On Sat, 24 Feb 2007 10:04:15 -0600, Tom Ring
wrote:

This came up last night while we were putting racks together for a new
data center. Someone mentioned that it was theoretically possible to
put an infinite amount of information through a fiber because you could
have an infinite number of carriers even though the total available
frequency response was finite.

Hi Tom,

His argument fails on the face of it. If something is finite, that
about ends the discussion. Your workmate is confusing his sense of
being unable to encompass a large number with the sense of infinity.

Given there are an infinite number of infinities, you could argue that
perhaps he is talking about one of the smaller, less significant ones.

That reminded me of something I
thought you mentioned months ago. You either named the effect, or gave
a link to an article about it.

It seems that google has changed their search engine and dumbed it
down significantly. I don't recall the discussion except in the most
general terms and searching for it was frustrating. Maybe the thought
will emerge later.

All I can offer that brings us back to 'tronics (and peripherally to
'tennas) is that to fit enough bits into the information content, they
need to be have a certain rise time (very small of course). Rise time
and bandwidth are inextricably related by a factor of about 3. An
infinite amount of information would thus require 3X infinite
bandwidth (see what I mean about infinities?). We encounter the same
limitation in listening to code. When CW is sent at a certain rate,
it defines the receive bandwidth necessary to recover the information.
And NA mismatch doesn't ring any bells.

As it shouldn't, it is a side topic but relatable to fiber
transmission. I struggled for years trying to get 5 joules of light
energy into a 1mm fiber. I was probably 5% efficient at best and
researching the topic didn't offer the prospects of my expecting much
better. That is why I find impedance matching debates here so amusing
when so much of was just armchair experience.

73's
Richard Clark, KB7QHC