But you just confirm my general statement that faster optics have larger central obstructions: F5 => 52%, and F7 => 48% obstruction. The faster F5 has a 4% larger obstruction. I've never come across any telescope design that from Newton to RC where the faster scope has a smaller obstruction. If such an optical design exists please let me know! I want to have such a telescope!
Concerning your first section. Now it becomes a problem with the CCD and not solely a telescope problem anymore. I was thinking about how this detection sensitivity could be solved only telescope-wise, but it obviously can't. So let's look at the pixels.
Generally faster optics have a larger "/pixel number; for a given pixel number, e.g. 2048 pixels, that increases the field of view FOV, therefore making fast corrected optics, wide-field optics. But, in principle, having a pixel scale of 2"/pixel instead of 0.5"/pixel means that more light falls into the same pixel. Take a look at page 39 of this link http://www.fcaglp.unlp.edu.ar/~observacional/papers/PDFs/phot_psf_merline-howell1995.pdf where two pixel scales are compared. You see that albeit much fewer pixels characterize the star, the light falling into the central pixel increases its ADU counts. Thus increasing the stars detectability for the telescope with larger pixel scale...meaning the telescope with shorter focal length, or in other words, the faster telescope (IF we assume a given unchanged pixel size without binning!). So despite the larger obstruction the detectability is not reduced in my opinion.
But where you would be absolutely right (not saying you're wrong above, but I still don't see why even a larger pixel scale would turn against a better detection sensitivity of a faster telescope) is, that the fewer pixel numbers characterizing the star, significantly reduces the precision in astrometry because the program has increased difficulty finding the correct centroid.