From the horse's mouth:  http://msowww.anu.edu.au/~rmn/C2006P1new.htm

--
Jeff R.

Just so there's some info on the thread (for lazy folk like me who
can't be bothered to click on links!)

I've been using exposures varying from about 0.5 seconds (when comet
was barely discernible through the dusk sky) through to 2-8 seconds
(late dusk), through to 30 seconds or more when the sky was almost
completely dark.
(Based on f5.6, ISO50-100).  It should now be dimming, and your milage
may vary, so bracket and (if digital) watch your results carefully.

To see all the filaments of the tail, exposures of several minutes or
more would be useful, but of course if you do this with any light in
the sky, it may get very washed out...  (O;

(apologies for top post, but in this case I don't think it matters
much..)

Settings depend on how early in the evening it is. When there is still a
fair amount of glow to the sky and the comet is just starting to become
visible to the naked eye, settings of about 1/2sec-1sec @ F5.6-F4 seem to be
ok (ISO 100). At this stage the meter in most cameras should be still able
to meter reasonably accurately so you can pretty much trust the camera.
As the light fades though, most cameras will not be able to meter, and so
experimentation, bracketing, and luck are the name of the game. Note that
anything I say is a _guide_  - you will have to experiment yourself
because...
1) not all lenses are created equal - just because it says F4 doesn't mean
it passes the same amount of light as another F4 lens (in theory it does, in
practice, no).
2) not all films/cameras are created equal - different brands of film react
differently, different digital cameras have slightly different responses at
different ISOs
3) not all nights are created equal - depending on your latitude, amount of
ambient light etc, your exposure values will change.
So with that out of the way - as the night gets darker, the best thing is to
bracket. Bearing in mind though that each night your viewing window is about
30 minutes (improves with greater latitude, so if your down south your in
luck), so when exposures get into the 5 minute range you won't have a lot of
chances. If you are using digital you do have the advantage of being able to
preview your shot on most p&s/prosumers or to examine the shot you just took
on DSLRs. I was using my S2IS and my film camera in tandem - I'd take a shot
on the S2IS, then use that as a guide to the exposure on the film camera.
For example, in the shot I posted under "Comets love the MZ60", I first used
the S2IS to take an exposure at 15sec (because that is the most the S2IS can
do), F2.7 ISO 400. It looked to be about 1 stop underexposed for the look I
was after, so working off that, I was shooting 100ISO film, so that makes it
1minute, my 28mm lens is F3.5 - 1/2 stop different, so we're now at 1.4
minutes. I estimated the S2IS to be 1 stop under, so we are now at 2.8
minutes. Add a bit for reciprocity failure and I figured 4 minutes should be
about right. As it turns out, I was pretty close - a bit more exposure would
have been useful, but 4 minutes @ F3.5 ISO 100 got a pretty good shot.
We've got cloudy skies here again tonight, and tomorrow and thursday I'm
working, so if by chance Friday is clear, my plan is to take 3 cameras out -
1 doing 5 minute exposures, and 2 doing 10 minute exposures with different
composition. Of course if the comet has faded considerably by then (and I
expect it will) I might have to revise those timings.