Northern
Utah WebSDR Receiving equipment Port frequency response plots
Figure 1:
Block diagram of the various RF signal paths at the Northern Utah
WebSDR from the TCI-530
omnidirectional antenna and from the VLF antenna system. Click on the image for a larger
version.
RF Distribution and
Filtering at the Northern Utah WebSDR:
Perhaps the most complicated aspect of the Northern Utah WebSDR is that
of the RF distribution system: It is not enough to simply take
the RF from an antenna and distribute it to multiple receivers, but it
is imperative - for both optimal performance and, less obviously,
lightning protection, - that the RF signal path be protected in some
way - specifically, via filtering. While individual-band
receivers may have their own band-pass filtering,
it has been determined that such filtering is generally inadequate on
its own to prevent overload of the receiver on the many, very high
power signals that often appear on adjacent bands when conditions are
favorable.
There are currently two systems in use - the TCI 530 omnidirectional
antenna, and the LP-1002 Log Periodic beam - and the philosophy for
both antennas' signal flows are the same:
Individual band-filtered receive ports.
The WebSDRs themselves have individual recievers for each of the
amateur bands covered, and each of these ports passes only the band of
interest - more or less.
Broadband receive ports. These ports are (more or less) un-filtered in terms of frequency passband and are intended for "DC to Daylight" SDRs (e.g. KiwiSDRs)
that might have use of continuous coverage across the HF
spectrum. This is the one place in the system where a two-way
splitter is implemented: Half the signal goes to the system of
band-pass filters mentioned above, and the other half goes to this
port. With less than 4 dB of loss, this amount of attenuation is
still low enough to provide plenty of margin to detect the noise floor
of even the quietest site at 10 meters, given that other losses (transmission line, etc.) are kept low.
These plots aren't perfect:
The discussion that follows relates to the frequency response of each,
individual port. These plots were taken using a VNA connected to
a 20 dB directional coupler at the antenna port, while the system was connected to an antenna and in use(I didn't feel the need to interrupt the RF path during the testing and inconvenience the users!) which means that at least some of the observed "ripple" is due to standing waves on the line due to reflection (mismatch)
on the input of the RF distribution network and the antenna feedline
itself. In a few instances, there are a few "spikes" in the plots
that are caused by very strong signals on those frequencies that
influenced the detector in the VNA - but, these can be generally
ignored.
The RF distribution system used at the Northern Utah WebSDR for HF is NOT based on splitters of the traditional sort where one incurs losses for each split (ideally 3 dB for two-way, 6 dB for four-way, etc.) but uses a series of filter networks to "pick off" individual segments of the HF spectrum
Comments about the plots:
These plots were generated using the DG8SAQ WVNA and importing the generated s2p ("Touchstone") files into a spreadsheet (LibreOffice Calc) and generating plots. Because the VNA's signals were coupled into the (live)
antenna system with a 20dB directional coupler, this factor has been
compensated in the plots. Some of the plots will also show a gain
in the signal path and this is due to inline amplifiers: Specific
details will be included in the discussion of each plot.
It is important to note: These plots to NOT tell the reader anything about the antenna or its performance, but rather the response of the RF signal path between the antenna input and the receiver.
Figure 2:| Plot of the 160 meter port of the omni antenna system. Click on the image for a larger version.
Omnidirectional antenna system:
The omnidirectional antenna at the Northern Utah WebSDR is a TCI-530, a 94 foot (29 meter) high tower with a maze of wires, rated to provide coverage from 3 to 30 MHz - but it is usable (receive-only) to about 400 kHz - and is used on all bands from 630 through 10 meters. The feedline from the antenna is about 350 feet (107 meters)
of buried 1-7/8" "Heliax", which provides a matched loss well under 1
dB across the design frequency range with no active devices (amplifiers) inline between the building and the antenna.
For the lower bands (below 30 meters)
the noise floor is high enough that significant signal losses can be
tolerated before signals are lost to this floor - even in "quiet" band
conditions, but the situation changes as frequency is increased and
it's for this reason that the Omni receive system has a single block
amplifier that operates on the "10 MHz and above" signal path.
160 meters:
Figure 2 shows the response plot of the 160 meter port on the omni signal path.
Of particular interest is the response across the 160 meter band (1.8-2.0 MHz)
which shows about 5 dB of ripple. In actuality, the ripple of the
filter is somewhat less than this, but because the TCI-530 is
mismatched at this frequency range - which is below its design limit of
3 MHz - we are seeing the effects of mismatch on the line.
Nevertheless, this represents less than on S-unit of variation and is
not really noticed in practice. The minimum attenuation is a bit
more than 3 dB, caused by the two-way splitter (described above) that splits this antenna's signals out to the "broadband" path.
Specifically, we can see that attenuation of this filter increases
dramatically below 1.7 MHz - the top of the AM broadcast band - and
without this amount of attenuation, many modern HF radios would simply
overload on the total RF power (about -5dBm)
that emerges from the antenna's feedline during the daytime when many
AM broadcast stations are operating at full power - including three
50kW, two 10kW, and two or three 5kW - one of them being the peak seen
at 1.6 MHz.
Figure 3:| Plot of the 80 meter port of the omni antenna system. Click on the image for a larger version.
Above 2 MHz the drop-off is more gradual as there are no strong signals
up there in the Western U.S. - although reasonable attenuation is
desirable as that reduced the total power of lightning static that can
momentarily overload even high dynamic range receivers.
80 meters:
The response of the 80 meter port can be seen in Figure 3.
This filter - like the 160 and the 40 meter filter - consist of a
cascaded high and low-pass networks, which simplifies design of a
band-splitting distribution network that produced multiple outputs for
different frequencies without simple splitting of signals, which can
incur high signal losses without providing any band-pass response.
There are two
antennas used for HF reception on site - and this portion describes
that associated with the TCI-530 omnidirectional log periodic antenna
which is designed for transmit coverage from 3 to 30 MHz, but is used
for reception down to about 400 kHz: Below this frequency, a
separate active antenna is used as noted below.
Across 80 meters (3.5-4.0 MHz)
the passband ripple is on the order of 2dB - a fraction of an
S-unit. One of the design goals of this filter was to minimize
the signal intrusion from the 60 and 49 meter shortwave broadcast bands
where - especially in the latter case - extremely strong signals can be
present, particularly during winter nights. The high-pass
response (below 3 MHz) is
also by necessity, as this is partly how the 160 meter signals are
separated from the 80 meter signals with mnimal loss. Like the
160 meter band, the minimum loss is 3 dB, caused by the two-way
splitter.
Figure 4:| Plot of the 40 meter port of the omni antenna system. Click on the image for a larger version.
40 meters:
The 40 meter band-pass response of Figure 4
shows the effects of the high and low pass filters. Originally,
it was intended that the output of this port might be split to provide
60 meter coverage, but this was later implemented from the "Broadband"
port, instead: In retrospect, it would have been better to
implement "stronger" filtering to better-attenuate signals from the 49
meter shortwave broadcast band (approx. 5.8-6.4 MHz) - but interference issues have not been noted with the current receivers.
As can be seen, the response is flat within 1 dB or so across the 40 meter band (7.0-7.3 MHz)
with significant attenuation below 5 MHz and above 9 MHz - the latter
nicely attenuating the 31 meter shortwave broadcast band.
At least through 40 meters, one need not worry too much about signal
path losses as the noise floors on this and the lower bands are very
high - even during "quiet" band conditions, so up to this point there
is no amplification preceding the filter, and doing so reduced the
probability of the generation of intermodulation distortion from the
multiple, strong signals that might be present because any such
amplifier following the filtering would be seeing fewer signals overall.
30 meters:
Figure 5:| Plot of the 30 meter port of the omni antenna system. Click on the image for a larger version.
At the lower frequencies, the width of each amateur band is a
significant portion of the frequency: For example, at 500 kHz, the 75
meter band consumes over 10% of allof
the frequencies below 4 MHz meaning that it is quite reasonable to
construct band pass filters that consist of a low and high-pass network
- but once you get higher than, say, 40 meters, this becomes
increasongly complicated. Similarly, on lower bands such as 80
meters, it's somewhat awkward to construct a true "band-pass" filter
that covers this same 10+ percent bandwidth - but it gets easier to do
so on the higher bands.
For these higher bands, the filters are of the series input type, which
means that each of the bands' filters can effectively be connected to a
common bus, but with the input of each filter being a high-Q series L/C
(coil-capacitor) network, the effect of each of these filters is (more or less)
limited to its frequency range, leaving the other frequencies alone -
and these filters can also be made somewhat sharper to
better-accommodate the comparatively "narrow" amateur bands that, in
terms of percent of the frequency, are closer together.
The down-side of this topology is that these filters have slightly higher losses - and these losses increase with "tigher" (narrower)
filters and, to a degree, frequency as well, so amplification has been
added. In the signal path is a separated "above 10 MHz" port (for which I neglected to take measurements with the VNA) to
which this amplifier is connected and its gain overcomes the loss of
these filters which assures that it will still be possible to hear the
noise floor at 10 meters.
Figure 5 shows the response of
the 30 meter filter, connected on the signal bus beyond this
amplifier. The rapid drop-off below 9.5 MHz is mostly due to the
high-pass nature of the "above 10 MHz" port itself with a drop-off of
about 20 dB above about 11.7 MHz. In the middle of the passband
the overall signal gain is about 14 dB provided by an amplifier based
on the 2N5109 which, itself, has a gain of about 18 dB.
Figure 6 and 7:| Figure 6 (left): Plot of the 20 meter port of the omni antenna system. Figure 7 (right): Plot of the 17 meter port of the omni antenna system Click on an image for a larger version.
20 meters:
The plot of the 20 meter port may be seen in Figure 6.
Using the same series-input band-pass filter as the 30 meter filter, it
provides flatness to within about 2dB across the 20 meter band (14.0-14.3)
MHz with the -20dB points being below about 12.2 and above 16.5 MHz,
providing reasonable attenuation of the 25 meter shortwave broadcast
band - but relatively little attenuation to the immediately adjacent 22
and 19 meter bands - but this has not proven to be a problem, yet.
Most of the lives of the world's WebSDRs has been during waning periods
of Solar Cycle 24 - which means that these same systems have yet to be
exposed to the expectedly high signal levels that will likely occur
during the rise and peak of the (current)
cycle 25 on the higher HF bands. Many of the receivers used on
WebSDRs are not particularly narrow - often being of the "octave" type
where the pass-band covers something short of a 2:1 frequency ratio
using cascaded low and high-pass filters. What this means is that
as the frequency increases, more and more spectrum is covered by a
given filter which, if the bands "open up", an increasing amount of RF
energy will impinge on the front end of these receiver, potentially
causing overload issues that have likely gone unnoticed at the time
that this page was originally written.
The band-pass filters used at the Northern Utah WebSDR are very much
narrower than those typically found built into receivers, but we are
prepared to construct even better ones if the need arises!
17 meters:
With the advent of the WARC bands, we now have HF bands that are no longer harmonically related to the lower bands (like 40 meter)
- and this also puts these bands, in terms of percentage of frequency,
quite close together. This can pose a bit of a challenge in some
situations, such as antenna design, where a designer must be careful to
minimize unwanted interaction between adjacent bands - and the
relatively "closeness" of the WARC bands makes this arguably more
difficult.
This is also the case for 17 meters where we must adequately isolate 20 and 15 meters - and the plot in Figure 7 shows this. Across the 17 meter band (18.068-18.168 MHz)
the passband is flat within about 1 dB, dropping off by 20 dB below
15.75 MHz and above 20 MHz - more than enough to afford isolation
between it an the adjacent bands.
Within the passband, the overall gain of the signal path is approximately 15 dB.
Figure 8 and 9: Figure 8 (left): Plot of the 15 meter port of the omni antenna system. Figure 9 (right): Plot of the 12 meter port of the omni antenna system Click on an image for a larger version.
15 meters:
The 15 meter band-pass filter (Figure 8) shows a response that is flat to about 1 dB across the 15 meter band (21.0-21.45 MHz)
with the -20dB points below 19.25 MHz and above 23.75 MHz. The
overall gain here is down very slightly compared to 17 meters at about
14 dB.
12 meters:
The 12 meter band (24.890-24.990 MHz) can be one of the most awkward to manage because it is so close (in percent of frequency) to the 10 meter band - with the top of the band (about 25 MHz) only three MHz away from the bottom of the 10 meter band.
The plot of Figure 9 shows the
response of this filter: It's flat within about 1 dB across the
12 meter band, but it is down by only about 15 dB at the bottom end of
10 meters, but it manages to be down by over 30 dB by the top of the 15
meter band.
The overall gain of the signal path at this point is the same as that
at 15 meters - approximately 14 dB. In theory, this filter could
be made a bit sharper and, with the amplification present, more loss
could be tolerated.
10 meters:
Figure 10:| Plot of the 10 meter port of the omni antenna system. Click on the image for a larger version.
Figure 10 shows the response of the 10 meter filter. Across the passband (28.0-29.7 MHz)
the passband shows a bit more ripple than the others - apparently a
result of the interaction between this and the 12 meter filter, but at
only about 3 dB across the entire 1.7 MHz span, it's reasonably good,
considering.
One saving grace is that the 10 meter band isn't particuarly close to
any of the commonly-used shortwave broadcast bands - the highest of
these being the 13 meter band, just above 15 meter which manages to be
attenuated by this filter well over 30 dB. For this filter, the
20 dB points are below 26.25 and above 32.25 MHz - surprisingly good.
Not shown in the diagram is the response into the FM broadcast band -
frequencies that may be strongly intercepted by this antenna (and having a transmitter within a few miles/km of the site doesn't help)
where the attenuation is on the order of 30-40 dB. No issues have
been noted thusfar related to this, but the addition of a low-pass
filter, likely set just above 55 MHz, is being considered as this will (experimentally) allow the reception of 6 meters, but quash the broadcast band signals.
"Other" ports on the Omni signal chain:
Already mentioned is the "Split" port on the Omni signal path - one
that, with no intentional filtering, splits the signal from the Omni
from the path that comprises the filters above to a "broadband" path
that goes to wideband SDRs such as the KiwiSDRs - and it allows
coverage of non-amateur bands (shortwave broadcast) as well as "non-traditional" amateur bands like 60 meters.
Figure 11:| Plot of the unfiltered split port of the omni antenna system. Click on the image for a larger version.
Unfiltered "Split" port:
Figure 11 shows the response of the unfiltered "Split" port on the Omni antenna showing a fairly flat response (within a few dB)
from well below the AM broadcast band to above 30 MHz. As
observed with the 160 meter plot, the "ripple" is more prominent below
3 MHz where the antenna itself is not matched at all to 50 ohms.
The spikes below 2 MHz are caused by the ingress of signals from strong AM broadcast transmitters nearby.
The "BCB Reject" port:
The output of the unfiltered "split" ports goes to a separate module
that consists of an AM broadcast band filter with amplification.
This filter - which is adjustable in both ultimate attenuation and has
at least six tunable notches for the strongest signals - leaves the
signals below approximately 510 kHz and those above 1.7 MHz (mostly) un-touched.
Figures 12-15:| Figure 12 ( upper left) shows the output of "BCB Reject Port" #1 while Figure 13 (upper right)
shows the output of Port #2. These are identical, aside different
amounts of amplification. The effects of the AM broadcast band
filtering and the high-pass equalization network - used to
preferentually attenuate lower-frequency signals - are apparent.
Figure 14 (lower left) Full bandwidth plot of the KiwiSDR port on the Omni antenna. Figure 15 (lower right) is the lower portion of the same port. Click on an image for a larger version.
Figures 12 and 13 show the
responses of these two ports - identical except for the amount of
amplification: Port #1 has a single 15 dB gain amplifier while
port #2 has another, identical amplifier - although internal splitter
losses cause these gain differences to be closer to 11 dB resulting in
an overall gain of about 9 dB and 20 dB for ports 1 and 2,
respectively. The increased attenuation at the higher end of the
range (above 25 MHz) are a result of roll-off of the amplifier and losses of the AM BCB filter itself.
Quite apparent is the attenuation of in the AM broadcast band which is
intentionally limited to 20-25 dB overall to allow the reception of AM
broadcast band signals - but knocking them down in amplitude enough to
prevent overload. Not obvious from the plot is that there a
number of steerable notches set to the frequencies of particularly
strong signals.
KiwiSDR port (omni):
In addition to the WebSDR servers, there are other web-enabled
receivers at the Northern Utah WebSDR site - the KiwiSDRs. These
are stand-alone Linux-based receivers that are capable of receiving
from nearly DC to 30 MHz - but they can support a very limited number
of users.
Because these are "direct sampling" receivers (e.g. an analog-to-digital converter that inhales the entire HF spectrum)
they are connected to the "broadband" port to allow reception over the
entire frequency range. One issue with these "direct sampling"
receivers is that the A/D converters used in these receivers - as with any
A/D converter - can accommodate definite maximum amount ot signal power
before it is driven into saturation. As it turns out (math is involved!) the dynamics of signals across the entire HF band (2-30 MHz)
are such that even an 18 bit converter is likely not to have enough
dynamic range to accommodate both the very strong HF signals present at
the low end of the HF range and the rather low noise levels found at a
quiet receive site at the high end of the HF range - at least not
without some additional measures such as AGC and/or filtering.
Many modern HF receivers/transceivers use direct sampling, and to allow reasonable performance both AGC (Automatic Gain Control)
and filtering are used to reduce the total signal level applied to the
A/D converter - but if the intent is to have a full-bandwidth, shared
HF receiver, filtering cannot be used. For example: If you have 5 people tuning randomly throughought the HF spectrum, how and where would you tune a filter?
The KiwiSDRs have 14 bit A/D converters - but only a modest amount of
gain in fronter of the converter which means that if one has a rather
quiet listening location for HF, the noise floor of the KiwiSDR is
approximately equal to the noise floor from a unity-gain antenna at a
quiet site above approximately 14 MHz, requiring about 10 dB of
additional amplification to be able to hear the noise floor at 10
meters. The problem with doing this is that adding 10 dB gain
will cause the A/D converter to overload on lower-frequency HF signals
- and that doesn't even include the effects from local AM broadcast
stations. To mitigate this effect, the KiwiSDR signal path
includes a high-pass network that attenuates lower frequency HF
signals, but has minimal effect on higher frequencies and in this way
it is possible to use amplification to attain the 10 meter noise floor
while reducing the probability of overload at lower frequencies.
The plot of Figure 14 shows
this slope with about 15 dB of attenuation between 7 and 15 MHz.
Ideally, the effect of this attenuation would be minimal above 15 MHz,
but there is a bit of roll-off intrinsic to the signal path (as seen in Figure 12)
- but there is also gradual signal attenuation between 20 and 32
MHz: This is likely due to roll-off of the amplifer, but its
cause will have to be investigated, which is one of the reasons why
these plots were taken. (Yes,
the KiwiSDRs' noise floor are currently higher than ambient on the 10
meter band - something that will be addressed during a future site
visit.)
Figure 17:| Plot of the <=500 kHz port of the omni antenna system. Click on the image for a larger version.
The "Below 500 kHz" port:
Even though the minimum frequency of the TCI-530 Omni antenna is specified as being 3 MHz, it is an effective receive
antenna on 160 meters, and it also works nicely as a receive antenna to
at least 400 kHz. As the frequency drops, the gain of the antenna
will drop significantly and its feedpoint impedance will be nowhere
near 50 ohms, but it is still quite useful, being able to hear the RF
noise at these frequencies. Built into the main filter system is
a means to separate these lower frequencies (below the AM broadcast band) to be fed to a receiver dedicated for use on the 630 meter amateur band.
The plot of Figure 16 shows the
response of this port. As expected, there is a fair bit of ripple
in the response - likely due to the antenna being something other than
50 ohms - but it lives up to its name, with the response dropping off
above 500 kHz and down by at least 50 dB at 1 MHz.
+The LP-1002 beam antenna system:
Another antenna at the Northern Utah WebSDR is a Hy-Gain (now U.S. Antenna Products) LP-1002 log-period beam. Pointed at 87 degrees (true north reference)
this antenna favors the Eastern U.S., but its beamwidth - being rather
broad - makes it useful for coverage as far north as extreme southern
Europe. With rated frequency coverage of 6 through 42 MHz, it is
used for coverage of the 40 through 10 meter amateur bands, and it was
designed using the same philsophy as the system used for the
Omnidirectional antenna.
Figure 18: Figure 18:
Plot of the 40 meter port of the beam antenna signal path. The
anomalies in the plot are due to strong, external signals. Click on the image for a larger version.
What is not included in these plots is the fact that there is amplification (about)- and some filtering - located at the base of tower on which it is mounted - the gain of this amplifier and this filter is not reflected in these plots.
When this antenna was first brought online, the amplifier was found to
be being badly overloaded by both AM and FM broadcast stations,
requiring the implementaiton of a 3 MHz high-pass filter to remove the
AM broadcast sigtnals and a 60 MHz low-pass filter for the FM broadcast
signals.
40 meters:
When the filtering for the 40 meter filter on the beam was designed,
consideration was given to the fact that with this antenna having
between 10 and 13 dBi of gain at a fairly low elevation, signals could
be quite high, so additional attenuation of 49 meter shorwave broadcast
band signals was implemented.
The passband of this filter is flat to better than 1 dB across the 40 meter band (7.0-7.3 MHz)
with the -20dB point below approximately 6.3 MHz. Interestingly,
this filter has a spurious response around 9.4 MHz, but it gradually
rolls of to greater than 20 dB of attenuation above approximately 10.5
MHz - just off the right side of the plot.
As it happened, the 40 meter receiver was overloading on strong
shortwave broadcast signals - but from those in the 41 meter shortwave
band which, in North America, starts at the top end of the 40 meter
amateur band (7.3 MHz) goes
up to at least 7.45 MHz - making it extremely difficult to
filter. Ultimately, another rather complicated filter had to be
constructed - one that was able to provide reasonable attenuation just
above 7.3 MHz while minimally affecting 40 meter amateur
reception: This filter is described HERE.
30-10 meters:
Figures 19-24:| Figures 19-24, starting in the upper-left down to the lower right, showing the responses on 30-10 meters on the beam. Click on an image for a larger version.
At 30 meters and above, the signal path is very much like that on the
Omnidirectional antenna, but with the benefit of better test equipment
such as a VNA, and larger-gauge wire which reduced loss and improved
filter "sharpness" somewhat. Because of this, the discussion of
these plots will be abbreviated.
As with the Omni antenna, there is a "Above 9.5 MHz" port that feeds
another module with the 30-10 meter filters, but unlike the Omni signal
path - which has an amplifier at this point - the signal path on the
beam antenna does not
have an antenn at this point. Because there is gain at the base
of the tower, near the antenna, we can afford several dB of loss
through the band-pass filters before
we amplify the signal, and because the amplification follows the
filter, it will be less prone to overload - particularly important
considering that there is already gain provided by the beam itself and
the earlier amplification stage.
30 Meters: Figure 19 shows the response of the 30 meter signal path - down by 20 dB at 9.25 and 11.25 MHz with 8 dB of gain in the 30 meter band..
20 Meters:Figure 20 shows the response of the 20 meter signal path - down by 20 dB at 12.75 and 16.0 MHz with 8 dB of gain in the 20 meter band.
17 Meters:Figure 21 shows the response of the 17 meter signal path - down by 20 dB at 16.0 and 20.0 MHz with 12 dB of gain in the 17 meter band.
15 meters:Figure 22 shows the response of the 15 meter signal path - down by 20 dB at 19.25 and 22.75 MHz with 8 dB of gain in the 15 meter band.
12 meters:Figure 23 shows the response of the 12 meter signal path - down by 20 dB at 22.0 and 27.75 MHz. Note: Because there is not currently at 12 meter receiver on WebSDR #4, this port is unused and there is no amplifier present.
10 meters:Figure 24 shows the response of the 10 meter signal path - down by 20 dB at 25.75 and 32.5 MHz with 7 dB of gain in the 10 meter band.
Two other ports on the beam's signal path:
There are two more signal ports found on the signal path of the beam antenna:
"Unbuffered" Split port:
Figure 25 and 26: Figure 25 (left): Plot of the "Unbuffered" split port on the beam. Figure 26 (right): Plot of the "Below 6.4 MHz" port on the beam. Click on an image for a larger version.
Just like that of the omni, this is the output of a 2-way splitter that
has no filtering of its own. As with the Omni, this is intended
for use with broadband receivers like the KiwiSDR and/or for use on
frequencies that do not include tranaditional amateur
bands. The response of this port may be seen in Figure 25. This plot shows a fairly flat response with occasional spikes that correspond with strong, off-air signals.
The "Below 6.4 MHz" port:
Although the beam antenna is specified for coverage down to 6 MHz, that
doesn't mean that it might not be useful at lower frequencies, and for
this reason a port was included to pass frequencies below the 40 meter
amateur band. As it happens, the antenna does
provide good reception down to approximately 1 MHz, but it loses
directivity very quickly below 6 MHz and becomes more or less
omnidirectional. Testing was done with the on-site WSPR receivers
on 160 and 80 meters, comparing it the results with those obtained on
the omnidirecational antenna, and it was found that it was comparable
on 80 meters and somewhat worse on 160.
The response plot of this port may be seen in Figure 26.
The high-end response is down by about 20 dB at 7.3 MHz with the low
end limited by the high-pass filter located at the tower. The
spikes seen below 2 MHz and above 9 MHz are from strong AM broadcast
and SWBC signals, respectively.
KiwiSDR signal path - beam:
Figure 27: Plot of the signal path feeding the KiwiSDRs on the beam
antenna. The spikes and ripple are due to on-air signals and the
varying impedance of the "live" antenna connection, respectively. Click on the image for a larger version.
The final plot to be discussed (Figure 27)
is that of the KiwiSDR signal path on the beam antenna. As with
the KiwiSDR signal path on the omni, there is a high-pass network that
attenuates lower-frequency signals along with the gain needed to allow
the KiwiSDR to hear the noise floor on 10 meters.
This plot shows that signals at 5 MHz are attenuated about 20 dB
relative to 30 MHz to reduce the probability of strong signals at the
low end of the HF spectrum overloading the KiwiSDR's A/D converter, and
despite this amount of attenuation - and the fact that the beam's gain
drops off rapidly below 5 MHz - the RF noise floor can still be heard
at that frequency.
Final comments:
As mentioned at the top of this page, a significant portion of "ripple"
depicted in these plots is due to the fact that these plots were made
with a "live", on-air antenna system with its varying impedances and
interactions. In several locations, spikes may be seen - and
these are strong, off-air signals that are being picked up by the VNA's
detector.
These plots are mainly for future reference, to establish the general
performance of the existing filters and give a general idea of how much
off-band rejection these filters might provide.
Pages about other receive gear at
the Northern Utah WebSDR:
Softrock Receivers
- This page describes the "High Performance" receivers that use
"Softrock" direct-conversion receivers and sound cards. These
receivers cover limited bandwidth (up
to about 192 kHz) but have excellent weak and strong
signal handling properties.
LF/MF reception using the
Softrock receivers
- The Northern Utah WebSDR's coverage includes the "new" 2200 and 630
meter amateur bands and these are received via modified softrock
receivers.
RTL-SDR Dongle-based receivers
- Described here are the "not high performance" receivers using the
ubiquitous RTL-SDR dongles. These receivers cover up to 2
MHz of
bandwidth, but their limited A/D bit depth (only 8 bits)
means that they can suffer from too much and/or too little signal input
- often depending on band conditions. Included on this page
is
information about how to make the most of these as well as helping to
manage when multiple RTL-SDR dongles are used on a Linux-based system.
RF Downconverter for RTL-SDR
receivers
- While there are RTL-SDR dongles that contain built-in upconverters to
allow reception across the entire HF spectrum, this may not be the best
way to do it. When receiving frequencies at or above the
Nyquist
frequencie(s) on HF, one can downconvert to lower frequencies and get
good results, all described on this page.
An AGC block for RTL-SDR receivers-
Because RTL-SDR dongles have only 8 bits of A/D, their dynamic range is
limited. While one can adjust the gain to fit their useful
signal
range "window", HF band conditions change constantly, making it
impossible to keep one of these receiver's limited dynamics optimized.
Preceding an RTL-SDR dongle with proper filtering and an AGC
circuit can make the best of these devices.
WebSDR #4 receive system - On a separate antenna, WebSDR #4 has its own RF infractructure that includes filtering and amplification.
For general information about this WebSDR system -
including contact info - go to the about
page(link).
For the latest news about this system and current issues,
visit the latest news
page (link).
For more information about this server you may contact
Clint, KA7OEI using his callsign at ka7oei dot com.
For more information about the WebSDR project in general -
including information about other WebSDR servers worldwide and
additional technical information - go to http://www.websdr.org
