Well at least we have the same problem.
In considering what component may have failed, I have attempted to study how the DLP unit works. ( I have a strong EE background and alot of hands on experience with various electronics). So here's my version of how this works. Maybe this will help determine what needs to be fixed. (See http://electronics.howstuffworks.com/dlp1.htm
The Ti micro-mirrors ( DMD- digital micromirror device sometimes called the DLP chip) points various pixels of light to the screen via DDR CMOS SRAM elements under each mirror to be in an "on" or "off" position. This mirrror control signal is generated by a microprocessor which is likely on the DMD board since it must be close to the CMOS elements to get the morrors to respond fast enough. It takes 16 microseconds to create the conditions to "paint" an entire screen, ie all pixels for one moment of time that your eye sees ( ie integrated over like 0.15 of a second). I do not know how the microprocessor logic works - if if has a logical elements for each pixel on the screen or processes a row at a time and holds it in memory until the next refresh of that row is calculated. The later would seem more likely given the wide use of parallel processing in such microprocessors. This micorporcessor data output is coordinated with the color wheel to produce the right color for each pixel at the correct time.
If elements of the the micro-processor does nothing and remains in the "on" position, then the micro mirrors will shine all colors from the color wheel on the screen to produce a small white spot on the screen. If the mirror is in the "off" position there is light missing and there will be a small dark spot.
So the sort of light painting failure that we see, could arise from either a CMOS controller under the mirrors is not operating, or the microprocessor that controls the CMOS positions is not processing the signal for that pixel element. If the microprocessor works by processing and storing the on off information for each pixel, then what ever stores this information is messed up. Given the likely use of parallel processing in this processor, a failure of a storage element in this processor would create some kind of repeat pattern in the white spots that should be seem - but the white spots do appear random. (However, I really don't know for sure how the processor works.)
Solid state electronic failures usually are caused by over heating and a breakdown of the electrical materials. CMOS devices are more robust than other solid state technologies, but can still fail. Functional the DMD gets alot of heat and must dissapate it effectively. That's why the fan and huge heat sink fins are mounted to it. Any over-heating could eventually cause the CMOS devices (which tilt the mirrors) to fail or solder in the pins connectons on the board to melt One of the large chips on the DMD board also has a small heat sink and is likely the microprocessor. The fact that Samsung redesigned the pin contacts in the version 2 DMD boards implies they were having heating problems of the DMD and needed the DMD board to better dissipate some of the heat better as well.
So my money is not on the DMD board itself but on the DMD unit- the digital micromirror array that gets mounted on the DMD board. Note that the micro-mirror arrays were invented, promoted and liscenced by TI so they may have further information about them....???