Mitigating Audio Hum

My station wiring is extensive and complex. I try to keep it reasonably well documented and the mass of cables reasonably tidy. However that's difficult to do in practice. The antennas and control systems are constantly evolving, and a lot of my home brew devices are not as pretty as I'd like. In short, it's a mess.

While I do have plans to gradually make improvements there are issues that I must either live with or troubleshoot and solve. One common problem that plagues many stations is largely absent in mine: RFI. This is because most of the towers and antennas are far from the shack. Some, though not all of the control lines have RF chokes and bypass capacitors to shunt RF to ground.

One of my challenges is grounding. I have ground rods at the base of all the towers, the coax and cables are almost all underground until they are near the shack, and I try to follow best practices for grounding where it is possible and practical. It is good enough that a couple of lightning strikes this year did minimal damage. They may have been secondary strikes, but I was either not home or not looking out the windows at the time to say for sure.

The grounding issues have more to do with hum than safety. The connections among RF, AF, control line, PC and equipment grounds are haphazard. For example, operating one of the rotators induces hum in the receive audio of one of the rigs. It is enough to impair reception of weak signals on FT8. Unfortunately it doesn't stop there: hum and related noise can appear on my transmitted audio, whether digital modes or SSB.

It came to a head during this year's sporadic E season. As the season progressed I noticed odd behaviour of the transmitted signal. The audio level drifted up and down and there were momentary spikes on my FT8 signal. I didn't know what it was for sure until a fellow ham sent me a spectrogram of my signal. There were weak spurs across the audio spectrum, and there was little doubt these were coming from my transmitter. In an email exchange we eliminated other possibilities.

Using the rig monitor function I confirmed there were hum and crackling sounds. The levels were well below the FT8 audio so that only stations receiving me strongly would notice. I inspected the spectrogram and noticed that the spurs, though not actual harmonics of 60 Hz, were separated by multiple of 120 Hz. It was time to dig deeper.

The first task was to reproduce the problem. There is no easy way to monitor your own transmit audio and I needed to do that. The rig's monitor function can give you a false sense of security since it monitors the input to the RF transmitter stages and not the output of the transmitter, and certainly not that of an external amplifier.

Sampling the RF requires a separate receiver. Since I am configured for SO2R in contests this was not too difficult to set up. I made use of the PC's two sound cards to transmit using WSJT-X on one rig and to receive on the other using JTDX. A wire stuck in the antenna jack provided a sufficiently attenuated signal that didn't overload the receiver.

I performed two tests at different audio frequencies (DF). One was positioned close to that in the original report. The spectrograms confirm that there is a problem and that it is as reported. There are a few unrelated signals that happen to be within the pass band that are from other sources in the house. For example, the heat pump controller is a little noisy when it is in air conditioning mode. Since there are no spurs at 50.313 and 50.323 MHz it is not worth spending time on it.

Notice that the spurious signals are mostly at intervals of 60 Hz and its harmonics, and especially 120 Hz. They shift along with the DF since I use WSJT-X in "split" mode, which offsets the transmit frequency to keep the FT8 audio within the SSB pass band sweet spot.

It was not only my transmit signal that suffered. There have been several hum issues on receive that have annoyed me for some time. I left those unfixed since they were not serious and I was too busy with station building. Here are a couple of examples:

In the upper pane are hum harmonics that appear when I release the brake on a Tailtwister rotator. Notice that the audio frequencies are odd multiples of 60 Hz: 300, 420, 540, 660, 780, 900, etc. These line up with the hum harmonics on the lower pane. These persistent lines wax and wane with no discernible pattern. They are too weak to interfere with FT8 decoding. The rotator hum does interfere with received signals and also appears on the transmit audio. I was careful not to turn the 6 meter antenna during transmit periods.

Finding and resolving the root cause of the hum is daunting. As already mentioned, my station is extensive and complicated. There are endless possibilities to explore. However, there are several simple things to try before diving deeper. I tried all of the following common cures, without positive effect:

• Checking and cleaning audio connectors, and then other connectors to the PC and rig
• Moving the rotator controller and other equipment, in case transformer proximity is a cause
• Moving power cords to different power bars and outlets
• Testing outlet polarity to ensure that neutral and ground pins are correctly wired
• Disconnecting and isolating equipment, including microphones, SM5000 monitor, headphones and PTT lines
  
• Adding and improving ground busses between equipment, and the service ground
  

Despite those failed measure, there were indicators of where the problem might reside. Disconnecting the transmit audio cable from the PC the problem could be replicated by "shorting" with cable with my fingers. This suggests that the hum source could be the rig itself (FTdx5000) since my skin resistance coupled the ground to the signal pin. 

Another indicator was that it was possible to alter the amplitude of the received hum by jiggling the audio connections to the PC. But, frustratingly, not always or not by a lot. This suggests that the PC power supply or internal wiring has degraded, possibly oxidation at a ground connection or a leaky capacitor. The effects are inconsistent and not persistent.

I disconnected the many interconnection cables and hauled the the FTdx5000 onto the workbench and opened it up. I poked around the connectors and audio stages to look for anything amiss. All the connections seemed secure and checks with an ohmmeter uncovered no obvious defects. I put it back together and reconnected the cable. It was no surprise that the hum was still there.

The next candidate was the computer. Modern PCs are notorious for poor and defective power supplies and I've had past trouble with them in my commercial signal processing work. Here I balked since time was running and the sporadic E season was at its peak on 6 meters. Perhaps I'll pursue this during winter when I have more time for indoor jobs. I instead chose to pursue mitigation rather than a cure.

I rooted through my junk box and located a miniature audio transformer. I mounted it and connectors on a small breadboard. I put it in series with the transmit audio line (PC headphone jack and rig audio in jack). It was a struggle to fit it under the fan-out PTT connector without wires getting pulled out of the breadboard.

It worked. My transmit audio was free of hum and all the snaps, crackles and pops. I had two friends check my signal, to be absolutely certain. They gave me a clean bill of health. 

I operated in this fashion for the remainder of the sporadic E season, confident that my signal was finally clean despite my failure to locate the root cause. Hum on the received signal continued, but I could live with that.

The transformer works because the direct ground connection between the PC and rig is broken. Common mode hum cannot cross the transformer. It is for the same reason we use common mode chokes on transmission lines. For small signal levels, such as for audio signals and for my Beverage receive antennas, two-winding transformers are very effective. However, if the hum is differential mode -- on the audio signal and not just conducted over the ground connection -- the isolation transformer won't help.

I needed a something better to replace the flimsy breadboard, and for both the receive and transmit audio lines. Less than confident that I could find the root cause of the hum without a great deal of effort I opted for a permanent mitigation measure. I chose to build a transformer-based isolator for both audio lines. Since I had only the one audio transformer I went online to shop for parts.

My original intent was to order more of these miniature audio transformers, which are very inexpensive. Sites like Amazon have good deals on packs of 10. It doesn't hurt to have a few extras in the junk box. Flea markets are also good sources but those have been few and far between during the pandemic. I already had an enclosure and connectors, and I chopped a piece from a small prototype board to fit inside. The enclosure should be plastic unless the audio connector grounds can be isolated from a metal enclosure. My remote Beverage switch shows one way in which this can be done.

In the list of items from the Amazon search results there were many commercial audio isolators. One or two of those (they're almost all for stereo lines) will work if wired for mono operation. Although they cost more than building an isolator from discrete components many might prefer those.

Further down the list there was a fully populated PCB with transformers (for stereo). Since it would save me the trouble of soldering small parts I decided that for $11 it was worth taking a chance. When it eventually it arrived from China I installed it in my enclosure with the connectors.

All the connectors are stereo, as they typically must be for the PC and rig jacks. I bridged the two channels on all the jacks for mono operation. Thus one side of the isolator is for receive and the other for transmit. Grounds on each side of the PCB are common so one side is for the PC and the other for the rig. Otherwise the unit is symmetric and there is no need to differentiate which side is which. The wires support the board inside the enclosure.

Initial tests with the front headphone jack were promising other than signal attenuation. The 8 Ω impedance is likely a poor match for a transformer designed for high impedance circuits. Signal levels were good with the FTdx5000's high impedance back panel jacks. PC sound cards are even higher impedance, typically 47 kΩ. Frequency response appeared to be reasonably flat from 100 Hz to 3000 Hz. I did not test the unit beyond that range.

One point worth noting is that the isolator is DC coupled. That is, there are no capacitors in series with the 4 transformer windings. Sound card microphone ports have a permanent or software configurable DC bias for electret mics. Depending on the transformer this may pose a problem. These miniature transformers typically have a DC resistance in the range of 100 to 200 Ω. It is unlikely that more than 10 ma of DC will flow due to the electret mic bias and that is unlikely to be a problem. I recommend checking the winding resistances to be sure.

I made a 2" (5 cm) cable from twisted Cat5 wire pair to temporarily connect isolator to the phono jack on the rig. The other jack was wired with a commercial 3.5 mm stereo cable. I plan to make short shielded cables for both rig ports to keep the installation tidy and hum free. Short cables are better than long ones.

FT8 testing was successful. Levels were good and there was no evidence of hum on transmit. There is still hum was on receive when I operate the rotator brake, but it is much weaker than before. It may be due to the short unshielded cable or the plastic enclosure.

With that job done I declared the problem solved. Some cleaning up is required and I will do that in the not too distant future. It is unlikely that I'll use FT8 until the winter 160 meter season. 

Many modern rigs reduce the cable count by combining CAT, audio and other functions on a single USB cable. That reduces the opportunity for noise and hum to sneak into the audio paths. I hope that will be the case when I eventually upgrade the shack equipment, and I can set aside the isolator. If the problem source is the PC that may not work. On the other hand, since PCs age faster than rigs it is probable that by then I'll have a new computer running the shack.

Finally, I strongly recommend that you monitor your RF or ask a friend to do so. Use the buddy system to help each other to keep your signals clean. Most hams are too shy or polite to tell you, so don't rely on unsolicited reports. Not only that, even if you are don't know I can guarantee that most hams in the vicinity know that your signal is dirty. They are likely miffed and silently cursing you for your negligence. Should someone reach out with a negative report, thank them and move quickly to confirm and correct the problem.