Fig. 1: A communications receiver tuned to 1440 kHz.
It reads like a Sherlock Holmes mystery novel.
Tim Laes, corporate director of engineering for Midwest Communications, received occasional listener complaints about an unusual noise heard on WNFL(AM) in Green Bay, Wis., in November and December 2013. WNFL is a DA2 with 5 kW day and 500 watts night into three towers on 1440 kHz.
However, AM radio constantly battles noise from fluorescent lights and a hundred other sources, and listeners often don’t know how to explain what they are hearing, even when it is a valid complaint.
Laes is a ham, call W9EE, so he got off to a good start in unraveling this mystery. Hams often think farther out of the box than most people, and Laes discussed the issue with both local engineers and other hams.
The WNFL signal in Green Bay was clean, so it was not a studio or transmitter problem. Then Laes determined that most listener complaints were coming from south of town. He lives north of Green Bay but was able to hear the interference, just barely, using a ham receiver and listening in the quiet environment of his ham shack. There it was, something like Morse code at about 15 words per minute, but it didn’t decode into letters and words. Instead, it was a 1 kHz tone of random dits and dahs, as heard in the AM mode of his receiver. That would be A2 modulated CW, which is almost never used.
(Hear a one-minute sample of the audio online at http://mwpersons.com/sounds/12-16-13.mp3.)
Was it a pirate station, deliberate interference or a switching power supply gone wild?
The plot thickens.
Driving tests confirmed that the interference became more audible as he traveled south. There didn’t seem to be a definite source, and it was nearly impossible to track because KFIZ broadcast on 1450 kHz in that area. It causes normal adjacent-channel modulation interference, since it is only 10 kHz up from WNFL. To make matters worse, the problem was only intermittently heard on the air.
In the spirit of cooperation, broadcast engineer Steve Konopka, N9FOY, got involved. He contacted Steve Brown, W9APL, chief engineer for Woodward Commutations in Appleton. Konopka had two radios going at once and noticed that the Morse-like sound seemed to happen in cadence with programming on 1150 kHz WHBY, some 24 miles at 171 degrees away from WNFL. Konopka listened at home while Brown briefly shut off his 25 kW transmitter on 1150 kHz and then on his 1 kW 1570 kHz station. Brown had recently installed a new STL system; it was interrupted as well.
All tests showed Woodward was not the problem.
Laes and Konopka decided to do some tuning and discovered the same tone was audible 9 kHz above and below 1450 kHz. They could only hear the unwanted signals by tuning receivers up and down in frequency to get the unknown signal inside the 3 kHz bandwidth of their receivers: KFIZ(AM), in Fond Du Lac, Wis., is 1 kW day and night into a single tower on 1450 kHz, a local channel. It is some 53 miles down the road, at 206 degrees from WNFL.
Now Laes and Konopka contacted Stu Muck, who does engineering for KFIZ and was in the process of replacing large electrolytic capacitors in the station’s Gates One transmitter.
Fig. 2: 4.5 kHz Bandwidth on NRSC Test Muck noticed that the transmitter’s occupied bandwidth was, as he said, “starting to work its way out of the NRSC mask.” Unfortunately, that repair did not fix the 9 kHz problem. The strange sound went away when he disconnected audio from the input on the station’s Orban 9100 audio processor. Muck discovered the noise issue was present only when program audio was passing through the processor. Further analysis showed the processor’s number four card was in oscillation at about 9 kHz, but only when there was programming audio present. Who’d ever heard of that before?
Remember how the 9 kHz tone seemed to come and go with modulation on WHBY? Well, that is because KFIZ was running the same nationally syndicated talk show as WHBY while some of the listening tests were being performed. Talk about confusing the detectives with a red herring!
So, the 9 kHz oscillations on 1450 kHz were heard as a Morse code-like 1 kHz note on receivers tuned to 1440 kHz, because it was within the receiver bandwidth when tuned to 1440 kHz. There is no nearby station on 1460 kHz to be interfered with, otherwise that would have been another clue.
Under current FCC rules, any station could legally transmit a 9 kHz tone at full 100-percent modulation and cause problems for RF neighbors. Muck does not have a record of how high the unwanted 9 kHz modulation was, but not much is required for people to take note when a steady tone, or worse yet a Morse code-esque tone, appears at 1 kHz, mixed with audio intentionally broadcast.
No complaints were received from KFIZ listeners. Why? Well, receiver manufacturers reacted to customer noise complaints years ago by restricting audio bandwidth in AM receivers to about 3 or 4 kHz. Very few AM receivers today can reproduce audio to the full NRSC 10 kHz audio bandwidth.
SO MUCH FOR 10 kHz
I did frequency response tests in my shop on a few AM receivers. They were a Sherwood RX4109 component stereo, a Sony CFD-10 boom box and a Bose AWR1-1W Wave Radio.
The test setup featured an AM stereo generator, set to monaural, with an NRSC pre-emphasis equalizer and an NRSC low-pass filter between an audio generator and the exciter. No dynamic audio processing occurred. The exciter has a sample transmitter output, which produced enough signal for receivers to hear it on the bench. Receiver tone controls were set for standard response with no bass or treble boost.
Test results were what we expected, or at least suspected. The surprising part is that the receivers had very similar response curves averaging 2 dB down at 3 kHz, 7 dB down at 4 kHz, 12 dB down at 5 kHz, 21 dB down at 7 kHz, and 32 dB down at 9 kHz. I understand that car radios are about the same. So much for 10 kHz audio!
On a side note, I recently sold an Orban 9300 AM audio processor to a client. Before it was installed, I optioned the software to allow only 4.5 kHz of audio bandwidth. The station sounded loud, in part because there was no audio power transmitted and wasted from 4.5 to 9.5 kHz. No complaints, just happy listeners who could hear that talk station better. Fig. 2 is a spectrum analyzer display of 4.5 kHz AM audio bandwidth.
Years ago, I remember troubleshooting a pure 1 kHz tone that listeners heard on an AM station. The cause was a second adjacent channel (20 kHz away) AM station that was rebroadcasting a stereo FM station using a monaural FM receiver. The 19 kHz stereo pilot from the FM went right through the FM receiver, audio processing, transmitter, and out the antenna. That was before NRSC 9.5 kHz low-pass filters were required. The 19 kHz tone was just 1 kHz from the station being interfered with and was plainly audible to many listeners in an area halfway between the two station locations.
I’d like to make it clear that the 9 kHz and 19 kHz signals that caused 1 kHz tones in receivers were not mixing products between stations. They were “signals on the band” heard by receivers.
Yes, some AM transmitters are capable of modulating to even 20 kHz, but the receivers are not there today to take advantage of it. With increasing noise levels on the AM band, I do not advocate increasing AM audio bandwidth. In my opinion, we must make the best of what we have to get viable programming to listeners. It makes perfect sense.
Comment on this or any article. Write firstname.lastname@example.org.
Mark Persons, WØMH, is a Certified Professional Broadcast Engineer who has more than 30 years’ experience. His website is www.mwpersons.com. Find past articles under the Tech Tips tab at radioworld.com.