Sofware audio bridge: Virtual Audio Control 22.214.171.12491
Software Com ports: vspMgr 1.0.3.01
miniwhip-pro active antenna: design of DL4ZAO (with
improvements of PA0RDT) @ 4m or 7m height indoors
Time synchronisation: NTP
Total eclipse on Antarctic
Total eclipse is at: 76.7833° N 46.1983° W
The Dec. 4th total eclipse will be between 05:29 and 09:37UTC.
Totality at 07:34UTC.
Totality happens in the middle of the continuous carrier of RWM
between 07:30 and 07:38UTC.
Transmitter at RWM Moscow (MSK), Russia
RWM is at: 55.7228° N 38.2049° E
MSK Sun rise is around 05:39UTC and MSK Sun set is around
Transmitter at CHU Ottawa, Canada
CHU is at: 45° 17' 47" N, 75° 45' 22" W
Midway CHU (in Atlantic) is around: 49.7° N 41.1° W
Midway CHU Sun rise is around 10:40UTC and Midway CHU Sun set is
Midway RWM (in Poland) is around: 53.5° N 21.9° E
Midway RWM Sun rise is around 06:32UTC and Midway RWM Sun set is
Receiver at Lochem (Lc), The Netherlands
Lc is at: 52.1614° N, 6.4156° E
Lc Sun rise is around 07:25UTC and Lc Sun set is around
The total eclipse happens close to the moment of Lc Sun rise.
Proces around measurements
A few steps were involved in the measurements:
If ones operational system (Windows 10 in this case) is set on
Dutch, it uses a decimal comma. The generated fldigi's csv files
on this decimal comma OS, meshes up the evaluation of these csv
files on a decimal point OS (such as English) in Excel. <This has been
discussed in the fldigi forum: best to have a user
definable csv separator in fldigi>
The doppler effect (frequency deviation) is proportional with
the station frequency. In the following deviation graphs the
measured deviation has been compensated towards a reference of
The first 4 seconds from each scan station frequency scan were
removed: this to stabilise the FLEX-1500 and fldigi
No GPSDO was used, so clock inaccuracies need to be
compensated. The FLEX-1500 radio frequencies (to get
1,000Hz+/-1Hz for fldigi) are 4,995,013, 9,995,027 and
14,995,040Hz. So the radio frequency is linearly changing due
to this inaccurate clock in the FLEX-1500.
It is important to make sure that fldigi sees a frequency
as close as possible to 1,000Hz. Otherwise the frequency
deviation will be skewed in value and it can have a larger
variation when locking happens:
At 14:30 the continuous carrier started at RWM. Until
around 14:34 the cursor frequency (in the bottom part of
fldigi's window) was set at 1,014 Hz, you see a large error
(around 10) and a large variation. From around 14:34 the
cursor frequency was put at 1,024Hz and you see a smaller
error (around 0) and a much smaller variation. From 14:38,
RWM transmitter is signed off.
between Dec 1st and Dec 10th:
FLEX-1500 scanned between three frequencies (interval
25sec): 4,995,013, 9,995,027 and 14,995,040Hz
between Dec 14th and Dec 20th:
FLEX-1500 scanned between three frequencies (interval
25sec): 4,995,012, 9,995,024 and 14,995,037Hz
Only RWM signals
that are a continuous carrier were analysed. These
continuous carriers are from xx:00 to xx:08UTC and xx:30
Outliners due to less successful fldigi's
Frequency Analysis (Vpk <0.2) were
Results have not been averaged
The frequency deviations for each radio frequency
have been shifted. For each radio frequency, the
shift has been determined in such a way that the
average frequency deviation over a day is zero.
Right click the below pictures to see an enlargement of the
picture in another window/tab.
In some cases no frequency deviation has been found during a 8
min time slot, this can be due to too much noise (low Vpk).
blue dots: 4,996kHz RWM
green dots: 9,996kHz RWM
red dots: 14,996kHz RWM
(dashed) purple line: MSK Sun rise/set
(dashed) orange line: Midway RWM Sun rise/set
(dashed) brown line: Lc Sun rise/set
dotted back line: (begin/end) eclipse
miniwhip-pro active antenna @
7m height indoors,
1st stable setup
<deviation between -1 and 1Hz>
Upto 13:30UTC startup problems with
hard- and software.
crashed between 01:00 and 09:00UTC
From 19:30UTC 4m height indoors.
miniwhip-pro active antenna @
4m height indoors,
2nd stable setup <deviation between -1.5 and 1.5Hz>
The antenna is not at an optimum height and place.
Before Dec. 1st and after Dec 11th
it was at 4m height indoors and from Dec. 1st
to Dec 10th it was at 7m indoors.
There looks to be more noise with the 4m heigth indoor
aerial. The 14,996kHz RWM signal is not really heard at
that lower aerial height.
The frequency stability of the FLEX-15000 looks to be
good. Though it is not understood why the radio
frequencies (proportional with 1Hz referenced to 5MHz)
are lower when using the 4m then for the 7m indoor areal
location. If people know a reason, let me know???
Differences in receiving power depending on frequency:
14,996kHz seems to be received only between Sun
rise and Sun set Midway RWM
9,996kHz is received from around Sun rise Midway
RWM and stops around Sun set at receiver
4,996kHz continues up to the Sun rise at the
receiver. And looks to start again after the Sun set
Differences in frequency deviation depending on
14,996kHz and 9,996kHz look to behave similarly.
At Midway RWM Sun rise they start high, then they go
down upto Lc Sun rise, and increased during the day
and then decrease and increases again and Midway RWM
4,996kHz There are changes around Midway RWM Sun
rise and Midway RWM Sun set and also a possible step
during the night
At what location do the layer
height change happen:
Smith (1951, page 256, 260): "The net electron
density [of E-layer] is seen to follow the
changing altitude of the sun without an
appreciable time lag.".
What should the Sun
rise/set dip angle be, that
changes the E-layer (around
110km) and the F-Layer (around
250km in winter). See here for some thoughts.
The critical frequency (fo) and height of the layers
during the day can be seen in below illustrative pictures
(so not necessarily correct for the location in this eclipse
research, but the form/behavior is similar).
A height and foF2 plot for the
F-layer is here (Smith, 1951, page
Here is an almost real time view of the fo and height
of the F2-Layer: IRTAM and GAMBIT
D-layer's absorption has similar (symmetrical) diurnal behavior as
the E-layer (Smith, 1951, page 260), but it is of course about
absorption (instead of reflection).
A vertical beam with critical frequency (fo) will just
be reflected at the mentioned layer (beams of higher frequency
will pass through).
In case the beam is not vertical a higher reflective higher
frequency (MUF: Maximum Usable Frequency) can be reached. This is
related to the MUF-factor.
MUF = fo * MUF-factor
<Poole (2004, page 44) looks to have a
typo: instead of the '*' it has wrongly a '/'>
For a path of some 2250km (which is the when looking at RWM and
Lc); the E-layer has a MUF-factor of
4.8 and the F-layer a MUF-factor of 3.2 (Poole, 2004, page 44).
So 4,996kHz might just use the E- and F2-layer behavior and the
9,996kHz and 14,996kHz will most likely only use the F-layer
behavior (for our case: DEC. 1936/left plot in above picture).
A height plot over the day for the E-layer and F2-layer can be seen above. Winter time
curves are important for this particular eclipse research.
The frequency deviation (doppler effect) will depend on the height
change. So one needs to differentiate these plots.
The speed (black line) of the E-layer after differentiating is
(Sun rise @ 07:40 and Sun set @ 15:40):
Before Sun rise and after Sun set the E-layer has vanished.
The speed (yellow line) of the F2-layer
after differentiating is (Sun rise @ 07:15 and Sun set @ 16:45):
Sun's dip angles
What does 'Sun rise/set' mean at the E/F-layers of 110/250km:
towards sea level (around 10/18° which are the Sun's visible light
dip angles as used by Verhulst&Stankov, 2014) or smaller
(related to UV) dip angles? A large dip angle would provide too much
absorption of the Sun's far/extreme UV rays through the air below
E-layer. <Sun's dip = - Sun's altitude>
So a Sun's dip angle closer to 4° is more likely when interpretating
the observations of Verhulst&Stankov (2014, Figure 5). The
following picture (the purple and green dotted lines follow the
first/last observability of the E-layer) shows this:
Based on the E-layer (at 110km): the layer
gets enough Sun rays from a Sun's dip angle of around 4°, which is
equivalent to a distant effective horizon of 90km. 90km is the bottom (this might be close the Kármán line [Verhulst&Stankov, 2014, page 7])
where the far-ultraviolet rays can reach (up to which the
influence of the E-layer ionisation
The extreme-ultraviolet that causes the F-layer has a bottom of
influence around 150km. This would result in a approximate
10° dip angle
for the Sun, when assuming an F-layer at 250km. This would be around 1 hour earlier that Sun rise
at sealevel, which maps somewhat the behavior seen in the
above F-layer height graph. <IMHO, a Sun's dip angle of around 10/18° due
to visible light, as used by Verhulst&Stankov (2014), does
not look correct. Let me know your ideas>
We need to combine, the time of day, the speed of the E- or
F-layer, MUF of E- of F-layer, the Sun's dip angle and the
absorption of D-layer, to provide a theoretical frequency
deviation over the day. The below is all about the qualative
An attempt has been made to simulate the behavior at the three
This has been derived from the above Layer-Heights, Layer-fo,
dHeight/dTime, D-layer absorption, dip angle and
adjusting for the Sun's position @ Midway. The deducted MUF-factors
are around the expected 3.2 (for F-layer) and 4.8 (for E-layer).
The simulation in general looks to follow the timing of the
measurements more or less. The F2-layer downward's trough just
before Sun rise is not visble in the 4,996kHz curve as that
F2-trough is partly in de 'shadow' of the E-layer (remember the
radio waves are under an angle of some 12°/18° when reaching the E/F-layer).
The F2-layer height (hmF2) has been traced from GAMBIT for 16-12-2021
at Midway RWM (red curve). To get the speed one needs to
differentiate the height (purple curve),
This speed can be transformed into a
frequency deviation (black curve in below graph) by using:
Frequency deviation [Hz] (doppler effect) = ((<light speed
[km/sec]> + <speed of layer [km/sec]>)/(<light speed
[km/sec]> - <speed of layer [km/sec]>) - 1) * <Station
frequency [Hz]> * MUF-factor
<the MUF-factor [3.2] is included, as the
radio waves do not fall perpendicular on the F-layer>
This gives the following picture:
This black curve looks to correlate with the frequency deviation
measured with the FLEX-1500. 4,996kHz measurements (green dots)
deviate near Sun rise/set, as that frequency is influenced during
those periods by the E-layer instead of F-layer. Need to check if this matching can be matched on other dates!
I think there migth be a different behavior when we swap the
receiver and transmitter locations; as the layers are
Influence of eclipse
The total eclipse in Antarctic is far away, does this have an
influence on the ionosphere around middle Europe?
Perhaps this can be compared with the 21 August 2017 total
eclipse and the related results.
The influence on the ionosphere of a total eclipse looks to be
around 45° (from the centre of the eclipse), as the Dec 2021
eclipse is at around latitude 76° S, its influence might extend to
30° S, so quite far away from Midway RWM (53° N).
measurements, no real effect can be seen in The Netherlands
from the total eclipse. But perhaps statistical analysis (by the Antarctic Eclipse
Festival) could proof differently. The measurements have
been shared with that group and Zenodo.
During two weeks the RWM continous carriers has been measured
using the SDR FLEX-1500. The radio frequency is not 100% the
station frequency. To get to hear the RWM station signals at 1kHz,
we need to put the radio on a slightly different frequency. But
this difference does not seem to change over time, so the
frequency stability of the FLEX-1500 looks to be good. It is
unlikley due to the height of the indoor aerial, but it could
difference in room temperature. Need to investigate further.
The E- and F-layers at the midway location between transmitter
and receiver is determining the effects of these layers.
The E- and F-layer are illuminted by the far/extreme UV rays and
this starts with Sun's dip angle of respectively 4° and 10° (so
not the same Sun dip angle if it were visible rays, which would be
10° and 18°).
At a qualitative level we can simulate the freqeuncy deviation
(due to the doppler effect) of radio ways based on modeled E- and
F-layers (by looking at [change of] fo/MUF, change of
layer-height, absorption in D-layer, time of day and location).
The measured frequency deviation did not really change
significant during the day of the eclipse. This is perhaps also
not expected so far away from the total eclipse location (distance
between Poland and Antarctic).
When results of Antarctic Eclipse Festival are available; a
reference to them will be included.
I would like to thank the following people for their help and
constructive feedback: Kristina Collins, Tobias
Verhulst and all other unmentioned people. Any remaining errors in
methodology or results are my responsibility of course!!! If you
want to provide constructive feedback, let me know.
Collins, Kristina et al.: Citizen scientists conduct distributed
Doppler measurement for ionospheric remote sensing. In: IEEE
Geoscience and Remote Sensing Letters (2021), pp. 1-5.
Poole, Ian D.: Radio propagation: Principles and practice Radio
Society of Great Britain 2004.
Smith, Newbern: Influence of the Sun upon the ionosphere. In:
Proceedings of the American Academy of Arts and Sciences 79
(1951), issue 4, pp. 254-265. Reijs Victor. (2021). Frequencty deviations
seen in RWM carrier wave (Version 1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.5774050
Verhulst, Tobias G.W. and Stanimir M. Stankov: Height-dependent
sunrise and sunset: effects and implications of the varying times of
occurrence for local ionospheric processes and modelling. In:
Advances in Space Research 60 (2017), issue 8, pp. 1797-1806.
Witvliet, Ben and Erik van Maanen: Impact of a Solar X-Flare on NVIS
Propagation Daytime characteristic wave refraction and nighttime
scattering. In: IEEE Antennas and Propagation Magazine (2016), pp.