wwv receiver

[Norm Dresner]

I'd like to build a Q&D receiver for the WWV time signal because I think
that it's a good, inexpensive way to get access to a very high accuracy time
standard. I'm not interested in the content of the signal, but in the
carrier itself which I'd use to calibrate several instruments and possibly
as a lab frequency standard too. WWV broadcasts on 5, 10, and 15 MHz and I
guess for most purposes either the 5 or 10 MHz signal would be most useful.

Could I use a simple dipole antenna, an ordinary 5 or 10 MHz crystal and a
simple one-three transistor RF amplifier as the basis for this?

If this will work, are there any real gotchas?

If this is stupid, impractical, etc, then what's the next best solution to
the frequency standard problem?

TIA
Norm

 

[Matthew Kendall]

If this is stupid, impractical, etc, then what's the next
best solution to the frequency standard problem?

An HP Z3801A GPS-based frequency standard, widely available on the surplus
market, liberated from CDMA cellular base-stations.

Lots of info available because they are popular with hams, e.g.
http://www.realhamradio.com/GPS_Frequency_Standard.htm

 

[Tim Shoppa]

I'd like to build a Q&D receiver for the WWV time signal because I think
that it's a good, inexpensive way to get access to a very high accuracy time
standard. I'm not interested in the content of the signal, but in the
carrier itself which I'd use to calibrate several instruments and possibly
as a lab frequency standard too. WWV broadcasts on 5, 10, and 15 MHz and I
guess for most purposes either the 5 or 10 MHz signal would be most useful.

Could I use a simple dipole antenna, an ordinary 5 or 10 MHz crystal and a
simple one-three transistor RF amplifier as the basis for this?

If this will work, are there any real gotchas?

You really, truly need a superhet unless you're right next door to Colorado.
(or Honolulu, WWVB). Especially on 10Mhz (but also true for the
other frequencies) there are big powerhouse SW broadcasters nearby
in frequency.

And keep in mind that you don't just get the WWV carrier, you also get
the modulation sidebands (extending several kHz in either direction).
The crystal won't know how to pick out just the carrier while excluding
the sidebands.

The traditional way of doing WWV frequency calibrations is to zero-beat. This
can be good to a Hz or so... varying propogation path-length on SW makes
it hard to do much better. But that's "only" 0.1 ppm or so. A good
GPS-locked OCXO can deliver a fraction of a ppb without any difficulty.

On the East coast, between 5, 10, 15, and 20 MHz, there's a usable WWV
signal for most but not all of the day. See my WWV usability page at

http://www.trailing-edge.com/~shoppa/wwv/

Tim.

 

[Norm Dresner]

An HP Z3801A GPS-based frequency standard, widely available on the surplus
market, liberated from CDMA cellular base-stations.

Lots of info available because they are popular with hams, e.g.
http://www.realhamradio.com/GPS_Frequency_Standard.htm

The only Z3801A on eBay right now are going for ~$450 That's a little
more than I'd envisioned spending for a Q&D standard. I guess I have to
reconsider how much it's worth.

Norm

 

[Norm Dresner]

You really, truly need a superhet unless you're right next door to Colorado.
(or Honolulu, WWVB). Especially on 10Mhz (but also true for the
other frequencies) there are big powerhouse SW broadcasters nearby
in frequency.

And keep in mind that you don't just get the WWV carrier, you also get
the modulation sidebands (extending several kHz in either direction).
The crystal won't know how to pick out just the carrier while excluding
the sidebands.

The traditional way of doing WWV frequency calibrations is to zero-beat. This
can be good to a Hz or so... varying propogation path-length on SW makes
it hard to do much better. But that's "only" 0.1 ppm or so. A good
GPS-locked OCXO can deliver a fraction of a ppb without any difficulty.

On the East coast, between 5, 10, 15, and 20 MHz, there's a usable WWV
signal for most but not all of the day. See my WWV usability page at

http://www.trailing-edge.com/~shoppa/wwv/

Tim.

I guess the next step is a SW receiver and a good PLL, huh?

BTW, thanks for the URL of your availability page. It's fascinating.

Norm

 

[Active8]

I guess the next step is a SW receiver and a good PLL, huh?

BTW, thanks for the URL of your availability page. It's fascinating.

Norm

The way I've seen it done in the old mags was to use super
regenerative. If you use heterodyne, the freq accuracy of the IF is
only as good as the local osc. The SR doesn't convert the freq.

 

[Ken Smith]

The way I've seen it done in the old mags was to use super
regenerative. If you use heterodyne, the freq accuracy of the IF is
only as good as the local osc. The SR doesn't convert the freq.

What you do is leak a little of your signal into the front end of the
superhet. The audio output contains the difference frequency.

How to calibrate:

It works best with 2 generators or a divider off of the 10MHz one. You
actually end up calibrating to 10MHz + X where X is something like 500Hz.
You get the beat note coming out of the radio and the ouput of a
500Hz generator on a scope at the same time. You then tune the 10MHz to
stop the motion between the two.

 

[Michael A. Covington]

At one time I took the 10-MHz output of my big, old HP frequency counter,
fed it to the mixer input of an NE602, and made a direct-conversion receiver
for WWV, using a microphone-level audio amplifier for the output. It
worked, and it enabled me to zero-beat the frequency counter with WWV. I
used a 40-foot antenna; I'm in Georgia. This was written up in RF Design
some time back in the early or mid-1990s.

 

[Michael A. Covington]

The way I've seen it done in the old mags was to use super
regenerative. If you use heterodyne, the freq accuracy of the IF is
only as good as the local osc. The SR doesn't convert the freq.

You can zero-beat a nearby, stable 10-MHz oscillator to WWV coming in on a
superhet; no problem. No direct connection is needed; the oscillator just
needs to be running in the same room.

 

[Michael Black]

At one time I took the 10-MHz output of my big, old HP frequency counter,
fed it to the mixer input of an NE602, and made a direct-conversion receiver
for WWV, using a microphone-level audio amplifier for the output. It
worked, and it enabled me to zero-beat the frequency counter with WWV. I
used a 40-foot antenna; I'm in Georgia. This was written up in RF Design
some time back in the early or mid-1990s.

I thought about that at one time, given that the problem with direct
conversion receivers with an AM signal is that the oscillator usually is
not exactly where it is. Considering that one often was trying to zerobeat
a local frequency standard to the carrier frequency of the WWV signal, it
makes a lot of sense to clear out some of the clutter and use the local
oscillator in the receiver as the standard, or at the very least something
to be measured by other equipment. I never got around to trying it, but
QST did run an article about such a receiver at some point, late eighties
or early nineties.

If going more traditional routes, it makes sense for a WWV-specific receiver
to be fixed tuned. At that point, a superhet doesn't make a lot of sense.
One might as well go TRF; after all one can get sufficient gain at HF
with available devices nowadays, unlike 1920. The trick would be to use
a crystal filter for the selective element, so tuning is indeed fixed, and
selective enough. This is especially suitable given that it's easy to
dig up 10MHz crystals; if one was receiving CHU, the crystal(s) would likely
have to be custom ground.

Other schemes seen were small AM portables, with a crystal controlled
converter ahead of them. Or knock out the front end of such radio, using
it merely as an IF strip, and then feeding it with a crystal controlled
converter. Of course, then on is back to non-standard crystal frequencies,
unless one uses the first scheme and tunes the radio to fit the available
crystals.

Michael

 

[Joe McElvenney]

Hi,

Here in the UK, I have often wondered about making a frequency
standard using the signal from MSF in Rugby (which like WWVB is
on 60kHz). The main problem using conventional circuitry would
appear to be that these time signal carriers are interrupted at
intervals and so would require very long time-constants to fill
the gaps. Further, to obtain a clean signal from off-air signals
such as these would require some regeneration - usually in the
form of a phase-locked loop.

One thought did occur to me once though and that was to use
the mechanism from a cheap 'radio-controlled' clock, which must
put out a once-a-second current pulse and so could be used, á la
Brooks Shera - W5OJM, to discipline a standard. Unfortunately I
found out that they sync-up only every so often which quickly put
the kibosh on that idea.

I haven't tried a standard derived from a signal using an
interrupted carrier but I have done this successfully with an AM
broadcast station such as the BBC's long-wave transmitter on 198
kHz. I wonder then if there are similar local stations in the
States that might also be locked to a national standard. Even one
that is not so may perhaps use a GPS-locked transmitter drive and
could be employed. BTW, the colour-burst signal on networked TV
transmitters is often traceable to a national standard.

On the AM-band a simple TRF in front of a hard limiter
suffices for the receiver (I used a single series resonant xtal
to strip the sidebands away) with the resultant divided down to
feed a PLL together with another input derived from a 10MHz VCO.
The output I found was accurate enough such that only the last
place in a 10-digit counter varied when it was locked to a
rubidium standard. Had it not been just bread-boarded at the
time, I expect it would have been an order better than that.

If I were to try this all over again though, I'd use one of
the relatively cheap GPS modules with 1-pps outputs now available
and go the W5OJM way as noted above.


Cheers - Joe

 

[Fred Bartoli]

The way I've seen it done in the old mags was to use super
regenerative. If you use heterodyne, the freq accuracy of the IF is
only as good as the local osc. The SR doesn't convert the freq.

You can solve this by first using a "flying" LO then locking it to the
obtained IF.
Obviouly you'll have to select the LO frequency as RF-IF. The other side
won't do any good


--
Thanks,
Fred.

 

[Joerg]

Hi Tim,

You really, truly need a superhet unless you're right next door to Colorado.
(or Honolulu, WWVB). Especially on 10Mhz (but also true for the
other frequencies) there are big powerhouse SW broadcasters nearby
in frequency.

And keep in mind that you don't just get the WWV carrier, you also get
the modulation sidebands (extending several kHz in either direction).
The crystal won't know how to pick out just the carrier while excluding
the sidebands.

You could do it with a few 15MHz crystals. I have done that with much
weaker signals, sometimes they were barely above the noise level. It was
pretty much up to par with my comm receiver which, after all, also has a
crystal filter in the IF path. A better shape factor, ok, but at WWV
that doesn't matter much because there isn't much interference within
10KHz or so around them. Problem is that the 15MHz crystals might be
expensive if Digikey or someone doesn't have any. Still beats a comm
receiver even if you have to shell out $20 each.

Regards, Joerg

 

[Tim Wescott]

You really, truly need a superhet unless you're right next door to Colorado.
(or Honolulu, WWVB). Especially on 10Mhz (but also true for the
other frequencies) there are big powerhouse SW broadcasters nearby
in frequency.

And keep in mind that you don't just get the WWV carrier, you also get
the modulation sidebands (extending several kHz in either direction).
The crystal won't know how to pick out just the carrier while excluding
the sidebands.

The traditional way of doing WWV frequency calibrations is to zero-beat. This
can be good to a Hz or so... varying propogation path-length on SW makes
it hard to do much better. But that's "only" 0.1 ppm or so. A good
GPS-locked OCXO can deliver a fraction of a ppb without any difficulty.

On the East coast, between 5, 10, 15, and 20 MHz, there's a usable WWV
signal for most but not all of the day. See my WWV usability page at

http://www.trailing-edge.com/~shoppa/wwv/

Tim.

1. I think you could do this with a well-built DC receiver. If you
measure your LO and audio tone you can get a good calibration of the
timebase in your counter.

2. You could do this easily with a single-conversion superhet. Choose
an IF that can use easily obtainable crystals, and once again measure
your LO and IF tones. If you're _really_ clever you can use this to
phase lock a single crystal oscillator.

 

[Tim Shoppa]

1. I think you could do this with a well-built DC receiver.

Not on 10MHz, not on the East Coast of the US. Within a few kHz of
WWV there are massive pileups of legit and pirate SW strong-signal
broadcasters.

When the bands are open 15Mhz and 20MHz aren't nearly as bad.

Tim.

 

[Tim Shoppa]

I guess the next step is a SW receiver and a good PLL, huh?

We don't really know your requirements, so it's hard to say.

If you only need a reference good to 100 or 200 ppm, then a cheap
10MHz crystal with no tuning may get you there.

If you need to tune a 10MHz reference crystal oscillator to a 10 ppm, say, then
zero-beating WWV once every few nights may be OK.

If you need to tune a 10MHz reference crystal oscillator to 1 ppm or better,
then a crystal oven and frequent zero-beating of WWV may work OK.

If you need accuracies in the parts-per-billion range then you have
to move to a different technology.

Except under the most perfect propogation conditions, you will not get
a usable 5/10/15MHz carrier frequency directly out of a TRF receiver.

BTW, thanks for the URL of your availability page. It's fascinating.

Yeah, I also have GPS-locked OCXO's but I still like to keep the WWV
refclock up and going as a backup source, mainly "just to be different" in
the homogenous world of GPS. Of course WWV is almost useless
during a geomagnetic storm or other ionospheric upset.

Tim.

 

[Norm Dresner]

We don't really know your requirements, so it's hard to say.

If you only need a reference good to 100 or 200 ppm, then a cheap
10MHz crystal with no tuning may get you there.

If you need to tune a 10MHz reference crystal oscillator to a 10 ppm, say, then
zero-beating WWV once every few nights may be OK.

If you need to tune a 10MHz reference crystal oscillator to 1 ppm or better,
then a crystal oven and frequent zero-beating of WWV may work OK.

If you need accuracies in the parts-per-billion range then you have
to move to a different technology.

Except under the most perfect propogation conditions, you will not get
a usable 5/10/15MHz carrier frequency directly out of a TRF receiver.

I've already got ~50ppm crystals and oscillators but I was hoping to
calibrate them. If I measured the frequencies of, say 4 different signal
sources, then simpel-minded statistics says that my error goes down as the
square root of the number, so that would halve it. Using, say 10, should
cut it to 1/3. That's not bad and probably good enough for an initial setup
of a reference TXCO. Does this sound like I've been smoking something or am
I making sense?

Norm

 

[Joerg]

Hi Tim,

2. You could do this easily with a single-conversion superhet.
Choose an IF that can use easily obtainable crystals, and once again
measure your LO and IF tones. If you're _really_ clever you can use
this to phase lock a single crystal oscillator.

3.58MHz color carrier crystals for the IF come to mind. Ain't going to
get much cheaper than these.

Regards, Joerg

 

[Tim Shoppa]

I've already got ~50ppm crystals and oscillators but I was hoping to
calibrate them. If I measured the frequencies of, say 4 different signal
sources, then simpel-minded statistics says that my error goes down as the
square root of the number, so that would halve it. Using, say 10, should
cut it to 1/3. That's not bad and probably good enough for an initial setup
of a reference TXCO. Does this sound like I've been smoking something or am
I making sense?

Do you need self-consistency or traceability to a national standard?

Indeed, the national standard is itself an ensemble of atomic clocks.
(Although the statistics there aren't quite square-root-of-N, because
the noise distribution isn't completely Gaussian).

All the self-consistency in the world won't do your typical shop any good
if it isn't traceable to a national standard. There are some applications
where self-consistency is all that counts, on the other hand.

To put in terms of something a little more pedestrian: if you buy 100 10%
resistors, it's not as good as a 1% resistor. Because the mean could still
be off by 5 or 10%. And typically a 1% resistor is gonna be cheaper than
100 10% resistors .

Don't make me send the NIST police after you (yes they really exist!)

Tim.

 

[Dbowey]

Tim posted:

I've already got ~50ppm crystals and oscillators but I was hoping to
calibrate them. If I measured the frequencies of, say 4 different signal
sources, then simpel-minded statistics says that my error goes down as the
square root of the number, so that would halve it. Using, say 10, should
cut it to 1/3. That's not bad and probably good enough for an initial setup
of a reference TXCO. Does this sound like I've been smoking something or am
I making sense?

Do you need self-consistency or traceability to a national standard?

Indeed, the national standard is itself an ensemble of atomic clocks.
(Although the statistics there aren't quite square-root-of-N, because
the noise distribution isn't completely Gaussian).

All the self-consistency in the world won't do your typical shop any good
if it isn't traceable to a national standard. There are some applications
where self-consistency is all that counts, on the other hand.

To put in terms of something a little more pedestrian: if you buy 100 10%
resistors, it's not as good as a 1% resistor. Because the mean could still
be off by 5 or 10%. And typically a 1% resistor is gonna be cheaper than
100 10% resistors .

Don't make me send the NIST police after you (yes they really exist!)
It is not necessary to be traceable to a "National Standard." In order to
interwork world-wide, a clock must be traceable to a Stratum 1 clock. There
are many ways to achieve this, such as ordering a digital channel from your
Telco if you can afford it.

In the long run it is probably most affordable to establish pleisiochronous
operation. Purchase a GPS receiver that is equipped with a clock output. That
gives you a signal you can use to lock your local oscillator(s). Many Cellular
Carriers use this method even though the 1.544 Mbit/s channels they order from
the Telco are, themselves, traceable to a Stratum 1 clock.


However, zero beating my 100 kHz reference oscillator occasionally with WWV,
gives me all the accuracy I really need.

Don