The LightBox Experiments... Comments please!
- Thread starter Quickbeam
- Start date
Help Support Candle Power Flashlight Forum
Doug S
Flashlight Enthusiast
[ QUOTE ]
Quickbeam said:
OK, OK... I know it may not be the most accurate device in the world... There may be an issue with wavelengths, etc... But on a budget of $5, whatdya expect?! /ubbthreads/images/graemlins/icon15.gif
[/ QUOTE ]
I think that you are paying too much for your milk /ubbthreads/images/graemlins/tongue.gif
Seriously, Doug, I am really excited by what you are doing. I am looking forward to whatever you do next.
Quickbeam said:
OK, OK... I know it may not be the most accurate device in the world... There may be an issue with wavelengths, etc... But on a budget of $5, whatdya expect?! /ubbthreads/images/graemlins/icon15.gif
[/ QUOTE ]
I think that you are paying too much for your milk /ubbthreads/images/graemlins/tongue.gif
Seriously, Doug, I am really excited by what you are doing. I am looking forward to whatever you do next.
dreamlogic
Newly Enlightened
- Joined
- Sep 22, 2003
- Messages
- 4
great job! the graph is really cool.
i'm wondering why the 2-2016 light scores so high. could it be that there's light coming out the side and hitting the sensor directly? i guess if that's the issue, it shouldn't affect the results of the real flashlights.
i'm wondering why the 2-2016 light scores so high. could it be that there's light coming out the side and hitting the sensor directly? i guess if that's the issue, it shouldn't affect the results of the real flashlights.
Quickbeam
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
Well, version 2.0 of the lightbox experiment is in progress. The milk carton has been cut open and is being coated with Red Devil Acrylic Latex Enamel RDHP2402 Satin White paint.
Why this particular paint? Because it's what I had laying around! /ubbthreads/images/graemlins/grin.gif
Special full reflectivity paint-on projector screen paints and lumen sphere paints are rediculously expensive, so fuggedaboudit! We're going CHEAP here! Flat, white, and already sitting on my shelf were the qualifications, and this can of paint fit the bill.
The idea is to improve the "whiteness" of the inside of the milk carton (it is rather yellow compared to even plain white paper) and to make the box more opaque to light from the inside.
Coat one is done, coat 2 comes in about an hour or two. Then hopefully, all the readings can be re-done this weekend or early next week.
Well, version 2.0 of the lightbox experiment is in progress. The milk carton has been cut open and is being coated with Red Devil Acrylic Latex Enamel RDHP2402 Satin White paint.
Why this particular paint? Because it's what I had laying around! /ubbthreads/images/graemlins/grin.gif
Special full reflectivity paint-on projector screen paints and lumen sphere paints are rediculously expensive, so fuggedaboudit! We're going CHEAP here! Flat, white, and already sitting on my shelf were the qualifications, and this can of paint fit the bill.
The idea is to improve the "whiteness" of the inside of the milk carton (it is rather yellow compared to even plain white paper) and to make the box more opaque to light from the inside.
Coat one is done, coat 2 comes in about an hour or two. Then hopefully, all the readings can be re-done this weekend or early next week.
milkyspit
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
Quickbeam, hope this doesn't discourage you, but I'm wondering about the bazillion lights you've already measured... doesn't this mean all those readings will be meaningless with the new measurement device? Put in programmer terms: are you breaking backwards compatibility by painting the inside of the carton? /ubbthreads/images/graemlins/icon15.gif
Also, I'm wondering about reproducibility of test apparatus by us lesser testers. /ubbthreads/images/graemlins/grin.gif Anything you can do to get specific about how to make an IDENTICAL apparatus for our own tests would be MUCH appreciated! You've already been giving us lots of juicy details, which is terrific, but for testing to be as comprehensive as possible, we need to find the best possible way that all of us can make devices yielding consistent results with yours. If that means some sort of specific template with exact measurements for hole placement, UPC code for the exact paint used, etc., so be it! None of us has ALL the flashlights we might want tested, not even you (though you seem to be close! /ubbthreads/images/graemlins/grin.gif ), so it would be great if anyone interested can get on board, and we can share the results in some sort of master list. For example, I'd like to see how my MR-X compares to some of these other lights. Any ideas on that? /ubbthreads/images/graemlins/confused.gif
Also, I'm wondering if high brightness laser printer paper (90+ rating) instead of paint might make a good lining for the milk carton, given that such paper is VERY white, widely available, would put a consistent "layer" over the carton every time, has a consistent texture (no funny streaks to bounce the light in weird directions), and is pretty much idiot-proof. Just cut and stick!
/ubbthreads/images/graemlins/icon3.gif
Quickbeam, hope this doesn't discourage you, but I'm wondering about the bazillion lights you've already measured... doesn't this mean all those readings will be meaningless with the new measurement device? Put in programmer terms: are you breaking backwards compatibility by painting the inside of the carton? /ubbthreads/images/graemlins/icon15.gif
Also, I'm wondering about reproducibility of test apparatus by us lesser testers. /ubbthreads/images/graemlins/grin.gif Anything you can do to get specific about how to make an IDENTICAL apparatus for our own tests would be MUCH appreciated! You've already been giving us lots of juicy details, which is terrific, but for testing to be as comprehensive as possible, we need to find the best possible way that all of us can make devices yielding consistent results with yours. If that means some sort of specific template with exact measurements for hole placement, UPC code for the exact paint used, etc., so be it! None of us has ALL the flashlights we might want tested, not even you (though you seem to be close! /ubbthreads/images/graemlins/grin.gif ), so it would be great if anyone interested can get on board, and we can share the results in some sort of master list. For example, I'd like to see how my MR-X compares to some of these other lights. Any ideas on that? /ubbthreads/images/graemlins/confused.gif
Also, I'm wondering if high brightness laser printer paper (90+ rating) instead of paint might make a good lining for the milk carton, given that such paper is VERY white, widely available, would put a consistent "layer" over the carton every time, has a consistent texture (no funny streaks to bounce the light in weird directions), and is pretty much idiot-proof. Just cut and stick!
/ubbthreads/images/graemlins/icon3.gif
Quickbeam
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
Milky,
Yes, I know they'll all need to be tested again.
As to the reproducablility - that's why it's RELATIVE brightness, not absolute. Relative brightness to the other lights tested with the same apparatus is the best we're ever going to do here.
Let me detail why an identical apparatus will be impossible:
Box construction could vary
Paint reflectance could vary between brands and even batches
Thickness of paint could vary
Texture of paint application could vary
Box paper reflection/spectrum selectivity could vary
Placement of the holes could vary
Meters could vary
Meter calibration could vary
Etc.
Etc.
Etc.
So is anyone going to exactly reproduce this setup? NO.
Can you do something similar and get somewhat similar readings? PROBABLY.
I tried going the "stick paper inside the box" route and it was just too much of a pain in the butt, so I just slapped white paint all over the inside of the box after cutting off the bottom. That will have to do.
"Your Mileage May Vary" is the recurring mantra of this project.
At least it seems to be working pretty well as-is! Most of the 2 cell 123A lights with a 1 hour runtime really do put out about the same amount of light, as would be expected. /ubbthreads/images/graemlins/thumbsup.gif
Milky,
Yes, I know they'll all need to be tested again.
As to the reproducablility - that's why it's RELATIVE brightness, not absolute. Relative brightness to the other lights tested with the same apparatus is the best we're ever going to do here.
Let me detail why an identical apparatus will be impossible:
Box construction could vary
Paint reflectance could vary between brands and even batches
Thickness of paint could vary
Texture of paint application could vary
Box paper reflection/spectrum selectivity could vary
Placement of the holes could vary
Meters could vary
Meter calibration could vary
Etc.
Etc.
Etc.
So is anyone going to exactly reproduce this setup? NO.
Can you do something similar and get somewhat similar readings? PROBABLY.
I tried going the "stick paper inside the box" route and it was just too much of a pain in the butt, so I just slapped white paint all over the inside of the box after cutting off the bottom. That will have to do.
"Your Mileage May Vary" is the recurring mantra of this project.
At least it seems to be working pretty well as-is! Most of the 2 cell 123A lights with a 1 hour runtime really do put out about the same amount of light, as would be expected. /ubbthreads/images/graemlins/thumbsup.gif
Quickbeam
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
OK.. This is weird. I picked a sampling of various lights to use with the new design (Painted white inside) and the results were... the same. I mean frikkin' identical.
The only one that really changed was the green KL1 which scored about 10% higher with the new interior, so really no big deal. Red LEDs scored the same, Luxeons scored the same, Incandescents scored the same, so the heck with it - the readings stand as-is!
I think that this also means that no matter what milk box you use, even if the "whiteness" is a bit off, you should get about the same readings if you try this at home!
OK.. This is weird. I picked a sampling of various lights to use with the new design (Painted white inside) and the results were... the same. I mean frikkin' identical.
The only one that really changed was the green KL1 which scored about 10% higher with the new interior, so really no big deal. Red LEDs scored the same, Luxeons scored the same, Incandescents scored the same, so the heck with it - the readings stand as-is!
I think that this also means that no matter what milk box you use, even if the "whiteness" is a bit off, you should get about the same readings if you try this at home!
Doug S
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
[ QUOTE ]
Quickbeam said:
OK.. This is weird. I picked a sampling of various lights to use with the new design (Painted white inside) and the results were... the same. I mean frikkin' identical.
[/ QUOTE ]
Doug, that's great news! BTW, could you confirm [or refute] my interpretation of your original post on this subject about your "+/-5% depending on focus" statement was the result of testing a focusable flashlight at the extremes of it's range of focus?
I was thinking that if enough people follow your lead and build their own boxes, it might be fun to come up with some type of light engineered to have a constant output to pass around to anyone interested to use as a "transfer standard".
[ QUOTE ]
Quickbeam said:
OK.. This is weird. I picked a sampling of various lights to use with the new design (Painted white inside) and the results were... the same. I mean frikkin' identical.
[/ QUOTE ]
Doug, that's great news! BTW, could you confirm [or refute] my interpretation of your original post on this subject about your "+/-5% depending on focus" statement was the result of testing a focusable flashlight at the extremes of it's range of focus?
I was thinking that if enough people follow your lead and build their own boxes, it might be fun to come up with some type of light engineered to have a constant output to pass around to anyone interested to use as a "transfer standard".
Quickbeam
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
Yea, the +/-5% or so seems to be about right.
I've updated the charts with more accurate readings on a few lights, including the 2x2016 light, 3-D mag (turns out mine has a little bit of a short - got a good reading this time) Added 4-D mag and corrected Green KL1.
Yea, the +/-5% or so seems to be about right.
I've updated the charts with more accurate readings on a few lights, including the 2x2016 light, 3-D mag (turns out mine has a little bit of a short - got a good reading this time) Added 4-D mag and corrected Green KL1.
Gransee
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
Excellent work Doug. A fine example of CPF ingenuity!
You asked for input, so here goes...
Aside from questions like:
- sample size (units, batteries) affects error percentage and accounts for variations in Vf, bulbs, batteries, etc (variations are especially large with LED products)
- Variation in accuracy of your intergrating box with light intensity (increased bleed through/loss with more intense light sources).
- Ambient temperature
A question I feel is particularly important is the effect of time. You may have noticed that your output readings usally dropped the longer you held the light over the sensor. Most flashlights have a "warm up" period that effects output. Sometimes this warmup period matches normal usage and sometimes the light is still warming up at the end of each normal use.
As a result, I would suggest that the light be tested in this fashion:
- Determine the average usage period of a flashlight (nMinutes) for the people who your review is targeted for and use this same number for each flashlight. Lets say for discussion that most lights are used for 5 minutes.
- With fresh batteries and the light off, allow each light to adjust to ambient temperature before testing
- Turn on the light and take an output measurement at 5 minutes of operation.
This will not improve accuracy from unit to unit but it will more accurately reflect normal usage. Some parts of the chart will not change much in their order, but other parts will.
Otherwise a disposable camera flash will beat the E2 every time. /ubbthreads/images/graemlins/smile.gif
I have seen some lights change in brightness by over 50% in 5 minutes! You may have noticed this effect yourself.
Just a thought. You asked for our input. Again, a great idea Doug!
Peter
Excellent work Doug. A fine example of CPF ingenuity!
You asked for input, so here goes...
Aside from questions like:
- sample size (units, batteries) affects error percentage and accounts for variations in Vf, bulbs, batteries, etc (variations are especially large with LED products)
- Variation in accuracy of your intergrating box with light intensity (increased bleed through/loss with more intense light sources).
- Ambient temperature
A question I feel is particularly important is the effect of time. You may have noticed that your output readings usally dropped the longer you held the light over the sensor. Most flashlights have a "warm up" period that effects output. Sometimes this warmup period matches normal usage and sometimes the light is still warming up at the end of each normal use.
As a result, I would suggest that the light be tested in this fashion:
- Determine the average usage period of a flashlight (nMinutes) for the people who your review is targeted for and use this same number for each flashlight. Lets say for discussion that most lights are used for 5 minutes.
- With fresh batteries and the light off, allow each light to adjust to ambient temperature before testing
- Turn on the light and take an output measurement at 5 minutes of operation.
This will not improve accuracy from unit to unit but it will more accurately reflect normal usage. Some parts of the chart will not change much in their order, but other parts will.
Otherwise a disposable camera flash will beat the E2 every time. /ubbthreads/images/graemlins/smile.gif
I have seen some lights change in brightness by over 50% in 5 minutes! You may have noticed this effect yourself.
Just a thought. You asked for our input. Again, a great idea Doug!
Peter
Quickbeam
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
Peter,
Thanks for the input! Yea, I asked for it - in more ways than one it seems! /ubbthreads/images/graemlins/icon15.gif
If I had all day (which, unfortunately, I do lately...) I could list about 100 things wrong with the experiment itself and the testing methodology. Being a scientist at heart, I probably know better than most other folks just what is "imprecise" with the apparatus and methods... But without unlimited resources and funding I'm afraid this will have to do...
Will keep your suggestion in mind. Thanks!
Peter,
Thanks for the input! Yea, I asked for it - in more ways than one it seems! /ubbthreads/images/graemlins/icon15.gif
If I had all day (which, unfortunately, I do lately...) I could list about 100 things wrong with the experiment itself and the testing methodology. Being a scientist at heart, I probably know better than most other folks just what is "imprecise" with the apparatus and methods... But without unlimited resources and funding I'm afraid this will have to do...
Will keep your suggestion in mind. Thanks!
Doug S
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
[ QUOTE ]
Gransee said:
Aside from questions like:
- Variation in accuracy of your intergrating box with light intensity (increased bleed through/loss with more intense light sources).
Peter
[/ QUOTE ]
Peter, most of your comments make sense to me but the above one has me scratching my head /ubbthreads/images/graemlins/thinking.gif
Your comment above seems to suggest material for with the % transmission is not constant with intensity. While I believe that there are some *highly engineered* materials for which this is true [think photoreactive sunglasses] I don't think it would be a concern with milkcartons.
[ QUOTE ]
Gransee said:
Aside from questions like:
- Variation in accuracy of your intergrating box with light intensity (increased bleed through/loss with more intense light sources).
Peter
[/ QUOTE ]
Peter, most of your comments make sense to me but the above one has me scratching my head /ubbthreads/images/graemlins/thinking.gif
Your comment above seems to suggest material for with the % transmission is not constant with intensity. While I believe that there are some *highly engineered* materials for which this is true [think photoreactive sunglasses] I don't think it would be a concern with milkcartons.
Gransee
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
The difference in accuracy with bleed through would be negligible, especially compared to the other factors. That is why I thought it was less important.
Bleed through would occur equally with all light intensities but due to quantization, the meter would not equally represent this. In most cases, the bleed through amount would be below the least significant digit accuracy of the meter. However, the bleed through at higher intensities could have an effect on that and other digits. This would mostly have an effect on the accuracy between widely different intensity samples.
However, most lights are compared to other lights similar in brightness. In those ranges, bleed through would be more consistent across samples. Since Doug's goal is to provide a cost effective method of comparing lights, some factors (such as bleed through) are less important to this goal. The 50% loss in brightness over 10 minutes is more significant IMO.
Again, Kudos to Doug's integrating box!
Peter
The difference in accuracy with bleed through would be negligible, especially compared to the other factors. That is why I thought it was less important.
Bleed through would occur equally with all light intensities but due to quantization, the meter would not equally represent this. In most cases, the bleed through amount would be below the least significant digit accuracy of the meter. However, the bleed through at higher intensities could have an effect on that and other digits. This would mostly have an effect on the accuracy between widely different intensity samples.
However, most lights are compared to other lights similar in brightness. In those ranges, bleed through would be more consistent across samples. Since Doug's goal is to provide a cost effective method of comparing lights, some factors (such as bleed through) are less important to this goal. The 50% loss in brightness over 10 minutes is more significant IMO.
Again, Kudos to Doug's integrating box!
Peter
milkyspit
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
The consistent results is great news, Quickie. /ubbthreads/images/graemlins/grin.gif We really need a testing methodology that others can replicate on their own. The sad truth is that we'll never have all the lights that need to be tested in one place. For instance, I've got a MR-X that I'd love to test, but I'm not crazy about the idea of shipping it to you just to do so, nor is it fair to you that the entire burden of testing everyone's flashlight would fall on your shoulders alone. So anything you can do to establish a specification for making an "official" milk carton test apparatus would be a HUGE win! Thanks again for moving us all forward with this, Doug.
The consistent results is great news, Quickie. /ubbthreads/images/graemlins/grin.gif We really need a testing methodology that others can replicate on their own. The sad truth is that we'll never have all the lights that need to be tested in one place. For instance, I've got a MR-X that I'd love to test, but I'm not crazy about the idea of shipping it to you just to do so, nor is it fair to you that the entire burden of testing everyone's flashlight would fall on your shoulders alone. So anything you can do to establish a specification for making an "official" milk carton test apparatus would be a HUGE win! Thanks again for moving us all forward with this, Doug.
Quickbeam
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
"Official apparatus?" Well, that's easy... here goes.
Equipment needed:
1 milk carton, white interior
1 small can flat white paint
1 xacto knife
1 ruler
Aluminum foil
Clear tape
1 piece of printer paper
Scissors
Procedure:
Lay milk carton on side.
Cut along three edges of one long side (two long edges, one short edge) to make a long hinged panel and open.
Paint 2 coats of flat white paint on interior of carton.
When dry, place a small pencil mark "x" in the exact center of the side and close up the cuts with tape.
Roll box over so cut portion of box is on bottom, pointy side away from you, small flat bottom of box toward you.
In the exact middle of the carton side facing up (long side), cut a 2" diameter hole for shining the light through.
In the small base of the carton (facing you) cut a hole appropriate for your LUX light meter in the center of the base so it fits snug.
Wrap entire carton in aluminum foil, securing the foil with the tape. Place shiny side in.
Cut the holes in the aluminum foil to match the holes already in the box.
Cut a 1" by 3" piece of printer paper and tape to the edge of the flashlight hole on the meter hole side of the box so that it forms a baffle, preventing light from directly shining on the meter sensor.
Take a reading by shining the light straight down onto the "x" with the edge of the bezel level with the top of the box.
That's it!
Quickbeam's lightbox experimental apparatus: patent pending, all rights reserved, available only through this exclusive TV offer, no purchase necessary, void where prohibited, (and my favorite and probably most appropriate...) RESULTS NOT TYPICAL! /ubbthreads/images/graemlins/yellowlaugh.gif
"Official apparatus?" Well, that's easy... here goes.
Equipment needed:
1 milk carton, white interior
1 small can flat white paint
1 xacto knife
1 ruler
Aluminum foil
Clear tape
1 piece of printer paper
Scissors
Procedure:
Lay milk carton on side.
Cut along three edges of one long side (two long edges, one short edge) to make a long hinged panel and open.
Paint 2 coats of flat white paint on interior of carton.
When dry, place a small pencil mark "x" in the exact center of the side and close up the cuts with tape.
Roll box over so cut portion of box is on bottom, pointy side away from you, small flat bottom of box toward you.
In the exact middle of the carton side facing up (long side), cut a 2" diameter hole for shining the light through.
In the small base of the carton (facing you) cut a hole appropriate for your LUX light meter in the center of the base so it fits snug.
Wrap entire carton in aluminum foil, securing the foil with the tape. Place shiny side in.
Cut the holes in the aluminum foil to match the holes already in the box.
Cut a 1" by 3" piece of printer paper and tape to the edge of the flashlight hole on the meter hole side of the box so that it forms a baffle, preventing light from directly shining on the meter sensor.
Take a reading by shining the light straight down onto the "x" with the edge of the bezel level with the top of the box.
That's it!
Quickbeam's lightbox experimental apparatus: patent pending, all rights reserved, available only through this exclusive TV offer, no purchase necessary, void where prohibited, (and my favorite and probably most appropriate...) RESULTS NOT TYPICAL! /ubbthreads/images/graemlins/yellowlaugh.gif
milkyspit
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
[ QUOTE ]
Quickbeam said:
Quickbeam's lightbox experimental apparatus: patent pending, all rights reserved, available only through this exclusive TV offer, no purchase necessary, void where prohibited, (and my favorite and probably most appropriate...) RESULTS NOT TYPICAL! /ubbthreads/images/graemlins/yellowlaugh.gif
[/ QUOTE ]
You forgot: "Sorry, no COD." /ubbthreads/images/graemlins/grin.gif
[ QUOTE ]
Quickbeam said:
Quickbeam's lightbox experimental apparatus: patent pending, all rights reserved, available only through this exclusive TV offer, no purchase necessary, void where prohibited, (and my favorite and probably most appropriate...) RESULTS NOT TYPICAL! /ubbthreads/images/graemlins/yellowlaugh.gif
[/ QUOTE ]
You forgot: "Sorry, no COD." /ubbthreads/images/graemlins/grin.gif
Quickbeam
Flashlight Enthusiast
Re: The LightBox Experiments... Version 2.0 !
LOL!
Well, the Version 2.0 has undergone official inspection and apparantly has passed with flying colors...
Here's the inspector caught in the act of checking the integrity of the apparatus as well as the impact of any residue milk inside the carton:
Perhaps a MILK carton isn't the best base for the test kit... /ubbthreads/images/graemlins/grin.gif
LOL!
Well, the Version 2.0 has undergone official inspection and apparantly has passed with flying colors...
Here's the inspector caught in the act of checking the integrity of the apparatus as well as the impact of any residue milk inside the carton:
Perhaps a MILK carton isn't the best base for the test kit... /ubbthreads/images/graemlins/grin.gif
