We are sliding off the topic but here is some more from me .
I enjoy Luminous Landscapes but sometimes their articles get a little dated ...
We are not allowed to provide links but the LL article ( do sensors outresolve lenses ? ) can be considered recent ; many of today's 15-25 Mp cameras are mentioned .
Even though most of the math is beyond a normal person's understanding, it is stated that the calculations are based on a diffraction free top class lens ( a CZ prime maybe ) and that results taken with other lenses will not reach these resolution figures . The rest of the assumptions are based on the accepted maximum capacity of the human eye / circle of confusion etc . They are all maximums !
So it can't really be argued that less diffraction & better results with other lenses are possible when the figures stated are theoretical & for best case scenarios.
Another link is from Nikon Rumors Guest posts - pls check http://nikonrumors.com/2010/11/27/guest-post-shooting-products-with-nikon-tilt-lenses.aspx/ . See how the images detoriate on a D3x at smaller apertures.
I couldn't find the tables but I see diffraction from my D50 6mp sensor at F/8 with my 50mm just as I do with my D300. ...
Your findings match exactly with the LL findings . Table showing Mpixel equivelant (table 3 )indicates that diffraction at f8 can only be detected in an APS-C size sensor camera at 13 Megapixels and over . Both the D40 and D300 are below that .
Sorry, but experience says "Ummmmm No". This is when you know the math is oversimplified and the article is dated. I like their newer examples taken with a 39mp Medium Format camera...
The math is not oversimplified and findings are not dated . And someone from LL has actually done his own diffraction test with a 39 Mp PhaseOne backed Linhof ( please look for "understanding lens diffraction , Luminious Landscapes " ).He has noted diffraction starting at f11 and that is with a Medium Format sensor ! His finding matches well with the Medium Format/ f11/ 31 Mp figure in LL diffraction table .
The fact that you are not picking any diffraction until f20 doesn't mean it is not there . Your work may not be requiring you to get too close into details & sharpness but make sure - it is there .
Maybe we need a diffraction subject :-)