My understanding is .66 arc sec/pixel was closer to ideal. Based on Nyquists sampling theory that it takes 3X higher accuracy to gain a decent sample (I may have dumbed that down but that is the basic concept).
So .66 as a target assumes then that .66 x 3 = 2 arc seconds seeing is available.
If your seeing is never that good then you can go for a higher than .66.
But 1.5 arc sec/pixel if that theory is true would mean you are expecting seeing of 4.5 arc seconds which is pretty poor seeing.
So following this idea along then the ideal setup would be in arc seconds/pixel 1/3rd of the best seeing you can expect from your location.
Also another rule is smaller pixel cameras for faster F ratios. I think because of how the above works out for this type of scope.
From what I have seen of images, those who apply the above seem to get very good results so it must be pretty close to being true.
Wodaski has a free CCD calc to make it easy to work this out for your setup and different scopes or cameras.
http://www.newastro.com/book_new/camera_app.php
This is a handy tool.
Greg.