The eBay listing no longer exists, but they are probably still being sold.

Been clicking away on the switch and still is working reliably. Requires a harder press than the stock setup, but it does work consistently, and not as hard a press required as my MM15.

I'll probably be adding another resistor to boost up the amps further. Came out a little lower than expected, so will aim for about 4A.

I probably got the same 100 pack of switches.. :D

I also got 20 of these:

Maybe they fit the L5, always nice to have some spare switches.. 0:) I have not used one yet, but I plan to replace a switch in a light that I suspect is not 100% well that is using similar switch..

Oh nice RacerR86 - looks the same as stock, including the 2 base pegs. I didn't do any searching, but these are definitely worth picking up for spares - a lot of side driver mounted switch's are like this. Weird this listing though says it doesn't ship to US, though it's from HK.

Nice light and mods. Thanks for sharing Tom E. :)

I just recently used 3 of my momentary switches fixing a coworkers expensive vape power pack that she accidentally washed with a load of laundry. My wife says one over here cruise control switches isn't working. That will be where another one goes. So I'm down to like 95 remaining.

Yep I have a hundred of those too, I also have a hundred of 2 other typed of SMT switched (including that exact one you needed). You can buy one for $3 or 100 for like $4 more.

I'm starting to see eBay sources for all sorts of deals on these small momentary switch's now. eBay's targeted ads had a whole bunch, once you are on the page of one of those deals. I copied the links and emailed to my myself to check out at home.

You have no idea how relieved I am bout this. I much prefer e-switch lights, but was always concerned about switch replacement and overall quality/reliability. Now that I know there seems to be fairly do-able replacements, this is great news. Maybe it's been mentioned arounf BLF and I've missed it, dunno. I'd like to re-explore some of this mod threads where guys converted lights to e-switch's, and also I'd like to look into adding a 2nd switch. Think the mechanical work may be beyond me, or at least very time consuming though.

To me, you can do a much better UI with an e-switch, or 2 e-switch's, and you eliminate the resistance of the mechanical switch for high amp builds.


thru-hole (used on the Roche F6DD, some others as well):

comfy - another couple of good ones - these are a bit pricey compared to the chinese knock-off's, but these seem to be quality for sure - makes sense coming from DigiKey. The "e-switch" brand is interesting... Gotta add a couple though in my next DigiKey order, or add them to the next order at work Smile.

I just saw these earlier today, might be worth a look.

Too big.

I saw that it was larger but figured it might be a useful addition to a parts box.

Been playing with the L5 as much as I can (people see flash's from my desk @work - fortunately they all know I have a "problem" Smile). Really liking the feel of this tiny 4x4 switch. Maybe conbined with the L5 button boot and the mount placement, etc., dunno, but it's real nice. Even though I find the driver's 2.5 sec OFF time to be a total PITA, I can click thru the 5 modes so quick, it's no big deal.

I'm thinking I should just go ahead and do this mod on other driver mounted e-switch lights, like the Y3, etc... I'm thinking the cheap plastic rectangular buttons on those switch's are prone to positioning/pressure inconssitencies.

Update Note on the L5:

I added another resistor, an R620, hoping to get about 4A, and again, only a minor boost to about 3.37A (from ~3.1A). It did boost measured output a little (~60 44 lumens). Dunno - may have to try adding something more signficiant. For some reason the resistor math is just not working (V=I*R, parallel resistos, etc.), but getting much lower results than expected.

Tom E - if you payed anything similar to what I payed for the 100x lot of those little buttons ($2 shipped I think)… what makes you have any confidence in them compared to the little rectangular right angle ones? I have no opinion, just asking.

EDIT: also, what resistor math are you applying here?

OP updated with a lot more pics I didn't originally cover, and some more details.

wight - totally good question bout the switch's. Think I paid under $4 w/free shipping... Well here's my thoughts:

  • the design, circular shaped as opposed to rectangular - the plastic rectangular ones can be pushed down on the left or right sides and seem to not make good contact
  • the actual little button is metal, not plastic
  • the depression seems more uniform
  • I think the distance the button depresses is less, so maybe less chance of it mis-hitting (maybe why it feels more uniform)
  • been using it as much as I could, and it's holding up well so far. Definitely more press's on it than the original one that got flaky

Pretty much speculation so far, but I'm really liking the feel of it in the L5. I've got limited experiences with buttons, but between all the products I've worked on the development of and had to support over the years, I've come across a lot of switch/button problems. The mil-spec buttons used in the mil-spec products I've been involved in, like nuc sub navigations systems, FAA ground radar systems, are of course extremely robust but they are pretty much to the extreme in size and quality, but can fail and do fail. In commercial systems, we've gotten burned in varieties of switch/button mechanisms, not to mention horror stories with connectors. Industrial grade of course is similar to mil - large size, hard to press, high quality.

Sometimes I think mil-spec parts suffer from low quantity and inexperience by the vendors, while the commercial guys know better how to keep things cheap and reliable - they have to for the high volume they do, and are more specialized.

I should spend some time tearing a couple of these apart and see the insides, but definitely would like to get more time in on testing them. Of course same ol' rules apply - could be one good switch I got, could be a good batch, who knows - chinese lottery...

Resistor math: V = I*R and this:

The 5 1R2's are 0.24 ohm, so adding another R240 in theory should double the amps? I think, so I thought? But adding an R500 should be 0.16 ohms. So, with V=I*R, I calc'ed it should raise the amps from 2.2A to 3.3A? Adding another R620 should be 0.13 ohms, so should be 4.06A?

Tom, you're working too hard. Just put an FET in there and be done with it. Resistance is futile!

(remember that the LEDs vF also comes into play here)

Yes - I know... Wastes so much time when I really wanted to put in a BLF17DD. Trying to help people out though that want to keep costs to a minimum, and "add more power". I got to build up some stock of boards - got all the parts laying around... Think that's what you meant by a FET? I haven't done any FET swaps in stock drivers - not feel'n that experimental yet Smile.

Hhmm - yes, LED vF --> law of diminishing returns. This is probably goin on. Could be just too much resistance in the driver design.

Just got my first "real" XM-L2 U3 1A's from Hank today - 3 mounted on 16mm Noctigons. Wondering what to put them in first... His bare U3 1A's on qty 10 sounds like a great deal Wink.

As RMM pointed out, you are only considering one factor. Vf, battery voltage sag, etc come into play… but the critical missing thing may be the cumulative resistance from the rest of the light and the tiny power components on the driver. It might make more sense to bridge the resistors, check current, and compare that with stock. Then you can fill in the blanks in between wherever you want. Maybe.

IIRC it will pull quite a bit of juice with the resistors bridged and springs over 6A. I bridged them on a stock light the first night I got them. I am still not sure whether I killed the stock emitter or the driver yet by doing that, but it lasted long enough for me to get some tailcap numbers.