Grounding and Rack Installations
In the data and broadcast world, virtually all equipment is mounted in upright metal cabinets or racks. By doing so, it is easy to establish safety ground and mechanically stable construction. When all equipments are mounted on rack cabinets which have grounded with a large wire running to a central ground point it is easy to establish safety ground.
Second, and most likely, conduit running in and out of therack will be the mains power cord. Virtuallyall professional audio and video equipments have a grounded (three-pin inUSA) power cord. That the gounding connection in power plug is connected inside that piece of equipment to the metal box or chassis of the equipment thus grounding the box and reducing the possibility of shock hazard to the user.
There are also other ground circuits going into and out of the equipments and equipment rack. If shielded signal cables are used for input or output (as almost always is), the shield of the cable is grounded to the metal box of that piece of equipment where the cable is connected.
Generally in the data world, shielded data cables are often attached to both the transmitting box and the receiving box. In video circuits, likewise, the ground is connected at both the source and the destination. In audio circuits, as well, shields are attached at both ends of each cable. (there are few special excpetions to those rules but they are quite rare).
Where the problems come
When the video and audio cables are wired between the equipments in the same rack cabinet, there is not problems to be expected because all the equipments are tightly grounded together.
But problems can be expected as audio, video ordata cable could be attached between two pieces of equipment in separate rooms, separate floors or even separate buildings. If the ground connection is good at the both of the ends of the cable and there is any voltage difference between those two grouding points where the equioments on the ends of the cables are connected, significant voltage current travel down the shields of the cables. This current flowing on the cable shields can induce humming noise to the signals inside the cable and inside the equipments connected to the cable.
You can see up to many volts of voltege on between the cable ends and up to amperes of current flowing on the shield which should only have micro volts, or millivolts of noise.
Even in buildings with excellent ground systems, the distances run can be so great that the natural resistance of the copper wire and metallic conduit materials guarantees that there will be voltage differences from one part of the building to another. Running a shielded cable between the ends virtually guarantees a ground loop and, there, hum or interference of some kind.
The effect of this huge amount of noise varies. In video circuits, it show up as "hum bars", buzzing slow moving bars which wipe across the screen diagonally. In audio, it can be heard as hum, buzzes or other constant interference. In data circuits, it can cause bit-error-rates which can shut down a system where nothing else is actually wrong.
Quick fixes for problems
Quick fixes are available for all these problems which attempt to eliminate the problem without addressing the actual problem of poor grounding.
In the video world, one can purchase a "hum-bucker". This is a torroidal transformer wound to reject 50 Hz or 60 Hz(the frequen- cy of the power line and most prominent frequency in the noise of a ground loop.). They are generally quite good in rejecting the mains induced frequency, but are generally a little bit bulky (somewhat large and heavy to be put in the middle of the cable connection), expensive (usually $100-200 per line) and hard to find.
For audio wiring, the insertion of an isolation transformer provides galvanic isolation which prevents ground loops. Transformers are passive devices which use two coils of wire wound on one core made of magnetic material (like iron or ferrite). While the signal can cross between the two coils magnetically, there is no physical contact. Thus, the ground loop is broken. Adding isolation transformer to each audio line is usually expensive and can make the sound quality somewhat worse than straight wire.
In the data world, they avoid the problem altogether by using unshielded twisted pairs (UTP). This kind of wires are isolated from one or both of the ends of the connection. This has proved to be a wholly acceptable solution and works well in practice. Most of the modern office LAN cable wirings are nowadays done using unshielded twisted pair wiring (and rest of them most often using shielded twisted pair wiring).
However, for the audio and videoworld, unshielded cables are not an option for any serious high quality signal transmission. There are some adapters available to transfer audio and video using unshielded twisted pair wiring, but those adapters are generally only suitable for not demanding dapplications like security cameras (CCTV) and similar applications where the absolutely best signal quality need not be preserved. In any case going from normal video or audio connections to unshielded cable needs special converters which cost money and there is no guarantee that such converters will solve the initial grounding caused noise problems (they are just designed to be wiring media converters, not groundloop problem stoppers).
Analog audio and video designers do not have an option of using unshielded twisted pair wiring as general media for signal transferring and must use shielded wire for optimum performance. The only option then , is to eliminate the difference in potential between any points wired with cable.
Star grounding solution
There are a number of ground schemes intended to eliminate ground loops. The most common is called a star ground. In astar ground, a point is chosen as the lowest potential. There are conductors radiating in as many directions as necessary from that point (hence the "star " name). Those points should end up in every room that will have equipment with signal cables.
When all the ground conductors to star central point are of equal length all theends of the star are, by definition, at the same ground potential and signal wiring between any equipment grounded to any of the points of the star will also be at zero potential. Therefore, no ground loops are created.
The idea that the ends of the star are at equal potential because they are equal length is critical to the success of this configuration. In order that the ends of the star are very low potential (low resistance), they must be large gauge wire.
Wiring between buildings
You can almost guarantee a s serious ground loop if you are running between two buildings. This is because each building probably has its own source for power from the power company and each building has its own central safety ground point. Since this point is not the same point for both buildings, and since the ground in one building is often at somewhat different potentia, there will automatically be a potential difference between any wiring run between the two buildings. Potential differences can be in some cases (power faults, lightning) can be dangerously high and the current flowing in wire shield can be dangerously high.
I do not recommend any direct galvanic connection using shielded cable between different buildings unless you absolutely know that it is safe do do so. Generally the connections between different buildings should be galvanically isolated (like for example telephone connections, the line itself is not grounded on the building where it enters) and they should be protected agains fault conditions (overvoltage spikes, excessive currents).
If you can get connection using fiber optic cable, I recommend using it instead of copper. One of the key advantages to fiber optic cable over copper cable is immunity to ground problems and overvoltages. Since fiber is glass,there is no metal contact and therefore, no ground problem. Unfortunately to convert to fiber means usually buying an extra box for each output and each input, a significant added expense.
This article is based on The Cherry Juice Issues Online: MAY CHERRY JUICE 1999 article GROUNDING and it is updated and made more accurate with my own knowledge and experiences.
Tomi Engdahl <Tomi.Engdahl@iki.fi>