Around the middle of February of 2009 I began experiencing bursts of pixellation on channels that had been fairly noise free for over a year. In the hope that someone may be able to associate what they see here with something they're familiar with, below are links to a couple of flash videos showing what I've captured on a Spectrum Analyzer.
This is about a 4 minute example but it should start playing fairly quickly.
This is about a 30 second example.
In my efforts to discover the source, I used a Spectrum Analyzer to "see what was out there", and quickly found previously unobserved channels, albeit at low levels. My excitement over the possibility that new stations were on the air quickly vanished when I realized it was not just a few, but dozens of channels. Eventually, by walking around the neighborhood with half of an Antenna Craft YA10-7-13 (directional VHF Yagi) antenna, I discovered a leak in the cable company's coax about 300 feet from me. It was what they called a "ring split", a complete rip of the foil outer shield in the form of a ring, causing electrical interruption of the ground conductor and thus broadband transmission of their complete cable line up. When I notified them, their response was immediate, and I commend them for their response.
Then I began focusing on steady almost random looking impulse noise that I could observe spanning all the way from the HF through UHF spectrums (at minimum). I began investigating the power lines, and in fact, I found I could reliably pass underneath a set of high voltage lines that go through the neighborhood, and observe (1) a peaking of the noise amplitude on a Spectrum Analyzer, and (2) a peaking of noise loudness on an AM radio tuned to a vacant frequency (as I passed under the wires). I then took my monitoring rig up an access road to the local transformer station, and there the amplitude blows my socks off.
So I called NV Energy, the local power company. I had asked that they investigate, and let me know when they were going to do it, so I could monitor any changes also. Alas, they came to my door only after they'd run their tests. They did say they'd confirmed there is something radiating from one of six possible houses, because they'd shutdown sections at a time to see if they could interrupt the noise, and indeed one particular section did make it go away. At least, it made whatever they were monitoring on their equipment go away. Unfortunately it did not make the offending noise go away. I confirmed this when NV Energy revisited the same section of homes and interrupted the power to replace a component, and I got my chance to monitor the noise with that power interrupted. There was absolutely no change on either the Spectrum Analyzer or the DTV signal "Quality" indicators when this section of homes was shut down. I later got a few of those neighbors to shut down one at a time, and indeed we found an old refrerigerator that made the AM radio go completely quiet when shutdown. But it did not make the offending VHF interference go away.
Further discussion with the power company did not yield any promising avenues. They did suggest I look at new Frequency Hopping Spread Spectrum technologies. But those I looked at operate over a fairly narrow band, and most in the GHz range. I honestly don't see how they could be producing energy across such a broad 2MHz through 700MHz spectrum (or more). Despite the fact that the power transmission plant is clearly conducting the enery, their responsibility "ends at the poles", and their hands are tied apparently; they are just not allowed to go investigating individual homes. When querying them about broadband spark gap noise, they felt that if something were arcing within their equipment they would know it. Their explanation for why the signal amplitude clearly peaks when in proximity to the transformer station was because that is where the best ground is. This actually raises additional questions. Currrent is current; you can't have 1 Amp flowing in a single wire at one end, and half that at the other end. But you can have branch circuits whose individual currents add up to that of a single leg associated with the current source. So if the current is highest at the transformer station, doesn't that suggest proximity to the source? Or, if the source were a single remote point, wouldn't the current on one of the branches coming into the station be equal to the current (going to the ground) at the station? It would seem so, but walking away from that transformer station in any direction results in fading of the interference. If you then start at a point 100 feet or so out from the station and move 360 degrees around it, the amplitude is always lower; it never reaches the same loudness under any of the branch lines as when standing immediately next to the station.
Having reached a dead end there, however, and after a number of hours watching the noise on a spectrum analyzer at a variety of bandwidth and sweep time settings, I eventually recognized that there is a steady regular pattern to the noise frequency. As observed on a spectrum analyzer, whose display has its own perodicity to it (its sweep reptetition rate), there is a regular movement of the peaks of energy across the frequency spectrum. I can only attribute this to something analogous to a good old "beat frequency". It suggests to me the frequency of the noise source might in fact not be so random. Rather it appears to regularly shifting across the spectrum by design.
Could, in fact, something like a spark gap transmitter do this? I mean, the kind that was always humming away in the lab where Frankenstein came to life? E.g., could the energy of that spark that travels from narrow base towards the wider top do this? Dunno. I'm stumped.
There is a keg of beer of choice to the first person that provides a clue that leads to identifying the source !!!