@Fraser, I've been reviewing this and your other FLIR teardown threads, as I completed my own teardown of my RS64 Thermosight (a rifle mounted thermal). This thread has helped point me in the right direction. The desire to disassemble started with what appeared to be battery failure, which can be corrected by plugging an external 12V battery into the device via USB. FLIR wouldn't service the unit since they discontinued that model, so I postponed any teardown, until I eventually dropped the device. Now, it's accuracy has reduced to about 5-7 minutes of angle (MOA), from its original 1 MOA accuracy (all other non-thermal components were evaluated/tested). So, the battery has failed, and now something internal is loose.
Regarding the battery, without an external battery, the unit will indicate fully charged, but will quickly indicate low battery and eventually shutdown. Multiple attempts to power the unit back on have mixed battery status levels and runtimes, but it always quickly drains and shuts down. I noticed my battery is exactly as yours, and was wondering if you or anyone has had any luck replacing the battery, whether as a whole, or rebuilding it from individual Li batteries? I've partially removed the wrapping from the battery pack to determine what it was made of, and I was surprised to find two additionally wrapped (metallic) battery packs, similar in size to 3- AAA batteries each. I'm not sure I should go much further into the batteries. I can tear down most things, and reassemble them, but batteries make me nervous. Any advice on their replacement? I searched a bit very late last night, and apart from a foggy eBay hit which is a gamble, I only saw speculations as to what the pack could be comprised of, all of which seemed incorrect.
Additionally, I found a very loose connector pin on the button input board, which passed continuity testing from board to cable-end while outside of the housing, but it should definitely have caused some type of problem. I have reseated it.
Regarding something loose... All major components of the thermal have been disassembled, and none of them were loose by any measure. That leaves the Tau 2 internals, or the LCD. Shaking the Tau only sounds like the multi-part shutter is loose as would be expected. I suppose I could dive into your Tau320 thread if I needed to dig into the core. There is a loose rattle in the LCD portion which I did not expect, but I would expect the display to show the video sent to it by the core in relation to the screen's overlay reticle. Any movement of the LCD (which is not apparent) should not affect the x,y coordinates of the displayed video. The screen rattle is where my gut tells me to look, but I assumed any error in accuracy should be introduced by the camera. What do you think about the origin of movement/loose component causing the accuracy issue?
Thank you for your threads and your time.