It is surprising to me how many times I have seen this done incorrectly in the field. Summing a stereo source, whether it is balanced or unbalanced is not simply twisting a couple of wires together. This will effectively reduce the impedance of the outputs by one half. With newer, active balanced outputs, this may cause damage to the output amplifiers.
The parallel resistance formula is thus:
Therefore a 600 ohm stereo output tied together would look like:
Rt = 1/(1/600+1/600) or 300 ohms.
It also creates an impedance mismatch with the next piece of gear, which will effect the common mode noise rejection of the circuit.
The best way to sum is through a resistive network. That way stereo separation is maintained, the impedance of the output circuits is maintained and the output amplifier will not current cycle. That looks like this:
resistive summing network
Pretty easy to fabricate in the field. It is good to do things the right way, it sounds better on the air too.
We are starting to work at a new client’s studios. It is a bit like stepping into a 1980’s time machine, as the newest console seems to be the Broadcast Audio console in the FM studio. I feel I should wear a wide colorful tie and part my hair in the middle when working there. There is also an older UMC console in the second production room.
A what?
Exactly.
It seems the UMC console (UMC was a Connecticut based console manufacturer that was later sold to Broadcast Audio) was having an intermittent hum problem on all the audio buses.
After poking around under the hood for a few minutes, I decided I should begin with the basics. Checking the power supply for ripple seemed like as good a place to start as any. This console has a 30 volt and a 12 volt power supply. The 30 volt supply checked out good, the 12 volt supply, not so much:
12 volts DC, 2.7 volts AC
12 Volt power supply
2.7 volts AC on the 12 volt DC power supply. That will put some hum on the audio, all right. I tried to replace the power supply main filter capacitor, but it had no effect. The regulator must also be bad and it is a Motorola part number which is likely not made anymore.
12 volt linear power supply
This is a pretty standard off the shelf power supply, I should be able to get one from Mouser for about $60.00 or so for a linear unit, which will be cheaper than us trying to trouble shoot and repair the old one. In the meantime, I took the 10 amp 0-30 volt bench supply and pressed it into temporary service. The console is working again, for now.
At some point, all this old, um, stuff needs to be replaced.
Audio Engineers will know this subject well. Grounding has many purposes, including electrical safety, lightning protection, RF shielding and audio noise mitigation. Although all types of grounds are related in that they are designed to conduct stray electrons to a safe place to be dissipated, the designs of each type are somewhat different. What might be an excellent audio ground may not be the best lightning ground and vice versa. Sometimes good audio grounds can lead to stray RF pickup.
The basic ground loop looks something like this:
Ground Loop schematic
Where RG should equal zero, in this representation it is some other resistance. This causes a different potential on the circuit (V1), which in turn causes current to flow (I1). It is that unexpected flow of current that creates the problems, causing voltage (V2) to be induced on another part of the circuit. In cabling applications, this will result in a loud, usually 60 cycle hum impressed on the audio or video being transmitted through the cable.
The resistance can come from something as mundane as the length of the conductor going to ground. This can often happen when using shielded audio wire in installations when the connected equipment is already grounded through the electrical plug.
There are two proven methods for eliminating ground loops, both of which are best implemented in the design phase of construction (aren’t most things).
Radio Station Common Point Grounding
The first is a single ground point topology, also known as a common point or star grounding system. A common ground system consists of one grounding point or buss bonded together so that it has the same potential. All grounded equipment is then connected to that point creating a single path to ground. All modern electrical equipment has a path to ground via the third prong of its electrical cord. Problems can or will occur when audio equipment is plugged into separate AC circuits, grounded via the electrical plug and then tied together via an audio ground. The longer the separate grounding paths, the more severe 60 cycle (or some harmonic thereof) hum can result.
To eliminate this problem, the shields should be broken at one end of the audio cable. Never cut the third prong off of an electric cord, which can create another problem called electrocution. Given the choice between a ground loop and electrocution, I’d stay away from electrocution, mine or somebody else’s.
For installations in high RF fields, the open shield or ground drain can act like an antenna. In those situations, the open end can be bypassed to ground using a 0.01 uf ceramic disk capacitor. Electrically, this will look like an open at DC or 60 cycles, but allow stray RF a path to ground. This problem can be a common occurrence when studios are co-located with transmitters.
Differential Signaling
The second is by using balanced audio or differential signals as much as possible. This poses a problem for those stations that use consumer grade components, especially in high RF fields. For shorter cable lengths, two or three feet, it is usually not a problem. Anything beyond that however, and trouble awaits.
It is relatively easy and inexpensive to convert audio from unbalanced to balanced. As much as possible, equipment and sound cards that have balanced audio inputs and outputs should be used. In the end, it will simply sound better to use higher quality equipment. Also, longer cable runs need to be properly terminated at both ends.
Installing equipment using good engineering principles and techniques will eliminate these problems before they start.
Old world and all that. I am of the impression that the European broadcast engineers are a more studied lot. There process involves much more deliberation, thoughtful analysis and planing than ours does. For example, when it comes to station loudness, most programmers and many engineers (myself not included) to the more is better. It is thus that we get the Omina 11 and other audio squashers.
EBU R128 (ed: Loudness Recommendation) is the result of two years of intense work by the audio experts in the EBU PLOUD Group
Aside from the above mentioned EBU R128, there are four technical papers dealing with implementation, meters, distribution and so on. The body of work is a recommendation not a requirement. I can’t imagine voluntary implementation of something like this in the US. Even so, there are advantages to having a single acceptable level of programming audio. It is interesting reading.
