Given that a 1dB change is NOT SUSPOSED to be noticed (without a meter, in
hearing, sight, ect. ), anyhow, Just what would be the Noticeable effect of
/ = .1 dB in the real world ? Would , say, 2/10's really kill you, or
1/100th dB extra get you that last DXCC country? As I say, am very cynical
when ANYTHING gets into these kinds of numbers! Jim NN7K




KB7QHC wrote:
-------------------------------------------------
50:12.5 Ohm with an insertion loss of around 0.1dB or less over the
interval of 100KHz to 30MHz.
AND THAT IS NOT THE BEST EXAMPLE OF LOW LOSS!
73's Richard Clark, KB7QHC

On Thu, 31 Jul 2003 17:07:10 -0700, wrote:

Given that a 1dB change is NOT SUSPOSED to be noticed (without a meter, in
hearing, sight, ect. ), anyhow, Just what would be the Noticeable effect of
/ = .1 dB in the real world ? Would , say, 2/10's really kill you, or
1/100th dB extra get you that last DXCC country? As I say, am very cynical
when ANYTHING gets into these kinds of numbers! Jim NN7K

KB7QHC wrote:
-------------------------------------------------
50:12.5 Ohm with an insertion loss of around 0.1dB or less over the
interval of 100KHz to 30MHz.
AND THAT IS NOT THE BEST EXAMPLE OF LOW LOSS!
73's Richard Clark, KB7QHC

Hi Jim,

Cynical? This 0.1dB corresponds to about 2% error from perfect where
too many think that 5% error is the worst they have to suffer from
making a power measurement with their Bird (which actually doesn't do
nearly that well in the first place due to these accumulations of
error). It doesn't take long for error to accumulate to the 10's of
percent where you couldn't convince the bench tech that he has too
many places of precision in his pronouncement of measuring 104.5W
(when it was in fact closer to somewhere between 85W to 115W).

I can well anticipate the "so what?" rebuttal. "Who needs 5%
accuracy?" being another. The general rule of thumb demands that your
standard exhibit 3 times the accuracy of the instrument being
calibrated (the Bird is already dead on arrival using this 0.1dB loss,
if it were not characterized already). With an out of whack Bird, you
barely qualify any power measurement to within 15% (and there are more
sources of error than the BalUn used to isolate the Bird). Again, I
am being generous with the 3 times rule (professionals generally seek
5 times and are more comfortable with 10 times). But this is all
really the provence of the professional Metrologist, not the Amateur.
For the Bench Tech that confidently made the 104.5W measurement (not
knowing it was in fact closer to 60W) would hardly know it through
contacts where they barely noticed the less than 1 S-Unit difference.

Returning to this 0.1dB, it also represents a heat burden of 20W (or
more, I am being generous) for each 1KW passing through. This is a
lot of heat for small packages carelessly regarded as being trivial
(after-all, who can see 0.1dB on their S-Meter?). There have been
more than single reports of Hams writing here in astonishment of their
BalUns exploding. Blame the BalUn seems to be a popular ballad played
to that audience. Others pronounce with hushed tones of reverence
remonstrating mankind for drifting from the true path of the air wound
BalUn (or choke, what will you) mindless of the same loss, but
gratified through ignorance of the greater heat mass.

This 0.1dB in the wrong hands is clearly an example of extravagant
dismissal or myopic attention. And speaking of hands, how long would
you consider it trivial if you had to hold onto the sucker for 20
seconds?

73's
Richard Clark, KB7QHC

I think it all boils down to signal to noise.
If you are trying to communicate with another station and he is putting out
100 watts and is not being copied and then he puts out 110 watts and you can
copy him that is what counts.

Bill N4WC

wrote:
Given that a 1dB change is NOT SUSPOSED to be noticed (without a meter, in
hearing, sight, ect. ), anyhow, Just what would be the Noticeable effect of

/ = .1 dB in the real world ? Would , say, 2/10's really kill you, or

1/100th dB extra get you that last DXCC country? As I say, am very cynical
when ANYTHING gets into these kinds of numbers! Jim NN7K




KB7QHC wrote:
-------------------------------------------------
50:12.5 Ohm with an insertion loss of around 0.1dB or less over the
interval of 100KHz to 30MHz.
AND THAT IS NOT THE BEST EXAMPLE OF LOW LOSS!
73's Richard Clark, KB7QHC

"Bill" wrote
I think it all boils down to signal to noise.
If you are trying to communicate with another station and he is putting
out
100 watts and is not being copied and then he puts out 110 watts and you
can
copy him that is what counts.

===============================

Bill, sorry to be so pessimistic.

If, because of bad signal to noise ratio you can't copy him when he's using
100 watts, then, as sure as eggs don't bounce off concrete, there's no hope
of any detectable improvement by increasing power to 110 watts or 0.4 dB.

Suppose when he's using 100 watts you can hear only 25% of words (or morse
characters). So you can't copy him.

If he doubles power to 200 watts you will still read only 40% of what he
says. So you still can't copy.

If he doubles power again to 400 watts you will be able to copy 60% of what
he says. You will still be in big trouble.

At 800 watts 80% of words (or characters) will be OK but it's not solid
copy. Requests to repeat will be common.

At 1600 watts 99% of words (or characters) will be OK and that's solid
enough.

There are many assumptions in the foregoing crude analysis. But as many have
experienced it is typical.

Claude E. Shannon's (of Bell Labs) original classical paper on the subject
of "Communication in the Presence of Noise", Jan. 1949 can be downloaded (I
have just discovered) by doing a Google on the title. Radio and phone
engineers had been trying for 40 years to describe in precise mathematical
terms the effects of noise and cross-talk in a communication channels. The
transistor had just been invented. So had PCM pre-war. But progress in the
design of the vast communication digital networks then envisaged and which
we now see was being impeded by the lack of understanding of the effects of
ever-present random noise.

It was basically a problem in Statistics. But Shannon went off at a tangent
back to Geometry where Pythagorus the ancient Greek had begun. He translated
the statistical problem into one of calculating the number of small spheres
which can be packed inside a much larger multi-dimensional sphere. The
calculating procedure acquired the everlasting name of "Ball Packing". It is
not difficult to understand. It was Shannon's dazzling multi-coloured flash
of inspiration which did the trick. His name has gone down in history. Think
of him the next time your electric light dimmer-switch goes faulty.
Following Shannon progress forged ahead. In-words such as signal-to-noise
ratios and error-rates became very popular.

A one-dimension sphere is a dot. A 2-dimension sphere is a flat circular
disk. A 3-dimension sphere is a ball. Followed by N-dimensions, all of which
have a surface area and and a volume involving Pi.
----
Reg.