The set of pictures below was taken in October of 2004. I used an HP141/HP8553B spectrum analyzer to look at the signal from my TH7DXX as I rotated the beam. The scan rate is such that the spikes are about 16 msec apart (60 Hz).
The center of the display is at 21.25 MHz.
The power line that was causing the problem runs North-South. The problem was isolated to a 'corner' where the lines made a junction with a new line running East-West. The corner was SSW from my location. The signal samples were made at time of day when signals on the band were minimum (band was closed). The noise spikes were more than 25 dB above the noise background.
This particular 'fault' was one that I used to train the local power company. I had been asked to teach them how to use the new equipment that they had purchased to locate power line noise. I agreed to do so for a fee of $1 and the right to borrow their equipment in the future. We chased down this problem in the middle of a snowstorm and the guys they had assigned to the job were quite impressed because we were able to get the same signature on their equipment as I was seeing and then they were able to find not only the pole involved but the faulty junction itself from the ground. The crew came out and did the repair and the noise from this source is virtually non-existent at this point.
Figure 1 below is looking SW.
Figure 2 below is looking West
Figure 3 below shows the signals looking NW.
Figure 4 below shows the view looking North.