The promised photos
Untouched
Bezel rand reflector removed
Removed the centering ring
MCPCB removed and thermal paste cleaned
Carefully removed the lens (I tried to get an intermediate photo but the lens popped loose)
Applied the thermal paste
Reseated the MCPCB
Resoldered the leads
Returned the centering ring
Reassembled
Very nice - thanks for sharing. 
The lens likely is not AR coated so causing transmission loss? Is that the reason for the gain?
I’ll have a closer look at it but it is supposed to be coated. Any lens or optic will cause some transmission loss, that’s one reason OTF lumens are different than emitter lumens (reflecors are also less than perfect).
AR coatings help reduce transmission loss due to reflection, but there will still be losses due to absorption, refraction, etc. They are also imperfect, drastically reducing reflection losses but not eliminating them completely.
Here is Vinh’s video if you are interested, it’s a bit long winded and doesn’t talk about the losses much, but it is a good walk through on dedoming and delensing various emitters. SBT90.2 De-Lens VS Shave Dome VS De-Dome - YouTube
The UCL lenses are here. The 51.3mm lens fits perfectly, 52.1 mm fits barely but does drop in without any pushing. Thicknesses of either are good.
Initial observation is that the coating makes the lens looks like it’s not there (same as what happens when eyeglasses are coated). It’s a strange (and beautiful) thing to see. As for the gain in brightness, testing for gain is more difficult as the gain is small and seems inconsistent; the result so far is at best *around 4*. Definitely, gain from AR coating is not nearly as large as the 15 gain of Samsung battery over Wurkkos battery.
If you are going to switch lenses, keep in mind there are two rubber rings: one below the lens “sitting” on reflector, and one above the lens “sitting” on bezel. Importantly, I have no way to confirm and not 100% sure, but it seems to me the one on the reflector is smaller. At least in my light, that’s the only way they fit nicely.
Awesome update, I broke down and ordered some 40T’s and a 52.1mm lens. I still have the Convoy lens coming from China, but I am weak and impatient. Hopefully the lens I receive is also slightly under sized and fits without issue. This light is less and less budget every day, but even after all the tweaks and batteries it is still a huge bargain. It’s hard to put a price on fun, and this light has provided me a lot of that so far.
If my measurements were anywhere close, then we get to start stacking some interesting gains. 4% from the lens when I get it, 5+% from delensing the emitter, and 15% from the 40T. And I will never know if these changes make a noticeable improvement, but it still makes me happy.
This thing is a beast too, lights up my 200m street end to end like it’s nothing!
Comparison of the tint of different lenses using light bulb reflection “method.” Basically I put the lens on dinner table and look at reflection of the chandelier’s light bulb above it, using the lens as mirror. This method “amplifies” the tint of the lens and makes it easy to see. In non-coated lens, there will be NO change in color. With AR coated lens, the light bulb usually looks green as with my Convoy lights and my coated eyeglasses, but there are exceptions as seen here with the UCL lenses.
Here’s a good article on why typical AR coated lenses (and AR coated eyeglasses) reflect green when looked directly down on it and purple when looked at from the side: http://rick_oleson.tripod.com/index-166.html . I don’t know why the UCL lenses take on different colors, but do know that they work perfectly and recommended as a must-have for the Wurkkos (hint: I have measurements coming 
) .
UCL Acrylic Lens
UCL Glass
Convoy S21D
I’ve struggled to get consistent results in comparing the lenses, made more difficult by the fact that gain in brightness from lens improvement is very small compared to battery gain by upgrading to Samsung 40T. For example every time I changed lens and repeated measurement, I would get a different number from the last trial. If the variation error, for example +/- 20, is larger than the improvement, say +10, then it’s impossible to reliably prove anything.
After several attempts and adjustments, two changes brought success with consistent/repeatable results:
Raw numbers are meaningless in ceiling bounce tests; just want to show example of one set of numbers I’m dealing with.
No lens 247
Coated UCL 243 (~1.6% output loss)
Non-coated OEM 233 (~6% output loss)
Impressively the UCL lenses achieved ~98.4% of max brightness, and ~4% improvement over OEM non-coated lens. BTW, the 2 UCL lenses run very close, in 243-244 range, but UCL acrylic hit 244 max a few times, while UCL glass got 243 max.
This is best I could do with my amateur skill and “equipment” lol, so pls take with a grain of salt 
.
Great job getting some repeatable results. On your advice I ordered a UCL glass lens and it’s very impressive. The AR coating seems to be very effective while bing rather subtle. I have a convoy AR lens on the way that just landed in the country, so it may be possible to compare these if I can get consistent results like you did (I expect to receive it in the next week or two, and there is no telling how well it will fit).
I have struggled to measure the improvement from the 40T using the Opple. The light heats up extremely quickly and I haven’t had a lot of time to mess with it. I did attempt to get CCT and DUV measurements with the stock and UCL lens, but it did not register a meaningful difference.
I wish I had taken pre and post de-lending measurements on something other than turbo, but I was excited to try to get max numbers. I just got a new puppy this weekend so I’m not sure how much time I’ll have to attempt this in the next few weeks, but it does give me more opportunities to use the light on walks at least.
Thanks and congrats re. the new puppy. Must be very cute but hopefully not chewing up things?
For battery test, yes light heats up very quickly and causes fast step-downs in brightness. The reading therefore must be immediate. In other words: one click to turn on, two clicks to Turbo, then read output immediately within a few seconds. Then turn off, recharge huntil indicator is green, then next reading after about 1 hour or so. Also, I mark positions of both light sensor and flashlight with masking tape so they’re in same exact location each time.
Unlike lens comparison, for me the battery comparison was easy since improvement is so large. After your post above, I re-run the test and just one run nailed ~15% gain from Samsung 40T over Wurkkos.
I may also have gotten some bad 40T’s. They came in with an initial voltage of 3.43v, I only charged one of them so far and it took 3600ma to get to a full charge, which seems like it was not necessarily healthy to start. It seems to have dropped voltage and capacity extremely quickly in my testing and didn’t seem to perform at the level I expected. I may call 18650batterystore and see what they have to say.
And yes, the puppy is amazingly cute and we are doing our best to set her up for success by not giving her the opportunity to chew up too many things.
OT - Tripod mount that I used for my beamshot shoot-outs in case anyone is interested.
For me this is better than the smaller rubber ones commonly seen as its diameter is about 5 mm bigger and the velcro strap provides additional safety.
Seller is legit, for me anyway: https://www.aliexpress.com/item/2251832688558005.html?spm=a2g0o.order_list.0.0.21ef1802x4tfGX&gatewayAdapt=4itemAdapt
They’re legit, been around quite some time. I’ve only ordered from them once. I keep them bookmarked as a store since they have a lot of doodads and cheap parts/older parts, so if my preferred stores are out I can look to this one or others (Wankedun, Red Hot, Dancelite, etc).
Thanks for sharing, that could make things a whole lot easier.
I tried to use my sbt90.2 as a lep with help from a 12x50 monocular lens. I took the bezel off of the flashlight as it was hitting the eyepiece and kinda cooking it but here’s some beam shoots that really cut the side scatter and show a lep like Hotspot.
Ts30s with 12x50 monocular zoom Ts30s with 12x50 monocular zoom - Album on Imgur
Thanks for sharing, very interesting. The monocular “converts” all output (hotspot and spill) into a single smaller hotspot? Not much output loss and I assume throw is a lot more?
I haven’t been able to check if the throw is increased, but it does seem more noticeable as there’s less sidescatter from the spill blocking the view.
One of the pictures with the ring around the side is from turbo. And just as always turbo still makes this light get very hott. So I was worried it would melt the tpu plastic eye piece on this lens.
There are two focusing adjustments and both were dialed to allow the utmost lep like beam pattern.
Every lens element you add will introduce a fairly significant loss in total output, particularly with plastic lenses. This may be offset by the increased focal distance, but I don’t think it’s guaranteed. Much has been written here about this, and I would suggest this is very similar to an aspheric lens (probably just a few aspheric lenses in a focusable assembly).
I found this interesting post on reddit, from what seems to me a very experienced and knowledgeable poster. With the 40% sale, the Wurkkos btw is the least expensive SBT90.2 light on market. Nothing comes close, not at 65 bucks.
Also note the CCT of my Wurkkos measures at 5163K, resulting in a surprisingly warm beam.
This LED is simply a must-have. It’s big, stupidly power-hungry, runs super hot, gets pretty green on low levels, is low CRI, and only available in 5700k, but is an absolute beast, being both stupidly bright, around 4000-5300 lumens (Though only a few lights hit over ~4800 lumens at startup), AND having incredible throw. It throws close to as well as a W2 but with 3.5 times the lumens. It’s nuts. It’s also one of the most expensive LEDs, so all the lights that have it are kinda pricey, but you can look for deals, and it’s 100% worth it to get an SBT90.2 light. Have a look at this list of SBT-90.2 lights and pick something. You basically can’t go wrong…
I’ve praised how wide, bright, and useful the Wurkkos TS30S’s beam is, and hopefully this beamshot explains it. It’s true the priority for a thrower is at the hotspot (brighter = more throw), but I also look at the spill because for me, a large spill makes the light more balanced and useful as the illumination area is larger.
BTW I have three M2RPro’s (pls don’t ask, my early flashlight experiments) and pictured are two of them, one has a nice warm color, the other has an out-of-control green tint, among the worst I’ve seen. Olight has beautiful finish and construction but LED consistency and color quality are not their strength. The beam is actually very nicely balanced between hotspot and spill, just the green tint that ruins that light for me (this was the moment I turned to Convoy ).
I actually do like the Nicecore a lot. It’s a nicely small, light EDC. The throw is lowest in this group but the bluish beam just makes it subjectively brighter than Olight’s warm colors. It also has a very good spill, better than all others except Wurkkos.
