Improving the DIY integrating sphere

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
I am truly amazed at the time and effort that many here have put into building integrating spheres. The cooperation seen in the past amongst the community in helping one another with calibration and builds has been one of CPFs high points. There has been a great deal of effort expended on deciding how best to build these spheres but I have found it intriguing that one of the most important aspects of an IS has completely been ignored. Sure there have been a few here and there that have made mention of the need for a proper coating but I know of no one yet that has actually done it. I know people here care about having as accurate a device as is possible so the only thing I can figure is that people just don't see the importance. Well hopefully we can change that.

If you do any amount of research into IS you will find how much they stress the importance of a good coating. Without the coating your IS is more of a S. than an IS. There are two main qualities you want in the coating. Diffuse reflectivity and it needs to reflect as much of the spectrum as possible. A standard flat white paint from the hardware store may fill the need for diffuse but it is doubtful it will faithfully reflect the spectrum of the light source. That would lead to incorrect readings.

Spectralon is the industry standard for coatings as it is an exceeding good diffuse reflector over a wide spectrum range. It is also exceedingly expensive and out of reach to the DIYer. So what to do?

In my research I found this article that mentions that barium sulfate powder alone on a surface actually performed better than Spectralon in the 425nm-600nm range which just so happens to be the sweet-spot for what we deal with here. Of course they rightly saw a need for a binder agent as just applying powder to a surface without having something to hold it all together and to the surface is not a good idea. So they mixed it in with a white latex paint. The problem is that the performance decreases as paint is added. This got me to thinking about the need to have the bare powder face out to get the best performance so I started to do a little experimenting. So far the best binder I have found is a white latex paint but I don't mix the BS into the paint but rather paint the surface and then while it is still wet dust it so that the powder adheres. This leaves a very flat(as in sheen) and diffuse reflecting surface. Perfect for an IS.

To show the difference in performance between the bare foam as is used in most of the DIY spheres I made a little demo and posted it over at my youtube account. I invite you to take a look at just what a difference it can make. Sorry that the sound seems out of step with the video. I have no idea how to fix that.:eek:

And here's the best part. It's incredibly cheap! Here is a link to a pound of the powder for only $6 plus shipping. This is an absolute must have in my opinion if you want your integrating sphere to be a true tool. So get cracking on those spheres.:thumbsup:
 
Last edited:

angelofwar

Flashlight Enthusiast
Joined
Nov 17, 2007
Messages
3,336
Location
South Carolina
Not to mention, there's an App to put a Lux-Meter on your Ipod Touch or IPhone...may be able to use this in the shoe-box method? Not sure how accurate it is, but it will definitely tell you which of the lights you have are brighter.
 

Glenn7

Flashlight Enthusiast
Joined
Jul 8, 2006
Messages
1,597
Location
Tasmania, Australia (the butt end of oz)
Mike,
I am so glad you did this as I have never really liked those Styrofoam sphere readings mostly because any photos I have seen of people's set ups show the light going/shining through onto the wall behind the sphere and that is light lost - even more noticeable with throwy lights and thats why some lights I know have not been given good ratings.
 

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
Not to mention, there's an App to put a Lux-Meter on your Ipod Touch or IPhone...may be able to use this in the shoe-box method? Not sure how accurate it is, but it will definitely tell you which of the lights you have are brighter.
Well I suppose it might be better than your eye but anyone who wants their measurements to be taken seriously should avoid using the iphone. It might tell you which one is brighter and then it may not especially when linked to a box. There is a reason they don't make integrating boxes.

Mike,
I am so glad you did this as I have never really liked those Styrofoam sphere readings mostly because any photos I have seen of people's set ups show the light going/shining through onto the wall behind the sphere and that is light lost - even more noticeable with throwy lights and thats why some lights I know have not been given good ratings.

Well don't get too caught up on the fact that some light escapes out the back. It does that with my setup as well although it is attenuated some. It would have in the test done in the link provided as well. A little light leakage will not change the fact that the powder is a diffuse reflector and faithfully reflects the spectrum. It does mean however that every effort should be made to have the layer of barium sulfate be uniform as possible. Otherwise it will make a little bit of difference in the consistency of performance between lights with very different beams.
 
Last edited:

angelofwar

Flashlight Enthusiast
Joined
Nov 17, 2007
Messages
3,336
Location
South Carolina
Yeah, I just got it for a quick reference, and don't plan to take "serious" lux readings...just thought I'd put it out there for others who may be interested :thumbsup:
 

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
Yeah, I just got it for a quick reference, and don't plan to take "serious" lux readings...just thought I'd put it out there for others who may be interested :thumbsup:
I have an iPhone so I looked it up. It doesn't get very good reviews. I'd do it if there was a free version but I'm not paying for it.
 

flashflood

Enlightened
Joined
Mar 9, 2011
Messages
608
I have an iPhone so I looked it up. It doesn't get very good reviews. I'd do it if there was a free version but I'm not paying for it.

The best one out there for iPhone is called Pocket Light Meter. Very good UI -- as in, there almost isn't one, it just works. I don't have a "real" light meter to compare absolute numbers, but it seems to be at least monotonic (brighter things measure higher) and consistent (same thing measured twice gives the same answer to within a few percent).

Please tell me I'm not the only person who's actually tried to get useful measurements from a cell phone and an Integrating Toilet Bowl.
 

QtrHorse

Flashlight Enthusiast
Joined
May 2, 2008
Messages
1,159
Location
Texas
That is a very interesting video. I knew the coating was important but I did not think it would make that much of a difference.
 

saeckereier

Newly Enlightened
Joined
Mar 22, 2007
Messages
138
I can't believe this thread hasn't gotten more response. I mean the video is a clear demonstration even without the lux meter it's just plain obvious what kind of impact the coating had. I would've guessed that all the DIY-sphere-makers would be congratulating you on the discovery..
 

jirik_cz

Flashlight Enthusiast
Joined
Jul 29, 2007
Messages
1,605
Location
europe
Interesting, I will probably build a new sphere with barium sulfate coating. Only problem is that I can find only 25kg or bigger packaging for sale locally :)

On the other hand I would like to see cross test between calibrated bare styrofoam sphere and calibrated barium sulfate coated sphere.

I did some tests with laser pointers and I was surprised how uniformly lit was the inside of the sphere. So if the styrofoam is diffuse enough for a laser it should be diffuse enough for flaslights.

spheregreenlaser3.jpg
 

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
Interesting, I will probably build a new sphere with barium sulfate coating. Only problem is that I can find only 25kg or bigger packaging for sale locally :)

On the other hand I would like to see cross test between calibrated bare styrofoam sphere and calibrated barium sulfate coated sphere.

I did some tests with laser pointers and I was surprised how uniformly lit was the inside of the sphere. So if the styrofoam is diffuse enough for a laser it should be diffuse enough for flaslights.

Using a laser is actually a really good idea to figure out how well the sphere works. Just remember that whether or not it is uniformly lit or not should be based on instrumented measurements not ones eyes. If the bare foam is in actuality a good performer than you should be able to move the aim of the laser around inside the sphere and have the light meter read the same. Have you tried that yet? What were the results?

But remember too that the diffusion is only half the issue. You still want a coating that faithfully reflects the spectrum.
 
Last edited:

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
I can't believe this thread hasn't gotten more response. I mean the video is a clear demonstration even without the lux meter it's just plain obvious what kind of impact the coating had. I would've guessed that all the DIY-sphere-makers would be congratulating you on the discovery..
Thanks. It may be that some have not seen it since I did not put it in the same forum as all the threads about IS(LED flashlights). I just don't feel the IS is so narrow in focus as to belong in that forum. Many people start their threads there because there is more traffic over there than there is here. Still undoubtedly many of them have read this thread. I don't require congratulations from them though. Hopefully at least some people will see the importance and make the needed changes.
 

MrGman

Flashlight Enthusiast
Joined
Feb 6, 2007
Messages
1,777
I certainly did not know of this thread until tonight. Very interesting video. But the differences in reflectivity of white styrofoam versus a coating of barium sulfate over white satin finish interior latex paint is not conclusive that the "calibrated" readings of the home made spheres versus the real integration spheres is that far off because its more reflective. Nor is Barium Sulfate by itself with a sweet spot of 425 to 600 nM the ultimate answer. Wbp and I were doing work with his very expensive spectral radiometer looking at the color spectrum of various LED sources. They certainly go well below 600nM especially the warm ones. Incandescents of course have lots of lot down into the red well below this range. A good integration sphere is supposed to reflect the entire spectrum of white light. I remember wbp was looking into this between the real sphere that he had and the styrofoam sphere, but to be honest, I don't remember the exact results. We were using various light sources that we had, what we didn't have was something to give a broad spectrum of "daylight" to compare the two. Point is there is nothing to say that the white reflection of a styrofoam sphere is any worse than the Spectralon used in real spheres as far as the color spectrum that its reflecting.

I had personally tried testing various individual colored lights in the professional sphere I had at work versus the home made spheres I put together, I had UV, blue, green and red. Now I wasn't using a $9K spectral radiometer at home to compare the 2, I was using the CA813 meter with photopic filter. Point is I got pretty close comparitive results. I had a lower cost meter and when I saw that the UV light source was reading more than the green one I knew something was wrong there. The UV light source should not show up through the photopic filter as visible light. It had about 0.1 lumens of visible light spectrum in the real sphere system at work and that's about all it had in my home made system with the CA 813 meter. But with the CA 811 meter it read about 5 lumens. I sent that meter back. The testing of the various color lights that I did was not conclusive either but it gave some level of confidence.

As to the question you raise about the foam being too reflective, creating a calibration "constant" number in the home made sphere by cross checking it with a light in the real sphere and then testing various lights I was able to get less than 5% error, or 95% accuracy. Going from a floody light to a thrower light in the professional sphere and then testing it at home in the stryrofoam sphere using the same "constant" number, I was able to get numbers that were within 95% or better agreement at home to the professional sphere. So correcting for the higher reflectivity in the styrofoam sphere just meant having a larger correction number. I think the key was to have a good baffle system to block direct light from entering the sensor port.

It would be very interesting to run a test of 2 sphere systems with the same sets of lights, one with the barium sulfate (since the cost of Spectralon is very prohibitive) and the other as is with just the original styrofoam, and after calibrating them with a known light source of a floody type, and coming up with a calibration "constant" number for each system (that number won't be the same for each sphere) to see if they both read the same lumens using the same correction constant for each system previously determined with a "thrower" type light. I would expect the pure foam unit would have higher divide by constant number than that of the barium sulfate coated unit, but that doesn't mean that it will work better when going from a soft flood type light to the thrower light. It might, it would definitely be with testing. Some one would have to do this with the exact same meter in two identical sized spheres with identical baffle systems in them but one with the barium sulfate coating.

Then after all of that is done, testing it with blue, green, red, Incan versus cool white LEDs in each system to see if the readings are the same or not, that would be a great test. Of course this is a lot of work, and that's part of the problem. Wbp and I at the time simply concluded that the plain styrofoam sphere was close enough.
 

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
I certainly did not know of this thread until tonight. Very interesting video. But the differences in reflectivity of white styrofoam versus a coating of barium sulfate over white satin finish interior latex paint is not conclusive that the "calibrated" readings of the home made spheres versus the real integration spheres is that far off because its more reflective.
No the test was not the be-all-end-all. It does dramatically show the inherent problems with bare foam and that being specular reflectance. This is the main problem I see with the bare foam. The surface must come as close to a diffuse reflector as possible and the bare barium sulfate blows the bare foam away in that regard. It is this quality or lack thereof that is the major contributor to lack of consistency between lights with a different beam shape.

Nor is Barium Sulfate by itself with a sweet spot of 425 to 600 nM the ultimate answer. Wbp and I were doing work with his very expensive spectral radiometer looking at the color spectrum of various LED sources. They certainly go well below 600nM especially the warm ones. Incandescents of course have lots of lot down into the red well below this range. A good integration sphere is supposed to reflect the entire spectrum of white light.

Keep in mind that what we care about is what we see. The closer you get into the infrared the less and less it matters. When I said "sweet spot" I was referring to the fact that bare barium sulfate(heretofore indicated as BS) is better than the industry standard across that region not that it just drops straight off either side of the range mentioned. In fact if you look at the chart you will see that at the very edge of a human's ability to see in the infrared the bare BS is still 97.5% the reflectance of the Spectralon. More weight however should be give to the part of the spectrum that we see more in and in that area the BS is superior.

For the price I'd say BS is the ultimate answer.


I remember wbp was looking into this between the real sphere that he had and the styrofoam sphere, but to be honest, I don't remember the exact results. We were using various light sources that we had, what we didn't have was something to give a broad spectrum of "daylight" to compare the two. Point is there is nothing to say that the white reflection of a styrofoam sphere is any worse than the Spectralon used in real spheres as far as the color spectrum that its reflecting.
It is hard finding info that is not buried in purchasable-only publications.
70d81b56.jpg



As to the question you raise about the foam being too reflective, creating a calibration "constant" number in the home made sphere by cross checking it with a light in the real sphere and then testing various lights I was able to get less than 5% error, or 95% accuracy. Going from a floody light to a thrower light in the professional sphere and then testing it at home in the stryrofoam sphere using the same "constant" number, I was able to get numbers that were within 95% or better agreement at home to the professional sphere. So correcting for the higher reflectivity in the styrofoam sphere just meant having a larger correction number. I think the key was to have a good baffle system to block direct light from entering the sensor port.
Yes you may have a baffle to block direct exposure but the first order reflections can bypass that baffle. The best test of whether or not the diffuse coating matters should be borne out testing with a laser and moving it around. Here is my point though. Why use a band-aid instead of just building it correctly? I am setting up my sphere so that absolutely no math has to be done. What the meter reads is what it is. I say that with one caveat. I may be setting up a system whereby I insert the light into the port and then use a separate light to measure the lost lumens to the inserted object. So I might have to do some very basic math but only to have extremely accurate measurements.

It would be very interesting to run a test of 2 sphere systems with the same sets of lights, one with the barium sulfate (since the cost of Spectralon is very prohibitive) and the other as is with just the original styrofoam, and after calibrating them with a known light source of a floody type, and coming up with a calibration "constant" number for each system (that number won't be the same for each sphere) to see if they both read the same lumens using the same correction constant for each system previously determined with a "thrower" type light. I would expect the pure foam unit would have higher divide by constant number than that of the barium sulfate coated unit, but that doesn't mean that it will work better when going from a soft flood type light to the thrower light. It might, it would definitely be with testing. Some one would have to do this with the exact same meter in two identical sized spheres with identical baffle systems in them but one with the barium sulfate coating.

Then after all of that is done, testing it with blue, green, red, Incan versus cool white LEDs in each system to see if the readings are the same or not, that would be a great test. Of course this is a lot of work, and that's part of the problem. Wbp and I at the time simply concluded that the plain styrofoam sphere was close enough.
I haven't coated my sphere yet so I can do some tests but like you I am very busy so don't necessarily have the time to do the full test before I need my sphere done. I will see what I can do.
 
Last edited:

MrGman

Flashlight Enthusiast
Joined
Feb 6, 2007
Messages
1,777
This is all good to know, and will be very interesting to see what you find out. I know that there are businesses that do test services for a fee. Last time we checked into it the cost was around $250.00 per sample. I would say that having at least 2 lights tested as standardized references would be a good start. Good Luck.
 

jirik_cz

Flashlight Enthusiast
Joined
Jul 29, 2007
Messages
1,605
Location
europe
If the bare foam is in actuality a good performer than you should be able to move the aim of the laser around inside the sphere and have the light meter read the same. Have you tried that yet? What were the results?

I've tried that, but my cheap DX lasers have very unstable output so I can not get proper results.
 

MrGman

Flashlight Enthusiast
Joined
Feb 6, 2007
Messages
1,777
to have a good stable laser do a test inside the surface of a foam sphere and see if the readings are fairly constant would be very interesting. Having seen the inside of the foam spheres I would concede right off the bat that they are not all that smooth in texture to give a consistent output reflection for a true pin point light source. There are tiny little recesses in the foam sphere. It would be interesting to get one of these really good focusing small flashlights that have a very tight beam and do the same test, moving it all around and see if the reading remains constant or not.
 

Got Lumens?

Flashlight Enthusiast
Joined
Mar 12, 2011
Messages
2,458
Location
Champlain Valley
I haven't coated my sphere yet so I can do some tests but like you I am very busy so don't necessarily have the time to do the full test before I need my sphere done. I will see what I can do.

Saabluster, I would suggest going the course you are with one deviation. That would be to do the first test on your unmodified(non-BS coated) IS with a couple of known sources, one floody one spot? Then as time permits finish the unpgrade and retest. The results should be almost identical to doing two consecutive tests with identical spheres.
GL
 
Last edited:

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
Saabluster, I would suggest going the course you are with one deviation. That would be to do the first test on your unmodified(non-BS coated) IS with a couple of known sources, one floody one spot? Then as time permits finish the unpgrade and retest. The results should be almost identical to doing two consecutive tests with identical spheres.
GL
Well that is not really a deviation as I was already planning on that. I just wasn't planning on doing a test with a bunch of different colors.
 

ti-force

Flashlight Enthusiast
Joined
Jan 14, 2009
Messages
1,266
Location
Georgia, U.S.
Saabluster,

Thanks for spending your time to do these tests. Do you still have the two small pieces of foam set up with your meter mounted to try one more test? Could you scuff the surface of another piece of foam with sandpaper so it takes some of the sheen off the surface of the foam, then test to see if that makes much difference? Try a fine grit paper.
 
Last edited:
Top