In article , "Joel Kolstad"
writes:

Avery Fineman wrote:
One of the most innovative in my estimation was the loop built on an
unused bicycle wheel rim (spokes removed).

Indeed. I saw one of those he http://lakeweb.com/rf/wwvb/ (may or may
not be the same one you found), and it is a clever idea. I'm thinking now I
might build one like that as well as a ferrite rod version to compare with;
since the ferrite will increase the effective area of the antenna by the
square of its effective relative permeability, I want to believe that a 1/2"
ferrite rod can compete with... well... at least an air core loop, uh... the
size of your hand? (I don't have various references here with me to start
looking up the effective relative permeability of a ferrite rod of a given
l/d ratio... hopefully in the ballpark of 20-50...)

Sounds good. I've not experimented with ferrite or powdered-iron
core antennas for my project. My loop was based more on the fact
that there was attic space above the interior workshop and the trap
door diagonal dimension set the maximum loop diameter to 2' 10".
#14 electrical wire was actually cheaper than purchaseable coil
wire for 500 feet and was a relatively calculable-before-build thing.
I could have put in a 5 foot diameter loop but would have to dismantle
a few things to get it there or spend time bent over while constructing
it. :-) [winding and wrapping took a LOT longer than expected!]

I also found that people did occasionally ask about really large diameter
(1") ferrite cores, and while the usual response was that you could, of
course, pack a bunch of smaller rods together to make a big one, apparently
no one has the considerable $$$ around to build, say, a coffee can sized
ferrite core antenna. Would be something to see though! Anyone want to
donate some fat ferrite rods to me? :-)

My own loop is 58 1/2 turns of #14 AWG THHN electrical wire self-
supporting with a mean diameter of 2 feet, 8 inches, then bound with
cheap twine that was well varnished with McCloskey's "Gym-Seal"
floor varnish. The electrostatic shielding was provided by heavy-grade
kitchen aluminum foil (with a gap, of course) that was bound with a
second application of twine, then varnished. Q at resonance is about
44, good enough for about 1.4 KHz BW by itself.

Not bad at all!

Worked out well
and I can't observe any funny spikes from appliances or other
non-WWVB signal things after the FET stage.

That's good to hear; noise seems to be a common concern.

I believe you said your receiver was a synchronous design, correct? I'm
still liking the envelope detector approach, but I've yet to hear about
anyone successfully employing this (simpler) method.

My project is specifically for deriving the 60 KHz carrier, as clean as
possible prior to mixing-dividing etc. to compare the frequency
counter timebase frequency of 10 MHz. With two commercial radio
clocks on hand, WWV over the Icom R-70 for time hacks, wasn't any
desire to decode the time signal.

To get a narrow receiving bandwidth "raw," the tuned loop handles the
initial selectivity. A differential FET with tuned transformer output both
gets rid of some common-mode RF pickup and the transformer Q of
just about 30 drops the far-frequency attenuation more (bandwidth of
the transformer is roughly 2 KHz, wider than loop alone). From there
on things are quite different.

The amplified, "wide" bandwidth WWVB signal is fed to an old MC1350P
video-RF differential amplifier with a local oscillator at about 246 KHz
into
the AGC voltage pin. That makes it an active balanced mixer. The
output of that goes to a single quartz crystal filter at 186 KHz...an old
surplus xtal acquired decades ago, presumably from a timebase or
marker crystal for a military radar set of ancient vintage. A second
MC1350P mixerf, fed with the same 246 KHz LO but the output tuned
to 60 KHz downconverts to the original frequency.

The two mixers need to have the LO injection with at least 30 db of
isolation between the two...takes some playing around and I got about
40 db. The up-conversion and down-conversion was to use a crystal
filter at a frequency different than 60 KHz (avoids re-radiation pickup in
the loop for one thing or other stuff in the breadboard stage). The crystal
filter is still being played with but the basic circuit is just series-
resonance at a single frequency, looking for the narrowest bandwidth
possible. Getting about 5 Hz of -3 db bandwidth right now and am having
to fuss with temperature stability of the L-C tuned LO. Gets a bit fussy
to test out that narrow a bandwidth...:-)

Note: The heterodyning (mixing) process doesn't change the relative
phase of a frequency-changed signal. A second heterodyne back to
the original frequency retains the original relative phase. The trick
there is to get isolation from one LO injection port to the other. That
avoids any feed-through which can spoil the crystal filtering action.
I use a couple more MC1350s to split out from the L-C transistor
oscillator with supply pins separately decoupled and gains set low
by using very low collector load resistors.

I might be satisfied with a 10 Hz bandwidth overall if the fussing-around
period eats up too much time. Right now it looks like a very clean
RF carrier waveform with annoying 1 second bounce up and down when
viewed on a scope. I'm planning on adding a limiter stage or two following
which will remove the modulation almost fully (!). The MC1350P (an old
Motorola part based on their metal can MC1590) makes an excellent
limiter when over-driven.

60 KHz signal voltage
across the loop at resonance is estimated at about 90 to 120
microVolts.

I don't suppose you have an estimate of your signal to noise ratio?

No. I'm looking for the carrier as clean as I can get it. There's zilch
front-end noise. There are some occaisional "bumps" from impulse
stuff from somewhere but nothing from horizontal TV sweep 4th
harmonic at ~63 KHz. Closest TV set is about 15 feet away diagonal.

Offhand, I'd say the S:N ratio is consistent with a 10 Hz bandwidth
at -3 db points. That would be fine if I were demodulating the WWVB
time code changing at a 1 Hz rate.

Location here is northern Los Angeles in the Verdugo
Hills (a mile of hills between here and Boulder, CO)..

I'm in Corvallis, Oregon, which -- eyeballing it on a map -- is perhaps half
again as far from Boulder. I have one of the inexpensive self-setting
clocks that does work pretty reliably _if keep in certain positions within
my house_. I've read that these also set themselves at night when the SNR
is significantly higher too.

Their automatic sensing varies with the manufacturer. NIST has some
info (in a FAQ page?) on their site. The little Oregon Scientific unit
here (older one) checks itself every 8 hours, the larger new one (by same
company) checks itself in the hour after midnight.

Ferrite/powdered-iron core "antennas" used in consumer market
radio clocks seem to work very well.

I was going to get some of the 1/2"x4" or 7" rods from Ocean State
Electronics he http://www.oselectronics.com/ose_p88.htm ... they also
have some inexpensive pre-wound rods meant for AM radios
(http://www.oselectronics.com/ose_p91.htm), but they're apparently tuned for
the AM broadcast band and therefore I'd have to re-wind the things anyway.

Couple of things there. Core permeability may be different at 60 KHz
versus the 550 KHz low frequency of AM BC band. If it is okay then
you can get away without rewinding the coil on it, just retune it with
more capacitance if your input circuit can stand the lower impedance.

At 60 KHz you can use a wide-range audio test oscillator for checking
resonance Q, measuring inductance, etc. I was using an ancient
Heathkit audio oscillator (Wien bridge type) with the 5-digit frequency
counter built into a French-made DVM newly purchased from Mouser.

This is hopefully going to end up as a class project and therefore the goal
of learning how to build your own antenna and receiver is the reason I'm not
intending to just go and use someone's "all in one" WWVB receiver IC (even
though colleges seem to push that approach these days... but then _someone_
had to design that IC, right!?).

Hans Summers has a nice section on his website in the UK that has
full particulars of his 1991 first-year university project of a 60 KHz
receiver-decoder for the Rugby station there. He used discrete TTL
packages for the entire decoder! [Rugby modulation code a bit
different compared to WWVB]

http://www.hanssummers.com/electroni...o/radioclk.htm

Hans (who appears in here from time to time) has _everything_ on
that project available there. Interesting!

If I type the link incorrect, just get www.hanssummers.com and
navigate from there. Interesting website with lots of different projects
well-described.

If I were going for a radio clock project, I'd get a reasonably-clean
signal at 60 KHz by analog means with a bandwidth of 10 to 50 Hz
on the output then use one of the many PIC microcontrollers to do
the decoding...and also readout control. Microcircuits has free
development software for download that works in any PC. Decoding
algorithms will be roughly the same whether done in software or
hardware and (in my estimation) your own software development can
be as much fun as handling lots of hardware. Your mileage may vary.

Have fun. Things are a bit different working with a very fixed frequency
station at LF with very slow modulation rates! :-)

Len Anderson
retired (from regular hours) electronic engineer person