My DIY Lightbox

[LilKevin715]

I wasn't sure whether to post this here or the Homemade & Modified lights section of the forum. Mods feel free to move this thread if it is not in the appropriate area.

Many of us guestimate lumen output of flashlights via methods such as a ceiling bounce test. While it does work somewhat it does leave alot to be desired in terms of accuracy. A integrating sphere with a properly calibrated light meter is the definitive way to measuring lumen output, but the cost(s) associated with the sphere itself makes it impractical for 99% of us. A relatively cheap, consistent, and accurate way of measuring lumen output (more accurately relative output) can be done with a lightbox. Here is a quick write up on how I created mine.

First we need a box. I wanted to use a box that was close to a cube shape with a volume of close to a cubic foot. I didn't have any readily available so I chose to use two rectangular boxes with dimensions of 9.75" x 11.5" x 6". The dimensions of the two boxes combined are 9.75" x 11.5" x 12".



Also as seen in the picture above I decided to round off the corner areas to reduce the amount of light that is lost and not reflected. Next up is to apply aluminum foil to in inside of the box with some glue. A thin layer of Elmer's glue-all multi purpose glue (the white stuff) worked just fine. The first purpose of the aluminum foil will be to prevent any external ambient light from entering the box. The second purpose is to help reflect back any light that passes through the diffusion layer (paper) to minimize light losses. Double-sided sticky tape was used to attach the paper to the aluminum foil.





I decided to make the cutout for the flashlight right in the center of the box. The cutout for the light sensor was made in the corner.



Lastly to eliminate any chance of light entering the box via the light sensor hole I sealed off the area with some aluminum foil wrapped with duct tape. The light meter can be easily removed for taking hotspot lux measurements.



I'm still working on fine tuning the design but for the most part the groundwork is finished. Suggestions, feedback, etc. are welcome.:wave:

 

[LilKevin715]

In order to estimate lumen output with a lightbox we need to calibrate it. When a light is placed in the hole opening and activated the light meter (a LX1330B) will measure the amount of indirect reflected light (light that is reflected off the inside walls) that hits the sensor . This amount of indirect reflected light will register a reading on the light meter that will be measured in lux. If enough accurate data points are obtained a trendline can be generated to estimate total output for any given lux measurement, which I will refer to for now on as Lightbox Lumens (LBL for short).

Here are the lights that I will be using to calibrate my lighbox. Please note that there will always be sample variance. It is always preferable to use lights that have been independently tested so any manufacturers claims/measurements of lumen output can be verified.

Eagletac P20A2 XR-E R2 (tested by selfbuilt)
Eagletac P20A2 XP-G R5 (tested by bigchelis)
Maglite 2D Rebel (ANSI)
Maglite 3D XP-E (ANSI)
Maglite with Terralux TLE-300M-EX (3 x XR-E R2) (tested by bigchelis)
Thrunite Neutron 2A XM-L T6 (tested by selfbuilt)

Here is a graph with the data points of the measured lux vs lumen output. The generated trendline (a power regression model) can be used to estimate LBL with a known lux reading from the lightbox. Trendlines will almost always never be the same for comparing different lightboxes as there will most likely be differences in lightbox construction, light meters, etc.

Here is a pciture of the Data & Graph, click on the image to enlarge it.



The trendline seems reasonably accurate. The only thing I can see to improve it would be to sample more lights in the 300 to 600 lumen range.

 

[reppans]

Looks awesome - can't wait to hear your results. Just curious if you can give a preview of lights in your collection that you will be testing.

I'm using the very simplistic bounce/sampling method and a DSLR as my light meter. Certainly can't claim the accuracy of a lightbox or lux meter, but I find it far more accurate and objective than my naked eye.

I'm honestly astonished at how far off certain so called "reputable" manufacturers are with their specs, even in comparison to their own lights, and modes on the same light. Also can not believe how much difference there is in ANSI interpretation between different manufacturers - really makes ANSI claims and spec sheet comparisons a joke. I mean seriously, 50% differences in tested output among reputable "ANSI" manufacturers??

Wish more people could test their own lights as Selfbuilt can really only cover a small portion of the market.

 

[LilKevin715]

The majority of the lights that will be tested are custom P60 setups and modded Mags. The only bad thing about having so many custom/modded lights is a reference point in terms of output.

I'm honestly astonished at how far off certain so called "reputable" manufacturers are with their specs, even in comparison to their own lights, and modes on the same light. Also can not believe how much difference there is in ANSI interpretation between different manufacturers - really makes ANSI claims and spec sheet comparisons a joke. I mean seriously, 50% differences in tested output among reputable "ANSI" manufacturers??

Tell me about it... I was counting on using some of my very few production lights (and their stated/claimed OTF and ANSI numbers) as a baseline for calibrating my lightbox and I'm running into problems already. I know there will always be sample variance but some of the numbers I've obtained so far make no sense at all. For example I have a Eagletac P20A2 MKII with two modules to compare. The first module is the stock XP-G R5 while the second module has been modded with a simple emitter swap to a XP-G2 R5 ("same" driver as the first module). Logic would dictate that the XP-G2 module should have higher output. The first XP-G module measured 2780 lux @ 30s in my lightbox, while the XP-G2 module measured 2660 lux @ 30s. WTH!?!? :confused::thinking:. I repeated the test several times and the numbers were close to the same. :banghead::hairpull:

I wish there was a company that had a service available where you could send in your lights and have them tested in a calibrated integrating sphere. At least that way I know I could trust the numbers of my personal lights.

 

[Cataract]

WOW! that's excellent work; seems like you thought of everything (then again this is coming from a guy whi probably would have thought about rounding the corners afterwards... You must have had fun gluing the foil in, though. Now I feel like making one... How long did it take?

 

[LilKevin715]

Total time to build the lightbox was probably around 5-8 hours. I could have probably finished it faster but I took my time to make sure it turned out good and not sloppy. Actually the aluminum foil was one of the easiest parts of the construction. Once I had the bottom glued down I simply rolled the foil off the tube to complete the perimeter. I applied glue to one wall, rolled the foil onto the glue, smoothed it out with my hand, and then repeated the same process for the next wall. The hardest part was attaching the paper to the foil with the double stick tape at the corners of the perimeter. I had to make sure the paper didn't get kinked/wrinkled at the bottom too much to keep it smooth, while at the same time making sure the paper sheets overlapped each other by just the right amount.

 

[LilKevin715]

The following is a bit of a rant, but in the end it can be useful for anyone trying to calibrate their own lightbox or integrating sphere.

I've figured out why some of my measurement results weren't consistent and it all boils down to the OTF & ANSI lumen measurement interpretation. One of the lights I was using to help calibrate my lightbox is a Eagletac P20A2 MKII XP-G R5. Eagletac states it is 230 lumens OTF and this has been verified by bigchelis (230 lumens @ 1s) in his integrating sphere testing thread (do a Ctrl-F to search for it). Another light with higher lumen output used for testing/calibration is a Eagletac T20C2 MKII XP-G "S2" (more on that in a bit). According to Eagletac it is rated for 375 ANSI lumens/500 lumens at the emitter. You can take a look at the specs of the various dropins for the T20C2 MKII in the screenshot below (click to enlarge) from Tod's site at illuminationgear (BTW great guy here of the forums).



Please take careful note that the percentage of ANSI lumens vs emitter lumens is exactly 75%. Nice precise percentage number eh:rolleyes:? Ok now lets compare the "S2" module to a older R5 module. The older R5 module is rated at 300 lumens OTF/400 emitter lumens. The percentage of OTF lumens vs emitter lumens is again exactly 75%. But wait a sec isn't the older R5 module rated for OTF lumens vs the newer S2 module rated for ANSI lumens?:confused: FYI if you lookup the details of the ANSI FL-1 testing in reguards to lumen output measurement the reading is taken at 30-120 seconds. Most manufacturers would probably choose a reading at the 30 second time minimum as luminous flux of LED's decreases with increased temperatures. In the first example when I mentioned the P20A2 MKII the OTF lumen measurement was accurate at 1 second, but output does sag down to 219 lumens @ 30 seconds. My T20C2 MKII XP-G S2 module measured 3570 lux @ 1s and 3270 lux at 30s in my lightbox, a decrease of 8.4%. So how can the OTF & ANSI lumen measurements Eagletac provides still have the same OTF/ANSI lumen vs emitter lumen percentage?:thinking:

In my opinion, at least for the XP-G S2 module, they simply changed the text from OTF to ANSI. If the ANSI testing was done in-house then they must have placed the light into a freezer before starting the test. That would be the only way to explain the same OTF/ANSI lumen vs emitter lumen percentage.:crackup: Also on a more alarming note, according to my measurements the output of my "S2" module is only around a estimated 318 lumens according to the trendline of all my data sampled (I will provide the data in post #2). This is closer to the output of a XP-G R5 bin module:wtf:. This could mean several things including:

Defective driver.
Using the wrong (R5) bin emitter on my module (labeled as a S2).
A R5 module with the wrong sticker.
Incorrect packaging for a older module/light.
Eagletac trying to get rid of old R5 emitters by selling them as S2.
A combination of the above, etc.

Even though I am dissapointed in the results of my "S2" module I still like Eagletac lights. Their build quality and the ability to swap out modules on certain models sets them apart from other manufacturers. To sum up my rant I have the following advice: Don't completely rely on manufacturers specifications in terms of lumen output. CPF and its members are a wonderful resource for independent real-world testing.:) If you are looking to build a lightbox or integrating sphere I would make sure to have independent test results to compare against the manufacturers specifications.

 

[Bigmac_79]

This is great, and this is the spirit I love to see here at CPF! I'll look forward to seeing how this goes for you. The only suggestion I might make is the possible addition of a baffle, something to prevent direct reflection off the first surface the light hits (position it between the sensor and the area where the hotspot shines). I did this for my box of similar design, and it really helped even out the discrepancies between throw-oriented and flood-oriented lights. Keep up the good work!

 

[LilKevin715]

Before building my lightbox and during my research I did take note of the possible need of a baffle from other members experiences. I'm glad to report no baffle is needed with my design. I verified this my using three different lights:

Maglite XM-L T6 @ 3.5A: chose low mode @ 150ma to minimize thermal related output sag.
Solarforce L2 Smooth reflector XR-E Q5 EZ900 @ 1A: 15.1k lux @ 1m
Solarforce L2 Aspheric XR-E Q5 EZ900 @ 1.05A: 29k lux @ 1m

First I tested the Maglite at its optimal focus point for throw and recorded a lux reading of 572. I then unfocused the beam so there was a donut in the middle of the beam and remeasured. With the donut in the middle of the beam I measured a lux reading of 570; pretty much the same. Next up is the L2 Smooth reflectored XR-E Q5 EZ900 @1A. This light measured 2100 lux @ 1s and 1980 lux @ 30s. The last light is the L2 Aspheric XR-E Q5 EZ900 @ 1.05A. This light measured 1010 lux @ 1s and 950 lux @ 30s. The lower output of the aspheric setup vs the reflectored setup isn't too suprising as a lot of light is lost due to not being reflected off of a reflector.

So whether its flood or throw my lightbox doesn't discriminate.

 

[Bigmac_79]

Great, it sounds like you don't need to mess with it!

 

[LilKevin715]

Before I list the LBL results of the lights in my collection I need to do some further testing. A lot of us guestimate how much light is lost from the emitter to OTF. This is one of the very few topics here on CPF that hasn't been fully tested. I frequently see members posting a inconsistent percentage number of emitter vs OTF lumens (anywhere from 66 to 85 percent, depending on the type of lens & setup used). Since I now have the ability to measure relative output its time to do some definitive testing and put these claims to the test. This testing will include light loss from components such as reflectors, bezels, and lenses. Here is a quick overview of how I would test a Maglite (specifically my XM-L Mag, build details in the link of my sig):

1) Measure output with the head removed (e.g. "candle" mode), light on low mode to minimize thermal related output sag.
2) Remove the bezel and lens from the head. Screw the head back onto the body tube without the bezel and lens attached. Take a reading noting how much output is reduced by only the reflector.
3) Screw on the bezel without a lens and take a reading, noting how much output is reduced by the bezel.
4) Unscrew the bezel and try various types of lenses (stock plastic, hardcoat acrylic, borofloat, UCL) with the bezel on. This final step can compare the efficiency of the various lenses.

In my opinion this testing is very important as it hasn't been done yet (to my knowledge), and more importantly if anyone wants to use my test data as a reference. Take for example the data in post #2. I claimed the 3D XP-E Maglite of having a output of 139 lumens @ 30s. The ANSI specs are actually 134 lumens @ 30s. Why is there a difference? I used a hardcoat acrylic lens (claimed ~95% efficient) instead of the stock maglite lens (~91.45% efficient) as tested by Roy here. The math can be done as shown below:

Reflector Lumens X Lens Efficiency = ANSI or OTF lumens

Since we know what the lens efficiency of the stock lens is and the ANSI lumens, we can calculate what I'll call Reflector lumens (light that hasn't gone through the lens yet). The resulting reflector lumens is 146.53 lumens. When using the hardcoat acrylic lens vs the stock lens output is increased to 139.2 lumens. Even though the difference is only 5 lumens, it would have a effect on a generated trendline. While the difference may not be huge at this level, it can make a noticeable difference at higher outputs.

When I release my LBL data I will include the lenses, reflectors, and bezels that are used. This can be important for using different lenses and bezels on a setup like a Solarforce L2.