Cables and Wiring

Home ] Up ]

Audio, Power, and Control Lines

RF Feedback, Hum, and Distortion

Improperly connected or configured audio and power supply lines can cause hum or distortion in audio systems. This includes microphones, headsets, computer sound cards, and nearly any other low level audio system or device.

RF feedback is characterized by anything from a little distortion causing a slightly rough sound on transmitted audio, to highly-distorted, unintelligible, raspy bursts of audio. The gateway for RF feedback is always an equipment design limitation.

RF feedback, and even hum and noise ingress, requires a few things. First, the equipment must have some sort of shielding or wiring error, inside or outside the equipment. The wiring error or imperfect shielding allows RF feedback, but does not actually cause RF feedback. Less-than-perfect shielding and wiring technique problems are called to our attention by excessive common mode RF currents, or by exceptionally strong direct fields from nearby transmitting antennas. The problems are not actually created by the excessive current or field levels, but by a mix of design problems. Since problems are level sensitive, anything that increases electric or magnetic field levels, voltage, current, or power can highlight wiring and equipment shortfalls or problems. The true cause is less-than-perfect shielding and/or wiring in the presence of a problematic field level, current, or voltage. An ideal design has a reasonable compromise between equipment design time and cost, and allowable common-mode RF current levels and/or local field intensity levels.

Audio systems and RF systems behave the same in many ways, and behave quite differently in other ways. For example, at audio frequencies shield dc resistance is critical. At very high radio frequencies, gaps in braid and the angle of braid weave, control leakage. At very low frequencies shield resistance and material can be critical, while small gaps or weave angle might be meaningless. At radio frequencies, aluminum works as well as (and often much better than) steel for shielding magnetic fields. At audio frequencies, steel might be required to stop magnetic flux leakage.

At radio frequencies, even a very thin foil can be many skin depths thick. This is because the shield is not thick compared to skin depth at low frequencies. Nothing passes through the shield, neither magnetic nor electric fields. This is true even if the shield is thin aluminum foil, and braid thickness or shield thickness is nearly meaningless.

To be most reliable with reasonable equipment cost and wiring complexity, designers must understand balanced systems, unbalanced systems, shields, audio, and RF system behavior.

SSB transmitters draw a very large portion of dc power supply current in direct proportion to audio input levels. The dc power supply negative ground lead can modulate a low-level audio line with the rough "brah waah braah" sound of SSB without a carrier. Many "RF-sounding" problems are actually caused by low frequency ac or dc ground loops from power supply negative leads, acting in concert with other grounding, wiring, or shielding shortfalls.

Treating Audio Lines Correctly

In the early 1970's through 80's, I worked at various broadcast stations. Much of the work was wiring studios and switching racks, which employed hundreds of line-level audio feeds. One very basic rule was followed by better, more-experienced audio and broadcast engineers. Never ground a shield at both ends of signal paths between different pieces of equipment. An audio shield, or any ground return lead with low-level audio, should only be grounded at the audio device's input port, or at one end of a path. This prevents inducting noise and hum from ground loops between device cabinets, which are almost always at slightly different AC potential. This rule is often ignored by amateur radio operators and amateur radio equipment manufacturers.

ground loop audio

Power Line induced hum

Harmful ground loops are caused by grounding sensitive leads or systems to different points. In this example current returning to the pole transformer causes a voltage drop along CABL1, from the pole to the house.

If we ground a low level audio source, like V1, to any external ground some AC current from the mains neutral will partially flow through shield resistance (R1) and modulate source V1 with hum.

Vac represents the AC loop caused by voltage drop in the power mains neutral system.

The ONLY correct place for an audio ground is at the microphone connector entrance to the shield enclosure of the radio. That ground has to be virtually zero resistance to the cabinet. Nothing else on the audio line should be grounded, unless it is grounded to that jack-to-chassis connection.





power supply ground loop



 Radios using mains powered 13.8 VDC power supplies suffer all the same ill effects from power mains hum, plus they have a new problem. The small resistance (R4) of the negative lead allows the power supply to modulate audio lines at a rate tied to the SSB envelope. This produces and audio sound much like RFI or SSB with a BFO turned off. This voltage is represented by Vdc.



Again the fault occurs because of the extra unnecessary ground at the audio source (V1).

Vac can occur along with Vdc, and usually does.


Once again the only place the audio line should be grounded is at the entrance to the radio housing.




Computer to Radio Interfaces

While we can on occasion get lucky and get away with breaking good engineering practice rules, we should use isolation transformers on audio lines. Many trashy PSK signals are created because users don't isolate the grounds, and often times this trash is out of the receiver's bandpass. Because the hum and noise is off-frequency from desired tones, many problems go unnoticed (except by people up or down the band who have noise or hum pop on their frequency when someone up or down the band transmits).

computer to radio isolation

T1 and T2 isolate the ground paths on the computer and the ground paths on the radio cabinet or chassis. Without T1 and T2 the resistance of the shields would drop voltage, and that voltage would effectively be in series with the audio inputs and outputs.

This can result in hum, or the dc power amplifier current of a power amplifier using an external supply modulating the audio system.

T1 and T2 can be small audio transformers of 150-1000 ohms impedance.





audio line isolation transformers

This is my house operating desk audio system. The transformers are normally behind the radios out of sight.

Pay particular attention to the nylon insulator system on my CW key jack. This is to prevent the CW keyer, which operates on 13.8 volts dc as well as connecting to the radio, from grounding to the headphone jack ground. If I short the key jack shell to the metal panel where audio lines are "grounded", hum appears in the headset.

The green audio line transformers are 42TU013-RC They are 8-ohm CT to 1000-ohm CT 1/2-watt transistor output transformers. I run the radio speaker audio output into the full eight-ohm winding. I use half of the 1000-ohm winding to drive the audio switching and buss system. The audio buss (gray cable) is terminated in  47-ohm resistors for each of the four audio channels.

The computer interface connects through the red transformers, which are 500:500 ohm Xicon 42TU500 transformers.






Audio bridging switching system

This is my basic headphone wiring in both my contest barn and house shack. It can be built out to handle as many radios as necessary, and will interface in stereo. It allows any radio to be on any ear, and multiple switch and jack units can be bridged.

The buss in my contest barn covers 4 operating tables in stereo.

The left half of the schematic drives the audio buss, the right half bridges the audio buss. With a low source impedance (like the speaker output of a radio) driving the buss, plus the intentionally heavy buss loading resistances R1 through R3(+), volume changes are minimal as headphones are switched in or out on any given radio.



Special note on T1-T3(+) . As many transformers can be added as the S1 and S2 will accommodate. The windings are center tapped. I use the full 8-ohm winding for the input from radios, and only half of the secondary to drive the audio buss. This makes the impedance ratio around 4 : 125 ohms. I load the secondary intentionally slightly heavy with no ill effects. The extra winding impedance extends low frequency response and limits the highs slightly, to slightly reduce hiss in headphones when using narrow CW filters.


Audio transformer test for you!

If I have an 1000-ohm CT transformer and I only use one-half of the winding, from one end to the center tap, what is the new winding impedance? 

Most people intuitively answer 500 ohms.


The impedance is 250 ohms! This is because impedance changes by the square of the turns ratio. The center tap to one end is half of the turns, so it has 1/4th the impedance.

An 8-ohm center tapped winding is not 4 ohms each side of center tap, it is 2 ohms each side of center tap. Double the turns ratio and we have four times the impedance. Halve it, and we have 1/4 the impedance.



More on Transformer Impedance

When a transformer has too little impedance low frequency response suffers. It also saturates easier.

When the transformer impedance is too high, high frequency response suffers.

It is the load that determines the "match" for transformer impedance, not the source. If we want a 600-ohm transformer to be flattest running into a 10K ohm radio input , we should connect a 680-ohm resistor across the secondary (load) side going into the radio. The 10,000 ohm radio can then bridge directly across the loaded secondary. If we needed to match the 10,000 ohms because the input of the radio is fussy (this is normally not an issue when a high-Z load bridges a low-Z source) we could use an L-pad. A suitable L-pad would be a 680 ohm resistor to ground, and a 10,000 ohm series resistor to the high impedance load. Power loss would be almost 20-dB.

As for the transformer going to the computer, most computers should have a line input impedance or be designed to bridge the source without issues so long as the source Z is less than load Z presented by the computer. If we wanted to get fussy we could load the secondary and bridge across it with the computer.

I very rarely find it necessary, unless I want high fidelity or maximum power transfer, to worry about getting an optimum match. All of my radios go into an 8-ohm to 250-ohm audio transformer to drive the audio busses in the shack. This transformer is an 8- to 1000-ohm transformer, and only use half of the 1000-ohm winding. There are separate isolation transformers on each piece of gear at the outputs, including the computer. Shields are ONLY grounded at inputs. Without this I have audio line noise and hum.

Wiring Blocks

To facilitate moving things around on my desks, I use movable wiring blocks. This minimizes the number of individual wires or cables running all over the room. In this example I have several power plugs connected to one supply line (fused at the source). The CAT5 cables connected to other items on the shelf, like antennas switches.

Every pin number, color code, and functions gets documented on a list!

portable wiring block













portable wiring block for antenna control functions






















 Page Under Construction









RF Ingress and Shields

Most shields