ZS6BTE
The
underlying mechanism is VHF and UHF tropospheric scatter. In Fig 1 aircraft
flying close to Rx will produce large Doppler shifts, with the characteristic
amplitude “flutter” and Doppler shift known to all VHF enthusiasts, but this
echo is of a transient nature and is not particularly useful. A QSO over this
distance will be possible under normal operation.
Of
much greater interest is working the path Rx to Tx, and provided the horizons
are not too obstructed distances of 600 km or more may be worked.
Detection
of aircraft uses bistatic radar theory, where the RX and TX are separated, and
there is an angle between the aircraft and both. Then amplitude flutter and the
Doppler frequency shifts are more easily understood, as well as the steady
received signal amplitude and lack of Doppler when aircraft are on the direct
path Rx-Tx.
The
scatter volume ABC has a profound effect on the link and defines the signal
strength received. BC represents the maximum height of around 10000m (33000 ft)
used by jet liners on long flights.
At
altitudes below Tx - A - Rx, aircraft produce no receivable echoes due to
terrain screening and remain invisible.
Fig 1: Tropo scatter path between
Randburg and
When
one of the sites has an inferior horizon, the scatter volume moves closer to it
and is no longer at the mid-way point. This increases the scatter angle and
reduces the overall path capabilities.
My
system monitors the tropo scatter path between Randburg and
Frequency
accuracy is confirmed with a GPS-locked external oscillator used off-air.
The
system is able to continually monitor the signals from this SABC site where the
EIRP is around 160 kW, despite non-optimum horizons only 5-7 km from my QTH.
Fig 2: UHF RX array – 16 phased reflex
dipoles gain about 16 dBi, with MOSFET preamp
Fig 3: SABC
Frequencies:
VHF
175.223998 MHz – on rubidium frequency standard (my measurement value)
UHF
495.249575 MHz – free running oscillator
TV broadcast beams have a
vertical beam width of some 2 degrees and are set up to cover an omni directional
target radius of about 50km, so are aligned at an elevation angle of about -1.5
degrees. Otherwise, if aimed at the horizon typically as much as 100 km distant
from these elevated sites, half the ERP would disappear uselessly into space.
This offset with respect to the horizon of 3 dB or more has not been
incorporated in the calculation for tropo scatter path loss in the Appendix.
Aircraft flying directly
above a TV TX, even only a few hundred metres, produce no aircraft scatter
reflections – they have to be some distance, typically a minimum of 50 km under
typical jet liner flight discipline involving height above local ground level,
to cause reflections except when landing or taking off.
A
waterfall type display with a FFT of 65k is useful in reducing the noise floor
by some 15 dB and transient events are easily displayed on a fast PC. USB is
used, with a convenient offset of about 1 kHz to make the carriers audible.
Using the UHF frequency from the Tx:
Fig 4: Barely detectable aircraft near
One
of the aircraft is visible for 22 minutes.
The
50 Hz ‘birdies’ are QRM from a free running power supply in the neighbourhood.
The TV carrier is barely visible at the noise floor despite the strong FFT.
The
traces arise from echoes off the aircraft bodies; with a very small Doppler
shift around 1-2 Hz at most. Aircraft, previously obscure, become strongly
visible in the vicinity of the direct line.
Note
both aircraft, jet airliners, fading at the top of the scan as they approach
Tx, are wobbling around a few metres as readily seen – the technique amplifies
the effect on the direct line of site Rx-Tx.
Fig 5: Flight radar showing the two
aircraft near
The
flight path from OR Tambo airport-RX-Kimberley
Aircraft
must fly closely along the direct line of site Tx – Rx and in the tropo scatter
volume (sometimes called the ‘common volume’) to produce the wanted enhanced
forward scatter effect. The signal level at Rx is lifted from noise level to
over 30 dB above noise under this condition.
Fig 6: First 50 Hz sidebands at least 25
dB down in a correctly set up
Amplitude
is comparatively stable +/- 2 dB versus the normal several cycles per second
10-20 dB fading on the tropo path. Frequency 495 MHz.
Fig 7:
In
Fig 7 there is no sign of the 50 Hz sidebands around the carrier so the
enhancement due to the aircraft forward scatter exceeds 28 dB based on the
level of the 3 rd pair of sidebands barely visible in Fig 6.
The
carrier at 1634 Hz is Nelspruit TV TX (distance 311 km) off the side of the
antenna.
Fig 8: Mild enhancement, the aircraft is
probably in the space RX-A-C in Fig 1
When
not in the tropo scatter volume, there is still some enhancement of the signal.
Here the aircraft has not entered the scatter volume and is flying in the air
space below A-C, but just above Rx – A (and also directly between Rx and Tx as
evidenced by the lack of Doppler frequency shift on the carrier - peak hold of
several seconds used in the spectrum analyser) so the signal benefit is more
limited, typically around 0 to 20 dB or so.
This
scattering effect greatly increases the virtual height of Rx, which reduces the
angle Tx – A – Rx and the tropo scatter path loss by that amount. The tropo
scatter path loss over such a distance is enormous (~238 dB), approaching that
of an EME return path (~254 dB) at the frequency 175 MHz
Fig 9a: The rise and fall of signal
level around enhancement is very abrupt
As
an aircraft enters or leaves the scatter volume ABC, the signal rises or falls
rapidly, in a second or two. In the figure above the aircraft is entering the
scatter volume. Sidebands of the TV signal, previously invisible, are brought
up strongly above the ambient noise level.
When
in the scatter volume the Rx-Tx path is enabled due to the strong echoing,
enabling QSOs of several minutes.
Fig 9b: Enhancement when possibly still
between Rx and A in Fig 1.
In Fig 9b, the aircraft
highlighted has apparently just taken off from the nearby (50 km or so) OR
Tambo airport as illustrated by the huge Doppler shift, gained some height,
possibly 3000m or so a.g.l, and then lined up with the direct path to
This might explain the
strange, alternating enhancement-slight fade-enhancement-fade shown in Fig 9b
and arises probably from the aircraft barely above the line Rx-A in Fig 1. This
form of enhancement is useless for QSO purposes because the aircraft is well
short of the scatter volume and invisible to a receiver located at Tx (with the
TX now at Rx). For QSO purposes the aircraft should be in the scatter volume at
the nominal mid-point on most tropo scatter paths.
Fig 10: The aircraft 5 Hz HF of the
carrier frequency shows some enhancement, but not enough to show the sidebands
In
the spectrogram above, three aircraft are flying directly between Rx and Tx.
However, one is 5 Hz or so high of perfect alignment and this displacement is
sufficient to nullify the full benefit. So sidebands from the other two
aircraft more closely along the path are visible, note the double sidebands and
the abrupt rise in signal strength as the first of these two aircraft enter the
scatter volume and is visible for 9 minutes.
This
is ample time for a QSO.
In
this approximately 13 minute sequence 10 aircraft echoes are visible.
Fig 11: Severe weather conditions, TV carrier
dispersed on tropo scatter link
When
weather conditions are turbulent on the tropo scatter path, there is
considerable dispersal of the radio carrier in both amplitude and frequency
‘width’. This is very visible in Fig 11. It has been suggested that the
turbulence created in the hot wake of an aircraft is responsible for the considerable
enhancement by refraction and the aircraft’s body is not the reflecting medium.
In Fig 11 there does not appear to be a displacement of the carrier necessary
to support the possibility. It is necessary to observe a disconnect between the
aircraft’s echo and the ‘echo’ from the refraction cone to prove this theory.
The aircraft, seen to be wobbling in the wind, does not have this signature,
despite severe spread of the TV carrier in the prevailing conditions. Compare
to typical conditions in Fig 12 featuring two aircraft near the direct line.
Fig 12:
Fig 13: Lighting flash shows apparent bistatic Doppler
effects
Lightning
discharges appear as semi-regular disturbances and are readily displayed
through 900 MHz. The duration of a decent lightning burst at UHF is as much as
0.6 seconds. A narrow band ham rig in USB is unlikely to capture the huge
Doppler shift, but will capture the residual ionization hanging in the air as
an echo from the carrier.
If
the ionized plasma path created during lighting discharges shows Doppler
effects as in Fig 13, presumably by displacement in the jet stream at
approximately the same altitude and distance as a jet aircraft en route, one
could reasonably expect an aircraft’s wake to be similarly displaced by the jet
stream particularly when across the direct path. This should become visible in
Fig 11, creating a dual echo, one off the body of the aircraft under autopilot
regime and another being the echo resulting from refraction in the displaced
wake which is obviously not under autopilot control. But I have never noticed
this possible refraction from the wake.
In
Fig 13 the antenna is pointed south and the jet stream is from west to east
across the path with a velocity of around 110 miles per hour (50m/s), and in a
period of 0.6 seconds would displace the ionized path about 30m, so the mostly
negative Doppler shift manifest on most of the carriers, except Bethlehem where
the weather is still arriving, would imply an increase in bistatic range being
a horizontal shift of 30m during the time the ionization lasts. This is
displayed in the frequency domain in Fig 13 as a shift of about 30 Hz, but I
have noted much more at times. Of course one never knows if the lighting is
discharging horizontally, or vertically, over possibly several km. Aircraft are
visible on some traces.
The
swing in the carrier frequency at
Rex Moncur, VK7MO published
very useful data on the effect of bistatic forward scatter on various aircraft
and concluded there is an enhancement of about 20 dB on the 2m and about 30 dB
on the 70cm and 23cm ham bands above the expected monostatic (back scattered)
echo level.
He found there is a direct
correlation between the calculated forward scatter radar cross section of the
aircraft and the observed results. In Table 11, experimental results across
these bands, amateur signals received under typical scenarios were between
0.035 and 0.6 µV (-136 to -111 dBm), i.e. from noise level up to solid signals
and this information will be of direct use.
I too, have found
enhancements above noise level around 25 dB at 175 MHz, and some 10 dB more on
UHF as depicted in the illustrations above and can concur with Rex’s
calculations and results.
Aircraft
Enhancement – Some Insights from Bistatic Radar Theory - Rex Moncur, VK7MO
Klein Heidelberg – a WW2 bistatic radar
system that was decades ahead of its time - Hugh Griffiths and Nicholas Willis
Russian / PLA Low Band Surveillance Radars - Technical
Report APA-TR-2007-0901 by Dr Carlo Kopp
Passive Bistatic Radar - Professor Hugh
Griffiths
A Little Flutter on VHF - RSGB Bulletin
Nov 1966 - G3BGL
Tropospheric Path Calculations
These calculations will provide an indication of the signal
to be expected from a distant VHF or UHF transmitter. Anything above this is
the scatter gain off the aircraft. For QSO usage, calculate the S+N/N ratio for
the path using the partner stations’ ERP and add about 25 to 30 dB (VHF vs.
UHF) to test the direct path feasibility – the aircraft must be on this direct
path.
A pure tropospheric scatter path will provide an
average signal around the calculated value, which will pulse up and down
several cycles per second. Note that an increase of 1 degree in the scatter
angle (Fig 12) due to horizon blocking adds 10 dB to the path loss. Changes in
the surface refractivity index will show up readily as day/night or seasonal
variations in link performance.
Figure 11: Bending
of radio beam due to refraction (a = true Earth radius)
Fig 12: Geometry of scatter path
The steps to follow are:
1. Scatter angle
2. Path loss
3. System noise temperature
4. Signal to noise ratio
Using the VHF Tx on 175.25 MHz and log periodic
antenna with no preamp:
Tropospheric Scatter
Angle
R = 4/3 earth radius (8446 km)
d = great circle path distance 428 km
h1 and h2 = respective antenna heights above sea level,1.65 and 1.385 km
h11 and h12 = height of radio horizons above sea level, 2 and 1.235 km
d1 and d2 = great circle distance between radio horizons and respective antennas,
5 and 100 km
scatter angle Ө = Ө0 - Ө1 - Ө2 radians
where Ө0 = d/R
Ө1 = ((h1- h11)/d1 + d1/2R)
Ө2 = ((h2- h12)/d2 + d2/2R)
So scatter angle= 428/8448 – ((1.65-2)/5 + 5/16896) – ((1.385-1.235)/100 + 100/16896)
= 0.050663 – (0.069704) – (0.007419)
= 0.112948 radians
= 6.5 degrees
Tropospheric Scatter
Path Loss
LP = tropospheric scatter path loss, dB
LFS = free space path loss, 92.4+20log dkm +20 log FGHz
LS = normalized all year scatter loss at NS, 57 + 10log(0-1) + 10log(F/0.4)
NS = surface refractivity index, around 300
in
So LP = LFS + LS – 0.2(NS – 300) dB
= (92.4+20log428+20log0.17525) + (57+10(6.5-1)) + 10log(0.17525/0.4)
= (129.9) + (112.0) + (-3.58)
= 238.32 dB
System Noise
Temperature
TSYS = α(Tα) + To(1-α) + T1 + Tm/gm-1
where
α = transmission line coefficient as a factor, 0.8 for 1 dB loss
Tα = temperature of transmission line, k, equal to To for uncooled lines
To = ambient temperature k, typically 290
T1 = temp of 1st RF amplifier stage, 2500k (noise temp of Rx-no preamp used)
Tm = temperature of 2nd RF amplifier stage
gm-1 = gain
of stage T1
since no masthead preamp was used the noise figure of the Rx is taken and Tm/gm-1 is set to = 0
so TSYS = 0.8(290) + 290(1-0.8) + 2500 + 0
= 2790k
Signal to Noise Ratio over
tropospheric scatter path
SNR =
where
Gt = gain
of Tx antenna, dBi (12)
Gr = gain
of Rx antenna, dBi (11.5)
LP = tropospheric
scatter path loss, dB (238.3)
Pn = noise power ratio of Rx 10log FkTB, -150.8, k is
Boltzman’s constant 1.38 x 10-23 and B = 2500 Hz in USB, T= TSYS
, F is Rx noise factor 2500/290 = 8.62
So SNR = 40+12+11.5-238.3-(-150.8)
= -24 dB
An FFT of
262k provides an improvement in SNR of about 30 dB:
So SNR = -24
+ 30
= 6 dB
This is in
agreement with the variable carrier received from the