ARRL

Noise Logging

A High Accuracy Method Of Logging Noise Occurrances

 

A High Accuracy Method of Logging Noise Occurrences
By Alan Robinson, W4LKE
  In view of the fact that other hams might want to use my method of recording noise, I think I should give you some more details about how I do it.

Equipment used: Icom 756PRO Receiver with Astron RS-35M Power Supply, Icom PW1 Amplifier, G5RV antenna, Radio Shack Model 22-805 Digital Multi-meter, and a PC compatible computer running Windows 98 SE. The PW1 is used, in this case, only as an antenna tuner. The Radio Shack 22-805 is connected to one of my serial ports. I have an 8 ohm non-inductive resistor plugged into the external speaker jack to replace my external speaker. The multi-meter is set to measure the voltage, in millivolts, across this resistor. Actually one can just connect the multi-meter across a jack plugged into the external speaker jack, but I prefer to have the resistor replace the speaker.

Equipment settings: I use the ScopeView feature of the Radio Shack software, not the logging feature. The ScopeView software has exactly all the features I need. The Icom 756PRO is set on AM with the AGC set to off. This is very important. With the AGC off the audio output voltage, in millivolts, across the 8 ohm resistor is directly proportional to the voltage, in microvolts, at the antenna jack of the 756PRO. This does disable the S meter. I set the receiver to 29.291 MHz. This was determined by doing a sweep from 1.8 MHz to 50 MHZ to see where the noise was the loudest. I got strong signals at many places across this frequency range, but decided that the 10 meter band would be a good place to record the noise. It is also important to pick a frequency NOT often used by hams. If a station comes on the frequency you are monitoring the data is of no value. I have found that there is very little activity at 29.291 MHz. The 756PRO also has a spectrum scope that I set to +/- 100 kHz to see if there are close by signals that might contaminate my data. The voltages recorded are, of course, dependent on the audio level. I set the volume control to produce 5 mV with no antenna connected. With that setting and a very quiet noise level the voltage level is about 10 to 12 mV with the antenna connected. An S9 noise signal reads about 200 mV. I set the ScopeView software to 40 mV per division (this gives a maximum of 320mV). I use the 1 minute sampling rate, and repetitive sweep. This will give three screens in a 24 hour period. This can be mounted on one 8 1/2 x 11 sheet of paper. Each day I stop recording and print the "scope" screens. Actually, when the 1 minute sample rate is used it is really a recording rate. The Radio Shack multi-meter samples at 1 second. So, if a peak noise occurs somewhere during the 1 minute recording, the peak will also be recorded and show at the right side of the ScopeView screen. I have also made a number of "calibration" runs. I set the sample rate of ScopeView to 1 second. Then, when the noise level is high, I momentarily turn the AGC on to get the S meter reading. Then turn the AGC off again and let ScopeView record for 30 seconds. This is one large division in ScopeView. Next I turn on 6 dB of attenuation in the 756PRO. Turn the AGC on to get the new S meter reading. Then turn the AGC off and collect data for another 30 seconds. This is repeated with 12 and 18 dB of attenuation. Note: the S meter on my receiver is NOT calibrated to 1 S unit for 6 dB. When I show the data to the utility I also give them the calibration plots so that the millivolt reading correspond to S meter noise levels. Everyone agrees that S9, or especially 20 dB over S9, is not acceptable and must be fixed.

Results from the recording system: I have recorded the noise at various times over the last four months. Recently I have been recording it continuously. During times when the utility has been working on the lines I have used the 1 second sampling rate in ScopeView. On one occasion the utility shut down the 69 kV transmission that runs 100 feet North of our house. I and another ham were recording the noise. The instant that the voltage was dropped from the line our noise dropped from S7 to S1. The recording system showed the exact times that the 69kV was off. Very convincing evidence!! On another occasion the utility opened a cutout on the under built 7.5kV distribution line. Again I used the 1 second sample rate in ScopeView. The noise quit when this cutout was opened. The lineman then cycled the cutout several times. The data showed the exact times for each cycling of the cutout as it was opened and closed. Again very convincing! The cutout was replaced. OK. So why are we still working on it? Well we haven't found all the sources yet. I have been using Marv Loftness' method and have identified poles that are sparking. I have recorded their sparking patterns and printed them out. During one particularly noisy time I recorded the sparking pattern at my station on 52 MHz. Then I went to the "problem pole" and recorded the same sparking pattern from that pole on 440 MHz with a six element beam. I gave copies of these plots to the utility. Again very convincing! I also prepared a map of the power line in our area with all the poles identified. I used a GPS receiver to get the pole positions. Also I obtained the position of W7PUA's 2 meter antenna. During several noise measurements Bob Larkin, W7PUA, used his highly directional array to take bearings toward the high noise area. I then plotted those on the same map. The bearing toward the greatest noise pointed directly at the problem pole in the previous paragraph. The map was presented to the utility. In all I gave the utility data covering nearly three continuous weeks and selected data before that, including the time they shut down the 69 kV line. I also gave them calibration curves, maps with bearing and pole positions, pictures of my spectrum scope, audio recording of the noise during S9 conditions, and the scope pictures of the sparking patterns.

Summary: I used this method for about four months now, refining it as I go along. I have been leaving the system on all the time and getting 24/7 recordings. Two days ago I gave the utility 2 1/2 weeks of continuous plots, with calibration curves. There is no doubt as to when the noise is "on" and when it is not. In our case hard rain wipes out the noise.

Comment: The utility knows the interference is their power lines. They are now in the process of making a plan to cure it. I hope this helps other hams. If I can be of any assistance please let me know. Email is best, but phone is OK too. Sincerely, and 73s, Alan, W4LKE Alan Robinson - Corvallis, OR