I use 135mm tube lenses with my Mitties; an MC Sonnar for marginally better image quality or a Vivitar/Komine to avoid the slight corner clip (vignette) of the Sonnar when running full frame. Both work perfectly on APS-C as the FoV doesn't stray too far outside each Mitty's specified image circle, whereas FF does cause noticeable image degradation toward the corners (albeit still eminently usable).
The down side of APS-C is a significantly reduced FoV compared to using FF. It's irritating when subjects don't quite fit and I either have to revert to the drudge of FF imaging, or drop to the next Mitty down (which is often too big a drop, wasting space in the frame). So I decided to find a shorter tube lens that would compress most of the full-frame FoV onto APS-C, but not quite all of it. My goal was to exclude the FF clipped corners of the Sonnar and not go quite as far into the "out of spec" territory where Mitty images degrade.
None of my existing lenses fit the bill as they either vignetted massively, or were simply too big for the rig. They were all too far from the "just over 100mm" focal length I figured was optimum anyway. So looking around, I came across a Pentax 105mm f/2.8 Super Takumar which is the ideal focal length and has the same filter thread as my existing tube lenses. Despite the one mention on this forum being negative, I decided to buy it anyway. The lens arrived today and I've only had time for a couple of test stacks, but it's looking pretty good.
First the negatives. This lens renders very flat, low-contrast images. Easily fixed in post of course, but it robs some dynamic range meaning contrast can't be pushed quite as aggressively to pull out more detail (when needed). There's a super-multi-coated version that's far better, but it's expensive and more than I wanted to spend on a speculative test. The lens exhibits quite strong field-curvature too, but that's irrelevant when stacking, and may be the cause of the "surprise benefit" (more on that later).
So I put each of my Mitties on it and visually checked general image quality across the frame and into the corners (APS-C mode enabled). Perfect with every one (5x, 10x, 20x and 50x)! Image scale is more to my taste too with far less over sampling at higher powers. Really nice crisp images. I used a random butterfly wing glued to some card as the test target, lit by a single flat panel lamp. It's not the ideal choice as the scales on this one are quite small and tend to be a bit "sparkly" - but good enough for the test.
Here's a stack done with the 50x Mitty - running at 26.25x onto the sensor for a FoV 0.91mm wide in APS-C mode. Sharp enough all the way into the corners - pretty much as expected. The image wasn't cropped at all.
Just for grins, I put the camera back to full frame mode, expecting to see vignetting and complete mush in the extreme corners. That was the big surprise! I couldn't believe how good the corners were - noticeably better than on the 135mm tube lenses! It's a guess, but I wonder if the field curvature is compensating in some way for the aberrations present in the outer (out of spec) image circles of the Mitties. No idea really, but I'll take it! Here's the same butterfly wing taken with a 50x Mitty on full frame (same mag of course, so just a larger FoV). Also not cropped.
More to test yet and I'll add some 100% crops tomorrow - but this is quite encouraging. Final image is just a short stack of the center of a daisy (a few grains of pollen) with the 50x Mitty. Cropped in for composition.
