[Elecraft] Spectrogram and frequency markers

[Don Wilhelm]

David,

Yes, the Spectrogram frequency resolution is a factor, but not nearly as
large as the one might expect - I believe Spectrogram and other similar
programs should come in at +/- 2 Hz or better.  It is all a matter of
percentages.  A 5 Hz error at 500 Hz is 1 part in 100 error or 1000 parts
per million - rather poor, but usable at audio frequencies.  When you get up
to the megahertz range, 1 ppm or better is what we need to end up with, at
10 MHz a 5 Hz error is only 0.5 ppm.

WWV received over the air reportably has an accuracy of 10E-8 or better, so
the limiting factor is our ability to properly zero beat with that signal.
If we could find exact zero beat, we could achieve 10E-7 accuracy in our
results, but if we are 2 Hz off in finding zero beat, we are still can
achieve 0.5 ppm accuracy.  The K2 limits its readout to 10 Hz, so there is
that +/- 10 Hz of uncertainty built in already when we tune to WWV, so if we
add a hertz or two, it is not that great a percentage.  The uncertainty
range is additive only when considering the worst case extremes - most of
the time we will operate inside the range of extremes - (maybe even hitting
the midpoint and be right on during the course of a few trials).  We do have
a double check on our final accuracy - simply zero WWV again and see how far
off it is.

73,
Don W3FPR

----- Original Message ----- 


| I note a number of people suggesting the use of Spectrogram's frequency
| markers as means of determining, say, 500 or 600 Hz.  Just how precise
| are these markers in Spectrogram?  Do they not depend on some
| frequency standard in the computer, and as such, are they not themselves
| subject to error and change somewhat from computer to computer?  What
| sort of precision is to be expected?
|
|