QUOTE (Mikau @ Oct 28 2005, 12:51 PM)

It would seem polishes don't work on my planet.

TAC is not the same material as Polymethyl methacrylate (Plexiglas) and polycarbonate (CD's and some anti-glare). You may have just confirmed that you have polycarbonate polarizer support films.
I wouldn't let the stuff dry and then polish it off. I would work with it in liquid form, it should cut quicker. As I understand it, dry polishing is meant for really smooth finishing. No matter what, if you work at it long enough you will get those scratches out. Of course, some materials don't polish (Lexan) but I don't think this is one of them.
Slightly off topic:Here is some idea of why so little light gets through the panel in the first place.The panel basically has these primary transmittance bottlenecks:
a. Reflections off the inside of the analyzer surface as light makes the transition from plastic to air.
b. Reflections off the outside of the polarizer surface as light makes the transition from air to plastic.
c. Total internal reflection caused by the anti-glare if oriented as the analyzer.
d. Single transmittance of the polarizer.
e. Parallel transmittance of the analyzer.
f. Dichroic efficiency of the color filters.
g. Transmittance of the color filters.
h. Transmittance of the various elements.
i. Masking.
Let's ignore total internal reflection as the removal of anti-glare, or standard panel orinetation should basically eliminate that. The parameter that is missing is diffusion in cenarios where the only useful light is light exiting at the normal (projectors). Removing the anti-glare largely takes care of this. I am also going to ignore the masking only because I am too lazy to work out what percent of the panel is masked (based on dot pitch)

.
So taking a simple slice of those parameters above that
really matter yields this estimate:
Air to polarizer interface: 96% transmittance (according to Nitto Denko).
Polarizer: Maximum 44.3% (Sanritz).
Color filters: 30% (deduciton based on 10 fairly equal parts to full spectrum white light, 3 of which we use).
Analyzer: Maximum 84% (3M).
Plastic to air interface: 96% (According to Nitto Denko).
Put that all together and you get:
100 * 0.96 * 0.443 * 0.30 * 0.84 * 0.96 = 10.2% transmittance.
So if a panel were to take choice of the optimal of each component we have found, and extract an exact cross section of each of Red Green and Blue (RGB), we would have a 10.2% transmittant panel. But I maintain that this is only the theoretical optimal. Without switching to less efficient color filtering, the transmittance can only go down from here.
But lets see what happens if the light that is not polarizeable can be 100% recycled. This is the case with commercial projectors:
The polarizer now takes on a theoretical tranmittance of 84% (same as optimal parallel tranmittance).
100 * 0.96 * 0.84 * 0.30 * 0.84 * 0.96 = 19.5% transmittance.
The light output has nearly doubled. This is one of the major reasons why a commercial projector is so bright. Just a breakdown of why I am excited about trying the light recycler roughly established earlier in this thread

.
Mark.