Alberto wrote:

You have to be careful about this -- the vast majority of computer sound
cards are designed to handle the audio spectrum up to around 20-22kHz.
Cards that have sampling rates in excess of about 48kS per second most
often DO NOT process input signals higher than ~22kHz.  The higher
sampling rate is used only for oversampling and noise shaping (look
these up if they are unfamiliar terms), and the digital signals are
noise-shaped and decimated on-card from 96kS, 192kS, 384kS, etc. down to
48kS and lower.

So, yes, a sound card designed for signals up to ~22kHz should handle
SAQ at ~17kHz, but sound cards that can digitize signals above 22kHz are
rare.  Some "professional" sound cards handle signal frequencies up to
~40kHz, but very, very few handle signal frequencies higher than that.
There are also digitizers designed more for instrumentation and data
acquisition than audio that may meet your requirements.

SO:  If you want to digitize frequencies > ~22kHz, read the
documentation for the cards you are considering VERY carefully.  I
haven't been shopping for extended-frequency audio cards recently, so
I'm not up to date on what is available. Perhaps others will have
particular suggestions for digitizing WWVB.

Finally, be aware that sound cards use sampling clocks that are
invariably MUCH worse in terms of jitter and drift than the WWVB carrier
(even as it is received over the air), so you need to deal with that if
what you are after is a 60kHz reference that is as stable as WWVB.

Best regards,

Charles

Hal wrote:

For any location near a city, the noise level (QRM and QRN -- mostly the
former unless there is storm activity within a few hundred km) is
shockingly high.  High enough to be clearly seen and measured with a
good spectrum analyzer.  So the simplest way (but not necessarily the
cheapest, depending on what is in your lab already) is to use a good
spec an with noise integration over the band of interest (e.g., HP 3585A
or B).  You get noise density readings in volts per root Hz.  Divide by
the antenna length and you have volts per root Hz per meter.

Lacking a suitable spec an, any receiver with a reasonably narrow rx B/W
and a calibrated, input-referred detector can be used.  Wave analyzers
(frequency-selectable voltmeters, e.g., HP 3586) are good candidates, as
are some commercial receivers with calibrated "S" meters (e.g., Ten-Tec
RX340).  It would also be pretty easy to design a simple "sniffer"-type
receiver (input op-amp, active filter, logarithmic detector feeding a
standard 1mA meter movement) that could be calibrated by design from
first principles and that everyone interested could build for, perhaps,
$25-30.

In the suburbs of a fairly large US city with aerial electric service, I
generally see noise densities measured in tens to hundreds of uV per
root Hz per meter below 100kHz.  In other, similar locations I have seen
as much as hundreds of mV or more per root Hz per meter.  It depends on
local factors (whether the electric service is buried or aerial, how
well the power utility maintains its equipment, how far away the nearest
industrial neighborhood is, how far between dwellings, how much noisy
technology the neighbors use, etc, etc.).

In order to compare with others, everyone needs to use the same antenna.
There are lots of possibilities, but for the sake of universality I
recommend a 1m vertical whip.  Everyone can make one of those.

Note that this sort of antenna is NOT the best type to minimize received
noise and maximize received S/N ratio.  For that, you generally want a
balanced, shielded loop.

Best regards,

Charles

Hi Hal:

You might want to check the orientation and location of the antenna before digging into more technical areas.
It's been my experience there's a lot of AC mains conducted noise at 60 kHz.
http://www.prc68.com/I/LF-Ant.shtml#Noise
http://www.prc68.com/I/Spec_0002.shtml - 0 to 200 kHz spectrum plot (PS when LORAN-C was on the air)
Another source of noise is an LCD screen.
Note Wellenbrook Communications suggests placing their loop antenna 100 feet from your house.

The loopstick antenna in the UltraLink is a single ferrite rod with nulls off the ends, so orientation is important, not
so much that you have to point the maximum at WWVB, but that you don't want to point the null at WWVB.  I'm in
California and have had to relocate WWVB clocks on walls 90 degrees to where I'd rather have them because of this.
http://www.prc68.com/I/Loop.shtml
http://www.prc68.com/I/Shadow-Clock.shtml#WT5360U

--
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html

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

Hi

First off, I don’t think there is an ideal antenna that “just works”. Maybe a proper set of EMAG
probes that come with calibration sheets come close. For a home built this or that …. there are
a lot of variables.

First up is very much part of receiving WWVB in the first place. Coax to an antenna can have currents
on the outer shield. If they meet up with everything else at the antenna, you are not just measuring
the antenna output. Equally, if reception is the goal, you may have a ton of noise that you didn’t really
want to have. Of course, the coax might act as a really good antenna …. who knows.

Something like a 6” diameter single turn  loop with a good choke at the antenna end of the coax would
be my first choice. Not super sensitive. It’s not the ideal reception antenna. For chasing down noise, smaller
is often better. As mentioned earlier we are after stuff that may be in the millivolts per meter range.

Classic data:

ftp://ftp.ngdc.noaa.gov/ionosonde/documentation/CCIR%20-%20Characteristics%20and%20Applications%20of%20Atmospheric%20Radio%20Noise%20Data.PDF ftp://ftp.ngdc.noaa.gov/ionosonde/documentation/CCIR%20-%20Characteristics%20and%20Applications%20of%20Atmospheric%20Radio%20Noise%20Data.PDF

puts the atmospheric noise at 120 db above KTB in the vicinity of 60 KHz. That would put it in the roughly
-54 dbm / Hz range. If your spectrum analyzer has a 1KHz bandwidth, that’s 30 db relative to 1 Hz. Your
SA should read about -24 dbm ( with an efficient antenna).  Coming up with a 1/4 wave vertical at 60 KHz
may make getting those numbers a bit difficult :). Bottom line is still — there’s a lot of noise at 60 KHz. Also
note that the report came out long before the modern era of 60 KHz switchers …..

Bob

Long ago I did some WWVB signal-to-noise measurements with an HP 3586C selective voltmeter (commonly used by the FMT-nuts).  I measured the signal power at 60.0 kHz with 20 Hz bandwidth.  Then I measured the power a small offset plus and minus  (100 Hz?  I don't recall), and took the mean of the two to get the noise power.  I used a voltage-probe antenna. 

Since all readings were taken with the same bandwidth I didn't bother normalizing to 1 Hz, and just used the dBm difference between the signal and mean noise as the result.  I took measurements every 5 minutes or so to capture the 24 hour cycle of SNR.

John

On May 6, 2018, 1:09 PM, at 1:09 PM, Charles Steinmetz csteinmetz@yandex.com wrote:

On 5/6/2018 6:39 AM, Charles Steinmetz wrote:

Bob and Charles,

  you are of course both quite right.
The card used for SAQ was the Delta 44 which is a semi-professional card.

For higher frequencies I have had quite good results with the E-MU 1212M, which
is a professional, mastering grade, sound card, used also in recording studios.

http://www.creative.com/emu/products/product.aspx?pid=19169

Not exactly inexpensive, but its 120dB signal-to-noise ratio cannot be overlooked.

73  Alberto  I2PHD