2026-07-04

In 1914 the Department of the Interior, through the Bureau of Reclamation, investigated the possibilities of developing the Columbia River. Thousands of arid but potentially fertile acres needed only water to become the Imperial Valley of the Northwest. Locked in the mountain ranges were valuable ores awaiting electricity to turn them into needed metals.

Two years later the State engineer of Oregon urged the development of the Bonneville site as a national-defense measure: he saw in the proposed power project a source of fertilizer in time of peace and nitrates in time of war. The dam also would completely drown out the Cascade Rapids and extend slack-water navigation some 40 miles eastward to The Dalles.

The Rivers and Harbors Act of 1925 directed the Secretary of War, through the Corps of Engineers, United States Army, to prepare and submit to the Congress an estimate of the cost of surveys, examinations, and investigations of all navigable streams and their tributaries where power development appeared feasible. (Q1)

It is difficult to succinctly explain why, exactly, the United States Army has spent much of its history involved in the construction of dams. It is partly an accident of history, partly the result of interagency federal politics, and entirely a product of American culture. In his book "Cadillac Desert," Marc Reisner examines the history of the American West's water control projects as a religious project, one animated less by practical needs than by a sense that domination of the West's rivers was destiny.

The Bureau of Reclamation, part of the Department of the Interior, was formed for that purpose. At the time, though, the Army had already been used to survey and improve rivers for nearly 100 years. They were not content to give it up. The result was a rivalry, one with several feints and blows before the two settled into their modern areas of control. For the Bureau of Reclamation, the Hoover Dam was their signature project. For the Corps of Engineers, the battle that would go down in history was the Columbia River Project.

The motivations for damming the Columbia were various. The Columbia was prone to flooding, which had caused damage and limited use of land along it. There was a great deal of land surrounding the Columbia that could be farmed, if the Columbia could be tapped for irrigation. Electricity, too, was a reason, although initially a somewhat secondary one. Perhaps the greatest reason, though, was simply economic: by the time that the major parts of the Columbia River Project were truly underway, the nation was in the throes of the Great Depression.

President Franklin D. Roosevelt was already a fan of hydroelectricity. As governor of New York, he was exposed to the pioneering Niagara Falls power plant and pushed for other similar projects in that state. As President, his "New Deal" naturally incorporated hydropower as well. By 1934, he had formed a Regional Planning Commission that sketched out a series of dams along the Columbia, two of which would become the Grand Coulee and the Bonneville. These dams would produce a tremendous amount of electricity, and unlike in other similar Corps of Engineers projects to date, that power would not all be consumed by irrigation pumping. There was power to spare. To distribute that power, the Regional Planning Commission suggested an independent government agency on the model of the Panama Canal or the recently chartered Tennessee Valley Authority.

As an interim measure, the loosely defined Bonneville Project coordinated the civilian side of the Corps of Engineers project until 1938, when the Bonneville Dam was complete and the Grand Coulee was much of the way there. The Bonneville Dam captures little water in its reservoir, so while it does have flood control value, electrical production is its primary purpose. The dam's two powerhouses produce up to 1.2 GW, an impressive number for the 1930s but one that pales in comparison to the Grand Coulee's eventual (1970s) full capacity of nearly 7 GW. The Columbia River dams increased the electrical capacity of the Pacific Northwest by orders of magnitude; the numbers were significant even at a nationwide scale.

The bumper crop of electricity triggered a predictable controversy: what to do with government power? One camp favored public control of the resource, with the government marketing the power on some sort of equitable basis. The other favored private control, arguing that the output of the dams should be contracted entirely to private utilities like Portland General Electric (itself the scion of an important early hydroelectric project at Willamette Falls). In the New Deal political climate, the first camp won: the Columbia did not quite get a TVA, but Congress did charter the Bonneville Power Administration (BPA), the first of what would come to be known as Power Marketing Agencies. Over the following decades, the BPA became part of the Department of Energy (DoE)—uncharacteristically, for the DoE, a part of it that actually generated and sold electricity. Well, technically, the Corps of Engineers generates it, and the BPA markets and distributes it.

In any case, starting in the late 1930s, the BPA was tasked with the construction of a network that could distribute power from Columbia River dams throughout the region—on an equitable, equal-rate basis often called the "postage stamp rate" that allowed rural coops to buy government-generated power at the same rate as the big city private utilities. The sudden bevy of power along the Columbia and the fair rates at which it could be obtained in great quantity led to an industrial revolution for the region, one that saw it as the seat of the American aluminum industry (with the Columbia Gorge producing something like 1/3rd of the nation's aluminum through to the 1970s) and that boosted the fate of hundreds of related industries (aerospace and, specifically, Boeing not least among them). BPA power has enduring influence today, with many towns on the Gorge (The Dalles, Boardman, Umatilla) disproportionately prominent on a map of the nation's data centers. AWS's us-west-2, for example, is a beneficiary of Columbia River dams and located near many of them—not just Bonneville, but the Dalles (1.8 GW), John Day (2.2 GW), McNary (1.1 GW), and more.

In marketing this power, the BPA faced a challenge: the dams are spread across a large area, as are the customers. Industrial customers, such as the Alcoa (Aluminum Company of America) smelter that opened in 1940 at Vancouver, Washington 1 were opening in rural areas where land was readily available, and an explicit goal of the Columbia River Project had been the extension of electricity to farmers and other rural industries. The concept of long-distance power transmission had been pioneered by an 1889 transmission line, the nation's first, between Willamette Falls at Oregon City and downtown Portland. Beginning in 1938, the BPA was tasked with expanding that concept across a region that would eventually span eight states.

The Master Grid

BPA's first administrator, J. D. Ross, presented a plan he called the BPA Master Grid. This ring-shaped network, made up of 230 kV long-distance transmission lines, would connect the dams not only to Portland and Seattle but to Pasco, Yakima, Spokane, Ellensburg, the Willamette Valley through to California, and the Oregon Coast. By 1945, the Master Grid covered three thousand "Circuit Miles" of transmission lines. It was the first integrated regional power grid in the United States, and would come to pioneer the market-based electricity pricing and distribution, independent system operators (ISOs), and pooling and wheeling agreements that form the modern US electrical infrastructure. The entire Western Interconnection, the unified power system that serves the US and Canada from the Rocky Mountains west, can be said to have crystallized outwards from the seed of the Bonneville Dam's switch yard.

Getting there required that the BPA solve formidable technical problems and develop many new technologies in power distribution. BPA transmission lines operated at higher voltages than any before them and, in the 1960s, introduced high voltage DC transmission to the Americas, connecting the Columbia system to the major demand centers of Southern California at 800 kV DC. BPA was only slightly behind the TVA on the installation of a remarkable analog computer called a Network Analyzer, in 1939, which simulated the behavior of the transmission network like a scale model. The rural nature of the BPA network put substations in remote areas, where they were minimally staffed, and the long stretches of high-voltage transmission line meant there was ample potential for damage by wind, weather, and trees, phenomena that the BPA came to better understand through research laboratories and experimental field sites.

This is not an article about the history of electrical distribution, or at least it wasn't supposed to be, so here we must exercise some discipline and narrow in on a topic. Telecommunications ought to do.

By 1940, as the Master Grid entered operation, its numerous substations already caused administrators a headache. Each had a small staff of technicians, but communicating with them was difficult. Coordinating changes across large areas, or quickly responding to faults, involved a flurry of telephone and radio calls. When Portland General Electric built the transmission line from Willamette Falls to Portland, they encountered the same problem, and by the 1910s had implemented a very early form of its solution: telemetry and teleoperation. Through a set of control wires strung along the transmission line, operators in Portland could see certain measurements from the Oregon City powerhouse and remotely throw switches to bring turbines on and offline in response to load. As the BPA built the Master Grid, they invested in the same technology.

Around 1939, the BPA commissioned a study of communications technology that could be used along the Master Grid. There were three main contenders: commercial telephone networks (which BPA called "land telephone" to differentiate it from the other two), "carrier current telephone" technology that superimposed telephone signals onto the electrical conductors of the transmission lines themselves, and radiotelephone equipment. A working agreement was reached with Pacific Telephone & Telegraph, the Bell System company that would later become US West, to share network information and analyze the cost tradeoffs between purchasing carrier current and radio equipment and leasing telephone lines. Ultimately, the diversity of the BPA network required some of all three.

Each of the BPA's substations had a building, called the control house, that contained control and monitoring equipment along with office facilities for the substation's operators. A room of each control house was dedicated to carrier equipment, devices that modulated multiple telephone circuits using frequency division multiplexing, and to a set of carrier frequencies that could be coupled onto the transmission lines to be received at the next substation. This equipment is similar to carrier equipment used in the telephone network, although specialized to power distribution applications by the choices of carrier frequency. I cannot say for certain, but it is very likely that BPA purchased their system from Lenkurt, a San Francisco-based communications equipment manufacturer that specialized in carrier current systems at the time 2.

The BPA's carrier system incorporated selective calling, meaning that users interacted with telephones that looked and felt much like conventional telephones, including a dial. An operator at one substation could dial the number for another and that phone would ring. The main difference from the telephones we use today is that these carrier current systems were interphones, more similar to intercoms or party lines than single-user telephone service. If you picked up a phone on a circuit, you would hear any conversations already underway. Of course, in industrial control applications, this built-in conferencing capability was generally considered a feature, and telephone circuits were assigned to shared use by departments or operating regions.

Radio was installed as well, primarily so that construction and maintenance crews in the field could get messages back to the administrative offices. Some substations, in strategic locations for coverage, had a radio site about a half mile from the substation for isolation from the powerful electromagnetic interference created by the high-voltage transformers. These radio sites were wired to remote heads located in the substation control house office, where substation operators relayed messages between mobile radios in the field and the carrier current telephone system. Bear in mind that these were still early days for mobile radios, and the HF units used by the BPA were proudly described as using only 1.5 cubic feet of space in the trunk of the vehicle, plus the microphone, speaker, and control head in the cab. Finally, while not the main purpose, it was already noted in the 1940 Annual Report that the substations could use the radio stations to substitute for the carrier current system in an emergency.

Complementing all of the above, the BPA leased telephone lines between major substations, agency headquarters in Portland, and the switch yard in Vancouver that was becoming the closest thing to a "main substation" in the network's distributed, ring-shaped design.

Immediately after the BPA's first Annual Report discusses the selection of communications equipment, it moves on to Protection. Here I must introduce a topic in electrical engineering that I have only a loose understanding of, even having spent the last few weeks in part on YouTube engineering tutorials. It's important that we get comfortable with the field of "protective relaying" because, as we will see, it became the most widespread application of private telecommunications networks after the railroads.

Protective Relaying

In your home, you are protected against certain dangerous scenarios by the over-current protection device that we Americans call a circuit breaker. The electromechanical contraptions in your service panel use a combination of methods to monitor the current that passes through them, and if they detect excessive current they open the circuit. The electrical transmission system has similar protections on a larger scale: on the distribution wires strung on poles outside of your house, for example, there are various fuses and circuit-breaker-like devices known as reclosers. High-tension transmission lines 3, like the 230 kV system built by the BPA, need similar protection for similar reasons—except that it is much more complex.

Electrical terminology can be complicated on a good day, and this situation is even more complex because of the historic bifurcation between terms and practices in building electrical wiring versus electrical distribution (which are governed, for example, by two separate electrical codes) and the fact that we are talking about a system that is nearly 100 years old. Relays were a newer technology in the 1930s, as was large-scale over-current protection, so transmission engineers viewed circuit breakers as just an application of the relay and over-current protection on power distribution is still referred to as "relaying" today. Since the whole broader field of supervising transmission lines for safety and reliability is called "protection," relays that open to protect generators, lines, or loads from dangerous conditions are called "protective relays."

Some of the protective relays used in transmission are very much like the circuit breakers in your home. Directional over-current relays, for example, monitor the current passing through them in one direction ("towards" the load) and open if it is excessive. Ground fault relays open when they detect, via various current transformers, that an excessive amount of power is leaking to the ground—just like the GFCI outlets or breakers installed in wet areas of homes.

Some of them, though, are much trickier. The first major problem is directionality. In your home electrical wiring, there is a clear sense of where power flows "from" (the service panel) and "to" (an outlet or fixture). Wiring thus only needs over-current protection in one direction. In a wide-area transmission network, this isn't true. The Master Grid was designed to incorporate a ring for much the same reason that SONET and other communications technologies favored rings: with a ring topology, you can lose the connection between two points and still be able to serve all points by sending power the "other direction." In general, it is common that electrical transmission lines can be "fed" from both ends, and have "load" on both ends. This flexibility to reach the same places by different routes makes the grid more reliable and responsive to changes. It also makes over-current protection more challenging.

Say that you have a span of transmission line, and that somewhere along it a tree falls and pushes one conductor against another. You now have a short-circuit fault at 230 kV (or more in later lines), a dramatic and dangerous condition. You also have thousands, if not hundreds of thousands, of customers that are depending on the power provided by that line. Current transformers can measure the enormous fault current, and indeed it will be detected at many points along the line. But what do you do?

Ideally, protective relays should open on both sides of the fault, and as close to the fault as possible. This cuts power to the dangerous situation while minimizing the number of customers who experience an outage. It's also difficult to achieve in practice: you cannot simply open a protective relay on over-current, or every relay on the line will open at the same time. Instead, analog circuits were used to measure the impedance to the fault as a proxy for its distance. This way protective relays could be carefully tuned to open only when a fault was near them.

Now, consider that a transmission line may be "tapped" and connected to load centers or generation at multiple points along its length. There may also be multiple parallel routes that current can take, with varying capacities. Both of these situations mean that it is often necessary to measure the current (to detect faults) at locations remote from the actual protective relays, which are large devices that needed to be located at substations. Further, the complications of parallel routes and possible ground faults on long lines required the use of a "differential protective relay" or "balanced current relay" that took a much higher level approach to the problem. In a differential system, you measure the current at every connection point to a given protection zone and sum them. The sum should be zero: the same amount of current goes into the line as comes out. If it's not, something has gone wrong somewhere, likely current that is escaping to ground or traveling on a parallel route not engineered for such abuse. But the points where you measure current may be many miles apart, and you still need to make real-time comparisons between them.

In 1940, the BPA was already planning the installation of "pilot relaying" over their carrier current telephone system. In an engineering context, a "pilot" is usually something small that controls something big. A pilot operated relief valve (PORV), for example, is an arrangement where dangerous pressure levels (of a liquid or gas) cause a small pilot valve to open which then triggers a pressure differential in a much bigger valve that causes it to open. There is a similar idea in protective relaying: a pilot-operated relay is a relay that disconnects a very big wire under the control of a smaller wire carrying a pilot signal. The simplest scheme works like this: a device, like a current transformer, monitors a safety parameter and produces a tone whenever it is acceptable. Elsewhere, a protective relay monitors that tone. If the tone ever goes away, the relay opens. The tone might go away because the pilot device detected an unsafe condition, but it might also go away because the line carrying the pilot signal was damaged, making it a fail-safe design. This is good for safety, but bad for reliability, and requires that the communications infrastructure used for protective relaying be very reliable.

Unfortunately, it was not: the carrier current systems installed by BPA through the 1940s were typical of the technology used in the industry at the time, but it was ill-equipped for the scale of the Master Grid. In part to alleviate concern of federal competition wiping out private utilities, and to improve general efficiency, the BPA formed the Northwest Power Pool in 1941. Initially made of about a dozen electrical utilities in the Pacific Northwest, the Power Pool formalized a set of arrangements by which utilities would buy and sell power among themselves, carried over the BPA's transmission network for a small "wheeling" fee. The Northwestern Power Pool would eventually become the Western Power Pool and a template for much of the nation's electrical industry. It significantly increased reliability and efficiency in the region by allowing utilities to sell their overproduction to utilities with high demand, and vice versa. It also brought the carrier communications system to its knees: a practical requirement of the power pool arrangement with wheeling over the BPA's transmission lines was that those transmission lines had to carry protective relaying pilot signals for all of the utilities involved.

A severe fault in a utility taking power off of a BPA line, for example, might require opening protective relays at a power plant operated by a different utility somewhere else on that line. New techniques for communications-aided protective relaying were under development, things like "permissive underreaching transfer trip" and "permissive overreaching transfer trip" that are difficult to explain briefly. These required that protective equipment at each substation have information about the state of protective equipment at all of the other substations, in order to make decisions that are not just based on local conditions but are globally optimal for the health of the whole line.

For example, avoiding a dangerous "islanding" condition on a transmission line might require opening relays at three or more locations along the line, but opening any relays beyond those required would simply cause unnecessary outages. Further, reliability is key and some types of faults on high-voltage lines are self-clearing (this is sort of a euphemism for the fact that a tree branch bridging sufficiently high-tension conductors will often be knocked off of them, if not vaporized entirely, by the resulting arc). Some protective relays are "reclosing" (and are often referred to in brief as "reclosers"), meaning that they will automatically reset (close) after a brief wait period. Some of the time, the fault will be gone and service is restored. Reclosing is potentially dangerous, though, and especially in transmission networks reclosing is only desirable in certain circumstances. Further pilot signals can be used to enable and disable reclosers based on the nature of the fault or the other locations at which it was detected, so that for example a fault that has taken a power plant offline does not lead to a recloser elsewhere "flapping" and stressing the remaining production capacity.

By the end of the 1940s, the BPA's carrier telephone system was overstressed to the point of failure. Power Pool utilities had connected their own carrier equipment, butting frequency bands so closely against each other that they began to interfere and degrade the quality of connections that could already be difficult to make out over hundreds of miles of high voltage infrastructure. The fact that a fault on a transmission line would generally prevent carrier current communications over that line working was a feature for simple fail-safe pilot relaying systems, but as protective relaying technology advanced it became the source of cascading failures. For example, by the early 1950s the BPA had found that lightning strikes on major transmission lines would cause enough interference to the carrier current system that protective relays all along the line would lose their pilot signals, open as fail-safe, and escalate what should have been a localized problem into a systemic one. Besides, many sections of the network had by that time become so congested by various carrier current circuits that there was simply no room in the spectrum for more—and so no capability to install protective equipment for new service lines.

Microwave

Meanwhile, the Second World War had had a profound impact in two ways: first, wartime demand for weapons, aircraft, and aluminum had driven Pacific Northwest industry to new heights. In 1949, power consumption in the Pacific Northwest had more than tripled, and population increased by 45%, compared to 1939. The Corps of Engineers built more dams, so the BPA's network carried more power. Engineers at the Administration's laboratory developed the techniques to operate transmission lines at record-setting voltages, 300 kV and beyond—making protection systems all the more critical. Second, wartime radar research had produced technological byproducts including high-bandwidth microwave transmitters and receivers.

So, the electrical industry adopted microwave communications technology alongside the telephone industry, and for much the same reasons. The high bandwidth of microwave links allowed for multiplexing huge numbers of channels, and the lack of wireline infrastructure (especially shared with the actual transmission lines) promised increased reliability. In my previous article on passive repeaters, I noted that they were especially popular with electrical utilities. Now, we learn why: utilities throughout the country built microwave communication systems that carried a fraction of the traffic of the Bell System, but often carried it to more remote locations and with higher reliability requirements.

The BPA's situation was different from that of more traditional utilities. Most private electrical utilities had started in a city and grown outwards, with a denser service network and fewer long-distance lines. They had mostly built private networks for communications, stringing their own open-wire telephone lines between stations. The BPA, with its far-flung network and huge distances, couldn't afford that kind of investment in stringing wires. In some ways, this proved an advantage: they could start from a clean slate. This network would be architected from the beginning for efficiency and performance, rather than to accommodate existing infrastructure.

While the BPA had initially built substations for a larger staff, the control technology was rapidly evolving and some of the more intensive activity BPA had expected at substations proved unnecessary 4. Despite the generously sized control houses, by 1950 most substations had only a single operator on staff, who would often be "out in the field" tending to the transmission lines. This created a problem if some sort of sudden problem required reconfiguring the network to restore service—a system operator might be left ringing a substation phone with no one there to answer.

That year, three substations had been equipped with "supervisory control" technology. This new system combined telemetering and teleoperation, so that a system operator in Portland could monitor voltage and current measurements at the substation and remotely operate the switchgear. The benefits of supervisory control were obvious, but the limitations of the carrier telephone system meant that those three substations were as far as the system could reach. Based on its promising experience with supervisory control of those three stations, and the clear need to continue expanding an already-stressed communications system, the BPA in 1949 committed to the construction of a completely new, completely modern control system for the Northwest Power Pool.

I put a lot of time into writing this, and I hope that you enjoy reading it. If you can spare a few dollars, consider supporting me on ko-fi. You'll receive an occasional extra, subscribers-only post, and defray the costs of providing artisanal, hand-built world wide web directly from Albuquerque, New Mexico.

The BPA's substation in Vancouver, already one of the largest, had a combination of ample space and proximity to Portland that made it a convenient location for support facilities. Construction yards, maintenance shops, and research laboratories had all been added onto the facility. The 1949 project launched with a symbolic gesture: the site was renamed, from simply the North Vancouver Substation to the J. D. Ross Complex in honor of the BPA's first administrator. A new building at the Ross Complex, the Control Center, became the nerve center of the system and the Rome to which all microwave routes led.

Keeping with our communications theme, one of the main features of what was then called the Ross Control Center was a "three position turret" (yes, a turret!) by which operators could call any line on the microwave, carrier current, or leased line telephone networks. The plan was to obsolete the turret, though. In 1950, the BPA awarded a half-million-dollar contract to the Philco corporation (originally Philadelphia Battery and later part of Ford, then GTE, then Philips, amusingly providing another expansion of the name) for its new microwave network. Philco had done extensive military work on microwave radar during the war, and was at the time one of the leaders in microwave technology. Philco's expertise would be needed, because the microwave network ordered by the BPA would be the largest electrical utility communications network in the world. While more difficult to conclusively state, I think it is likely that it was either the second largest microwave network of any kind after AT&T's, or the third largest after those of AT&T and the Santa Fe Railroad (both of which were also Philco customers).

At this point, the BPA's transmission network had extended to the Hungry Horse Dam in northwestern Montana, adding customers along the way. There were some 14,000 circuit miles of transmission lines and 400 substations in the system, and Philco's contract specified that they would build out the initial operating network in just 360 days. Work got underway: BPA signed the contract in February of 1950, in September testing and demonstrations were conducted, and in October the BPA officially activated the first leg: a 200 mile route from the Ross Complex to Snohomish, Washington.

Snohomish

This first route, built at a cost just under a million dollars, used repeaters at Mt. Rainier, Chehalis, Olympia, Squak Mountain. The longest single jump was Olympia to Squak Mountain, about 55 miles, an unusually long distance for microwave facilitated by Squak Mountain's prominence and a 150' tower at Olympia. With multiplexing equipment, this route carried 23 channels from the control center to the Puget Sound area.

This first microwave link was quickly put to work for one of the most interesting new applications of utility telecommunications: fault locating. When a fault occurred on a long transmission line, the BPA's first step was to search the whole line for the problem. Many lines were in difficult terrain, so a helicopter or airplane was used to speed up the process. The faults were sometimes minor and not easy to see from the air (say, a broken insulator), so the survey aircraft would take photos for processing and analysis on the ground. This process was expensive and, moreover, it was time consuming. BPA engineers realized that sudden open circuits or shorts in transmission lines created electrical signals that propagated back through the line and could be observed on test equipment—so you could presumably locate a fault by calculating its time of flight to the substations at each end. The problem was obtaining a measurement of when the fault signal arrived at two different locations, in precise synchronization.

Requiring high speed and, more importantly, consistent latency, this was exactly the kind of problem that microwave lines were well suited for. ITT 5 designed the system to BPA specifications, including devices that detected fault waves and reported them over a microwave channel, and a machine that compared the timing of the received reports and calculated a likely fault position (in miles) relative to each substation. FTL seems to have estimated that the system was accurate to 600', BPA to 1,000'.

The Ross-Snohomish route carried other important traffic as well: as part of its inaugural celebration, Washington State Representative Henry M. Jackson used the new internal telephone at Snohomish to call Ross over the microwave link, congratulating BPA's administrator and chief engineer on the accomplishment. Although I am unclear on the exact criteria being used, newspaper reports consistently identify the link as the "first of its kind in the world" 6.

1950 was still early for microwave technology; AT&T's first commercial microwave link had only gone into service in 1948, and that was experimental. The transcontinental telephone "skyway" wouldn't be completed until a year later. As a result, microwave technology was unfamiliar to the public, and the appearance of parabolic antennas on BPA facilities—and at repeater stations on mountaintops and out in the woods—was conspicuous. One reporter called them the BPA's flying saucers, another explained the repeaters in terms of "pitcher" and "catcher." Every paper ran photos.

Over the next two years, Philco completed a second microwave route up the Columbia Gorge connecting each of the dams through to Spokane, and a third that linked Beverly, on the route to Spokane, to Snohomish—forming a ring like the original Master Grid that provided redundancy and direct protection channels for transmission lines on that route. The 1952 microwave network included fourteen primary terminals and 21 repeaters; it connected the dispatch telephone system at Ross with dams and substations along the microwave routes as well as seven mountaintop HF stations to reach field crews.

As was typical at the time, the microwave equipment at each repeater and terminal ran directly from battery power. The batteries were charged (normally floated) from two different power supplies, one from the utility and the other from an on-site propane generator with a two week fuel supply. BPA initially used prefabricated aluminum shelters for the equipment at repeaters, although many were originally built or later rebuilt as cinderblock. This was, in part, due to the weather: mountaintop repeater stations in Washington coped with severe winters, and BPA went through several rounds of modifications to their building and tower designs. Towers were reinforced against ice accumulation, and repeater stations in particularly snow-prone parts of the Gorge and the Snoqualmie Pass were made two story. These buildings had a "balcony" entrance on the second floor with a ladder to reach it, allowing access even when the first floor was completely buried in snow. Towers were rated for 100 MPH winds, and some for thousands of pounds of ice. BPA designed ice shields for the antennas, and a heated cover to keep ice from covering the reflector surface.

BPA also took advantage of passive repeaters to relocate microwave sites to more accessible locations. At Rockdale, on the edge of the Cascade Mountains, a repeater was installed just next to the highway. Its antennas aimed more upwards than sideways, at reflectors at the top of the mountain ridge. Microflect passive repeaters were manufactured in Oregon, and the BPA was one of Microflect's first large customers, contributing design changes for mountainous service.

By 1955, the BPA microwave network had reached Hungry Horse Dam in Montana and connected all of the major substations of the system. When a decision was made, in 1955, to relocate the power dispatch office from the Ross Complex to the new Portland headquarters, the capacity and expandability of the microwave system made the process much easier. The BPA's new HQ building looked more like a telephone exchange than a federal building: one of its most prominent features was a rooftop microwave tower. The 1957 annual report inventoried 61 microwave radio sites covering 1,300 miles of route, by then using a mix of Philco, ITT, and Motorola equipment.

BPA's network had, by this time, achieved many feats of rural service. One of the most impressive was service to the southern Oregon Coast: this area was very remote and faced terrible weather throughout the winter. BPA's 115 kV South Coast line was the only electrical service into the region, and it was regularly disabled by ice storms and flooding. Because of the area's mountainous terrain, line crews working in the area were only infrequently able to reach a substation near Eugene on their mobile radios, and that was their only way of talking to dispatchers to coordinate repairs. In an effort to improve service reliability, the South Coast became the pilot for a new model of field communications. Relatively closed spaced microwave repeaters, each with a VHF radio, would bridge radio channels onto the microwave telephone system. To stand up to the Oregon Coast, each of these stations used an aluminum enclosure with heating, fire suppression, and a generator. Some of the enclosures were cabled to the ground, to better hold them down against the wind.

During the 1960s, the value of the microwave system had been proven but it was once again facing the limits of its capacity. Microwave technology had improved tremendously in the post-war decade and BPA's 23 and 24-channel multiplexers were obsolete. A $1.6 million contract was let to Lenkurt to upgrade much of the microwave network to the modern Lenkurt 76C microwave radio and 46A or 34A multiplexers, capable of up to 600 channels at around 8 GHz (the previous system varied from site to site but operated at around 2 GHz).

This was equipment designed and built under GTE ownership, and was also typical of the long-distance microwave links in the GTE telephone network. As part of the project, Lenkurt also installed VHF radio relays throughout the system. At some sites, Lenkurt installed dual-polarized antennas to allow simultaneous operation of the old and new multiplex systems and, later, increased capacity. Further contracts expanded the microwave network to Bellingham, Washington and to Corps of Engineers projects on the Snake River. In 1966, BPA dispatchers in Portland could monitor production at 21 dams, receive alarms from 250 substations, and completely remote control fifteen of the network's key switchyards.

When the BPA built the Pacific Intertie, a combination of two 500 kv AC circuits and one 800 kV DC circuit stretching 900 miles from the Columbia River to near Los Angeles, it was the largest transmission line project in US history. Under construction from 1965 to 1970, the line's standard bearer came in microwave form. Collins Radio built microwave routes parallel to the Intertie transmission lines, a $2 million project with 22 new radio stations on the 600-channel Lenkurt system. Most of the work was finished by 1967, a prerequisite for some construction the transmission line itself, since the microwave channels were used for testing and commissioning.

The Computer Age

Microwave was not the only arena in which the mid-century had brought new technology. The BPA was not new to computers; various forms of electromechanical computation had been part of their engineering works since the 1930s. By the 1960s, though, projects like SAGE (a military air defense system) and SABRE (a commercial airline reservation system) demonstrated the potential of combining computers with telecommunications for real-time control. The BPA had a telecommunications network, and it had a real-time control system... and they decided to add a computer.

In 1966, the BPA announced that its control center would once again move, from the Portland headquarters building back to the Ross Complex, where a new building would be designed from the ground up for centralized, computerized control of the power system. Named the Dittmer Control Center after a previous BPA power manager, the low-slung building had a prominent concrete microwave tower and was partially sunk below ground level for hardening against attack (it was, after all, the Cold War).

Much of the Dittmer center's lower-level floorspace was devoted to equipment rooms, which would soon be the home of the computer system that BPA contracted to Rockwell. Among other equipment, Rockwell installed a PDP-10 computer that received, analyzed, and logged telemetering data from throughout the system for display to operators. From its commissioning in the early 1970s, BPA continued to enhance the computer center into an integrated power dispatching system that supported operators in monitoring the network, predicting future demand, switching transmission lines, and ordering changes in production at power plants throughout the Pacific Northwest.

The ultimate manifestation of the PDP-10's software was called RODS, the Real-Time Operations, Dispatch, and Scheduling System. RODS was one of the first systems of its kind, initially contracted to Rockwell in part due to their experience with control computers for the Apollo program. Features of RODS included a time-synchronized data acquisition system to support differential current monitoring (ensuring that the computer compared current measurements taken at precisely the same time), which used an atomic time standard at the Dittmer Control Center to distribute high-precision timecode through the microwave network. As RODS matured, it established the 15-minute scheduling loop used by dispatchers to configure power plants and transmission lines for a constantly changing electrical load. DEC themselves, in an internal sales meeting whose minutes fortunately made it into the historic record, noted the Dittmer PDP-10 as a critical early sale in their efforts to break into the utility industry.

Despite the technical firsts of the Dittmer Control Center and RODS, this chapter of BPA history is best known for its incidental brush with the Pacific Northwest's most famous chapter of computer history: as temporary employees of TRW, the aerospace contractor brought on for RODS software development, Bill Gates and Paul Allen both spent time at Dittmer. They were still in high school, it would be years before they moved to Albuquerque to found Microsoft.

Many parts of the BPA's microwave network are still in use, although improving radio technology and the adoption of fiber optics (including fiber embedded into the neutral conductors of transmission lines) have allowed for elimination of some repeaters. As microwave technology continues to fall obsolete (in comparison to fiber, commercial terrestrial radio networks, and satellite), many of the remaining sites are likely to be demolished. Some of them—sites like Chehalis, Squak Mountain, and Rainier—have been in service for 76 years.

The BPA microwave network is not unusual. It was the first of its type, but the transmission and power marketing concepts pioneered by the BPA are now used nationwide. Wherever power goes, protective relaying follows, and the telecommunications networks that shadow the electrical grid from coast to coast. Few enterprises outside of the communications industry itself have ever operated communications networks on the scale of electrical utilities. They find themselves in the company of railroads and oil pipelines, ventures that must span deserts, climb mountains, and cross rivers.

Unlike many of those operations, though, the BPA is a federal agency, subject to the National Environmental Policy Act and National Historic Preservation Act. BPA's extensive Section 106 NRHP compliance program has produced extensive historic documentation of its transmission lines, communications infrastructure, and research facilities. While sometimes onerous, these federal policies have created the best-documented historic electrical utility communications system in the nation.

An inventory of historic microwave stations still under BPA ownership turned up 28 in Washington, 22 in Oregon, one in Idaho, and two in Montana, and many are eligible for nomination to the Historic Register. From an aluminum shed at the Ross Complex to a squat concrete building and utility pole at Mary's Peak, they are monuments to American history—a singular, almost megalomaniacal vision of the Columbia River tamed; the rise of Pacific Northwest industry under wartime demands; a technical approach to environmental preservation and economic equality. Infrastructure for the public benefit: a great American achievement. We might, one day, find the will to do it again.

1. This Vancouver is just across the Columbia from Portland, and should not be confused with the big one in British Columbia. Having grown up in Portland, I will probably not disciplined enough to put Washington after it each time, so just remember that we aren't talking about Canada. Yet.↩

2. Lenkurt would later merge with GTE, becoming something like the Western Electric to AT&T's principal competitor.↩

3. This is another example of the difficulty of electrical terminology. "High tension" in this context is a result of the historic use of "tension" (and enduring use in some languages) to refer to what we now usually call "voltage" or "potential." Specifically in electrical distribution, "high tension" is older language but still in use to refer to 100 kV and higher transmission lines.↩

4. BPA was navigating many firsts in the construction of the Master Grid, and the high-voltage transformers used to step up and down from the 230 kV lines were new technology. They were filled with a heavy oil for cooling and electrical insulation, and the BPA had at first expected that they would need to drain the oil and open the transformers on a regular basis. Early BPA substations featured a network of rails and low-slung metal carts that would be used to pick the transformers up off of their pads and move them to a specialized building called an "untanking tower," where they could be drained of oil and the covers lifted off by a gantry crane. In practice, transformers were virtually never opened, and the untanking towers and rails became vestigial.↩

5. This company changed names from Federal Telecommunication Laboratories to Federal Telephone and Radio and was then acquired by International Telephone and Telegraph (ITT), all during the time period covered by this article. I will refer to it simply as ITT for readability.↩

6. You know that I am a little bit obsessive about this kind of superlative, and since it seems to have originated with Chief Engineer Sol Schultz, I tend to think he had a good definition in mind. My best guess is that it was the first microwave link in the world (at least that they were aware of) to carry telemetering and teleoperation signals over such a long distance.↩