From Dave G3YMC

I was about to reply direct to Alan about his receiver bandwidth suggestions, 
but other mails now demand a reply to the reflector.

Nick G4WHO sums it up very nicely when he says that the theoretical improvement 
is for noise considerations only.  This assumes you are listening to a single 
signal on an otherwise unoccupied band.  In this case there will be little 
difference between the 500Hz signal and 350Hz.  However we live in a real 
world, with often several strong in-band signals and the weak one we want to 
copy among them.  For this purpose the bandwidth and skirt response is the most 
important thing. Frequently at weekends it is impossible to find a clear 
frequency in the normal CW part of the band when using normal bandwidth CW 
filters.

My 350Hz Yaesu filter is rather lacking in slope response, such that for 
instance when G3XTZ's signal is on 136.0 I have to tune up to 136.5 to totally 
clear it.  Most HF transceivers are configured for upper sideband CW.  An audio 
filter may help (I have not tried one) but AGC effects will still reduce 
sensitivity for copying weak signals in the skirts of stronger ones.

As a long time HF CW operator, both in contests and DX pile up chasing, I could 
never go back to using an SSB filter for CW, it is just literally as wide as a 
barn door.  Like on HF when it is the norm to have 500Hz or less, it should 
also be the norm on 136.  However though I know some operators have much 
narrower filters than that, when I transmit I always try to avoid transmitting 
too close to others, assuming if it is too close with my filter it will be too 
close for the other amateur's receiver as well.  Not that I am a big enough 
signal to cause much QRM!

Note that most good CW operators tend to tune for a note of 600Hz or below, as 
copyability and human ear filtering is better for lower notes.  Those of you 
attending the HF convention next weekend may like to try your hand at the CW 
pile up competition, which will give you a feel about what real CW operating is 
like!

I am also rather puzzled by the emphasis given to DCF39.  With a normal CW 
filter it is so far out of the band that it simply doesn't come into the 
equation (unless you have receiver overload problems that is).

73s Dave G3YMC
[email protected]
http://ourworld.compuserve.com/homepages/sergeantd