floating lumens! my integrating sphere project

!!

[quote=Illuminaria]
!![/quote]

LOL, I was wondering already if anyone noticed that I stopped working on it. I still have to do some tests on it before I start using it but started doing other things. Thanks for trying to keep me on the job :-)

Light leaking through the wall of the sphere is not of much concern… your calibration constant for converting lux to lumens will take care of that. And the same goes for external light getting into the sphere. Just subtract the “LED off” reading from the “LED on” reading. Even less of a concern with your sphere-in-a-box that blocks external light.

In fact the Sphere-o-Many-Mysteries has a ultra-bright light testing mode where I mount the lux sensor to the outer wall of the sphere and use the styrofoam wall as a light filter. It can handle around 600,000 lumens that way. Normally the lux sensor (a Taos TCS3210 color sensor green channel) looks into the sphere through a 1/8” diameter hole in the sphere wall. The sphere wall can act as the baffle, but I do have a piece of white cardstock as a baffle. The sensor is mounted around 45 degrees from the light port.

I saw some tests that showed the just sanding the gloss off the inner surface of a styrofoam sphere performed pretty much as well as a barium coating.

Any updates? I really like seeing these projects. It's a kick in the butt for me to finally getting around to making one. I really should build one since I don't have a convenient space to measure throwers with my lux meters. I even went to the hardware store to buy the parts to build an integrating plumbing pipe, but they didn't have the parts that other folks here have been using. It looks like I can get 30cm styrofoam hollow spheres online though, so I'll go that way. That part is easy though. It's calibration that I'm not looking forward to.

I have precious little time for the hobby at the moment (writing posts is a bit easier than actual hobbying, I can only really work on things when the family has gone to bed and in the last couple of weeks I have been too tired to go hobbying late at night)

Concerning the integrating sphere, I'm still measuring/troubleshooting to get within the 1% accuracy that I aim for. The latest problem is that the output readings of the same flashlights and leds are slowly declining over the last couple of weeks. I already found out that it is not a battery problem of the luxmeter. Leftover is 1) declining reflectivity (yellowing?) of the coating on the inside of the sphere, which would be disastrous but I do not consider this likely, and 2) slowly peeling off of the piece of tape with the hole in it that I used to decrease the sensitivity of the sensor. So next step is to dismantle the IS and have a look at the piece of tape, I really hope that that is the problem because that makes for a relatively easy fix (not use tape but something that stays in place better).

By the way, if I would make another sphere I may not bother with a coating (just stick with careful sanding of the styrofoam inside of the sphere with fine grit paper to decrease direct reflections), the integrating properties and angle insensitivity of the sphere without coating appear to be ok already without one. Also I would perhaps start with a bigger ball, in my sphere I really need the adjustment light to compensate for the with every different flashlight varying reflectivity of the hole area of the sphere, but with a considerable bigger sphere (like: bigger than 50cm diameter), the effect of the hole may be more negligable.

Real calibration will still be a problem, also for me. The only way for me is measuring as many known lightsources as I can, and make a sort of 'average best conversion factor'. I already found that the ceiling bounce lumens that I collected for the several lights relative to each other do not exactly match with the integrating sphere readings, the reason may be that the ceiling bounce method is more or less sensitive to the detection angle of the luxmeter (in other words: the beam profile of the light source), in the IS the angle is obviously not supposed to matter.

Nice build! Can get barium sulphate on Ebay. I will try to use some if I make a sphere, though I’d like to see before and after figures.

Was the meter really overwhelmed at 200000 range without the duct tape?

I will post some before and after coating figures, worrying is that I actually measured a bit less after doing the coating. I like to think that is because before coating the reflection was less 'lambertian' and relatively much of the first reflected light reached the sensor (dream on, djozz), but unfortunately the easiest explanation is that the reflectivity of my coating is worse than that of the bare styrofoam :-( (but hopefully at least more equal across the wavelength spectrum)

Before the piece of tape the maximum measurable amount of light was about 720 lumens. I was able to do some awfully precise moonlight measurements before taping the sensor :-)

the various posts, the large variations on calibration methods and results is suspect. A common BLF calibration LED or source would help make the various lumen claims a more accurate.

A known common binned LED at an agreed exact voltage or using the port size method for an across the forum calibration method if you will.

I like the idea: so we want a BLF version of a calibrated light source (cheaper, and a bit less accurate is no objection ;-) ). I tried to make one ( https://budgetlightforum.com/t/-/25593 ), but for cross-BLF use an stock light would be better. During the measurents I found that the Sunwayman D40A (mine is cool white) on the high setting (not Turbo) is extremely flat regulated, as good as the reference light that I built myself. And many BLF-members own one. So if we assume that SWM uses as identical emitters as they come (not true probably), that could be a good source?

I will also have a look into a cheaper light with flat output.

You really can’t use bin ratings to get decent accuracy. LEDs from the same bin can vary over 7%.

My reference lights are an ensemble of LED light bulbs from reputable makes (Philips, Sylvania, etc). Their lumen numbers tend to be quite accurate, but I have noticed that bulbs made later than others (with the same part number) can have higher lumen outputs than the published numbers… you can get more than you paid for due to ongoing improvements in LED manufacture. The bulb makes don’t upgrade the bulb specs on older model bulbs as they improve their products. It is best to use bulbs models that have been recently released.

I updated post #2 with a report of several initial measurements and alternations that I had to do before the sphere could be put to use.

bump

Hey, I spent two and a half hours last night writing post#2 of this thread. Has anyone even read it???? ;-)

(EDIT: it is officially #1 I see now, technically the first reply)

Yes…

BTW, I spent last night adding a 8 channel 16 bit A/D converter to the Sphere ’o Many Mysteries. Before I was using the 10 bit A/D’s on the processor chip. Now I can measure 50 volt and 20 amp signals with around 100 micro-volt resolution (averaging readings over a second) and around 0.01% accuracy…

BTW, pasting up a board with parts that have a 0.5mm lead spacing is a royal pain.

Once coated in latex the barium sulfate is now encapsulated?

I read where this was sprinkled over wet paint for 100% barium sulfate top layer if possible.

I’m about halfway on my dome until work let’s me return home in two weeks…

cheers

Thanks for reading texaspyro and Illuminaria !!

tp: I am not sure if I understand it well (I am more an analog person), but it sounds very accurate :-)

Ill: good luck finishing the dome! I read about the sprinkling technique too, I believe that was Saabluster who wrote that. I am not sure if what he did would be much better than mixing the BaSO4 with the paint and then sanding it, that would expose the BaSo4 particles just the same. And as I understand it, you need a fairly thick layer (like: more than 2 mm) of BaSO4 to get maximum reflectivity (in my sphere the layer of paint/BaSO4 is at least a mm and it still clearly lets some light through), which implies that much of the light is not directly reflected from the very surface, but the light enters the layer, bounces around, and eventually comes out again. So sprinkling a thin layer of BaSO4 on the wet paint will not make maximum use of its properties, in fact the reflectivity of his layer comes mostly from the white paint under the BaSO4 layer (the reflectivity of flat white paint is already very good so you will not easily notice the merit of the BaSO4 without accurate spectrofotometric measurements).

The barium sulphate is not meant to increase the reflectivity (it will actually decrease it for a lot of surfaces). It is meant to provide a diffuse reflective surface with a constant reflectivity coefficient over a wide range of light wavelengths and eliminate reflective “hot spots”.

For styrofoam spheres, just sanding the inner surface of the sphere with fine sandpaper (I used 220 grit) to get rid of the surface sheen works very well. I have some barium sulphate, but have seen no need to use it. You also need to be concerned with the reflectivity/aging characteristics of the paint that you use as a binder. For most people the barium is more trouble than it is worth…

I tend to agree, I found the 'angle-independance' a tad better after the BaSO4/latex coating was applied but it was already very good with just the styrofoam surface (after careful sanding with 800 grit paper). About aging of the paint: time will tell but in my house the latex wall paint still looks very white after years, while terpentine based white paint has turned a bit yellow.

One issue with paint in the sphere could be heat induced yellowing of the paint… particularly if testing high power lights over time. A high power beam can be quite warm.

Hi
I am interested in integrating sphere reflectivity. I am using integrating sphere of 20 cm diameter (aluminium) and coating of BaSO4 (0.4 mm thickness). Is coating thickness count for reflectivity? For one silicone photo detector what should be the size of sphere?Can u help me please to find the percent reflectivity of sphere?
best regards
Saira Saleem

Knowing the reflectivity of a sphere does not really help. There are lots of other variables that come into play. You really need to calibrate your sphere against a known light source.