My Search for Power Line Noise in Early 2007

On December 12, 2006 very significant line noise appeared at my Idaho amateur radio station location. For brief periods the preceding two years there had been line noise of the same intensity but it tended to appear and disappear rather than being present 24 hours per day every day.

I had characterized this line noise using my HP141T-8552B-8553B Spectrum Analyzer during the periods that it was present in 2005 and 2006. These periods were between December 15 and late April of each of those years. Also, the noise would start and stop periodically -- without apparent reason. Sometimes it would disappear for an hour and other times it would be gone for a week or two. This year it appeared in early December [happily, after the ARRL160 Meter Contest] and continued steadily through all of February [when this was drafted] totally wiping out the Stew Perry Contest, the CQ 160 Meter CW Contest and to a large extent the ARRL CW DX Test in January. The CQ 160 Meter SSB Contest in February was affected but the conditions were so bad {QRN} that it made little difference.

I decided that I needed to do something quickly to locate the specific noise source {or sources} and get the Idaho Falls Power company to fix the problem. Idaho Falls Power had already accepted responsibility to fix the problem but they had no trained people who could locate the noise so their crews could take action.

They had a Radar Engineering 240A receiver and a log periodic antenna for doing such locating but no member of their staff had ever used the equipment or gone through the process of finding an actual noise source using the equipment. I was asked to assist them in learning to use the equipment and locating my specific noise source. I agreed to do so with the understanding that they still owned the problem and I was just assisting them to deal with it.

I had used my rotary beam antenna to gather an idea of the direction of the noise from my location. In addition, I had a pair of K9AY loops that were switchable and I could get a  good idea of the quadrant that the noise was maximum on 160 meters.

I went to the ARRL web site to see what information they might have to help me. The have a series of web articles dealing with power line noise location and fixing which were written in a way that would be useful to a power utility. I downloaded these articles, printed them and gave them to the Idaho Falls Power staff. In addition I read them carefully and sent an email to ARRL HQ to Ed Hare, W1RFI, posing a question. Much to my surprise Ed called me to discuss what I was doind and made some suggestions that seemed to dramatically improve my search for the noise location.

The article includes the search principle that  Ed articulated and my own implementation of that principle.

The basic principle is:

Determine where the noise source is not.

In essence, eliminate all of the potential source locations until you have only one remaining. The find the specific cause at that location. I had told Ed that I really felt that there was more than one source based upon my spectrum analyzer photos. He looked at them and was not ready to immediately agree. We finally did agree that if we found and eliminated the most serious source them we might find a secondary source that needed treatment also. But it would be critical to get rid of the primary source whether it was the only source or not.

I had done some searching for noise during December using my car radio tuned to 560 kHz. That was below all the Broadcast station frequencies used within a 200 mile radius of Idaho Falls and was essentially a 'clear channel'. With this approach I had found a great many places in and outside of Idaho Falls within a couple miles of my station location that did NOT have noise. I had isolated the noise to an area NW of my station which was about one square mile. The most prominent noise was at a location about 0.3 miles away where there was an intersection that had a stop light. One of the roads leading to that went by the front of my property and tow distribution lines ran along the road off to the intersection where they split into additional distribution line running at right angles to my road. The lines that went past the front of my property continued through the intersection and went another 1.5 miles to various business and a Zoo. There were no businesses on that road -- the businesses and the Zoo were about 0.25 miles off the road and fed by the distribution lines.

A the time I first spoke with Ed I was quite certain that there were several problems because that was what my drive by sampling on 560 indicated.

Ed instructed me to switch to the amateur bands and use 10 or 15 meters in a mobile rig to do the locating rather than the 560 kHz. I had no mobile rig and there was no local person I knew who had an HF rig in the car so I jury rigged a setup using my IC-706 and a mag-mount with a 20m whip. I installed a switched value attenuator in the feed line and then drove around the area. I started with no attenuation and then added attenuation when the noise level increased. I kept the S-meter close to S3 or less. Before long I found I could drive in many areas where there had been 560 kHz noise and there was no noise on 20 meters. I could copy CW signals even with the attenuator in place so I knew that the antenna and radio were functioning correctly. Just to be certain I turned on the AM radio and listened to 560 kHz. Sure enough there was a lot of noise on 560 kHz and none whatsoever on 20 m.

I expanded my range and began to eliminate large segments of area that I had previously suspected. At the end of the first hour or so I had really narrowed down the locations.

After the first pass I went to the Idaho Falls Power company and borrowed their Radar Engineers 240A. This nifty unit operates on self contained batteries and covers the range 1.8 to 1000 MHz. -- all AM. I used this as the receiver and it had an RF gain control that could do the signal attenuation. I repeated the 20 meter drive around and everything was the same as the IC-706 run.

Since I had a 2m-70cm antenna mounted on the car for use with a mobile VHF-UHF radio I decided to try that as the antenna. I could really tell when I was tuned to the 2m or 70 cm band -- the noise went way up.

I drove around again and this time the area grew even more restricted. I was down to one intersection and the roads leading to the intersection. As I went away from the intersection the noise dropped. About 250 feet from the intersection it was essentially not there -- at least at 2m with that antenna.

Now I was having a problem with the route. The speed limit into and out of the intersection was 40 mph. If I timed things right I could end up stopped for the light at the intersection. But if I took to long to leave the intersection because I was trying to make readings, people felt compelled to help me become more socially aware. I finally gave up the driving on week days and did some more driving on Sunday afternoon when the traffic was only modest.

By Sunday afternoon it was clear that I needed to get out of the car and make some measurements on foot. Unfortunately there was only one sidewalk located along one side of one the streets and it stopped just before the intersection. The intersection had an impressive number of poles -- about 12. Happily, four were simply  supports for the stop light and its wires. But the other eight had lots of "stuff" on them.

Running N-S there were two distribution lines, one above the other with a cable system running below them. Running E-W through the intersection there were three distribution lines, one above the other and the cable line running East and located at the lowest level on the poles. At the corners of the intersection the two N-S lines had feeders up to the E-W lines running above them. The top line also had drops to the line below it at the intersection.  Everywhere I looked there was evil looking power line hardware -- all parts ready to emit radio noise on the bands I wanted to use [plus all frequencies in between].

I got out of the car and used the log periodic antenna that came with the Radar Engineers unit. I tune to 250 MHz and then in the vicinity of the intersection started aiming at poles. While I got increases at each of the poles I was surprised to find that as I turned the antenna away from the intersection there was another peak. It was south of the intersection whichis the direction toward my house. I walked back away from the intersection and found a pole that was just as 'hot' or hotter than the ones I had been looking at. I moved the frequency up to 500 MHz and tried it again. Now the south pole seems to be clearly more of a generator than the poles at the intersection. I went back to that pole and went to the one south of it. The signal was there but much less. I went back to the pole and as I got within about 75 feet of the pole the meter pinned -- even with the gain set almost off.

I moved the yagi up and down to see if I could tell it the bottom line or the top line was the ones with the noise. At 500 MHz I could not tell. I moved up  to 800 MHz and tried again. This time it was clear that the major noise was coming from the top line and the lower line might not be contributing any noise.

I took some pictures of the pole and then retreated to my house. During this outdoor walkabout the actual temperature was 28 degrees and the wind was blowing at about 30 mph which made things feel much, much colder than the reasonable 28 degrees.

While drinking hot chocolate at home I got to thinking that I should not tell the power company where the problem was exactly, but rather get them to commit people to me to train by going through the very same process that I did. If I did that the next time (I hope there is not a next time) they can do the searching out in the cold and I can cheer them on from the inside.

It still remains to determine what element exactly is making the noise. The 'unusual' parts are the fuses and i suppose one should suspect them first.