Happy to report that with the help of rockstarbill I've tested a reducer that works very well with my 12.5" AGO iDK and APS C QHY268M. Enter the Takahashi .73X ReducerCR! The reducer is designed to work with the corrected Mewlon 250 CRS and the FSQ106. Nothing fancy needed to make it work. I threaded it on the NiteCrawler and bam! The 2132mm f6.7 iDK now solves to 1563mm for .733X reduction, and a very pleasing f4.9 focal ratio. This yields a image scale of 1.0" (.99" actually). So now the reduced fov with APS-C is very close to native focal length with a full frame chip. Still a bit smaller, but a nice fov gain for sure. Effectively this added a new telescope to my lineup for $700.
Having previously tried the 27TVPH with disappointing results, I'm very satisfied with the field produced by this combo! Pixel-Peepers like myself can find some very slight error in the corners, but IMHO its nothing of concern, and I probably have a small amount of tilt I could work out if so inclined. I'm not sure how well it would hold up to a full frame chip, though. You might get a little extra, but I doubt complete full frame coverage would be in the cards. I imagine, being a reducer and not a flattener, that you can vary the spacing between reducer and chip to vary the amount of reduction. So perhaps by adjusting the spacing to achieve for example .8x reduction, you may improve field coverage for a larger chip.
As for me I'm happy as a clam with this combo and have started my first project with it. The out of the box results were great, and the reduction revealed I had a small collimation adjustment to make that I didn't notice before. The post-collimation adjustment 120s red frame I'm uploading is excellent imo. I'm linking a folder with raw data and stacks for your perusal. I do not intend to use bin1x1 with this camera/scope combo because the pixels are so small they resolve small reflections/halos that are (arguably) a result of central obstruction optics, good seeing, and tiny pixel low noise cameras. I think the winning recipe for me is to bin2 and then drizzle if more resolution is desired. I've included a few drizzle integrations to compare to the standard integrations. JFYI guiding was horrid with the fish data. I think the seeing took a dive. Guiding was good for the cygnus data.
Full datas here: https://1drv.ms/u/s!...PjUo4Q?e=bYu90d
The 120s star frame is red filter bin 2. The Cygnus wall is about 1.5 hours of red data bin 2. The Fish is 2.5 hours of Ha data bin 2. No processing other than calibration of the fish and cygnus data. 120s sub is raw. Look forward to hearing everyone's thoughts. This could prove to be an option for CDK owners that don't need full frame coverage.