Byrd's idea sounds good to me. It's been a while since I took one apart but I think there is a thick semi clear plastic light spreader panel, a thin plastic diffuser sheet, a thin plastic polarizer, then the LCD guts, from back to front. That polarizer sheet needs to be at the correct angle relative to the front bonded polarizer in the LCD to get good extinction (blackness) at all points excluding where the pixel field voltage is playing with the twist of the LC material. Perhaps the plastic sandwich was put together with the polarizer slightly cocked, or even worse if mechanical keying does not prevent, in the correct layer order but reversed front to back so that any angular error is doubled. And it would be really bad if the diffuser and polarizer were order interchanged! If you decide to do some investigation, you can just view each individual sheet with polarized sunglasses on to see which one does the polarization function.
That doesn't sound right to me. Two linear polarizers that aren't aligned correctly cause an apparent dimming of the light passing through them, they don't create a grid pattern. So if the polarizer behind the LCD were somehow out of alignment by rotation (not merely translated - that should have no effect), I would expect to see an overall dimming of the display brightness, not this grid pattern.
I suppose that could be a possibility, but I don't understand how that would cause this crosshatching pattern. In that case, shouldn't the display appear to be washed out as too much (unpolarized) light is getting through?
Depends on whether the particular LCD uses a black masking matrix everywhere but over the pixels. If it does, only the pixels can light up, and a grid pattern around all good pixels is impossible, only pixels misbehaving in vertical and horizontal patterns surrounding apparent good pixels but uniformly across the entire screen.
