In article , "Joel Kolstad"
writes:

Thanks for that most informative response, Avery, and also the link to Hans'
project.

Hopefully next week I'll get around to constructing my antenna (need to
visit a bike store and ask for busted rims this weekend... :-) ).

The circularity shape of the loop has little relation to its ability to
intercept the magnetic field of the signal. There's a slight variation
from perfect circle to octagonal to square shape. Sensitivity is
related to the loop's total area and, most especially, to the number
of turns of wire in the loop.

From the
signal strengths you were quoting (100uV), it seems as though it's iffy
whether or not you'd see anything at all taking the output of the antenna
(with resonating capacitor) and feeding it directly to a spectrum analyzer
(since we're looking at, oh, -70 to -80dBm into 50 ohms, which is starting
to push the noise floor of at least one spectrum analyzer I have available
to use). If there's nothing visible at that point, however, hopefully after
a differential FET amplifier as you suggested the signal will be visible.
(I find it very reassuring to be able to see you actually have some signal
present at each point of a porject's development...)

I don't "see" anything directly at my loop in the way of signal from
WWVB. A resonated, relatively high impedance loop (mine is about
93 KOhms at 60 KHz) will lose everything in the way of sensitivity
if loaded down with 50 Ohms input Z of any other device. The purpose
of the differential FET input amplifier is manyfold: Present a
differential,
balanced short feed to reduce common-mode pickup; provide a
coupling from a high-impedance source (resonated inductor of
moderate Q) to the lower impedance of later amplifiying circuitry; to
enable tuning for resonance remotely when the loop is 8 feet over-
head and inaccessible.

"Signal strength" was calculated based on measured gains of the
input stage, loop impedance simulated by a series resistor, all done
at higher voltage levels. The loop voltage and field strength would
then be equal to the differential amplifer output voltage divided by its
voltage gain. One CAN see amplified signals of a milliVolt with most
higher-end oscilloscopes...even if the WWVB has a slow, dull, not-
directly-interesting RF envelope. :-)

The number of amplifying stages following a moderate-gain input
stage is really a sort of arbitrary thing, depending more on what
one DOES with the output. Since WWVB is going to about as
fixed-tuned and stable on 60 KHz as anything can be, there should
be no problem with fixed tuning of those following stages. Even if
the Q of those tuned circuits are low (15 to 25, maybe), they have
an advantage of providing even more skirt attenuation on either side
of 60 KHz. The 4th harmonic of NTSC horizontal sweep is 3 KHz
higher than 60 KHz...some TV receivers spritz out a lot of that while
others have minimal output. Attenuation at around 63 KHz is
probably desireable for household locations in North America.

I'm going to eventually use some moderately-high input Z comparators
after extreme narrowband filtering, that to square it up for CMOS TTL
digital levels. That isn't a necessary, just a convenience for me. If
the amplified signal level can get above about a half Volt p-p, then
an ordinary digital Schmitt trigger gate or inverter will act just fine to
square up the signal, make it compatible with other digital packages.
Did that before for other applications and no problem.

Squaring-up the received signal doesn't offer much for improving the
signal to noise ratio. That may make it LOOK like it is "noise free"
but that is only because the existing noise is transferred from the
vertical direction of a scope picture to the horizontal. It's harder to
observe the horizontal display noise due to the very fast transition
of a comparator or trigger output. Squaring up is absolutely necessary
to couple the signal to any kind of digital circuit for time decoding.

LEN Anderson
:-)
retired (from regular hours) electronic engineer person