I repeated Steve's test with my K200D and 50mm SMC Takumar. As it's dark here, I settled for my white ceiling, illuminated normally by the various lights in the house. The camera was tripod-mounted, so the composition was identical between trials. ISO was set to 400 throughout.
The exposures in manual mode all seemed pretty reasonable to me, but due to the somewhat strange setup, this test is more about consistency between modes than exposure per se. (In my defense, exposure in manual mode for this lens has been fine so far for "real" pictures.)
In unmodified Av mode, there was severe underexposure.
Av mode:
f/1.4 - 1/750 second
f/2 - 1/750
f/2.8 - 1/500
f/4 - 1/250
f/5.6 - 1/125
f/8 - 1/60
f/11 - 1/20
f/16 -1/10
Manual mode:
f/1.4 - 1/180
f/2 - 1/180
f/2.8 - 1/125
f/4 - 1/60
f/5.6 - 1/20
f/8 - 1/10
f/11 - 1/6
f/16 - 1/2
Av mode with tinfoil shorting all contacts*:
f/1.4 - 1/180
f/2 - 1/180
f/2.8 - 1/125
f/4 - 1/60
f/5.6 - 1/20
f/8 - 1/10
f/11 - 1/6
f/16 - 1/2
The conclusion seems to be, shorting the contacts makes M and Av mode metering the same; without the short, Av underexposes quite a lot.
I wouldn't want to use tinfoil regularly, as it shredded significantly during the lens install process. I wouldn't want small conductive bits of metal floating around inside my camera.
I worry that the aluminum tape trick would have the same problem -- small, conductive, sticky bits would seem rather worse.
On the other hand, southy says that this doesn't seem to be be a problem.
Reid
* Some people have wondered whether this lens is in fact wide enough to short the contacts -- it covers about 1/2 the width of them. On my cheap Vivitar which I did remove the paint from, and which seems to cover them about the same, the shorting was successful.