[Lowfer] Medfer freq measurement

[John Davis]

>I am building a receiver to pick up 2.5MHz from WWV(?).  I want to feed
this
>to the counter, then calibrate it, then read my medfer frequency.
>
>I have heard some people mention frequency drift due to propagation
changes.
>How accurate will my reception of 2.5MHz be?  I want to measure with an
>accuracy of 1Hz or better.
>
>I can listen to WWV on my voltmeter, and it sounds good after my
>amplifier/receiver.  Before I spend more time finishing it, I wanted to
>check if this idea is real or not.


Hi Mick,

If the ionosphere is reasonably stable, and if your counter allows for a
10-second counting interval, any short-term errors will average out to well
under 1 Hertz.  The real problem, though, will be in getting a signal out of
the receiver that is stable enough in terms of amplitude for your counter to
work reliably.

Remember that WWV is an amplitude-modulated signal, so for significant parts
of each cycle of audio (negative modulation peaks) the envelope of the
signal may drop below the threshold of your counter.  Noise bursts will also
affect the count rather badly.

The generally preferred method of using WWV as a frequency standard is to
buy or build a stable crystal oscillator, and zero-beat the oscillator's
output to WWV's carrier.  Then, use the oscillator as a secondary standard
to calibrate the counter.  The crystal can be at the carrier frequency, or
you can use a harmonic of the oscillator.  Five MHz is a good choice, as it
can be beat directly against WWV at 5 MHz, or the harmonics can be used at
10, 15 and 20 MHz, depending on the time of day.  The oscillator needs to be
smoothly adjustable, its output needs to be well buffered so that external
load changes don't affect frequency, and ideally you should have a way of
adjusting how much of its signal gets picked up by the shortwave receiver
(you achieve the clearest beat note when the WWV carrier and the local
oscillator are matched in level at the receiver input).

Zero beating with an off-air signal is a handy skill to develop.  You'll
find it helps to have either a speaker or headphone with extended low
frequency response, because otherwise as the oscillator gets within a couple
hundred Hertz of the carrier, the beat note will diminish in amplitude
rapidly.

When you start tuning, you might be anywhere from hundreds to a couple
thousand Hertz apart.  You need to tune the oscillator quite s-l-o-w-l-y.
As you get to within a few hundred Hertz of the carrier, if there is tone
modulation on the signal, you will first beat with that sideband.  If you
continue, you will eventually beat with the carrier, and perhaps the
opposite sideband of the tone.

Of course, the carrier is the one you want.  It's better to avoid the double
sideband issue by working during the segments when there is no tone, if
possible.  Even then, however, you have to take into account the
time-code-modulated 100 Hz signal, which is always inconveniently close to
WWV's carrier.  As you get this close to the carrier, that's when you need
extended low frequency response to help distinguish the beat note from the
100 Hz subcarrier sidebands.

Once you're close enough that you can be sure you're beating with the
carrier itself, that's when you start listening not for the direct beat
note, but for modulation of the background noise.  As the oscillator adds to
and subtracts from the carrier, the residual background will take on a
fluttering quality that you can use until you get within a Hertz or two of
the carrier.  Then--assuming propagation is stable and you're not
experiencing fading at about the same rate--you'll be able to watch the
frequency difference as swings of the S-meter.

Accuracies of one part per million are not difficult to achieve this way,
and one part in ten million is possible on a good day.

73,
John