After reviewing a hundred nearly identical images I took during the May 15 lunar eclipse, I decided none cleaned up better than this 5-second exposure:
This image was taken with a 3-inch APO refractor at F5, 100 ISO and cropped to 1920x1280 pixels. The mount is a Losmandy G-11 and I used its lunar tracking rate. I believe the tracking rate made a difference in quality of fine detail, because I see motion blur in the background stars that is consistent across many images. If not for the lunar tracking rate, the motion blur would be in the moon image. The motion blur of the stars is detectable in my 5-second exposure but obvious in this 15-second exposure:
I've been seeing a number of eclipse composites, where the images of the entire eclipse event are superimposed on a scene. Most bother me, for the scale looks wrong. I prefer that composites be an honest depiction of scale and motion, or contain a caveat.
Here's the eclipse with a 28mm lens.
This shows the size of the moon against the roof line of a two story house about 60 feet away.
A half hour earlier, the moon was in the trees:
The distance the moon travels between my wide-field images and it's apparent size are hard to squeeze into these composites from other sources:
jg