Walt,
The setup I'm using is almost identical to what's shown
here. Motion control is using a machinist's mill table, which moves 0.1" per turn of the handwheel, marked as 100 tick marks of 0.001" each. When I write 0.00025", what I really mean is 1/4 of a tick mark, which I eyeball through a 2.5X plastic magnifier mounted to focus on the index mark. I adjusted the slides on the table to be looser than they would be for machining, so there's essentially no stickiness in the movement. Individual steps probably vary from 1/3 to 1/5 of a tick mark. This is at the edge of the equipment's repeatability, but it's good enough that I've never gotten around to doing anything better.
At this magnification, vibration is the biggest problem. I run with mirror lockup (1.5 second delay from mirror up to shutter open), and I turn down the illumination and ISO setting far enough to make a shutter time of at least 1 second. These shots were at ISO 100, 3.2 seconds. The long exposure gives time for the high frequency vibrations from curtain movement to die out. Turning down the illuminator intensity also turns down its fan, which reduces vibration from that source.
One might think that the long exposure time would greatly slow down the whole process. But it doesn't, because there are so many other things that take even more time. I rechecked the original images. Overall rate for the stack was right at 7 frames per minute. So the actual shooting time for this image was about 9 minutes. That's a very small portion of the overall couple of hours that it took to prepare the specimen, set up the focus, framing, and lighting, process the stack, and post-process the image for posting.
About the extension, what I'm quoting here is from sensor plane to mounting shoulder on the objective. 150 mm is the design spec for this objective, which is designed for "160 mm tube length" but that's with the image formed 10 mm down inside the tube.
The extension is not critical. In theory there's some loss of image quality due to added aberrations if the extension is not right, but low NA achromats turn out to be surprisingly tolerant. (See the discussion
here.) If I want a little more magnification -- most of which is "empty" -- then I'll happily stretch the bellows to full extension, which gives about 240 mm and something like 16X onto the sensor.
--Rik