Sorry to hear that!
The instrument I made during my time at Philips also had problems with repeatability. The team that made the separate measuring head (with suction cup) started with placing the three sensors at an angle so that they pointed at the same pixels on the CRT screen to measure but they soon found out that a rotation of the head gave a different result. Eventually they ended up using a piece of milky, acrylic plastic in front of the sensors, just as you have in mind. The calibration took care of the slight miscoloring from the plastic. I hope that a piece of plastic will solve it for you.
I should probably try that too. 
I had been thinking of trying to attach a really white piece of paper or something at a 45 degree angle and bounce light in, but it’d be physically a lot easier to just stick a translucent sheet of something on the front of the measurement device.
Ok, thanks for the honest evaluation!
Guess its time to start trying to convince myself I could actually get a decent ROI on the Passport.
Thanks Again,
Michael
I have some stuff packed away somewhere that might work well. It’s a diffusion filter used in iPhone touchscreen/LCDs. Helps to equally disperse the light from the 4 or 5 LED’s mounted on the perimeter. Works really well at diffusion, will turn almost any light, even those with fairly tight beam, into more of a mule style beam pattern, all spread.
If you think that might help I’ll try to dig through my packed away stuff at storage and find you some.
-Michael
Ok, I bought a bottle of PVA (it is sold as a mould release agent for polyester and epoxy) and did a first experiment: I made a mixture of 20ml PVA and 15 grams of BaSO4, which appeared a good thickness to apply with a brush (about the thickness of latex wall paint), and applied two layers on a flat test-piece of styrofoam, drying the first layer in the sun before applying the second layer.
Observations:
-it applies really easy in a thick layer, much better than the BaSO4-latexpaint mixture that I used before. The second layer applies well too without disturbing the first layer.
-the result feels very durable, it does not scratch off at all with my finger nails.
-it is even somewhat flexible, does not crack when the styrofoam is bended a little.
-the two layers do not let much light through: compared to the bare styrofoam (2cm thick), the light through the styrofoam+2layers is reduced by 80-90% (estimation by eye).
-the coating looks really matt, like you want it to be for an integrating sphere coating.
-for an actual integrating sphere coating I would use 3 layers, also to get a smooth finish.
-a very rough first estimate is that when applied like I did, for a 30cm inner diameter styrofoam sphere, you need about 500gram BaSO4 and 650ml PVA.
I have not measured the big questions yet: if this coating improves the reflectivity of a sphere (my former latexpaint/BaSO4 coating even reduced reflectivity compared to bare styrofoam), I have to test that in an actual sphere, but when shining a flashlight at various angles suggests that reflectivity may just improve a bit.
So far so good!
Thats fantastic news. I am excited to see how it performs in the sphere as well.
djozz,
When you are using this setup and you take a light and move it in and out of the ball deeper does it effect the measurement? For example if you stuck the bezel of a convoy L6 in there like an inch deep?
Yes it does matter how deep it is in, but there is a maximum and that is what I take as the correct measurement. If the light is not deep enough in some light is lost that is not entering the sphere (thus a lower measurement) and when it is too deep, the flashlight body/bezel (usually black) is absorbing light and lowes the reflectivity of the sphere (thus a lower measurement).
These are not huge differences by the way, if you are in about the correct position and move the flashlight a few mm in or out, the reading stays well within a percent.