This is a multi-test:
First light Deep Sky for my Canon EF 400mm/5.6L
First light Deep Sky for my Hutech-modded Canon 40D
First test of my Kenko Skymemo mount at 400mm focal length

The Kenko Skymemo is a tracking mount at its simplest; no guiding, no computerised go-to, no setting circles. Consequently at night it is nearly impossible to point a 300mm or 400mm focal length lens without some kind of finder scope, hence I taped a laser pointer onto the lens hood. But you really need something visible by naked eye to point the laser at. Even though my intent was to do all this testing on M33, I could not see it by naked eye, and it would probably have taken me a couple of hours just to achieve satisfactory framing. Consequently I decided to have a quick go on M31, easily visible at this dark site. The skyfog was Mag 21.3/sq arc-sec (Visual Limiting Magnitude dimmer than 7).

Lens: Very satisfactory. The above was a simple one-click autofocus on one of the brighter stars near M31. I put the camera on Live-View, centered on Mirach (I think it was) and clicked the autofocus button, beep :-) Since this lens is designed to cover a full 35mm-format there are no issues regarding the performance in the corners of a 1.6x crop DSLR like the 40D. Fully satisfactory for astro use wide open.

Camera: The temperature was rather warm, fairly close to my cut-off of 25deg C for using a DSLR on long-exposure DSO imaging. In fact when I viewed the Raw frames in Canon Digital Photo Pro I nearly panicked that I had wasted a whole night because there was a huge amount of hot pixels. But calibration with darks, bias and flats (about 15 of each) cleaned up the frames very satisfactorily, as may be seen above. I also followed up the field test with a direct comparison of 30-minute darks at ISO 1600, 40D vs 20D. The back-of-camera luminance noise-cliff histogram showed that its toe reaches about 13% of the X-axis on the 40D vs 20% on the 20D, both at 24deg C. The conclusion has to be that the 40D has considerably less thermal signal/noise than the 20D.

Kenko Skymemo at 400mm Focal Length: Note that this is quite a tough test, considering that the pixel pitch in the 40D runs at a quite demanding 5.7microns. Examining the stars on my desired 4-minute exposure at 10x magnification on the back-of-camera LCD showed that they were slightly oblong. A 4-minute sub at a dark site and at f5.6 puts you clearly into the desirable skyfog-statistics-limited regime. The oblong stars made me cut back the length of the subs, reluctantly, to 3 minutes each. At 1:1 the stars are still oblong, but at normal web presentation sizes, the tracking can be deemed as quite satisfactory, as seen above. No attempt has been made to use star-rounding software on the above image. My conclusion, also incorporating other test data, is that the Kenko skymemo is satisfactory to use unguided at up to 400mm focal length, but not if one is going to crop the image such that it is displayed at 1:1 with the tiny pixels of the 40D. From other tests I am inclined to conclude that the oblong stars are most likely due to gear noise, rather than due to periodic error, since the stars remain oblong even with one-minute subs. The issue of oblong stars goes away with a shorter focal length, say, 300mm, and/or use of a camera with a larger pixel pitch.

For other stuff peruse my primary website: http://www.samirkharusi.net/index.html