I do not believe you, remove everything 
The proverbial skeptic ... there is always one in the audience. Here's a nice story ... Two weeks ago when I gave the talk it was actually not at a space conference that I am more used to, but rather, a medical conference (a large gathering, mostly pathologists), where I demonstrated the applications of my instrumentation in cancer research for the high-speed detection of malignant cells. My talk was the last one, but during the day I noticed one fella who would challenge the speakers with curve-ball questions. When I finished, I got mine too. But mine was a bit more special because it was less science and more what I felt between the lines as him being threatened by technological advances and the potential of his job being replaced: "I can't see a use for this", something seemed to me to as automated checkout machines replacing cashiers. I knew straightaway what to say to pacify and appease him, and after I was done, I got a nice smile and a very gracious, "Ok, we'll see what you have for us next year." For whatever reason, the bottom line is folks like you and him keep us honest.
So for you here is my answer/comment ... The MECHA connector from the ORIGINAL MECHA (still mounted on the frame) is disconnected and has been since the beginning. The CCD that is in that MECHA, has always remained connected, and still is. So before I can remove the ORIGINAL MECHA from the camera frame, the first issue is how to handle the CCD upon *its* removal. Since a 1-micron dust particle can easily cast a show on a pixel, a small completely contamination-free enclosure has to be thought-out and fashioned, and then the CCD can get transferred immediately upon removal in a zero-dust environment (Class-A).
The SACRIFICIAL MECHA (the one sticking up on top), still has its MECHA connector in place, with the hair-thin breakout wires soldered as we did a month ago. The reason is that during development, the mainboard needed electrical access to *both* the SACRIFICIAL MECHA or the MCU, in alternation. So when you disconnect the Canon components, you have to plug in the MCU, and vice-versa. Thus during solution development, the code for each component was handled one-at-a-time, until all Canon components ended up disconnected leaving the mainboard only connected to the MCU // the present state. That means at present the MCU is borrowing the MECHA's connector to communicate with the mainboard, but all MECHA components are disconnected. To have both connected at the same time would corrupt signals and nothing would work. So yes it may appear from the setup illusion in photo that I'm jiving!
Next, is to cut SACRIFICIAL MECHA ribbon cable near its connector, leaving only the connector, and make sure that no traces short from one side to the other. Tricky job. Following that, the design I am thinking of will scavenge the IS ribbon cable (~4 cm) and its female connector to bring signalling from the MECHA connector to the MCU.
The biggest challenge is not to make a mistake in the wiring (MECHA connector -> ribbon cable -> female conmector -> MCU (and putting SMT level-shift resistors on the bottomside). If I make a mistake, and in the worst case blow the mainboard, it does not invalidate the solution from the last three months of work. I have a backup camera for that, just in case.
So the next time you see photos and an improved movie, you will see a match-box-size (literally) camera, tethered to the USB 2.0 machine, CCD hanging off on the side, and a movie of everything working together as a system, including CHDKPTP live on the computer screen.
In addition, I have to attend to electron microscopy prep this week, that I have neglected along the way. So "removing" is not a simple matter. Does that help mitigate your disbelief to some degree for now?
