No, it should be white and as flat as possible, ideally a photon hitting the surface from any angle has an equal chance to leave again at any angle (so-called Lambertian reflection), that is key to the integrating properties of an integrating sphere.
So, why not paint the interior with a flat white acrylic paint to mimic ceilings vs an ideal laboratory condition? I get that companies want the best possible ratings for their products.
You could do that. For a discussion on integrating spheres you could follow the links in my sigline to some posts that I wrote on my journeys towards a working sphere. A very nice read on integrating spheres in general, if you are into that, can be found here: https://www.labsphere.com/support/system-product-brochures/tech-guide-a-guide-to-integrating-sphere-theory-and-applications/
So I ran those checks on “light leakage” between my old unpainted sphere and the new painted half round foam. The pained foam (4 coats now) is definitely blocking light! I took my Convoy M1 XM-L2 (high at 3.04A) pressed it right against the foam on the inside and placed the luxmeter sensor right against the foam on the outside directly across from the light. Here are the results:
Unpainted foam
Low - 4,280 lux
High - 12,920 lux
Painted Foam
Low - 118 lux
High - 337 lux
Some pics (painted foam is the one where the light looks yellow):
My dilemma now is how to mount the luxmeter. This time I’m going to put the sensor hole 90º away from the light input hole. I’m going to open up the sensor hole completely to the inside (the wall is now painted, so I pretty much have to). I can either mount the sensor directly at the hole (the face of the sensor will be set back from the inner sphere wall since I can’t get it fully into the foam without some extreme cutting), use PVC pipe to “couple” the sensor to the sphere wall (inserting the PVC until it’s flush with the inner wall and painting the inside of the PVC matte white also) using either straight pipe, a 45º bend, or a 90º bend. With the bends I’m guessing I may get away without using a baffle inside the sphere. I’m concerned though about light reflecting off that PVC even with it painted matte white except maybe in the case of using a piece of really short straight piece of PVC. Any thoughts on this? Some mock up pics:
90º Bend:
45º Bend:
Straight Section:
By the way, anyone looking for glass like I have been, search for “clock glass” in order to find various size round flat glass that is readily available and doesn’t cost an arm and a leg.
-Garry
Still working on my new sphere (slowly). I happened to take my company truck to get a new windshield installed and they were able to cut me glass to size, so I got (2) 4 1/2” diameter pieces for $5. Now I’m struggling with what to do about a baffle. I think I read that it’s best if the baffle is made of the same material as the sphere, or at least has the same coating. So I’ve picked up a flat piece of this same “smoothfoam” and have painted it matte white on all surfaces. This material is about 23mm wide, so it’s kinda big (thick). I was thinking to mount it using small wood dowels glued into the baffle and into the sphere (dowels painted matte white as well). My biggest problem is coming up with the proper shape/size. I think I read somewhere that light should reflect at least twice before reaching the sensor. Well if this is the case it seems you’d need quite a large baffle, because you’re not only blocking direct light from the entrance hole, but also beams from that first reflection. Also with the sphere surface being round (and hence a “sphere” - lol), you’d have to block from numerous directions.
What about this idea for a baffle?
Or is this layout preferred? This seems to be the more common approach. And is the 2nd baffle necessary?
Will it matter how close the baffle is to the sensor? Obviously I won’t put it extremely close, but is there a point where it’s too close and affects the luxmeter sensor’s readings adversely? Any help/guidance is appreciated!
Thanks,
-Garry
Aarrgghhhh! I wonder how widespread this issue is (percentage of people that can’t view my photos). I finally settled on using Google to host my pics (I wasn’t happy with alternatives I found and Google is super simple to utilize) and now run into people not seeing them. This might be better suited for a new/different thread rather than being discussed here.
-Garry
I can’t see any pics in #115 and #116.
The baffle is just to prevent direct light and should be opaque.
Ok, I finally see the issue by using another browser on my phone. I’ve looked in Google Photos for sharing settings, but don’t see anything to change. And googling the issue doesn’t lead to anything helpful. Guess it’s back to the drawing board for photo hosting.
-Garry
Ok, any ideas on a material to use that’s stiff enough to stay in place? And I guess styrofoam is not a good choice.
-Garry
The baffle position is not critical but should be somewhere in the middle between entrance hole and sensor, and as Jason pointed out it just prevents direct light from light source to sensor so that detected photons have undergone at least one reflection (not 2). Ideally, the baffle should be as small as possible while still blocking all direct light. This is all discussed in my integrating sphere threads before, and in the sphere theory article from Labsphere.
If styrofoam is convenient, use it. Just glue a solid material over it, like posterboard, thick plastic or metal painted white. Just something to keep light from going through.
Thanks guys. It’s a lot easier to block direct light only. Poster board painted matte white sounds simple enough. I’ll definitely err on the side of making it too big (but not way too big).
-Garry
I just made one of these using a 11.9” OD hollow half balls from Amazon. Made a 70mm flashlight entry hole. I want to reiterate the importance of rigging it so the luxmeter is in the exact same position each time you use it, for repeatability. This was the most difficult thing to get right for me. Thanks djozz!
With the right paint coating, maybe you could make a sphere out of two of these 14” bowls?
Pretty cheap from ikea 
Should work nicely and will not melt from high power flashlights, but start with a good undercoat that sticks to SS.
Right, good point.
Wasn’t there some CPF post about a special powder that was good for integrating spheres?
Maybe a white/clear epoxy coated with the powder would be enough to stick to the SS if you sanded it first?
There is a whole science to that, and the knowledge level on cpf was not so high. The powder is bariumsulphate, which is widely used in professional integrating spheres. The suggestion on cpf was to “sprinkle” a layer on top of a still wet latex coat. But the literature says that BaSO4 is actually quite transparent for light, and that you need a layer of the stuff that is several mm thick to obtain enough reflectivity. So you need a very clear substrate to mix it in and I found that PVA (polyvinylalcohol) is sometimes used for that. I tried several layers of this mixture in a test sphere and does produce a fairly solid coating, but upon drying some fine cracking does occur (which I reckon does not significantly affect the reflectivity properties).
But no coating that I checked sofar leaves the tint of the light intact, the CCT drops at least 500K, which is mainly caused by reduction of the blue peak. This is a cause for inaccuracy: cool tints have a high blue peak and will be more affected (read lower) than warm tints. This is also dependent on how well your sphere integrates: the better integration of the sphere, the more reflections, the more tint shift. Integration works against preservation of tint.
But that is also true for professional spheres, they cause tint shift just as well. The way to compensate that is to make your detector a spectrometer instead of a luxmeter (which is maukka’s setup), then you can produce a spectral difference characteristic for your sphere and calculate the tint shift out of the result.
From what I’ve read, PVA doesn’t work well with thick/several layers as it will crack, if acrylic is suitable optically then it is preferable for thick layers.
I think that the main thing to look for is something that does not yellow over time. All acrylic white paints that I used in the house do show yellowing. But that could be other ingredients too, no idea. Latex paints do not show visible yellowing but that is not a clear substrate, you can mix it with BaSO4 but the white latex itsef also affects the reflectivity and does not reflect as uniform over the spectrum as BaSO4 (there have been measurements of that).