Passive Microwave Repeaters
One of the most significant single advancements in telecommunications
technology was the development of microwave radio. Essentially an evolution of
radar, the middle of the Second World War saw the first practical microwave
telephone system. By the time Japan surrendered, AT&T had largely abandoned
their plan to build an extensive nationwide network of coaxial telephone
cables. Microwave relay offered greater capacity at a lower cost. When Japan
and the US signed their peace treaty in 1951, it was broadcast from coast to
coast over what AT&T called the “skyway”: the first transcontinental telephone
lead made up entirely of radio waves. The fact that live television coverage
could be sent over the microwave system demonstrated its core advantage. The
bandwidth of microwave links, their capacity, was truly enormous. Within the
decade, a single microwave antenna could handle over 1,000 simultaneous calls.
Microwave’s great capacity, its chief advantage, comes from the high
frequencies and large bandwidths involved. The design of microwave-frequency
radio electronics was an engineering challenge that was aggressively attacked
during the war because microwave frequency’s short wavelengths made them
especially suitable for radar. The cavity magnetron, one of the first practical
microwave transmitters, was an invention of such import that it was the UK’s
key contribution to a technical partnership that lead to the UK’s access to US
nuclear weapons research. Unlike the “peaceful atom,” though, the “peaceful
microwave” spread fast after the war. By the end of the 1950s, most
long-distance telephone calls were carried over microwave. While coaxial
long-distance carriers such as
L-carrier saw
continued use in especially congested areas, the supremacy of microwave for
telephone communications would not fall until adoption of fiber optics in
the 1980s.
The high frequency, and short wavelength, of microwave radio is a limitation as
well as an advantage. Historically, “microwave” was often used to refer to
radio bands above VHF, including UHF. As RF technology improved, microwave
shifted higher, and microwave telephone links operated mostly between 1 and 9
GHz. These frequencies are well beyond the limits of beyond-line-of-sight
propagation mechanisms, and penetrate and reflect only poorly. Microwave
signals could be received over 40 or 50 miles in ideal conditions, but the
two antennas needed to be within direct line of sight. Further complicating
planning, microwave signals are especially vulnerable to interference due to
obstacles within the “fresnel zone,” the region around the direct line of
sight through which most of the received RF energy passes.
Today, these problems have become relatively easy to overcome. Microwave
relays, stations that receive signals and rebroadcast them further along a
route, are located in positions of geographical advantage. We tend to think of
mountain peaks and rocky ridges, but 1950s microwave equipment was large and
required significant power and cooling, not to mention frequent attendance by a
technician for inspection and adjustment. This was a tube-based technology,
with analog and electromechanical control. Microwave stations ran over a
thousand square feet, often of thick hardened concrete in the post-war climate
and for more consistent temperature regulation, critical to keeping analog
equipment on calibration. Where commercial power wasn’t available they consumed
a constant supply of diesel fuel. It simply wasn’t practical to put microwave
stations in remote locations.
In the flatter regions of the country, locating microwave stations on hills
gave them appreciably better range with few downsides. This strategy often
stopped at the Rocky Mountains.
In much of the American West, telephone construction had always been
exceptionally difficult.
Open-wire telephone leads
had been installed through incredible terrain by the dedication and sacrifice
of crews of men and horses. Wire strung over telephone poles proved able to
handle steep inclines and rocky badlands, so long as the poles could be
set—although inclement weather on the route could make calls difficult to
understand. When the first transcontinental coaxial lead was installed, the
route was carefully planned to follow flat valley floors whenever possible.
This was an important requirement since it was installed mostly by mechanized
equipment, heavy machines, which were incapable of navigating the obstacles
that the old pole and wire crews had on foot.
The first installations of microwave adopted largely the same strategy. Despite
the commanding views offered by mountains on both sides of the Rio Grande
Valley, AT&T’s microwave stations are often found on low mesas or even at the
center of the valley floor. Later installations, and those in the especially
mountainous states where level ground was scarce, became more ambitious. At Mt.
Rose, in Nevada, an aerial tramway carried technicians up the slope to the roof
of the microwave station—the only access during winter when snowpack reached
high up the building’s walls. Expansion in the 1960s involved increasing use of
helicopters as the main access to stations, although roads still had to be
graded for construction and electrical service.
These special arrangements for mountain locations were expensive, within the
reach of the Long Lines department’s monopoly-backed budget but difficult for
anyone else, even Bell Operating Companies, to sustain. And the West—where
these difficult conditions were encountered the most—also contained some of
the least profitable telephone territory, areas where there was no
interconnected phone service at all until government subsidy under the Rural
Electrification Act. Independent telephone companies and telephone
cooperatives, many of them scrappy operations that had expanded out from the
manager’s personal home, could scarcely afford a mountaintop fortress and a
helilift operation to sustain it.
For the telephone industry’s many small players, and even the more rural Bell
Operating Companies, another property of microwave became critical: with a
little engineering, you can bounce it off of a mirror.
James Kreitzberg was, at least as the obituary reads, something of a
wunderkind. Raised in Missoula, Montana, he earned his pilots license at 15 and
joined the Army Air Corps as soon as he was allowed. The Second World War came
to a close shortly after, and so, he went on to the University of Washington
where he studied aeronautical engineering and then went back home to Montana,
taking up work as an engineer at one of the states’ largest electrical
utilities. His brother, George, had taken a similar path: a stint in the Marine
Corps and an aeronautical engineering degree from Oklahoma. While James worked
at Montana Power in Butte, George moved to Salem, Oregon, where he started an
aviation company that supplemented their cropdusting revenue by modifying
Army-surplus aircraft for other uses.
Montana Power operated hydroelectric dams, coal mines, and power plants, a
portfolio of facilities across a sparse and mountainous state that must have
made communications a difficult problem. During the 1950s, James was involved
in an effort to build a new private telephone system connecting the utility’s
facilities. It required negotiating some type of obstacle, perhaps a mountain
pass. James proposed an idea: a mirror.
Because the wavelength of microwaves are so short, say 30cm to 5cm (1GHz-6GHz),
it’s practical to build a flat metallic panel that spans multiple wavelengths.
Such a panel will function like a reflector or mirror, redirecting microwave
energy at an angle proportional to the angle on which it arrived. Much like you
can redirect a laser using reflectors, you can also redirect a microwave
signal. Some early commenters referred to this technique as a “radio mirror,”
but by the 1950s the use of “active” microwave repeaters with receivers and
transmitters had become well established, so by comparison reflectors came to
be known as “passive repeaters.”
James believed a passive repeater to be a practical solution, but Montana Power
lacked the expertise to build one. For a passive repeater to work efficiently,
its surface must be very flat and regular, even under varying temperature. Wind
loading had to be accounted for, and the face sufficiently rigid to not flex
under the wind. Of course, with his education in aeronautics, James knew that
similar problems were encountered in aircraft: the need for lightweight metal
structures with surfaces that kept an engineered shape. Wasn’t he fortunate, then,
that his brother owned a shop that repaired and modified aircraft.
I know very little about the original Montana Power installation, which is
unfortunate, as it may very well be the first passive microwave repeater ever
put into service. What I do know is that in the fall of 1955, James called his
brother George and asked if his company, Kreitzberg Aviation, could fabricate a
passive repeater for Montana Power. George, he later recounted, said that “I
can build anything you can draw.” The repeater was made in a hangar on the side
of Salem’s McNary Field, erected by the flightline as a test, and then shipped
in parts to Montana for reassembly in the field. It worked. It worked so well,
in fact, that as word of Montana Power’s new telephone system spread, other
utilities wrote to inquire about obtaining passive repeaters for their own
telephone systems.
In 1956, James Kreitzberg moved to Salem and the two brothers formed the
Microflect Company. From the sidelines of McNary Field, Microflect built
aluminum “billboards” that can still be found on mountain passes and forested
slopes throughout the western United States, and in many other parts of the
world where mountainous terrain, adverse weather, and limited utilities made
the construction of active repeaters impractical.
Passive repeaters can be used in two basic configurations, defined by the angle
at which the signal is reflected. In the first case, the reflection angle is
around 90 degrees (the closer to this ideal angle, of course, the more
efficiently the repeater performs). This situation is often encountered when
there is an obstacle that the microwave path needs to “maneuver” around. For
example, a ridge or even a large structure like a building in between two
sites. In the second case, the microwave signal must travel in something closer
to a straight line—over a mountain pass between two towns, for example. When
the reflection angle is greater than 135 degrees, the use of a single passive
repeater becomes inefficient or impossible, so Microflect recommends the use of
two. Arranged like a dogleg or periscope, the two repeaters reflect the signal
to the side and then onward in the intended direction.
Microflect published an excellent engineering
manual with many
examples of passive repeater installations along with the signal calculations.
You might think that passive repeaters would be so inefficient as to be
impractical, especially when more than one was required, but this is
surprisingly untrue. Flat aluminum panels are almost completely efficient
reflectors of microwave, and somewhat counterintuitively, passive repeaters can
even provide gain.
In an active repeater, it’s easy to see how gain is achieved: power is added.
A receiver picks up a signal, and then a powered transmitter retransmits it,
stronger than it was before. But passive repeaters require no power at all,
one of their key advantages. How do they pull off this feat? The design
manual explains with an ITU definition of gain that only an engineer could
love, but in an article for “Electronics World,” Microflect field engineer
Ray Thrower provided a more intuitive explanation.
A passive repeater, he writes, functions essentially identically to a parabolic
antenna, or a telescope:
Quite probably the difficulty many people have in understanding how the
passive repeater, a flat surface, can have gain relates back to the common
misconception about parabolic antennas. It is commonly believed that it is
the focusing characteristics of the parabolic antenna that gives it its gain.
Therefore, goes the faulty conclusion, how can the passive repeater have
gain? The truth is, it isn’t focusing that gives a parabola its gain; it is
its larger projected aperture. The focusing is a convenient means of
transition from a large aperture (the dish) to a small aperture (the feed
device). And since it is projected aperture that provides gain, rather than
focusing, the passive repeater with its larger aperture will provide high
gain that can be calculated and measured reliably. A check of the method of
determining antenna gain in any antenna engineering handbook will show that
focusing does not enter into the basic gain calculation.
We can also think of it this way: the beam of energy emitted by a microwave
antenna expands in an arc as it travels, dissipating the “density” of the
energy such that a dish antenna of the same size will receive a weaker and
weaker signal as it moves further away (this is the major component of path
loss, the “dilution” of the energy over space). A passive repeater employs a
reflecting surface which is quite large, larger than practical antennas, and
so it “collects” a large cross section of that energy for reemission.
Projected aperture is the effective “window” of energy seen by the antenna at
the active terminal as it views the passive repeater. The passive repeater
also sees the antenna as a “window” of energy. If the two are far enough away
from one another, they will appear to each other as essentially point
sources.
In practice, a passive repeater functions a bit like an active repeater that
collects a signal with a large antenna and then reemits it with a smaller
directional antenna. To be quite honest, I still find it a bit challenging to
intuit this effect, but the mathematics bear it out as well. Interestingly, the
effect only occurs when the passive repeater is far enough from either terminal
so as to be usefully approximated as a point source. Microflect refers to this
as the far field condition. When the passive repeater is very close to one of
the active sites, within the near field, it is more effective to consider the
passive reflector as part of the transmitting antenna itself, and disregard it
for path loss calculations. This dichotomy between far field and near field
behavior is actually quite common in antenna engineering (where an “antenna” is
often multiple radiating and nonradiating elements within the near field of
each other), but it’s yet another of the things that gives antenna design the
feeling of a dark art.
One of the most striking things about passive repeaters is their size. As a
passive repeater becomes larger, it reflects a larger cross section of the RF
energy and thus provides more gain. Much like with dish or horn antennas, the
size of a passive repeater can be traded off with transmitter power (and the
size of other antennas involved) to design an economical solution. Microflect
offered as standard sizes ranging from 8’x10′ (gain at around 6.175GHz: 90.95
dB) to 40’x60′ (120.48dB, after a “rough estimate” reduction of 1dB due to
interference effects possible from such a short wavelength reflecting off of
such a large panel as to invoke multipath effects).
By comparison, a typical active microwave repeater site might provide a gain of
around 140dB—and we must bear in mind that dB is a logarithmic unit, so the
difference between 121 and 140 is bigger than it sounds. Still, there’s a
reason that logarithms are used when discussing radio paths… in practice, it
is orders of magnitude that make the difference in reliable reception. The
reduction in gain from an active repeater to a passive repeater can be made up
for with higher-gain terminal antennas and more powerful transmitters. Given
that the terminal sites are often at far more convenient locations than the
passive repeater, that tradeoff can be well worth it.
Keep in mind that, as Microflect emphasizes, passive repeaters require no power
and very little (“virtually no”) maintenance. Microflect passive repeaters were
manufactured in sections that bolted together in the field, and the support
structures provided for fine adjustment of the panel alignment after mounting.
These features made it possible to install passive repeaters by helicopter onto
simple site-built foundations, and many are found on mountainsides that are
difficult to reach even on foot. Even in less difficult locations, these
advantages made passive repeaters less expensive to install and operate than
active repeaters. Even when the repeater side was readily accessible, passives
were often selected simply for cost savings.
Let’s consider some examples of passive repeater installations. Microflect was
born of the power industry, and electrical generators and utilities remained
one of their best customers. Even today, you can find passive repeaters at many
hydroelectric dams. There is a practical need to communicate by telephone
between a dispatch center (often at the utility’s city headquarters) and the
operators in the dam’s powerhouse, but the powerhouse is at the base of the
dam, often in a canyon where microwave signals are completely blocked. A
passive repeater set on the canyon rim, at an angle downwards, solves the
problem by redirecting the signal from horizontal to vertical. Such an
installation can be seen, for example, at the Hoover Dam. In some sense, these
passive repeaters “relocate” the radio equipment from the canyon rim (where the
desirable signal path is located) to a more convenient location with the other
powerhouse equipment. Because of the short distance from the powerhouse to the
repeater, these passives were usually small.
This idea can be extended to relocating en-route repeaters to a more
serviceable site. In Glacier National Park, Mountain States Telephone and
Telegraph installed a telephone system to serve various small towns and
National Park Service sites. Glacier is incredibly mountainous, with only
narrow valleys and passes. The only points with long sight ranges tend to be
very inaccessible. Mt. Furlong provided ideal line of sight to East Glacier and
Essex along highway 2, but it would have been extremely challenging to install
and maintain a microwave site on the steep peak. Instead, two passive repeaters
were installed near the mountaintop, redirecting the signals from those two
destinations to an active repeater installed downslope near the highway and
railroad.
This example raises another advantage of passive repeaters: their reduced
environmental impact, something that Microflect emphasized as the environmental
movement of the 1970s made agencies like the Forest Service (which controlled
many of the most appealing mountaintop radio sites) less willing to grant
permits that would lead to extensive environmental disruption. Construction by
helicopter and the lack of a need for power meant that passive repeaters could
be installed without extensive clearing of trees for roads and power line
rights of way. They eliminated the persistent problem of leakage from standby
generator fuel tanks. Despite their large size, passive repeaters could be
camouflaged. Many in national forests were painted green to make them less
conspicuous. And while they did have a large surface area, Microflect argued
that since they could be installed on slopes rather than requiring a large
leveled area, passive repeaters would often fall below the ridge or treeline
behind them. This made them less visually conspicuous than a traditional active
repeater site that would require a tower. Indeed, passive repeaters are only
rarely found on towers, with most elevated off the ground only far enough for
the bottom edge to be free of undergrowth and snow.
Other passive repeater installations were less a result of exceptionally
difficult terrain and more a simple cost optimization. In rural Nevada, Nevada
Bell and a dozen independents and coops faced the challenge of connecting small
towns with ridges between them. The need for an active repeater at the top of
each ridge, even for short routes, made these rural lines excessively
expensive. Instead, such towns were linked with dual passive repeaters on the
ridge in a “straight through” configuration, allowing microwave antennas at the
towns’ existing telephone exchange buildings to reach each other. This was the
case with the installation I photographed above Pioche. I have been
frustratingly unable to confirm the original use of these repeaters, but from
context they were likely installed by the Lincoln County Telephone System to
link their “hub” microwave site at Mt. Wilson (with direct sight to several
towns) to their site near Caliente.
The Microflect manual describes, as an example, a very similar installation
connecting Elko to Carlin. Two 20’x32′ passive repeaters on a ridge between the
two (unfortunately since demolished) provided a direct connection between the
two telephone exchanges.
As an example of a typical use, it might be interesting to look at the manual’s
calculations for this route. From Elko to the repeaters is 13.73 miles, the
repeaters are close enough to each other as to be in near field (and so
considered as a single antenna system), and from the repeaters to Carlin is
6.71 miles. The first repeater reflects the signal at a 68 degree angle, then
the second reflects it back at a 45 degree angle, for a net change in direction
of 23 degrees—a mostly straight route. The transmitter produces 33.0 dBm,
both antennas provide a 34.5 dB gain, and the passive repeater assembly
provides 88 dB gain (this calculated basically by consulting a table in the
manual). That means there is 190 dB of gain in the total system. The 6.71 and
13.73 mile paths add up to 244 dB of free space path loss, and Microflect
throws in a few more dB of loss to account for connectors and cables and the
less than ideal performance of the double passive repeater. The net result is a
received signal of -58 dBm, which is plenty acceptable for a 72-channel voice
carrier system. This is all done at a significantly lower price than the
construction of a full radio site on the ridge [1].
The combination of relocating radio equipment to a more convenient location and
simply saving money leads to one of the iconic applications of passive
repeaters, the “periscope” or “flyswatter” antenna. Microwave antennas of the
1960s were still quite large and heavy, and most were pressurized. You needed a
sturdy tower to support one, and then a way to get up the tower for regular
maintenance. This lead to most AT&T microwave sites using short, squat square
towers, often with surprisingly convenient staircases to access the antenna
decks. In areas where a very tall tower was needed, it might just not be
practical to build one strong enough. You could often dodge the problem by
putting the site up a hill, but that wasn’t always possible, and besides, good
hilltop sites that weren’t already taken became harder to find.
When Western Union built out their microwave network, they widely adopted the
flyswatter antenna as an optimization. Here’s how it works: the actual
microwave antenna is installed directly on the roof of the equipment building
facing up. Only short waveguides are needed, weight isn’t an issue, and
technicians can conveniently service the antenna without even fall protection.
Then, at the top of a tall guyed lattice tower similar to an AM mast, a passive
repeater is installed at a 45 degree angle to the ground, redirecting the
signal from the rooftop antenna to the horizontal. The passive repeater is much
lighter than the antenna, allowing for a thinner tower, and will rarely if ever
need service. Western Union often employed two side-by-side lattice towers with
a “crossbar” between them at the top for convenient mounting of reflectors each
direction, and similar towers were used in some other installations such as the
FAA’s radar data links. Some of these towers are still in use, although
generally with modern lightweight drum antennas replacing the reflectors.
Passive microwave repeaters experienced their peak popularity during the 1960s
and 1970s, as the technology became mature and communications infrastructure
proliferated. Microflect manufactured thousands of units from there new, larger
warehouse, across the street from their old hangar on McNary Field.
Microflect’s customer list grew to just about every entity in the Bell System,
from Long Lines to Western Electric to nearly all of the BOCs. The list
includes GTE, dozens of smaller independent telephone companies, most of the
nation’s major railroads, electrical utilities from the original Montana Power
to the Tennessee Valley Authority. Microflect repeaters were used by ITT Arctic
Services and RCA Alascom in the far north, and overseas by oil companies and
telecoms on islands and in mountainous northern Europe.
In Hawaii, a single passive repeater dodged a mountain to connect Lanai City
telephones to the Hawaii Telephone Company network at Tantalus on Oahu—nearly
70 miles in one jump. In Nevada, six passive repeaters joined two active sites
to connect six substations to the Sierra Pacific Power Company’s control center
in Reno. Jamaica’s first high-capacity telephone network involved 11 passive
repeaters, one as large as 40’x60′.
The Rocky Mountains are still dotted with passive repeaters, structures that
are sometimes hard to spot but seem to loom over the forest once noticed. In
Seligman, AZ, a sun-faded passive repeater looks over the cemetery. BC
Telephone installed passive repeaters to phase out active sites that were
inaccessible for maintenance during the winter. Passive repeaters were, it
turns out, quite common—and yet they are little known today.
First, it cannot be ignored that passive repeaters are most common in areas
where communications infrastructure was built post-1960 through difficult
terrain. In North America, this means mostly the West [2], far away from the
Eastern cities where we think of telephone history being concentrated. Second,
the days of passive repeaters were relatively short. After widespread adoption
in the ’60s, fiber optics began to cut into microwave networks during the ’80s
and rendered microwave long-distance links largely obsolete by the late ’90s.
Considerable improvements in cable-laying equipment, not to mention the lighter
and more durable cables, made fiber optics easier to install in difficult
terrain than coaxial had ever been.
Besides, during the 1990s, more widespread electrical infrastructure,
miniaturization of radio equipment, and practical photovoltaic solar systems
all combined to make active repeaters easier to install. Today, active repeater
systems installed by helicopter with independent power supplies are not that
unusual, supporting cellular service in the Mojave Desert, for example. Most
passive repeaters have been obsoleted by changes in communications networks and
technologies. Satellite communications offer an even more cost effective option
for the most difficult installations, and there really aren’t that many places
left that a small active microwave site can’t be installed.
Moreover, little has been done to preserve the history of passive repeaters.
In the wake of the 2015 Wired article on the Long Lines network, considerable
enthusiasm has been directed towards former AT&T microwave stations, having
been mostly preserved by their haphazard transfer to companies like American
Tower. Passive repeaters, lacking even the minimal commercial potential of old
AT&T sites, were mostly abandoned in place. Often being found in national
forests and other resource management areas, many have been demolished for
restoration. In 2019, a historic resources report was written on the Bonneville
Power Administration’s extensive microwave network. It was prepared to address
the responsibility that federal agencies have for historical preservation under
the National Historic Preservation Act and National Environmental Policy Act,
policies intended to ensure that at least the government takes measures to
preserve history before demolishing artifacts. The report reads: “Due to their
limited features, passive repeaters are not considered historic resources, and
are not evaluated as part of this study.”
In 1995, Valmont Industries acquired Microflect. Valmont is known mostly for
their agricultural products, including center-pivot irrigation systems, but
they had expanded their agricultural windmill business into a general
infrastructure division that manufactured radio masts and communication towers.
For a time, Valmont continued to manufacture passive repeaters as Valmont
Microflect, but business seems to have dried up.
Today, Valmont Structures manufactures modular telecom towers from their
facility across the street from McNary Field in Salem, Oregon. A Salem local,
descended from early Microflect employees, once shared a set of photos on
Facebook: a beat-up hangar with a sign reading “Aircraft Repair Center,” and in
front of it, stacks of aluminum panel sections. Microflect workers erecting a
passive repeater in front of a Douglas A-26. Rows of reflector sections beside
a Shell aviation fuel station. George Kreitzberg died in 2004, James in 2017.
As of 2025, Valmont no longer manufactures passive repeaters.
Postscript
If you are interested in the history of passive repeaters, there are a few
useful tips I can give you.
- Nearly all passive repeaters in North America were built by Microflect, so
they have a very consistent design. Locals sometimes confuse passive repeaters
with old billboards or even drive-in theater screens, the clearest way to
differentiate them is that passive repeaters have a face made up of aluminum
modules with deep sidewalls for rigidity and flatness. Take a look at the
Microflect
manual for many
photos. - Because passive repeaters are passive, they do not require a radio license
proper. However, for site-based microwave licenses, the FCC does require that
passive repeaters be included in paths (i.e. a license will be for an active
site but with a passive repeater as the location at the other end of the path).
These sites are almost always listed with a name ending in “PR”. - I don’t have any straight answer on whether or not any passive repeaters are
still in use. It has likely become very rare but there are probably still
examples. Two sources suggest that Rachel, NV still relies on a passive
repeater for telephone and DSL. I have not been able to confirm that, and the
tendency of these systems to be abandoned in place means that people sometimes
think they are in use long after they were retired. I can find documentation of
a new utility SCADA system being installed, making use of existing passive
repeaters, as recently as 2017.
[1] If you find these dB gain/loss calculations confusing, you are not alone.
It is deceptively simple in a way that was hard for me to learn, and perhaps I
will devote an article to it one day.
[2] Although not exclusively, with installations in places like Vermont and
Newfoundland where similar constraints applied.
