Re: phantom bypass caps

On Sun, 29 Apr 2007 21:44:40 -0400, "Soundhaspriority"
<nowhere@xxxxxxxxxxx> wrote:

More can be done, trivially by paralleling ADC's, or also
by higher power consumption and dissipation. There is no wall,
per se, except by commercial convention and cost practicalities.

Sure, but doubling the number of converters increases the s/n by only about
1 dB. The correlated voltages (signal) add by simple addition, while the
uncorrelated noises add by the Pythagorean rule: up x 1.4. A 10 dB
improvement would require 2^10 converters. So something more subtle would
have to be done :)

I really would be surprised if this were true. Of course,
lots of coolstuff surprises me.

My current understanding is that paralleling A/D's is just
like paralleling resistors, and that uncorrelated noise
would decrease 3.whatever dB per doubling.

Why is there an exception in this case?

We're assuming that driving stages are sufficiently
quieter and that switches are sufficiently low resistance,
yada, yada, of course?

Cool beans,

Chris Hornbeck
"But of course, when you need it, it ain't headroom any more."
- Don Pearce
.