I had smoke in the air and a very bright moon when doing these tests tonight, but they got the job done.
I took single 1 minute subs of M20, through the FLT98, with a Canon 40D.
The attachments are in the order of DG-L, Flat III, Flat IV.
The verdict:
The DG-L is just very slightly under-correcting on the FLT98, so I suspect I need to change the spacer ring that I am using to compensate. On the Megrez 90, I was unable to detect any under-correction when I last tested.
The Flattener III has, as always, lots of astigmatism. If you crop the centre 80% you probably would be fine, so if you are using a smaller chip than this it may work for you.
The Flattener IV was indeed flat to the edge when set to one end of its range. However, the magnification that it gives and the resulting loss of f/ratio has a dramatic impact on brightness.
For pixel peepers, I have these unresized images (converted to .jpg from raw)
Borg DG-L
WO Flattener III
WO Flattener IV
Edit: The fully processed image combining 10x 60 second subs from the Borg DG-L
can be found here. You can tell which flattener I will be keeping
Regards,
Eric