What did you mod today?

Thanks! Here are some pics:

And yep I flashed a customized Anduril to get all the fancy effects. Though I already had it, because I love Anduril lol

The two people, (manxbuggy and his brother) did their best and built a contraption that produced miles better results than what our eyes can do, but they simply did not have the knowledge and equipment to make it measure lumens very accurately. If we want to use the lumen unit to describe the amount of light coming out of a flashlight, I think it is a good idea to approximate the lumen as closely as we can, and maukka has the equipment (professional sphere and verified calibrated tungsten light source) to get closer to the truth than with the tube-style light measuring devices and an average over a number of flashlight manufacturer’s claims (whose measurements can hardly be called independent, and I pretty much doubt how serious their equipment and method is).

I really think that your equipment is adequate as a way to measure light output of most types of flashlight, and as long as your calibration (whatever calibration in fact) is constant over time, it is an useful way to measure differences between flashlights, but if the aim is to approximate the lumen (which is a S.I. unit), maukka’s approach is simply way more accurate.

Remember though that I matched factory ANSI ratings for years, factories that showed a massive Lumisphere that put Maukka’s to shame. Just saying. (Yes, remember the huge green sphere that appeared to be at least 3’ in diameter? Professional as run by the engineers that design the products.) (Maxtoch has a huge Military and Police contract, as well as Firefighter and Search and Rescue)

Interesting how assumptions are made, without even knowing the people involved…

It could even be argued that the very style of the calibration tungsten bulb is SO different than our actual flashlights that the methodology is skewed… but whatever, it’s a hobby not a company and even Maukka is limited by the ever present budget monster.

The original intention of the home light box was not targeting accuracy, merely giving a tool to indicate if the mods performed were worth the gains or if something was missed or not functioning properly. Simple then, made so much more complicated now.

It took a few years for me to discover that manufacturer claims (also Fenix, Nitecore, SWM etc.) of flashlight output were considerably higher (up to 30% in some cases) than what the german TLF members found in their own pretty serious attempts to measure flashlight output. Maukka’s contribution to this forum confirmed this, so I’m pretty sure that he is on the right track and many manufacturers are wrong, even if they claim ANSI-conform measurements (is there even a body that checks all those claims?) and show a huge sphere to impress their readers.

The calibration with a tungsten calibration lamp in combination with a sphere with a professional coating is perfectly correct, no skewed methodology here.

You can dismiss a method for it not being perfect and then you miss that it is a big step forward to previously existing methods on the flashlight forums.

As I just said, originally it was never about accurately measuring the light output but a tool to aid the modder in getting more out of a budget light. A check point out of the box, and continuing check points along the way during the mod to ensure the work done produced viable results. EXACTLY how many lumens the light makes is really, in the end, irrelevant. All the variables knock the numbers down in a constantly descending curve anyway. Ball park figures, for the end user, are all that is truly necessary. Remember, the percentage of difference between the home built PVC P-Trap light box and the very expensive professional sphere are most of the time simply not visible to the eye anyway.

Measuring a side emitting low output tungsten light with the same equipment used to measure forward emitting LED light, apples and oranges. Science says light is light, of course, and Photography favors the golden hour, but what are all these millions of flashlights actually used for? Finding the keys dropped in the grass? Science application overkill. Using NASA grade equipment to go for a walk. And yes, I know, I could well have donated the funds required for me to play with all these flashlights the past five years… the incremental decension into a bad habit is rarely recognized as the wrong path.

An interesting read… Integrating Sphere

Notice the bullet points, especially interesting as we’ve been finding this out all along…

New lamps should be seasoned for 5% of their rated life.
A statistical analysis of a large number of samples is imperative for accuracy.
The ambient temperature in the laboratory should be controlled to 25±1 ˚C.
Temperature within the sphere must be monitored due to its effect on both the photometer head and the source.
The mains power must be monitored and controlled.
The baffle shape should be optimised for every source’s geometry.
Humidity must be controlled as moisture affects the reflectance of the barium sulphate sphere paint.
Exposure to radiation changes both absorption and transmission causing filter fade.
Sphere dirt changes spatial uniformity.
Detector hysteresis – overloading may cause temporary or permanent changes.
The effects of gravity on incandescent and fluorescent lamp orientation.
Sphere off-axis response. Exact placement within the sphere should be repeatable.
Stabilisation of light source may take up to 30 minutes.

However I do believe that with every sphere that you, djozz, maukka, and other experts build, the results will become a bit more accurate. And maybe even trending towards the same outcome. Until that time I have faith in what (all of) you report as the percentual changes of the before- and after- measurements of your mods.

The problem with light measurements is that if you do not do it precisely right, before you know it the errors are sky-rocketing. So even for flashlight measuring it is a good idea to do an attempt at accuracy. To measure voltage you buy a 10 dollar DMM and you have a 0.1% accurate result. Light is different, to measure light you buy 100K-dollar of equipment and are still stuck at 7% error. Nevertheless I see light output numbers posted on BLF, measured with homebuild equipment, that suggest 0.01% accuracy (“the output is 11.348 lumen”) while the actual error range is higher than 10% (so even “the output is 11k lumen” is actually exaggerated).

But I agree fully that measuring light as a tool for the modder does not require fancy stuff at all, as long as people do not claim to be measuring actual lumens (I call my lumen consistently “djozz-lumen” for that reason), and do not suggest a precision that is not there.

Yep, problems problems problems, all causing measuring errors. But if you do not at least attempt to do it right, I guarantee that the errors will be miles worse than what these sphere users encounter.

Some parts are interesting, but the text as a whole is a bit pedantic, they should have kept to what they actually measured themselves.

That was quite interesting but now………………….back to modding!

Actually I’m hijacking this from Djozz as he did most of the work.
I’m good at working out ideas and doing the rough calculation; Djozz has the skills and equipment to make it actually happen.

What did we do?

Well, I had a Thrunite TN31 that was already converted to an aespheric, doing 649Kcd with a dedomed XP G2.
I quite liked it but the led deteriorated after a while and at 549Kcd I decided it was time for further development.

The light moved to my storage and, as soon as a very limited number of wavien collars came available (early 2018 I believe) I bought
two of them.
The size I bought nicely fits on the shelf of the TN31, with a millimeter or so clearance on all sides.
This is quite important because a wavien collar not only needs to be in perfect focus to achieve optimal results, it needs to be perfectly
centered above the led as well.

I got a friend to drill the old led & ledboard out on a laithe and Djozz glued an old style XP G2 on a copper dtp board to the shelf.
After that Djozz made a copper ring to fit around the dtp board.
It’s function is to get the wavien collar on the correct height to provide maximum foton recylement. So to speak.
It then was fixed in place with three generous dots of heat glue.

The light color is not great - as expected - a dedomed XP G2 with a wavien produces a nice green/yellow tint but the throw numbers are
excellent: we measured 1,308.000 CD @ Djozz.
Translation: 1,3Mcd in a handheld size light; I call this a success!
It’s a pencil beam of course, but that’s ok. It’s a fun light, not meant to have any practical use. And fun I do have with it!

In the pictures you can see the various states of development, a comparison with other aespheric lights and a few beamshots.

The closest high voltage pylon is at 134 meters (146 yards) and the one farther away is at 713 meters (780 yards).
Below a small light, above to the right BTU Shocker (greenish due to dedomed XM L2’s), above that to the left Djozz’ BLF GT, on top TN31.
The beamshot pictures are quite accurate; but they do not show that we had a slight fog just above the ground, causing the light to spread out a bit.

Enjoy!

Edit: sorry but I can’t somehow get the pics uploaded. Will ask Djozz for help.

Top work! My board just arrived today, looking very forward to installing it shortly myself.

Did I see that you have the double click in lockout set to a higher momentary level?

Sounds like a fun time! I wish I lived near djozz, he keeps liking the same LEDs as me.

Any reason you didn't opt for a White Flat (or other newer LED besides the legendary "old good" XP-G2)?

I agree that accurate measurement is probably not all that critical. However, I think Maukka’s calibration lights add something very useful that you’re overlooking - they make it possible for two people in different countries, with different supplies available and different homebrew integrating devices, to exchange numbers. Previously, every “DB Custom Lumen” may have been perfectly repeatable and totally sufficient for your purposes, but they wouldn’t help in the slightest if I was building something similar and wanted to compare results.

People griping and complaining don’t actually care if you have accurate values for your own use, they just want to have a way to compare without needing to own the same “very large collection of ANSI rated lights” for comparison. It wouldn’t be sufficient to just own one light in common, because of variation between units and batches. But with the unique calibration printoff Maukka provides, his two lights can more or less fill the same niche as your “very large collection”.

Apologies for the unstable thermal response on that one. I wasn’t actually involved in making that particular light, and didn’t get a chance to calibrate it properly. It shouldn’t really be an issue during normal use, but it still bugs me a bit.

Anyway, I hope you enjoy it! :slight_smile:

When I post the decimal numbers it’s not to imply a degree of accuracy, it’s to show I’m not rounding up (or down). That’s the only reason I post it exactly as it came out, just to show no liberties were taken. :wink:

And um, sorry, I don’t know why I even posted all that this morning, really rough night, pulled back muscle spazzing and for whatever reason my stomach decided to do a reboot at around 2AM. Sorry.

This isn’t really a mod… more of an accident… but I found out how a XHP50.2 looks when it’s completely stripped so it has no dome and no phosphor layer. Someone apparently tried to mod this and messed it up, and then returned it as “defective”.

First, in normal lighting:

And here, it’s the same thing except I shined another flashlight into the bezel to brighten things up. The bare diode reacts to white light by making rainbows:

Purdy ain’t it? :smiley: And it’s output color is an amazing blue! :smiley:

I did that test when the XP-G3 came out, MEM and I were wondering… :wink: I actually installed the bare emitter in a zoomie and used it some, pretty neat really but you have to be careful about seeing too much of it or looking directly at it of course. It’s more just blue than it is UV but nonetheless I was careful with how I used it. Got tired of the potential problems after a while and swapped emitters in that zoomie, tossed the die in the can.

I got to thinking about this diffusion panel from the Samsung monitor, there were two of them in there. So I started cutting discs and testing again, 7 turns out to be too much but after some careful placement, placing 3 discs on each end of the P-Trap, with all the satin sides facing the flashlight, my meter now reads almost spot on with the calibrated Convoy S2+! No multiplier! I’m charging up both cells , the small light and the larger one, to test again but it looks like it’s going to work out well right here, 6 panels from the monitor inside the box.

Yep, still bored. :stuck_out_tongue: