This page contains some of the more technical details of the
temporary radio
system used at Panorama Point during the Friendship Cruise.
Receiving signals at
Panorama Point:
Fig. 1 - This is the receive site
for the Panorama
Point repeater system showing the three Yagis, weatherproof box with
the equipment, and the solar panels. Click on the picture for a larger version.
When it comes right down to it, while Panorama Point is in a
geographically favorable location to have a fairly straight shot up
both the Green and Colorado rivers, there are several strikes against
it when compared to the "ideal" radio site:
It can't see much river. If you are in the
neighborhood of Panorama point, there are only a few very small
stretches of the Green River that are even visible: The river is
mostly
hidden in the deep gorge below and the Colorado is never visible at
all..
North of Mineral Bottom, the "lay of the land" works against
radio propagation to boats on the Green River. For most of its
course north of this, most of the river is in a
deep gorge - except for those areas near Green River, where the rising
land to the south blocks the view toward Panorama Point.
Most of the Colorado River north of the confluence is blocked by
very
large land masses - namely the Island In the Sky and Junction Butte
masses.
For these reasons, one can't simply throw a portable repeater on the
air
and expect it to work very well. From the initial survey done in
1997, it
was clear that any "upriver" signals that arrived at Panorama Point
would likely
be extremely weak and "multipathy" - not surprising
considering that the signals will have had to find their way out of
deep, narrow river gorges some distance from the receiver site at
Panorama Point. It was decided that, from the first moment, that
extraordinary efforts should be implemented on the receiver to maximize
performance.
Duking it out with Boltzmann:
For any receive system,
there is one fundamental limiting factor that imposes an absolute limit
on how sensitive it can be: Thermal Noise. This
thermal
noise
comes
from two places: The equipment itself, and
the environment in which the receive system is used.
The "Noise
Environment":
As it turns out, everything gives off noise - assuming
that it is warmer than absolute zero - and the warmer it gets, the more
"noise" it puts out. For an example, consider a piece of
metal. At room temperature, it does not "glow" visibly, but if
you were to heat it, it would begin to glow - very dull red at first,
but as it got hotter, it would become closer to being "white
hot." As it cools again, its glow disappears once more to our
vision. Just because it may have cooled off to room temperature,
don't think that it isn't still "glowing" - because it is! As it
turns out, any object that is above absolute zero does glow -
not only at infrared wavelengths (as you may have seen in footage of
police finding criminals in the dark using "heat sensitive" cameras)
but also at plain radio wavelengths where it manifests itself as noise.
Keeping in mind that all warm surfaces "glow" at radio wavelengths, one
can complete the analogy by likening a radio signal to a light source
on this glowing surface: If the surrounding surface is "glowing"
more brightly than the radio signal, that radio signal is simply lost!
The magnitude of this noise can be calculated using this equation:
N = k * T * B
Where:
N = noise power in watts T = temperature in Kelvin B= bandwidth in Hz k = Boltzmann's constant - which is approximately 1.38x10-23
(expressed
in
Joules per kelvin, or J/K)
When numbers are crunched, assuming a 300K (about 80 degrees F)
temperature and a 15 kHz receiver bandwidth, we can calculate that in a
50 ohm system, as is used on a typical receiver, this is equivalent
to a noise voltage of 0.056 microvolts when using an isotropic antenna
immersed in a 300k environment. (Yes, that also means that
there is an equivalent noise voltage of 0.056 uV coming out of a 50 ohm
dummy load that is at 300K: It is through the use of the
intrinsic noise of dummy loads the EMEers can check the relative
performance of the antenna/preamp system!)
This example assumes that we are receiving a signal using an isotropic
antenna, surrounded on all sides by matter that is at 300K: While
this isn't exactly the case in a "real world"
scenario, it is a reasonable approximation of actual operation of a
receive system in which the signals are emanating from the surface of
the Earth and the numbers calculated above illustrate that our receive
system is, in fact, limited by more than the fact that the
signal is just getting weak: The signal is, in fact, being
lost in thermal noise being radiated by the earth itself!
Fig. 2 - Top: One of the
preamps, in-use and
connected directly to the receive antenna with a minimum of loss.
Center: The inside of the preamp showing simple "dead
bug"
construction with the input on the left side. Bottom:
The
schematic
of
the preamp. Click any picture for a larger version.
Comment:
On a narrowband FM receiver, 12 dB SINAD correlates
approximately with a noise voltage at 60% of the signal voltage for a
typical receiver. This
means that in a 50 ohm receive system that is terrestrially based
(that is, the antenna receives a signal originating from the Earth's
surface) or in a test setup involving dummy loads/attenuators that are
at room temperature
the maximum sensitivity possible - no matter how good
your receive system might be (using a standard 15 kHz FM voice channel)
is approximately 0.09 microvolts for 12 dB SINAD!
Without
cryogenic cooling of the entire planet, better sensitivity than this is
impossible to obtain with the aforementioned receive
system! Figure
3 shows the approximate relationship between noise figure and the
sensitivity of a typical narrowband FM communications receiver.
Minimizing receiver
system
noise figure:
Now that we know the magnitude of the noise emanating from the receive
system's environment, we should consider the other major source of
thermal noise: The receive equipment itself. While we can't
really control the noise from the environment, we can design a
receive system that does not add to it and unnecessarily
degrade the system.
When determining "receiver performance" it is worth little to consider
just one or two individual components, rather it is absolutely
necessary to take into account the entire signal path and hence, we
calculate it all as a receiver system. There are
several things that degrade the performance of any receiver
system from one that is "perfect" and these include:
Antenna noise temperature. For the most part this is
dictated by what the antenna is "looking" at (e.g. outer space or the
surface of the Earth.)
Antenna losses. Every antenna has losses of its own
-
typically in the matching network and the coaxial connector.
Feedline losses. There's likely some sort of
feedline
connecting the antenna to the preamplifier stage and its losses
contribute directly to the system's performance.
Preamplifier noise figure. Any signal amplifier will
have
its own intrinsic noise that it contributes to the system. Modern
GaAsFET amplifiers typically have a noise figure of less than 1 dB.
Preamplifier gain. In order for a preamplifier to be
useful, it must boost the signal. The more it boosts the signal,
the more its output signal can overcome the losses and noise figure of
those elements of the receiving system that follow it.
Post-preamp losses. Usually, one will have a length
of coax
and some connectors after the preamplifier in order to connect
the preamplifier to the receiver.
Receiver noise figure. The receiver itself will have
a
noise figure that will degrade the system performance.
Typically, a high-performance receive system will have a
low-noise preamplifier in front of the receiver and for
the most part, the performance of the receive system is set by the
performance of this preamplifier and the losses that are between the
preamplifier and the antenna.
One common misconception is that a low-noise preamplifier will cure all
woes if it is installed in the system. What is often not realized
is that with modern preamplifiers (those that have noise figures of
under 1 DB) it is not necessarily the
performance of the preamplifier that limits receiver system
performance, but the performance of the system is limited by those
losses in front of the preamplifier.
Take, for
example, a
hypothetical preamp that has a 0 dB noise figure. If that
preamplifier is connected to the antenna (assuming a "perfect" antenna)
via a piece of coax with 0.5 dB of loss through a bandpass cavity that
has 1.0 dB of loss (for a total loss of 1.5 dB) then we have just
created a system that has an equivalent noise figure of 1.5 dB:
That is, every dB of loss preceding the preamp contributes to a
noise figure increase of a dB!
What's worse is that once
this "1.5 dB" is lost forever, the increase in noise
cannot be recovered - no
matter how good the preamp
may be or how high gain it is! Why is this? All of these
losses are subject to the laws of thermodynamics and are resistive
and/or dielectric and therefore "generating" noise energy.
The GaAsFET
preamplifiers:
The preamplifiers used on the Panorama Point receive system are
homebrew, built in 1997, using the Mitsubishi MGF1302
GaAsFET - a now-obsolete device capable of a typical noise figure of
0.4-0.7 dB on 2 meters. These preamplifiers are designed to have
about 18 dB of gain (an
"amplification factor" of 63) so we can tolerate the inevitable
intrinsic
noise of our receiver as well as additional losses after
the preamplifier and it is assumed for the purposes of these
discussions that this preamp has a conservative noise figure of 0.8
dB. For
the Panorama Point receive system, the
losses in the coax and connectors following the preamp is about 1 dB,
yielding an equivalent total
system gain of 17 dB (an amplification factor of 50) prior to the
receivers.
These preamplifiers are built into small die-cast aluminum cases with
BNC connectors and a feedthrough capacitor for power for a
cost of
$20-$25 each in materials. The original circuit found in a
late-80's ARRL Radio Amateur Handbook was modified to reduce
gain (from the original design's 24 dB) and to improve stability by
using a simple 100 ohm
resistor (R2) to decouple the drain circuit from the power supply
rather
than the traditional tuned circuit or 4:1 balun. This gain
reduction was done primarily to prevent overload of the receiver as
the gain
exhibited by this amplifier (18 dB) is more than
adequate to override
the intrinsic noise of the receivers enough to assure that the received
antenna noise and preamp performance was the limiting factor in system
performance. In later testing, it was observed that it takes
about 6 dB
of additional attenuation after the
preamp
before the system sensitivity begins to measurably
degrade.
Performance of the Panorama
Point receive system:
For the reasons mentioned above, the preamplifiers on the Panorama
Point
receive system
are mounted directly to the antennas. It is expected that the
losses incurred by this configuration are below 0.25 dB (a typical
antenna value) taking into
account losses in the matching network of the antenna itself plus those
of the adapters used.
Comment: We know that the losses of these
antenna/matching networks are quite low owing to the fact that one of
these same type of antennas is used for transmitting and when we
dump 300 watts into them, they don't burst into flame or otherwise get
hot!
The receiver itself has its own noise figure, but the sensitivity
of FM communications receivers is typically measured in terms of 12 dB
SINAD for a given number of microvolts. While it is not
particularly easy to quantify the performance of an FM system in terms
of noise figure, typical values can be determined and it turns out that
a modern receiver with a 15 kHz IF bandwidth and a sensitivity of
0.15 uV
for 12 dB SINAD (3 kHz deviation of a 1 kHz tone) has an equivalent
noise figure of approximately 5.7
dB - and
this noise figure adds slightly to the overall noise figure of
the system. See figure 3 for a chart that approximately
equates receiver sensitivity with noise figure.
So, taking into account these numbers:
Antenna losses: 0.25 dB (Note that this does NOT take
into account thermal noise from the environment!)
Preamp noise figure: 0.8 dB (59K noise temperature)
Preamp gain: 18 dB
Post-preamp cable losses: 1 dB
Receiver noise figure: 5.7 dB (788K equivalent noise
temperature)
Now, to crunch numbers:
0.25 dB (ant losses) + 0.8 dB (preamp noise figure) = 1.05 dB
(equivalent "Preamp + antenna" noise figure) This 1.05 dB noise
figure equates to a "noise temperature" of 79K.
Overall preamp gain + cable losses = 17 dB, yielding an
"amplification factor" of 50.
Effect of preamplifier gain on the "dilution" of the receiver's
noise: 788K (receiver noise) / 50 (amplification
factor) = 15.76K contribution of the receiver itself (about 0.23 dB in
terms of "noise figure")
Total receiver system noise: 79K + 15.76K = 94.76K - let's
round that up to 95K which is equivalent to a noise figure of about
1.23
dB.
For details of how noise temperatures, amplifier gains, and system
losses add up, or for a link to an online calculator that converts
to/from noise figure and noise temperature, see the links at the bottom of the page.
Fig. 3 - This table shows
approximate equivalents
of receiver sensitivity (12 dB SINAD, 1 kHz modulation, 3 kHz
deviation, 15 kHz bandwidth) with receive system noise figure.
(This
information derived from data published by E.F. Johnson Corp. as well
as my own empirical testing of radio gear.)
12 dB
SINAD
Sensitivity
Approx.
Noise Figure (dB)
12 dB
SINAD Sensitivity
Approx.
Noise
Figure (dB)
0.1 uV
3.4
0.3 uV
10.7
0.125 uV
4.6
0.4 uV
13.1
0.15 uV
5.7
0.5 uV
14.9
0.2 uV
7.6
0.75 uV
18.4
0.25 uV
9.2
1.0 uV
20.8
As can be seen, the fact that the preamp (plus coax losses) has
significant amplification (50x amplification of the signal) means that
the intrinsic noise of the receiver itself contributes only about 10%
to the overall system noise and therefore it is the preamplifier's
noise figure that plays the major part in setting overall system
performance.
It should also be noted that if our preamplifier had even more
gain (let's assume that it had an infinite amount
of gain) OR if our receiver had a zero
noise figure of its own instead of the 4dB mentioned, we would only
lower the "noise temperature" of our receive system by only 15.76K - or
about 0.23 dB!
If our system noise temperature is, in fact, 95K (a noise figure of
about 1.23 dB) that would equate to a
thermal noise voltage of about 0.031 uV. Because we already know
that the intrinsic sensitivity of any terrestrially-based system can
be, at best, 0.09 microvolts (because of the fact that the Earth is, in
our
example, at 300K) we now know that our receive system's sensitivity
exceeds this by
about 9.3 dB and this means that the 0.031 uV noise will have
relatively little effect on a 0.09 microvolt signal. Bench
testing with signal generators and attenuators (which are at about 300K
and are contributing their own thermal noise) shows that the actual
sensitivity of the
receive system is, in fact, 0.09 uV for a 12 dB SINAD.
Comparison of other
hypothetical receive systems:
Case #1 - Using a
typical receiver with coax and cavity losses but no low-noise
preamplifier:
Let us compare the above system with what one might see if one used the
same receiver, but added 2 dB of cavity losses and didn't use
the preamplifier:
Coax loss: 1 dB
Cavity losses: 2 dB
"Typical" receiver with 0.15 uV 12 dB SINAD sensitivity:
5.7 dB noise figure
To crunch numbers:
1 dB (coax loss) + 2 dB (cavity loss) + 5.7 dB (receiver noise
figure) = 8.7 dB receive system noise figure (1860K equivalent noise
temperature)
From the above noise figures alone, one would have a system noise
figure of 8.7 dB (about 1860K) - a noise voltage of equivalent of 0.14
uV - more than enough to completely bury our "Earth limit" weak signal
of 0.09 uV and would effectively limit the receive sensitivity to
signals above 0.23 uV - more than 8 dB worse than the Panorama
system.
Case #2 - Case
#1 with a low-noise preamp:
Supposing that one were to "improve" the system by placing a low-noise
preamp right at the receiver - but after the coax and cavity
losses:
Coax loss: 1 dB
Cavity losses: 2 dB
Preamplifier noise figure: 0.8 dB (59K equivalent noise
temperature)
Preamplifier gain: 18 dB (63x gain)
Post-preamp losses: Assume 0 dB and that the preamp is
connected directly to the receiver.
"Typical" receiver with 0.15 uV 12 dB SINAD sensitivity:
5.7 dB noise figure (788K equivalent noise temperature)
To crunch numbers:
1 dB (coax loss) + 2 dB (cavity loss) + 0.8 dB (preamp noise
figure) = 3.8 dB noise figure of the "preamp system"
Effect of preamplifier gain on the improvement of the receiver's
noise temperature: 788K (receiver noise) / 63 (amplification
factor) = 12.5K - or about 0.18 dB noise figure
Total receiver system noise: 3.8 dB (preamp) + 0.18 dB
(added by receiver) = approx. 4.0 dB noise figure (438K equivalent
noise temperature)
From these numbers one would expect an equivalent thermal noise voltage
of about 0.07 microvolts - about 2 dB above the 0.056 uV intrinsic
noise voltage that one would experience on a terrestrially-based
system, implying a maximum receive sensitivity
of approximately 0.12 microvolts at 12 dB SINAD. This figure is
about 2.5 dB worse than the Panorama system - but is 4.7 dB better than
case #1 - a significant improvement!
It should be pointed out that with FM (unlike amplitude-modulated modes
such as AM or SSB) is affected by noise in a nonlinear way once
the signal drops below what is called the "limiting" threshold:
On AM or SSB, the addition of 3 dB of noise into the receive system
will, in fact, degrade the signal to noise of the received audio by 3
dB no matter how strong or weak the signal is. FM, on the other
hand, is not affected in this way once signals fall below this
"limiting" threshold.
When signals fall below the "limiting" threshold of the FM receiver
being used, the noise abruptly gets worse: Depending on the
receiver's design characteristics and the exactly where on the "curve"
we happen to be, a 1 dB decrease of the signal to noise of signal
that is already at or just below the limiting
threshold can cause a 3-10 dB increase in the amount of noise in the
recovered audio and the received signal rapidly falls into the
noise. In light of this effect, "2.5 dB" deficit in Case #2
above can amount to a significant degradation in the quality of
a received signal: When this 2.5 dB difference straddles the
limiting threshold of an FM receiver, it can easily turn a somewhat
noisy but perfectly copyable signal into one that is completely lost in
the noise.
Comments on preamps,
mismatch and coax losses:
One factor that is not often realized is that high-performance
preamplifiers are tuned at the factory for best NOISE FIGURE
- something that does not usually coincide precisely with
maximum gain and is seldom associated with anything resembling a good
50 ohm
match. What this means is that many preamps will terminate a
piece of coaxial cable with a rather poor SWR - something that will
exacerbate losses of the interconnecting coaxial cables, degrading
system more than just the coaxial cable loss alone.
For a system in which the preamplifier is wholly integrated with cable
and cavity assemblies, the preamplifier may be tuned in situ to
mitigate some of these problems - but it is also true in many of
these cases that with the additional losses of components in "front" of
the
preamplifier, the ultimate system sensitivity is more limited by the
losses of these additional components than the performance of the
preamplifier itself. Knowing this, one can further understand why
it is that preamplifiers
mounted at the antenna are the preferred configurations for
those systems where preamplifier noise is the limiting factor - most
notably in Satellite and EME ("Moonbounce") operations.
Fig. 4 - This shows the voting
controller (in the
blue box) with the two modified Tactec receivers and the UHF link
transmitter (a Yaesu
FT-470 HT). Click on the picture for a larger version.
How to beat Boltzmann:
With the existing system, we now know that we are nearly
(within less than 0.25 dB) of the theoretical maximum
sensitivity possible for an Earth-based receive system:
Any further improvements in the performance of the
receiver itself and the preamplifier would NOT offer any
more than a very minor (and operationally insignificant) improvement
that could be measured only with sensitive instruments. How is it
possible to obtain further improvements in system performance?
One obvious answer is with the use of directional antenna rather than
omnidirectional antennas.
Using directional
antennas:
Because our major limiting factor is thermal noise from the Earth
itself it would, therefore, make sense to limit the degree that the
Earth's thermal noise "dilutes" the signal that we are trying to
receive. Consider, for example, a system with an omnidirectional
vertical antenna that has gain
over 360 degrees of azimuth and compare that with, say, a 12 dBi Yagi
that typically has a 3 dB beamwidth of about 60 degrees. Assuming
that the
vertical beamwidths of these two antennas are comparable, that implies
that the Yagi is receiving a noise
contribution of less than 1/6th (360/60) of the Earth as compared to
the
vertical. This (theoretically) corresponds to at least a 7.8 dB
of
improvement. Note:
The improvement is actually greater than this owing to the fact that
the beamwidth of the Yagi is roughly circular whereas our 1/6th
"section" of the vertical's beam pattern is a rectangular piece of the
cylindrical shape.
At first guess, it might be assumed that the fact that our antenna has
a gain of 12 dB (an "amplification" factor of about 16) that the
increased signal voltage is the source of improved performance in our
system - but it is not: In the case above where our receive
system's
noise figure is already about 1.2 dB, the effect of the
"gain" of the antenna on increasing the signal's voltage doesn't help
us as much as our system's sensitivity already allows us to easily
"hear" the Earth's thermal noise - and the antenna's gain effectively
increases amplitude of that thermal noise. What does
affect performance is the fact that the "spotlight effect" of the
antenna's
beam pattern is rejecting the thermal noise from directions other
than that of the source of the signal that would otherwise "dilute" it.
If, however, our
sensitivity was limited by the thermal noise of our receive system
rather than that of the Earth itself (as
in "Case #1" and "Case #2" above)
then the "gain effect" would
improve the system
performance as it would help overcome the intrinsic receive system
noise. In "Case #2" where we already have a preamp inline, the
use of the Yagi would overcome much of the thermal noise generated by
our "3 dB loss in front of our preamp" losses caused by the cavity and
coax: In "Case #1" the improvement would be even more dramatic.
Voting receiver
system using Space Diversity:
Fig. 5 - These pictures showed the
pair of
"married" Mocom-70 amplifiers that were used in 1998. Using a
Wilkinson power divider on
the input to split drive power - and another on the output,
the 110 watt output of each amplifier was combined to provide over 200
watts of RF output with only a watt or so of RF drive. Click on either picture for a larger version.
There is yet another technique that can be exploited to further improve
performance of the antenna system: A voting receiver system using
space diversity
- that is, having independent receiver/antenna systems spaced a short
distance apart from each other. Under multipath
conditions and at any
given instant, one antenna is likely to have a better signal than the
"other" and furthermore, the likelihood of both
antennas
experiencing weak signal conditions at exactly the same time is
reduced. Although the amount of improvement is difficult to
quantify,
published literature indicates
that a typical 2-receiver voting system with fairly closely-spaced (1/2-1
wavelength
separation)
antennas improves the probability of signal intercept such that
the improvement of the receive system is between 6 and 10 dB when the
major cause of degradation has to do with multipath.
Using both directional (gain) antennas and voting receivers, it is
reasonable to expect an apparent receive system improvement of well
over 10
dB as compared to just a single omnidirectional antenna under the
majority of conditions encountered. It is likely that
further improvement could be obtained by the addition of more voting
receivers - possibly placed further away from the existing receiver
site - but that still doesn't address the problem that the user must
still be able to hear
the repeater that he/she is using, but in this case the user has the
advantage of being able to move location to find a usable signal.
The receivers:
The receivers used for the voting receiver system are a pair of
identical Tactec transceivers that have their transmitters
disabled. A number of modifications have also been done to these
receivers - mostly to minimize current consumption (such as
replacement of the
programming EEPROM with a CMOS version which alone saved about 100 mA)
as well as modifications to allow the disabling of the audio amplifier
and front-panel LED displays. With all of these modifications,
the receivers consume only about 180 mA each, compared to about 360 mA
in their original, unmodified state.
One important feature of the Tactec receiver that makes its rather
high power consumption worth the trouble is that it has an excellent
receiver on at least two fronts: The receiver's synthesizer is an
unusual design that has very low phase noise - something important in a
high-dynamic range receive system. Secondly - and more important
- is that the
receiver unconventionally (for a "mobile" radio) uses a passive
diode-ring doubly-balanced
mixer (DBM.) The use of the DBM is important in that it gives the
receiver a much higher dynamic range than more conventional receivers
with active mixers, allowing the receiver to function in the presence
of other strong signals. Both the low phase noise and the high
dynamic range allow the receiver to function well in the presence of
the very strong signal from the transmitter itself. ("Barefoot"
these
receivers
have
an average 12dB SINAD sensitivity of about 0.15 uV
- equivalent to a noise figure of approximately 5.7dB.)
The Voting controller:
The voting controller originally used a PIC16C84, but because this
processor is now obsolete - and I needed to add more functions - it now
uses the pin-compatible PIC16F819. This newer processor not only
cheaper, but it has a lot of on-chip peripherals (such as A/D
converters and comparators) that would have made the original
design
simpler. With the original source code written in C, it was a
fairly simple matter to "port" it over to the newer processor.
Wanting to retain compatibility with the original firmware and being
somewhat reluctant to simply toss already-proven hardware and software,
I chose to
retain the use of an LM339 outboard comparator rather than use the
chip's own comparators and A/D converters for signal-quality voting.
Another of the modifications to the Tactec receivers is to bring out
discriminator
audio and the "noise voltage", a signal generated inside the
radio used to determine now noisy the signal is and thus
determine whether or not what is received should be
squelched. The voting controller looks at this
squelch voltage to see which receiver (or both) is receiving a
signal. If both receivers "see" a signal, another comparator is
used to determine which one has the lowest "noise voltage" and selects
that receiver so that, at that instant, it is used to
provide audio from the receiver with the best signal.
When a signal is received from either receiver, the processor keys the
link transmitter (a
modified Yaesu FT-470 HT) and routes audio from
the receiver with the best signal to
the transmitter. This FT-470 has been modified such that an audio
path is
available on a jack that directly feeds the modulator. In this
controller, "discriminator" audio is used throughout and passed through
to the transmitter without any de-emphasis or pre-emphasis: The
only processing done is a 3 kHz lowpass filter to prevent
high-frequency noise inherent in weak FM signals (the same noise that
is used for squelch detection) from being transmitted and causing the
UHF link receivers to
experience "squelch clamping" when the 2-meter input signals are
weak. In doing this, the audio response from the 2 meter receiver
and through UHF transmitter is flat
to +/-1dB from 300 Hz to 2.5 kHz and flat to +/-3dB from 75 Hz to 2.75
kHz, thus avoiding any discernible "coloration" of the link
audio. The controller also monitors the UHF link activity by
"looking" at the
transmit/receive LED on the FT-470 using a CdS photocell to reduce the
chance of the controller IDing atop a transmission originating from
another repeater in the system.
Fig. 6 - Front and rear views of
the Vocom 300 watt
RF power amplifier. This amplifier requires about 50 watts of
drive to produce full output. At 13.8 volts - the designed
voltage - it will produce about 350 watts, but when powered from a
partially-discharged battery bank under full load the voltage drops to
the 12.2 volt neighborhood and the output power is in the 200-225 watt
range. Click on either picture for a larger version.
A more-recent addition (which required the use of the newer
processor) is a
power management function. When the receivers have been idle for
a certain period, they are power-cycled to reduce the average current
consumption and extend the charge life of the battery. Under
normal conditions, a "fast" power save function starts one minute after
the last transmission in which the two receivers are alternated between
each other at of about
3 times each second. The instant either receiver detects a
signal, both receivers are powered up and the "power save" cycle timer
is reset. After 5 minutes of "fast"
power save, a "slow" power save kicks in where each receiver is
alternately powered up for 1/5th of a second, each second, further
reducing power
consumption. As in the "fast" power save function, both receivers
are instantly awakened should a signal appear, restarting the "power
save" cycle.
Additional code was written to protect/prolong battery operation:
If the battery voltage exceeds 14.5 volts - possibly due to charging by
the solar panels - the power save functions are inhibited (to increase
current drain) until the
voltage drops below 13.9 volts in order to prevent battery damage due
to overcharging. If the battery voltage drops below 11.7 volts,
the "fast" power save mode begins immediately after a received
transmission ends and the "slow" power save mode begins 1 minute after
the last transmission: This more-aggressive power saving remains
in effect until the voltage rises above 12.5 volts. In testing,
the "fast" power save function causes a negligible "wake
up" delay while the "slow" power save function might, at worst, cause
the loss of the first word in a transmission. For continuous
monitoring of the "health" of the power supply, the controller
sends the battery voltage just after each ID.
In addition to canceling the power save features at a high battery
voltage, an N-channel MOSFET is used to disconnect the negative lead
from the solar array: If the voltage exceeds 14.5 volts, the
solar panels are disconnected until the voltage drops below 13.5
volts. In actual practice, the results in the battery voltage
averaging about 13.8 volts under full sun.
In testing, it
was noted that this receive system will function without degradation
down to 10.5 volts and continue to operate with sensitivity degradation
down to about 9 volts (at
which point the receivers start to fail and begin to show serious signs
of performance degradation)
allowing
for a significant margin in battery voltage. Originally, the
receivers and controller consumed nearly 700 mA when idle (about 350 mA
more when the transmitter was keyed) making power management of the
receive site somewhat of a challenge - particularly overnight following
cloudy days. Now, with the "fast" power save mode, the average
current is about 300 mA and it drops to less than 120 mA when in the
"slow" power save mode.
Telemetry:
The controller also provides a bit of telemetry as well. After each
transmission, there is a "courtesy beep" that indicates which
receiver(s) were used and to what extent with the "A" receiver, having
a low-pitched tone and the "B" receiver having a high pitched
one. The beep - which is always the same duration - will sound
high or low pitch in proportion to how much one receiver or the other
was used: In other words, if both "A" and "B" were used equally,
half the beep would be high-pitched and the other half would be
low-pitched while if only one receiver were used, the beep would be a
single pitch. When it is time to ID, the pitch of the ID will
reflect that of the last-used receiver.
Following the ID is a readout of the battery voltage in Morse using the
letter "R" as a decimal point as has been long tradition. This
"R" also has significance: If the "R" in the battery voltage is
of a different pitch than the rest of the ID/battery voltage,
that indicates that since the last ID, the battery bank has achieved
full charge and that the solar array has been disconnected at least
once.
These bits of telemetry have proven to be invaluable as the receiver is
located some distance away from camp and keeping tabs of the battery
voltage provides a good indication that the solar charging system is
working. In at least one instance, high (>50 MPH, or 80kPh)
winds have flipped over the solar panels and inhibited charging:
It was by monitoring the telemetry - and noting that the voltage had not
been increasing under full sun as expected - that we knew to make the
0.83 mile (1.3km) walk to the receive site and investigate!
Receiver placement:
You may have already noticed that this receiver system does not
have any bandpass or band-reject cavities in it: Having them
would most certainly degrade system performance. Typically, a
cavity filter is required on a repeater system owing to the fact that
the receiver will be overloaded by the transmitter's signals and/or
that the transmitter itself can radiate a low-level masking noise that
will effectively degrade the receiver's performance.
At Panorama Point geography is used to allow such a system to be
implemented without the use of any additional filtering as the
transmit and receive sites are about 2900 feet approx. 900 meters)
apart (yes, we have
to carry armloads of gear to/from the receive site - but with three
people, we can make it in just one trip...) from each
other with the receive site being "downrange" of the transmitter's
antenna pattern. This spacing alone provides significant
transmit/receive
isolation (about 75 dB) but there is also some terrain blocking the
direct view between the two sites as well as the fact that the receive
antennas are directional, having side and rear pattern rejection.
See Figure 8 at the bottom of the page for a map and picture showing
the relative locations of the sites.
Transmitters:
The first year that the Panorama Point receive system was deployed (in
1997) it used the receive system described above, but its transmitter
produced 50
watts into vertical antenna, yielding an EIRP of, perhaps, 150
watts. It soon became apparent that while we could
hear the fixed stations and th boats on the river quite well, those in
farther-flung locations (e.g. near Moab and at the northern end of the
Green River) were having
trouble hearing the Panorama Point repeater in some locations - that
is, our repeater was an "Elephant" (e.g. big ears, small mouth...)
The next year (1998) I decided to improve things a bit. Because
the river course is located in mostly one general direction from the
Panorama Point, there was
absolutely no need to use an omnidirectional antenna: One could
use a modest-gain Yagi to beam the signal in the direction of the
river course, "tweaking" the pointing of transmit antenna as necessary
to
accommodate the crowd of boats as it moved along, or if a particular
station had difficulty in hearing the signal.
In addition to the Yagi, more transmitter power was used. Having
a bunch of old Motorola Mocom 70 radios kicking around, I "married" a
pair of 110 watt power amplifier modules together for a combined
total power of over 200 watts of
RF. In conjunction with the 5-element Yagi, this produced well
over 2500 watts of effective radiated power - more than a 12 dB
improvement in signal from the previous year. Unfortunately,
after a day or two of
operation, one of the Mocom 70 amplifiers failed - but a simple
jumpering allowed us to continue with just one of the amplifiers and
over 1200 watts of EIRP: No-one really noticed...
Was the extra power worth it? Reports from those on the river
indicate that it was: At this power level it is possible to hear
the Panorama Point repeater nearly everywhere along the rivercourse -
although, as before, it is sometimes a bit spotty near the beginning
(just south of
Green River) and nearer the end (near Moab.)
Fig. 7 - Top: A view of the
transmit
site showing the old 2kW generator and the two antenna masts. Bottom:
A
view
of
the equipment used for the
transmitter. Click on either picture for a larger version.
A few months after the 1998 Friendship Cruise I was given a broken
Vocom 300 watt amplifier. As it turned out, only "half" of the
amplifier (which consists of a pair of 150 watt amplifiers operating in
tandem) was bad so, in 1999, I operated this "kludged" amplifier at 150
watts. By the time the 2000 Friendship Cruise
rolled around I had completely repaired the amplifier and operated it
at
its full 300 watt output power (when the battery bank was at full
charge) yielding a transmitted EIRP of over
4500 watts.
Even with this much transmitted power, users still
report that the repeater is still somewhat of an "Elephant" - but it is
unlikely that either logistical practicalities would allow us
to provide a dramatic increase in effective transmitter power:
Going from 300 watts to full-legal power of 1500 watts would be in
increase of 7 dB -
just over 1 S-unit! On one occasion, I used a 7-element Yagi
instead of the usual 5-element Yagi, but it soon became obvious that
while the 7-element had higher gain, it had a noticeably narrower beam
pattern and did a poorer job of covering the entire breadth of
the rivercourse at once.
The radio used for transmitting is a Kenwood TM-733A dual-band mobile
radio. This radio can be configured to provide cross-band
repeating and it is this mode that is used to receive the UHF signal
from the Panorama Point receive site - or from other repeaters in the
system.
Because repeater operations often involve long periods of transmitting,
the TM-733A itself cannot be used to directly drive the Vocom power
amplifier - which requires 50 watts of drive: While the '733 can
output 50 watts, it cannot do so for extended periods without
overheating. Instead, the '733 is set to a 5 watt output (a power
level that it can handle easily) and an outboard 50 watt
amplifier (a VHF Engineering unit that is, fortunately, an exact
copy of the 75 watt amplifier in Motorola App. note AN-791) is used to
boost the power. This outboard
amplifier, with its larger heatsink, is placed in the airflow of the
300 watt amplifier's fans and can sustain continuous-duty operation.
A look at the pictures reveals two antenna masts. The shorter
mast is one that is erected immediately upon arrival and is used to
hold not only the UHF link Yagi antenna, but a "backup" VHF transmit
antenna and until the main mast is set up, this "backup" antenna is
used for transmitting (at 50 watts) - or if some
maintenance/adjustments are required on the taller mast during normal
operations. The taller mast is a military-surplus sectional guyed
mast with an added pipe section on top for total height of about 27
feet and it allows adjustment of the beam heading from the base.
With 300 watts of RF being used, prudence dictates that such a strong
RF field be moved up in elevation atop the mast to minimize any RF
exposure hazards.
Linking the sites:
Because the transmit and receive sites are separated by some distance,
it is necessary to link them together - and the most convenient way to
do this is via a UHF link. For the Panorama Point receive site,
an old Yaesu FT-470 Handie-Talkie is used, operated by the voting
receivers' controller. Having a UHF link such as this
conveniently supplied the means by which other sites - such as the
Canyonlands Overlook site - could be linked together with this one.
When a signal is received on the Panorama Point receiver, it is
transmitted via UHF. A crossband repeater at the transmit site at
Panorama Point (over 1/2 mile away) receives this and simply
retransmits what
it hears. Simultaneously, a UHF receiver at one of the other
sites (such as Canyonlands Overlook) receives this signal and
retransmits it, on its output frequency and provides a simulcast of the
transmissions
received at Panorama. Conversely, any transmission received at
Canyonlands Overlook is also transmitted on UHF, received by the same
receiver at Panorama that receives the signals from the Panorama
receive site, simulcasting the signals received at Canyonlands Overlook.
Even though there is not line-of-sight between the two primary repeater
sites, Panorama Point and Canyonlands Overlook, the path is quite
reliable as it seems to "knife-edge" over the Island-in-the-Sky
district of Canyonlands. The UHF link transmitter (the FT-470) is
set at just 1/2 watt output and uses an 8-element Yagi pointed in the
direction of Canyonlands Overlook and provides a very strong
signal. It has been noticed that on those occasions where both
Canyonlands Overlook and the "local" Panorama Point UHF link
transmitters are active, it is not uncommon for the more-distant
Canyonlands Overlook signal to override the "local" Panorama signal.
It is worth mentioning that each of the linked repeaters operate on
their own 2-meter frequencies, requiring the users to select the one
that
covers best in their respective locations. Having all of the
repeater systems linked to each other allows the users to stay in
contact no matter which system they might be on.
Fig. 8 - Top: A topographical
map showing the relative locations of the transmit and receive sites. Bottom: A composite "Google Earth" (tm) image of the
terrain highlighted on the topographical map, showing the locations of
the radio sites. Note that the darker mass toward the top is
actually at a lower elevation than the narrow isthmus on which the
Panorama
Point receive site sits: It's the shadowing and coloration that
makes it
appear differently. Click on either picture for a larger version.
Power supplies:
As mentioned before, the Panorama Point receive site is self-sustaining
in its power requirements, using batteries and solar panels for
charging. The transmitter site, however, is a different
story: At the full
300 watt output, the Vocom amplifier, the 50 watt driver, and the
TM-733A pull nearly 70 amps of current. To supply this current,
several large lead-acid rechargeable batteries are used.
Because of the high current consumption and frequently heavy repeater
use, the batteries need to be charged - typically twice each day - to
maintain the system. To do this, a generator is used along with
several high-current power supplies along with several high-current
switching power supplies to fully-load the generator and effect as
rapid charging as possible.
In the past, 3-4 hours of run time per day have been needed to maintain
the batteries. While the generator is being used to charge the
transmitter's battery bank, other batteries - such as those used for
running an HF rig or tabletop VHF rig or other gadgets that we might
bring along - are also charged.
Earlier, I used an old 2 kW generator that had been refitted with a
very large noise-suppressing muffler, taking the sound from a deafening
din to more of a distant hum. Typically, the generator was placed
at the end of an extension cord, behind rocks from the campsite so its
noise was just "there" and not at all offensive. After the
battery bank it was brought to full charge, it was shut off and
re-fueled in preparation for the next time.
Several years ago, I bought a 2kW "inverter" type generator - a "Kipor"
model (essentially a Chinese "knockoff" of the similar Honda and
Kawasaki generators) that varies its engine speed depending on load and
is thus more efficient. With this generator there is more of a
tendency start it up and walk away, allowing it simply to run through
an entire tank of fuel at its own rate: In the first two or three
hours, it is fairly heavily loaded as the battery bank is quickly
charged, but once the charging current drops off, it would simply idle
along for another 4-6 hours - kicking its speed up only when the
transmitter gets keyed up and the load increased momentarily. As
it turns out, keeping the battery bank fully-charged with the newer
generator requires that we use 2.5-3 gallons (10-12 liters) of gasoline
over the weekend while the older "conventional" generator required
about 6 gallons (24 liters) for the same task and duration.
For more information, history, pictures, etc., go to the Friendship
Cruise Communications Site. This page is maintained
only in
the
month or so prior to the cruise so don't be surprised to see some
terribly
outdated information.
The "Official" Friendship cruise page may be found
at: http://www.ecso.com/friendship/index.html.
Please
note
that
this page may or may not be up-to-date: Use the
email address on that page for current information!
