AUTOVON: A Phone System Fit For The Military

It’s a common enough Hollywood trope that we’ve all probably seen it: the general, chest bespangled with medals and ribbons, gazes at a big screen swarming with the phosphor traces of incoming ICBMs, defeatedly picks up the phone and somberly intones, “Get me the president.” We’re left on the edge of our seats as we ponder what it must be like to have to deliver the bad news to the boss, knowing full well that his response will literally light the world on fire.

Scenes like that work because we suspect that real-life versions of it probably played out dozens of times during the Cold War, and likely once or twice since its official conclusion. Such scenes also play into our suspicion that military and political leaders have at their disposal technologies that are vastly superior to what’s available to consumers, chief among them being special communications networks that provide capabilities we could only have dreamed of back then.

As it turns out, the US military did indeed have different and better telephone capabilities during the Cold War than those enjoyed by their civilian counterparts. But as we shall see, the increased capabilities of the network that came to be known as AUTOVON didn’t come so much from better technology, but more from duplicating the existing public switched-telephone network and using good engineering principles, a lot of concrete, and a dash of paranoia to protect it.

Parallel Lines

The world’s militaries have always been early adopters of technology, eager to put the latest scientific and technical advances to use for their dual missions of national defense and projection of political power. This is particularly true of communications technologies, with telegraph, telephone, and radio communications being quickly adopted by militaries between the mid-1800s and the early 20th century, and the US military was certainly no exception.

Most of the connections between US military bases before the 1930s were dedicated lines, with some leveraging of the growing public switched telephone network. But with mobilization for World War II and the subsequent Cold War period leading to a massive expansion of the geographic footprint of the military as well as an increased number of users, military planners saw the folly of relying on the civilian telephone system that was even then experiencing its own growing pains. By the early 1960s it was clear that a phone system dedicated to and specialized for the use of the military was a national security priority.

The designers of what would become known as AUTOVON, or Automatic Voice Network faced a conundrum. The civilian long-distance phone network, which was undergoing a massive buildout at the time, was almost exactly what they needed to provide for the military. But they also knew the limitations of the public telephone network, a system that military planners were certain could collapse under the weight of a panicked populace rushing to the phones to contact loved ones during a national emergency. That precluded leveraging the existing network for the military, so the decision was made to build a second long-haul telephone network, dedicated to military use but mirroring the civilian network in architecture and adding features needed by the military.

Map of AUTOVON stations in the continental US, c. 1970. Note the healthy distance between AUTOVON stations and major population centers. Source: A Secret Landscape: America’s Cold War Infrastructure.

Like the AT&T Long Lines system that it essentially duplicated, AUTOVON was laid out as a series of long-haul connections that hopscotched across the country. But while the civilian system was designed to hop from city to city and provide service to the maximum number of paying customers, AUTOVON used what was known as “avoidance routing” to build a hardened network.

They weren’t fooling around. Blast doors on the AUTOVON bunker in Pottstown, PA. Source: Long-Lines.net.

AUTOVON switching sites were placed safely outside the range of a nuclear attack on any major population center, to ensure their survivability. For instance, Pottstown, Pennsylvania was chosen as an AUTOVON switching site on the transcontinental route from New York to Los Angeles. It was linked to the important naval facilities in Philadelphia, about 35 miles to the southwest, by coaxial lines and microwave links.

Connections between AUTOVON sites were mainly via buried coaxial cables. Much of this cable was directly in the ground rather than encasing it in concrete, as that was deemed sufficient to protect the cable against anything short of a precision attack. Concrete conduits were, however, used to protect sections of the AUTOSEVOCOM, or Automatic Secure Voice Communications network, a parallel military communications network that was designed to pass only encrypted traffic.

Many of the AUTOVON main stations, like the previously mentioned Pottstown station, were equipped with massive underground spaces to house all the telephone switchgear. Burying the buildings was an attempt to protect them from all kinds of disasters, although the details of the structures make it clear what the designers had in mind. The buildings were linked to the surface by shafts with reinforced blast doors, there were massive ventilation fans and filters to provide positive pressure, and diesel generators could provide half a megawatt of power to keep the facility running. Add to that the fact that all the equipment was mounted on shock-absorbing springs and the AUTOVON sites would have been a great place to ride out a nuclear attack.

Flash, Flash, Flash

Aside from its avoidance routing and the reinforced bunkers at its main stations, the AUTOVON network was almost identical to the civilian telephone network of the time. Like its civilian counterpart, it was a circuit-switched network, meaning that it was designed to establish a fixed physical path between sender and receiver and maintain it for as long as required. As such it used much of the same switchgear used by the civilian system, initially using the same crossbar switching gear that had been in use since the 1940s. Later, the crossbar switching gear was swapped for electronic switches, which enabled some of the special features that AUTOVON became known for.

“Get me the president.” An AUTOVON phone in “Hotline Red” with the extra column of Touch-Tone keys. Source: AUTOVON.org.

One of the shortcomings of the circuit-switched civilian network is that once a connection is established, it stays connected until one or the other party ends it. Outgoing circuits are a limited resource, though, and once all the circuits are in use, calls cannot go through. In a time of emergency, it wouldn’t do to have a soldier’s call home to his sweetheart prevent a base commander from receiving an order from the Pentagon, and so a system to prioritize and override calls was devised.

Dubbed multilevel precedence and preemption, or MLPP, the scheme provided five levels of precedence. The lowest precedence level was Routine, which comprised the bulk of AUTOVON traffic. Above that was Priority, followed by Immediate, Flash, and finally Flash Override, theoretically to be used only by those with direct permission from the president. AUTOVON phones accessed the MLPP system using a fourth column of buttons on special 16-key Touch-Tone keypads. The four extra buttons were used to assign one of the higher precedence levels to the call by pressing the corresponding key prior to dialing the rest of the number. There was no key for Routine-level calls, since that was the system default.

If a precedence level was assigned to a call and there were no outgoing circuits available, the MLPP system searched for a circuit occupied by a lower precedence call and immediately disconnected it. Call participants would hear a characteristic fast busy tone to let them know they had been bumped. If an outgoing priority call was made to a number that was engaged in a call with lower precedence, the original connection was broken to let the higher precedence call through.

The Packets Are Coming

Construction of AUTOVON began in 1963, and the system went into service at military bases throughout the country in 1966. Three years later, AUTOVON switching stations had been built in Europe, Japan, Korea, the Middle East, and in the Caribbean. Eventually there were about 70 sites in the system, providing the bulk of non-secure communications for the US military and allied forces around the world.

But even as the initial system was still being rolled out, a little project in the Advanced Research Projects Agency was in the works that would plant the seeds of AUTOVON’s eventual demise. The project was ARPANET, the first serious packet-switched network, which would lead to the Internet and make almost everything about AUTOVON obsolete. It would still take almost another 30 years for the final pieces of AUTOVON to be retired, though, which is a testament to the staying power not only of circuit-switched networks but also of great engineering.

50 thoughts on “AUTOVON: A Phone System Fit For The Military

    1. Yah, I was kind of thinking something similar. It would be so hard with those buttons just sitting there.

      But I suspect that just means you and I would have been weeded out long before we reached a position with access to one of those phones and not that people were pushing them when they shouldn’t have.

    2. When I worked at a place, we used the FO button every day to make sure that it worked as intended. In the current system, the level of priority that can be used is assigned line by line. We even programmed certain lines and speed dials to always dial with certain priorities.

    3. When I worked at SAC HQ in 1973 I had a regular (dial?) telephone on my desk. I was a computer programmer with no special need for communications. My phone had no special buttons. I could still get priority Autovon access by using the dial. A straight number got someone on the base. 9 got a local outside line. 8 got a routine Autovon line, and 7 got a priority Autovon line. (I hope I got the numbers right after some 50 years). At the time I had no idea Autovon was so new but this article changed that. We also had Autodin which had the same precedence codes for TWX messages which were similar to email but had a different addressing scheme. TWX messages had to be signed by your superior officer and were sent from the communications office rather than directly from your desk.

    4. I worked on this system. When you were assigned a phone line, it had a maximum capability so not everyone would have access to use Flash Override. Flash Override is reserved for executing nuclear authorizations and National Command Authority communications. Flash could be things of national importance like the command of forces during a major attack, think of strategic kinds of communications or something the White House/Pentagon needs to know right now. Immediate and Priority can be used for day to day high priority stuff passing orders around the military.

      If you made a Flash Override call it would definitely trigger alerts and monitoring in the switching centers and you could expect a nasty call if abused.

      Very interesting that Bell designed DTMF signaling with the priority system in mind. If you ever wanted to know why your DTMF pad is a 3 x 4 matrix, the answer is because the far right would contain four more buttons for precedence on a military phone keypad so they end up with a four by four matrix which is easy to handle with a total of 8 DTMF tones. Each row is a tone and each column is a tone. The two tones combined are your DTMF.

  1. I love that the author thinks that a soldier would be using this system to “call home to his sweetheart”. Every soldier and their girlfriend was given an AUTOVON line as standard issue, right?

    1. Back when an airman’s pay was low and long distance fees were high, yes it was used to phone home. The trick was you had to call from your base to another base that was local to where you wanted to call. When connecting to the distant base you asked the operator to connect you to an outside number.

      1. Ditto that! When I was working on a non-classified research project on the DEWLine in northern Canada the station members showed me how to dial to a local military base in the US and ask for an outside line which was close to a family member or friend and have the charges reversed. They were happy to do so and quite used to doing it. DEWLine stations were about 300 miles apart and there were NO landlines for them. They used forward scatter communication system to go that far. Another communication method not many know about.

        1. I third that and that was the only way to accomplish “morale calls” when stationed at any of the many remote sites overseas. Like you said, operators were happy to do that for you. I never worked the DEW line which would have been interesting, but was comm maintenance at a remote site in mainland Europe for the standard remote tour length. That was definitely interesting in a good way, but I would have liked to have experienced the far north to add to my long list of the many interesting places I’d been and interesting things I did in the USAF which I could never have seen or done if I hadn’t been in the USAF.

          DEWline Adventures

          http://www.dewlineadventures.com/

    1. 1980, and I was calling my girlfriend (now my wife of 40 years) who was stationed 250 miles away. We both worked mid shift (24:00 to 07:00) and spent several hours most nights talking while doing our work. I learned early on that I could use AUTOVON to call her base operator and ask for a patch to her civilian number – I was never denied.

      And, it was not very difficult to get access to those phones with 16-key pads. I ended up in a tactical comm squadron, and everyone in the unit (plus many other units) had daily access.

  2. So I guess there were “peering” arrangements with the PTSN? I guess there were maybe instances where you had to hold up a PTSN receiver up to the left ear and the AUTOVON receiver up to the right ear to act as teh peers between the networkz.

    1. When I was in grad school at UMass, we got a tour of the Chesterfield facility. 3 floors, underground. Civilian Long Lines switch on the top floor, AUTOVON switch on the next floor down, and batteries on the bottom floor. The air intake was a huge vertical shaft, 3 stories down, and at the bottom were blast doors, kinda like mushrooms mounted horizontally, so a blast would force the covers onto the air intakes.

      There was at least one “nuclear detector” outside on the lawn, connected into the AUTOVON system to indicate that this site had detected a nuclear blast, and the usual 2 ft thick blast doors, underground fuel tanks and generators. Filtered blast-proof air intakes, etc. Coaxial cable connection, I believe and a disused microwave tower on the hill nearby.

      Long abandoned, and bought by a local contractor, renovated and used as his work location, I believe.

      http://coldwar-ma.com/ATT_Chesterfield.html

  3. While the system did have dedicated switching, and capacity, most of it was still transported over the regular microwave network or cable system. The use of only concrete ducts for Autsevocom or Autodin is fantasy. Most of the sites were colocated with regular Long Lines equipment for commercial use.

  4. The article missed an important part: AUTODIN. AUTODIN was the digital communications network that worked over AUTOVON lines. It was the direct precursor to ARPANET. The equipment looked like an IBM keypunch/verifier/card reader with a phone attached. I would use the AUTOVON number specifically assigned to various similar stations ranging from Florida to Washington, DC to Ankara, to who-knows-where. I would reach the operator at the other end and we would place our handsets in the modem receptacle cups, and, after the handshake, I would initiate transmission of data contained on Holerith cards. If I received the call from someone else, they would be the ones to initiate the transmission. (The modem had a switch we had to set for “Receive” or “Send”) A full hopper of cards took about 55 minutes, so I studied a lot. I learned to program (IBM 1401 AUTOCODER), got many college credits, practiced karate and fencing… In other words, it was very boring. It also required exacting attention during the setup to transmit; any out-of-order cards required a re-transmission and counted as a “gig” against me in my log. Worse, sometimes I had to activate the cyphony equipment. Any out-of-order cards and the communication would be hopelessly scrambled, so the whole thing had to be regenerated and re-encrypted before re-sending.

    BTW, the civilians DID have a similar network. It was called a WATS line. The only difference between the two was that the military had some hardened connections and all AUTOVON was dedicated military.

    And I was using this system in late 1965. Mine might have been in one of the first installations, which might have made sense seeing that I was stationed in Alaska at the time and we were always playing games with the Soviets. There must have been more than 70 sites; just the connections between Ft. Richardson and Elmendorf AFB would have accounted for 2 of them, and I had connections all over the USA, Canada, and NATO bases.

    1. This is necessarily a hazy memory, since it’s from 1976, but as I recall, the phone network in Alaska was unified – Autovon and commercial phone networks were not just co-located, but interconnected as well. You couldn’t call Autovon numbers outside of Alaska, but from any phone in Alaska, you could directly call any Autovon phone in Alaska – to the commercial system, Autovon numbers were ordinary numbers in the Alaska area code. I demonstrated this to someone once, by pulse-dialing (using the hook switch) the operations center at Cold Bay Air Force Station from one of the phones at the airport in Anchorage, which didn’t have dials, but were set up to auto-dial local hotels to arrange for pick-up. My own low-grade phone-phreaking experiment.

      There were also a few of us who had been stationed at Cold Bay, who ended up scattered around the globe after our tour there. We used to make conference calls among us, by arranging to call Cold Bay ops at a certain time, and have them transfer us each to one of the tie lines that were set up for doing conference calls within the site.

      1. White Alice. It was a combination microwave troposcatter system. Built by Western Electric, so I assume they wangled a deal so they could handle civilian traffic when the military wasn’t using the bandwidth.

        1. Indeed. I worked alongside White Alice technicians at Cold Bay.

          Basically, White Alice was Western Electric’s (and later, RCA Alascom’s) way of piggybacking commercial telephone service (which would otherwise have been impractical) on the military communications system that they had won the design and maintenance contracts for. This pre-dated Autovon, Autodin, and Autosevocom, but it was incorporated into those systems when they appeared. Point is, this wasn’t like the Autovon system in the “lower 48” that kept the military system completely isolated from the commercial system, even where they were co-located. There was a firewall between White Alice and the U.S. telephone system (we had to go through an operator to connect to make personal calls to the lower 48, for example, and had to dial a prefix number (9?) to call off-site Autovon numbers), but I’m pretty sure we could call any number in Alaska (and vice-versa) directly. In Alaska, we had ordinary dial telephones, not the 16-button touch-tone Autovon phones we had in the rest of the Air Force.

          1. Most of the autovon access in CONUS was through your standard handset. Dial 61 and you got a dialtone, then dial the AV number. Only locations that may have a need for priority access (and there were not that many, even on my SAC base) got the 16 button phones. I wish I had hung on to some of the Autovon phone books that I tossed when we went to the DSN.

    2. Everyone overlooks AUTODIN for some reason. There were terminals (Teletype based) that were completely automatic, with message traffic that any reasonable person today would call “email” or “file transfer”. I think there are a few good videos on Youtube going over the system.

      1. Yes, AUTODIN was a pretty cool routing system for any kind of digital data. Could handle everything from paper tape, mag tape, OCR forms, etc. Essentially your message went with routing indicators to AUTODIN switching centers for distribution according to the routing indicators. The AUTODIN system interfaced with an even harder system called SACDIN or SACCS that allowed Strategic Air Command to access this network from their missile sites and command posts. Think of the network as the predecessor to email. You would have a base communication center with AUTODIN capability and could bring messages there to be sent anywhere in the world. Things like payroll and supply data could also be sent over that network. You could receive a payroll mag tape that gets carried over to the base mainframe and loaded up. Often that base communications center and mainframe computer centers were located in the same building for easy of transport.

    3. When I was working on both networks in the 80s and 90s, they did not share communications lines. The AUTOVON was TDM / FDM voice circuits or analog phone lines where AUTODIN was digital circuits ranging anywhere from 75 baud TTY to 56k circuits. There were some capability to use a modem to dial up a 9600 bps modem into AUTODIN over an AUTOVON voice circuit but those were mainly for backup or contingency use (difficult to use because encryption systems in the day did not play well with analog modems, bit errors would drop crypto sync). They were pretty much two separate networks with separate switching centers. The only thing in common between AUTOVON and WATS is direct dialed long distance calls. There were many many more sites on AUTOVON than 70 but that is probably the count of switching centers. Many bases in the same region went to the same switching centers. All AUTOVON really was is a separate telephone network that worked just like the civilian network and actually interfaced with the civilian network. Both networks were Bell System creations. The point was the AUTOVON could not be congested or destroyed by traffic or damage to the civilian network. Other than the precedence system, the network were identical. The AUTOVON network essentially tied together base phone switches than in turn tied together building PBX systems. Just like today’s phone network. Overtime the AUTOVON network evolved into the DSN that exists today.

    1. When my son was deployed as an enlisted specialist to a base in Iraq (2006-ish), he’d call home when on night duty in the command bunker. Sometimes the call quality was…low. But when you figure it was probably bouncing off a satellite, meh. We were glad to hear from him.

      Also, he said the night sky was amazing, the best he had ever seen. He’s now a Captain, and stationed in Hawaii, and you bet we’re going out at Christmas :-)

    1. I’m wondering where (specifically) the Glendive, MT, or Wheatland, ND stations are/were located, as I have a better chance of being in the area of one of those locations someday.

      Are any of the old bunkers available for sale?
      B^)

      1. Be aware that I have noticed, over the years, that the name of the facility does not always correspond to its actual physical location.

        e.g.: This summer, I was in Winter Harbor, ME, on a vacation trip. So I stopped by the National Park office and asked where the old Naval Security Group Wullenweber DF antenna site was. All the documentation I have ever seen, says “Winter Harbor”. It’s actually a few towns over, in Corea (pronounced “Korea”, which was the cause for a little humor). Just an anonymous driveway off the main road. The antenna’s gone, the buildings are still there, and the site was up for auction when I sneaked in. Sold, I think, to an aquaculture company just after I saw it. The wife, for some reason, was not interested in owning it.

    1. The network was in flux at times. They would move a switch office with the rest of the network taking up the slack. An interesting aside, not all the switching centers were AT&T, some were local Telcos. And no, they did not lay in new cable in hardened ducts.

    1. I don’t know of any that are “abandoned”, all have owners. Up for sale yes. Be aware though – not all the sites were actual bunkers. Many of them had a basement with the battery plant, and a ground level with equipment. Still pretty tough buildings, but not “bunkers”. Mounds, Jasper, Ennis, all come to mind right off.

      1. Many were collocated with Bell System Central Offices. They would take a Bell System CO and the government would essentially pay the difference to beef it up. Made more economic sense and improved the interfaces between the AUTOVON and the civilian PSTN. Sites that were primary entry points for critical locations or backbone sites would be much harder. Look up the 465L backbone system to learn more about the hardened cable plants. A lot of AUTOVON sites corresponded with that backbone.

        1. Mounds was a microwave site, then they added the the Autovon component, ending up with a 1ESS. It was again expanded. When the area for Autovon was built, they took out one of the walls of the microwave repeater, and opened it up to the main room. The basement was built just under the additions of course. Almost all of the plant was on the main above ground level.

          It was a semi-hardened site, just like most of the AT&T junctions, the personnel shelter was in the basement next to the battery plant.

          It also had a SAGE component, an area that was dedicated to SAGE communications, even with phones labeled “SAGE”.

          Mounds was in the directory as a MACC site.

          I have a Autovon directory online, I will post a link. It also has a bunch of pictures of Mounds.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.