On Sun, 21 Jun 2009 09:03:51 +0100, Jim Lesurf
wrote:

I got my latest copy of 'Stereophile' yesterday and started to read it. I
came across comments by Paul Messenger about some work that Russ
Andrews and Ben Duncan have recently put onto the web. This seems to be
taken by Paul Messenger as showing that Russ's claims re some of his
products are "now supported by proper scientific analysis".

But having looked at

http://www.russandrews.com/downloads...estPremRes.pdf

[above file size 700K]

I can't say I agree with that belief simply on the basis of what the above
contains. But that may in part be because I've examined a past set of
measurements by Ben Duncan and come to rather different conclusions to the
ones he and a co-author asserted about them at the time.[1] I would
therefore like to know all the measurement systems/proceedure details that
are sadly omitted from the above.

I thought others here might be interested to read the above pdf and
consider it for themself.

I am curious to know if the reactions of others agree with my own. In
particular, if others can spot 'The dog that did not bark in the night'.
:-)

Enjoy,

Jim

[1] See
http://www.st-andrews.ac.uk/~www_pa/...eshift/cp.html

Interesting. The big problem here is that they were measuring the
wrong thing. They should have been measuring effects at speaker
terminals, not on power rails. My intuition tells me that the audible
difference between 80 and 90 dB of attenuation at the power rails is
going to be close to zero. After all, you must add to that the CMRR,
which is already going to be the right side of 100dB, so effectively
we are talking the difference between -180 and -190. Both of these are
altogether huge compared to what is actually needed.

Add to that the idea that 1000V spikes are common enough occurrences
that they impinge on your day to day listening, (rather than being a
"bugger me, what was that?" moment as half the fuses in the house
blow), and require dealing with for listening pleasure.

Of course, if this were a single ended valve amp with no intrinsic
power supply rejection, there might be a case to be made.

d