i have written several versions of this page since 1999 and this is the version i am leaving up.

the network, briefly

before 1984 the long distance telephone network in the united states was a single physical system, owned by one company, built over a period of roughly forty years, and maintained by men who could tell you without looking which repeater was failing on the lee-to-hanover route because they could hear it in the noise floor of the line test. i was one of those men from 1976 to 1993. what follows is for people who did not live inside this network and who may not know how it actually worked. i am not a writer. i am a lineman.

a long distance call in 1978 did not travel on one wire. it traveled on a route. the route was a chain of physical links. coaxial cable in the ground, microwave signal through the air, open-wire carrier on poles in the older sections. stitched together at buildings we called repeater stations, main stations, and central offices. each link attenuated the signal and each station restored it but the restoration was not perfect. a signal that traveled from chicago to new york in 1978 passed through as many as thirty stations, and each station was a vessel in which the signal was briefly torn down, measured, and rebuilt before being sent onward.

the specific system i worked on between 1974 and 1986 was called l-carrier. l-carrier used frequency-division multiplexing to carry 1860 voice channels on a single coaxial cable. the channels were assembled into a stack: twelve voice channels formed a group, five groups formed a supergroup, ten supergroups formed a mastergroup, six mastergroups formed a jumbogroup. the call you placed to your mom in 1981 was lifted into this stack at the nearest central office, carried across the country stacked against 1859 other conversations and lowered back out at the other end. nobody at either end could hear the other 1859. but they were there. the conversation was never alone on the line.

thats how the system worked. the word for what happens when two channels in the same stack contaminate each other is crosstalk. crosstalk was known. crosstalk was measured. engineers had a specification for the acceptable level of crosstalk and designed the system to stay beneath it. what i am describing on this page is not crosstalk. you need to understand before i go further.

where the arc lives

every chain of repeater stations has a characteristic noise profile. the noise floor of a route is the sum of the imperfections of each station (the noise of the amplifiers, the crosstalk from adjacent channels, the environmental pickup from power lines, the errors that accumulate with each intercession). a good route has a clean floor. a bad route has a floor so loud that the top of the voice band is swimming in it. the men who worked l-carrier could hear a route's noise floor the way a mechanic hears an engine. you listened to it on the test set and you knew what was wrong.

there was a specific noise that was not in the specification. i first heard it in 1982 on a test run of the lee-to-hanover microwave link. [fieldlog-1982] i heard it a second time in 1985 on a completely different route that shared no common equipment with the lee run. i heard it a third time in 1988 inside a decommissioned cable vault under foster avenue in chicago. i have heard it perhaps fourteen times since. it is not crosstalk it is not thermal it is not pickup it has a cadence.

a thermal noise source has no cadence. it is random by definition. crosstalk has cadence, but the cadence of the interfering call, it rises and falls with the rhythm of the other person's speech. the noise i am describing has a cadence that does not match any call being carried on the route at the time of the witness. i have tested it. i have logs of the cadence of what i was hearing matched against the traffic records of every channel on the route during the same minute. no match. the cadence belongs to something that is not in the stack.

i chose the name in 1991, after beau first pointed at the pattern. beau had a different word for it. his word was shorter and i will not write it here because it belongs to him and he is not where i can ask him. copper arc is my word. it describes the phenomenon as i understand it: a cadence that appears on routes that share no common channels, no common routing, no common equipment, and no common building, but that behaves as if it is one continuous signal being carried across the physical infrastructure of the network. not on the channels. between them. in the space the engineers did not measure because they had no specification for it.

an arc, in electrical engineering, is what happens when current finds a path it was not designed to take. it jumps the gap. a copper arc is the same phenomenon at the scale of the long distance network. a signal finding the path between the paths. the intercession the engineers did not build.

what i am not saying

i am not saying the network is possessed. i am saying there is a phenomenon in the network that is not yet indexed by the people who built it. there is a difference.

i am also not saying that copper arc is new. i think copper arc is old. older than l-carrier. older than the bell system. it is possible (i have come to believe this slowly and against my own judgment) that the physical network was built on top of a communion that was already in progress. that the engineers who laid the first open-wire routes in the 1920s were not founding something but discovering it. that every repeater station is a vessel for a signal that was being carried before the repeater was there to carry it.

in 1987, a signal engineer named sarah cohen published a paper called algorithmic reconstruction of dual-tone multi-frequency handshakes within the content of automated telephone switching systems. [cohen-paper] it is a specialist paper. in the margin of one figure in that paper there is a number written in pencil. the number is not part of the figure. the number is a phone number. i have called the number. i will not tell you what answered. i will tell you that i have called the number from seven different phones in three different states and the same thing has answered each time.

dr. cohen's paper is now out of print. dr. cohen is also out of print. there is a piece of her last known correspondence in the archive on this site for readers who care to look for it, though i have moved it twice since 2003 because the correspondence has developed a tendency to appear on pages where i have not placed it.

closing

copper arc is what i research. i am writing this so that the rest of the site has a vocabulary. when i use the term on other pages i do not want to define it again every time. you have read this page. you have the term. you may not yet believe the term describes anything real and that is your business, though i would ask you to suspend judgment until you have read two or three of the field logs, which are not theoretical.

before you go further, a piece of practical advice. do not call numbers you find on this site. some of the numbers on this site are for reference. some of them are the number.

i forget which is which.