Utilizing a commercial SWBC station for time signals
54 views
Skip to first unread message
Ethan Miller K8GU
Jun 7, 2026, 3:18:51 PM (10 days ago) Jun 7
to ham...@googlegroups.com
All,
There have been a lot of good ideas rapidly put forth to address
different parts of the WBCQ opportunity including the time broadcast
signal design. Phil, W1PJE, and I (K8GU) have tried to take a big
picture, engineering design flavor view of this effort, including the
scientific and regulatory/license considerations.
Some categories of thoughts about issues/considerations:
1. SAMPLE ALIGNMENT AND COHERENCE: The stability and timeliness of the
broadcast itself is a function of the carrier and every other
oscillator or sample clock involved in the chain being aligned. When
the "science waveform" was added to WWV a while back by HamSCI, we
learned on discussions with station staff that apart from ticks, time
code, and tones, all of the other "messages" on WWV are provided via a
"vintage" Windows computer feeding the samples from a sound card into
the modulator (through appropriate conditioning circuits). Because
this is a standard broadcast transmitter, all of that will again
apply, except for the fact that we may (emphasis on "may") have the
opportunity to control
the sample coherence and lock a sample clock to the
carrier...eventually, if we "build a box" that provides the message to
be modulated. The time-of-flight capability using some kind of
defined edge ticks (e.g. 1 sec), for example, would depend on this
box.
(Note that we wouldn't expect any message being sent from a computer
with an unlocked sample clock to be useful for anything absolute,
perhaps even if playing back a WAV file.)
2. STABILITY: How much stability is necessary should be flowed down
from the science we want to do. I was at a 10U baseball tournament
yesterday and spent the night at the Zoo with Cub Scouts last night;
so, I didn't make much progress on that. :)
However, we can make a heuristic argument based on something that is
already known: the "GPSDO" quality waveform that Hans Summers G0UPL
(QRP-Labs) synthesizes with U3s systems is known "good enough" to see
a lot of Doppler signatures of interest in QRSS CW. Furthermore, our
receivers will be GPS-disciplined TCXOs at best. Thus, if the errors
are uncorrelated, they can be summed RMS... and a
reasonably-carefully-designed GPSDO is probably better than the U3s
and quite good enough for this application. I'm confident in John,
N8UR, and others to do that.
3. THE WAVEFORM: This is one that we've spent the most time thinking
about. There are lots of things we would _like_ to do. However, we
need a "minimum viable product" that will tell us enough system
qualities to do some early versions of the message and to design a
better product when we're ready. (This was not too feasible with WWV,
due to its Federal operational status.)
Allan and WBCQ appear to be willing to move quickly, and possibly with
multiple iterations on the message. This is a valuable, and unique,
aspect that we can maximize. Thinking further, we (at the moment)
will be a primary customer of this message; so, to the extent that we
can tolerate a changing format aimed at iteratively improving the
diagnostics provided by the signal, we should embrace that. The key
qualities of selected waveforms are: (A) It must tolerate distortion
by the transmitter; (B) It cannot exceed the licensed bandwidth of the
station; (C) It has to allow us to measure something scientifically
useful.
Now, the suggested approach:
Task #1: FEASIBILITY ASSESSMENTS
- It is somewhat urgent to get something onto the transmitter
while there's some interest...and it doesn't even really matter how
stable the transmitter oscillators are at this stage. This
"placeholder" waveform can be an audio file, preferably lossless, that
we use to assess the ability of the transmitter to support different
waveforms.
Here we propose to use a revised version of the WWV science waveform
that has been updated to correct some known deficiencies and elucidate
details. Features of the modified waveform would include
a) time ticks of some flavor that are easier to analyze than the
5-cycle 1 kHz burst-tones WWV used as ticks;
b) assessment of transmitter modulation power fidelity properties
through use of a large number of stepped tone input power levels
c) (potential) use of multiple simultaneous single-frequency
waveforms, added together in the voltage domain, to check
Peak-to-average power ratio (PAPR) and nonlinear products of the
transmitter's transfer function compared to an ideal square power law
(e.g. linear voltage response) AM modulator.
Executing the waveform output jobs could probably be timed initially
by a "cron" job, running on a Linux stack within a single-board
computer, that has appropriate audio outputs. These are widely
available and not hard to program.
- Assessment is needed, given the offered transmitter hardware, of
whether it's possible to inject a direct exciter signal (with
oscillators we control) into the WBCQ transmitter that's available.
- If we can't get the carrier disciplined, we should assess whether
the effort is worthwhile at all. Allan seems to indicate that carrier
disciplining is a possibility through direct replacement of the
current crystal-based master oscillator. However, how DIY is this
suggested MO (in master-oscillator power amplifier / MOPA) replacement
in reality? Note also that if it's a modern transmitter with a
reference input, we can probably suspend the need for the modulation
samples to be completely coherent with the carrier, but non-coherence
in modulation inputs compared to carrier coherence may have some
processing consequences for certain approaches.
- To rapidly evaluate these tests for development of an empirical
transmitter transfer waveform, we will require someone with a
high-quality SDR system (I/Q capable) that can be GPS-disciplined
(e.g., USRP) to make ground-wave/direct-path measurements of WBCQ
transmitting HamSCI's test waveform over the test transmission period.
The recorded data would initially be used to evaluate the linearity of
the transmitter, as well as its realistic passband. WBCQ's location
in Monticello - very eastern Maine in Aroostook County - could make
that challenging, or perhaps not.
Task #2. EXCITER DEVELOPMENT
In parallel with #1, we would develop the GPS-disciplined transmitter
(exciter) with the idea that we need to be able to also clock "audio"
samples in a coherent fashion. If it's possible to do this with a
beefy microcontroller (e.g., ARM Cortex, etc) and high-quality audio
DAC (e.g., a TI PCM5101), we could load the "audio samples" of
waveforms into it as static buffers, or alternately could do some
generation on-chip based on GPS time or other inputs. The chipsets
mentioned are older and well-known; but, we would not be married to
specific chips if someone has justifiably better choices. (This is
not a job for a precise-timed FPGA application after all.). The system
should be made flexible to enable the audio message to be picked off
directly if we can't use the exciter approach.
Task #3. WAVEFORM DESIGN
Once we know what the system's transfer function is from Task #1 and
with a capable transmission source in hand from Task #2, then we can
design waveforms. I don't think there's a problem with spitballing
things we'd like to have in the interim; but, they're going to be
bound by WAVEFORM issues (a)-(c) in the realistic analog hardware and
it would be important not to get too far out in front of one's skis,
so to speak. Again, this leverages the advantage of having Allan's
flexibility and willingness in testing for system characterization.
Respectfully submitted...73,
--Ethan, K8GU, and Phil, W1PJE.
--
http://www.k8gu.com/
Repair. Re-use. Re-purpose. Recycle.
There have been a lot of good ideas rapidly put forth to address
different parts of the WBCQ opportunity including the time broadcast
signal design. Phil, W1PJE, and I (K8GU) have tried to take a big
picture, engineering design flavor view of this effort, including the
scientific and regulatory/license considerations.
Some categories of thoughts about issues/considerations:
1. SAMPLE ALIGNMENT AND COHERENCE: The stability and timeliness of the
broadcast itself is a function of the carrier and every other
oscillator or sample clock involved in the chain being aligned. When
the "science waveform" was added to WWV a while back by HamSCI, we
learned on discussions with station staff that apart from ticks, time
code, and tones, all of the other "messages" on WWV are provided via a
"vintage" Windows computer feeding the samples from a sound card into
the modulator (through appropriate conditioning circuits). Because
this is a standard broadcast transmitter, all of that will again
apply, except for the fact that we may (emphasis on "may") have the
opportunity to control
the sample coherence and lock a sample clock to the
carrier...eventually, if we "build a box" that provides the message to
be modulated. The time-of-flight capability using some kind of
defined edge ticks (e.g. 1 sec), for example, would depend on this
box.
(Note that we wouldn't expect any message being sent from a computer
with an unlocked sample clock to be useful for anything absolute,
perhaps even if playing back a WAV file.)
2. STABILITY: How much stability is necessary should be flowed down
from the science we want to do. I was at a 10U baseball tournament
yesterday and spent the night at the Zoo with Cub Scouts last night;
so, I didn't make much progress on that. :)
However, we can make a heuristic argument based on something that is
already known: the "GPSDO" quality waveform that Hans Summers G0UPL
(QRP-Labs) synthesizes with U3s systems is known "good enough" to see
a lot of Doppler signatures of interest in QRSS CW. Furthermore, our
receivers will be GPS-disciplined TCXOs at best. Thus, if the errors
are uncorrelated, they can be summed RMS... and a
reasonably-carefully-designed GPSDO is probably better than the U3s
and quite good enough for this application. I'm confident in John,
N8UR, and others to do that.
3. THE WAVEFORM: This is one that we've spent the most time thinking
about. There are lots of things we would _like_ to do. However, we
need a "minimum viable product" that will tell us enough system
qualities to do some early versions of the message and to design a
better product when we're ready. (This was not too feasible with WWV,
due to its Federal operational status.)
Allan and WBCQ appear to be willing to move quickly, and possibly with
multiple iterations on the message. This is a valuable, and unique,
aspect that we can maximize. Thinking further, we (at the moment)
will be a primary customer of this message; so, to the extent that we
can tolerate a changing format aimed at iteratively improving the
diagnostics provided by the signal, we should embrace that. The key
qualities of selected waveforms are: (A) It must tolerate distortion
by the transmitter; (B) It cannot exceed the licensed bandwidth of the
station; (C) It has to allow us to measure something scientifically
useful.
Now, the suggested approach:
Task #1: FEASIBILITY ASSESSMENTS
- It is somewhat urgent to get something onto the transmitter
while there's some interest...and it doesn't even really matter how
stable the transmitter oscillators are at this stage. This
"placeholder" waveform can be an audio file, preferably lossless, that
we use to assess the ability of the transmitter to support different
waveforms.
Here we propose to use a revised version of the WWV science waveform
that has been updated to correct some known deficiencies and elucidate
details. Features of the modified waveform would include
a) time ticks of some flavor that are easier to analyze than the
5-cycle 1 kHz burst-tones WWV used as ticks;
b) assessment of transmitter modulation power fidelity properties
through use of a large number of stepped tone input power levels
c) (potential) use of multiple simultaneous single-frequency
waveforms, added together in the voltage domain, to check
Peak-to-average power ratio (PAPR) and nonlinear products of the
transmitter's transfer function compared to an ideal square power law
(e.g. linear voltage response) AM modulator.
Executing the waveform output jobs could probably be timed initially
by a "cron" job, running on a Linux stack within a single-board
computer, that has appropriate audio outputs. These are widely
available and not hard to program.
- Assessment is needed, given the offered transmitter hardware, of
whether it's possible to inject a direct exciter signal (with
oscillators we control) into the WBCQ transmitter that's available.
- If we can't get the carrier disciplined, we should assess whether
the effort is worthwhile at all. Allan seems to indicate that carrier
disciplining is a possibility through direct replacement of the
current crystal-based master oscillator. However, how DIY is this
suggested MO (in master-oscillator power amplifier / MOPA) replacement
in reality? Note also that if it's a modern transmitter with a
reference input, we can probably suspend the need for the modulation
samples to be completely coherent with the carrier, but non-coherence
in modulation inputs compared to carrier coherence may have some
processing consequences for certain approaches.
- To rapidly evaluate these tests for development of an empirical
transmitter transfer waveform, we will require someone with a
high-quality SDR system (I/Q capable) that can be GPS-disciplined
(e.g., USRP) to make ground-wave/direct-path measurements of WBCQ
transmitting HamSCI's test waveform over the test transmission period.
The recorded data would initially be used to evaluate the linearity of
the transmitter, as well as its realistic passband. WBCQ's location
in Monticello - very eastern Maine in Aroostook County - could make
that challenging, or perhaps not.
Task #2. EXCITER DEVELOPMENT
In parallel with #1, we would develop the GPS-disciplined transmitter
(exciter) with the idea that we need to be able to also clock "audio"
samples in a coherent fashion. If it's possible to do this with a
beefy microcontroller (e.g., ARM Cortex, etc) and high-quality audio
DAC (e.g., a TI PCM5101), we could load the "audio samples" of
waveforms into it as static buffers, or alternately could do some
generation on-chip based on GPS time or other inputs. The chipsets
mentioned are older and well-known; but, we would not be married to
specific chips if someone has justifiably better choices. (This is
not a job for a precise-timed FPGA application after all.). The system
should be made flexible to enable the audio message to be picked off
directly if we can't use the exciter approach.
Task #3. WAVEFORM DESIGN
Once we know what the system's transfer function is from Task #1 and
with a capable transmission source in hand from Task #2, then we can
design waveforms. I don't think there's a problem with spitballing
things we'd like to have in the interim; but, they're going to be
bound by WAVEFORM issues (a)-(c) in the realistic analog hardware and
it would be important not to get too far out in front of one's skis,
so to speak. Again, this leverages the advantage of having Allan's
flexibility and willingness in testing for system characterization.
Respectfully submitted...73,
--Ethan, K8GU, and Phil, W1PJE.
--
http://www.k8gu.com/
Repair. Re-use. Re-purpose. Recycle.
Phil Karn
Jun 7, 2026, 4:57:23 PM (10 days ago) Jun 7
to ham...@googlegroups.com
What do we know about the transmitters? Shortwave broadcast transmitters tend to run a LOT of power, so the electric bill is a major consideration. Is it solid state? Tube? How is the modulation generated? High level plate modulation is considerably more efficient than linear amplification, but it has many more artifacts. Several high-efficiency AM transmitter designs have been around a long time, eg, Doherty, inphasing and envelope elimination and restoration (EER), but I don't know if they're in common use on shortwave. Many modern AM broadcast transmitters are all solid state with arrays of power amplifiers configured essentially as a big stepped RF digital-to-analog converter; conversion efficiencies are in the 90s, so they've largely relegated tube transmitters to standby duty.
WWW's 5, 10 and 15 MHz transmitters (10 kW) use plate modulation, and they have substantial AM-to-PM conversion: the phase advances as the carrier power increases, drawing a banana-shaped trace on an IQ display instead of a straight line. This causes significant intermodulation distortion, particularly between the 100 Hz timecode subcarrier and the 500 and 600 Hz audio tones, and has made it difficult to separate WWV from WWVH. WWV's 2.5, 20 and 25 MHz transmitters use linear amplifiers and are much cleaner, but at lower efficiency and only 2.5 kW output. The AM->PM conversion could probably be reduced by careful neutralization of the final stage, but some will always remain. There is also the response of the modulation transformer in the plate supply circuit to consider.
We could probably correct transmitter distortions by carefully characterizing them with a nearby monitor receiver and applying equalizing responses to our received signals. One of my one-of-these-days projects is to do this for WWV with the nearby WW0WWV ka9q-radio receiver (which sees SNRs of 40-50 dB on most frequencies). A local monitor receiver could also track an unstabilized carrier's frequency in real time, depending on just how much phase noise there is. Most AM broadcast transmitters I look at seem quite stable in frequency, even if they're not exactly on frequency. I understand some BCB DXers actually use these frequency offsets to identify individual stations when receiving several co-channel stations at night.
Phil
--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hamsci/CAFkosZJOB-k%3Dj0LOeZBGzgr8DHbMYBjYrHx%3D%3Dp8QcMeW%2BY62vg%40mail.gmail.com.
Dave Doler
Jun 7, 2026, 6:06:21 PM (10 days ago) Jun 7
to ham...@googlegroups.com
Maybe Bill Hoger, Research Council Officer at the National Research Council Canada who was responsible for maintaining CHU might be willing to answer some technical questions.
To view this discussion visit https://groups.google.com/d/msgid/hamsci/CAMwB8mjHRbqAd-5JUB0ZAO6CfzftDXjx4dxA%2ByrZyfYtaFvgfQ%40mail.gmail.com.
googl...@cox.net
Jun 7, 2026, 8:39:39 PM (10 days ago) Jun 7
to ham...@googlegroups.com
Phil,
I would suspect that the co-located monitoring receiver idea would be the best approach in determining how much distortion is occurring in the transmitter. We would also want to know how much the transmitter waveform varies in time due to temperature swings in the hardware.
The co-located monitoring receiver could be used to generate “correction data” used in any TOF measurements. I think we’re stuck with whatever transmitter is provided.
73,
Mark Braunstein WA4KFZ
To view this discussion visit https://groups.google.com/d/msgid/hamsci/CAMwB8mjHRbqAd-5JUB0ZAO6CfzftDXjx4dxA%2ByrZyfYtaFvgfQ%40mail.gmail.com.
googl...@cox.net
Jun 7, 2026, 9:07:18 PM (10 days ago) Jun 7
to ham...@googlegroups.com
Ethan,
"- If we can't get the carrier disciplined, we should assess whether the effort is worthwhile at all. Allan seems to indicate that carrier disciplining is a possibility through direct replacement of the current crystal-based master oscillator. However, how DIY is this suggested MO (in master-oscillator power amplifier / MOPA) replacement in reality?"
Can Allan provide us with the make and model of the transmitters they use, or intend to leverage for this experiment? If the manufacturers are still around, we could engage in a discussion with their design team to find out if they would be willing to assist in this citizen science effort.
"make ground-wave/direct-path measurements of WBCQ"
There is a high school (Houlton High School) that is about 23 miles away from the WBCQ transmitter site. In the opposite direction is Central Aroostook Junior-Senior High School (Bridgewater) about 14 miles away. Maybe they would be interested in hosting a STEM, citizen science experiment?
73,
Mark Braunstein WA4KFZ
"- If we can't get the carrier disciplined, we should assess whether the effort is worthwhile at all. Allan seems to indicate that carrier disciplining is a possibility through direct replacement of the current crystal-based master oscillator. However, how DIY is this suggested MO (in master-oscillator power amplifier / MOPA) replacement in reality?"
"make ground-wave/direct-path measurements of WBCQ"
73,
Mark Braunstein WA4KFZ
googl...@cox.net
Jun 8, 2026, 1:46:25 PM (9 days ago) Jun 8
to ham...@googlegroups.com
Ethan,
As we consider the use of shortwave broadcast transmitters, it dawned on me that we would also need to consider their program scheduling and antenna selection. Examples of these schedules can be found at other shortwave broadcasters such as WRMI and WWCR.
Unlike WWV and CHU ("always on", omnidirectional antennas), these stations only broadcast in specific directions, at specific frequencies, for specific times, to reach specific audiences. We wouldn't have the luxury of being able to watch one single transmission vary throughout a 24-hour period. Instead, propagation would have to be analyzed from a series of "snapshots" from the various frequencies and azimuth directions available across an entire day.
One question: If CHU stops transmitting later this month, what happens to their frequency allocations on 3.33MHz, 7.85MHz and 14.67MHz? Does the Canadian government retain licensure or are the allocations reverted to the ITU? Could these frequencies be reallocated to other stations and allowed to run from more modest antennas and transmitters operated by (possibly) universities engaged in ionospheric research?
73,
Mark Braunstein WA4KFZ
-----Original Message-----
From: ham...@googlegroups.com <ham...@googlegroups.com> On Behalf Of Ethan Miller K8GU
Sent: Sunday, June 7, 2026 15:19
To: ham...@googlegroups.com
Subject: [HamSCI] Utilizing a commercial SWBC station for time signals
As we consider the use of shortwave broadcast transmitters, it dawned on me that we would also need to consider their program scheduling and antenna selection. Examples of these schedules can be found at other shortwave broadcasters such as WRMI and WWCR.
Unlike WWV and CHU ("always on", omnidirectional antennas), these stations only broadcast in specific directions, at specific frequencies, for specific times, to reach specific audiences. We wouldn't have the luxury of being able to watch one single transmission vary throughout a 24-hour period. Instead, propagation would have to be analyzed from a series of "snapshots" from the various frequencies and azimuth directions available across an entire day.
One question: If CHU stops transmitting later this month, what happens to their frequency allocations on 3.33MHz, 7.85MHz and 14.67MHz? Does the Canadian government retain licensure or are the allocations reverted to the ITU? Could these frequencies be reallocated to other stations and allowed to run from more modest antennas and transmitters operated by (possibly) universities engaged in ionospheric research?
73,
Mark Braunstein WA4KFZ
-----Original Message-----
From: ham...@googlegroups.com <ham...@googlegroups.com> On Behalf Of Ethan Miller K8GU
Sent: Sunday, June 7, 2026 15:19
To: ham...@googlegroups.com
Subject: [HamSCI] Utilizing a commercial SWBC station for time signals
Reply all
Reply to author
Forward
0 new messages