Individually, your effects processors, mixers, recorders and MIDI instruments may function perfectly, but connect them together and the chances are you'll hear at least some background hum. If you're lucky, this will be quiet enough to live with, but at worst, it may be so intrusive as to make your system unusable. Those unfortunate enough to be bothered by this problem often start disconnecting the earth cables from various mains plugs, in the hope that the hum will go away. Although this frequently works, it's not a good idea from a safety point of view. The hum is generally caused by ground or earth loops, and there's no instant cure. However, once you understand what brings them about, it's not too difficult to track them down and eliminate them.
Most home studios rely largely on unbalanced audio connections, where the signal travels along screened cables, each one comprising a single insulated core surrounded by a screen. The screen is grounded, so as to prevent outside electrical interference from reaching the signal on the centre conductor, but this isn't a foolproof arrangement. An audio signal is really the voltage difference between the centre (hot) conductor and the outer screen, so if the screen isn't held firmly at zero volts, any audio frequency voltages that find their way onto the screen will end up superimposed on the audio signal. But if the screen is earthed, how is it that hum interference from the mains can still get into our systems?
All cable has an electrical resistance, and though low, it does nonetheless exist. Referring back to school physics for a moment, if you pass an electrical current through any material that has an electrical resistance, a voltage will be produced between the two points of contact, its magnitude depending on the strength of the current and the resistance of the material -- as stated by Ohms law. It follows, then, that if you pass a current through the screen of a cable, there will be a difference in voltage between one end of the screen and the other. If this all sounds a little academic at this point, bear with me, because all ground loop hum problems stem from this simple fact, and the same knowledge can be used to cure the problem.
As mentioned earlier, a typical home studio includes numerous mains-powered pieces of equipment joined to each other via unbalanced screened cables. All the screens and mains earths interconnect, and because cable does have a finite resistance, there's a real danger that interference signals will cause current to flow in the cable screens, resulting in audible signal contamination. Most interfering signals, such as those from distant radio transmitters, are pretty feeble, but the 50Hz mains supply feeding your studio is a different matter. If you were to place a closed loop of wire inside a studio, you'd be able to measure a 50Hz current flowing in the wire, because the loop acts exactly like a transformer. Of course, real-life transformers have more than one turn of wire, but the principle is the same, and a very small percentage of the current flowing in the mains system gets inductively coupled into our wire loop. Because audio signals are measured in millivolts rather than volts, even the most inefficient coupling of the 240V mains supply into our wire loop will produce enough current to generate a voltage, which, when added to a typical audio signal, will be audible as hum.
While the loop of wire in our test example is purely hypothetical, the earthing arrangement of our studio is very real. Figure 1, above, clearly illustrates how the earth and screen connections between just two pieces of equipment can form a closed loop, which will be affected by induced mains hum. In reality, the wiring in a typical studio is likely to create many ground loops, all of which interact with each other.
In Figure 1, the circuit is completed by the mains lead earths and the signal cable screens to form our single-turn transformer. The resulting 'hum' voltage is effectively in series with the signal path, and is sometimes known as 'series mode interference'.
To reduce or eliminate the effect of ground loops, we have to follow just one simple rule: each piece of equipment should have only one earth current path between it and the rest of the system to which it is connected.
To comply with this rule, it is necessary to locate any ground loops and break them in some way. And this creates a dilemma; we either have to disconnect a signal screen at some point to break the loop, or we must remove the mains earth and keep the signal screens connected. The latter usually works, but then there will be no protective earth other than via the signal leads, which won't stand the kinds of currents that occur during serious fault conditions. Furthermore, if the signal lead is unplugged, the earth protection is completely removed. From a safety point of view, removing mains earths is not a good thing -- don't try it at home!
Note that equipment operating from external mains adaptors is designed to be used unearthed, and so may be less susceptible to ground loop problems. However, if the unit is bolted to a metal rack, a ground loop may be created via the casework of the unit.
In professional studios, where everything is balanced, disconnecting the screen at one end of a signal cable will usually cure any hum problems, because the screen isn't used as a return path for the signal -- it's purely a protective screen. In an unbalanced system, disconnecting one end of a screen can cause difficulties, because you're then relying on the mains cable earth to act as return path for the audio signal. This can lead to RF (Radio Frequency) interference problems, and if the mains cable is removed as well, the signal has no return path at all, and you're greeted by a monitor-shattering hum!
A simple dodge is to connect a small resistor, in series with the screen, at one end of the cable, as shown in Figure 2a, below. In a typical audio system, a resistor of around 100 ohms will be high enough to significantly reduce any induced hum currents, while still being low enough not to affect the level of the signal passing through the cable. Using a resistor alone slightly increases the risk of RF interference. Usually, this isn't a problem, but if you do experience high frequency whistles or breakthrough from radio stations, a 100pf capacitor connected in parallel with the resistor should help. Because the current flow we're dealing with is very small, low wattage resistors may be used, and a quarter (or even eighth) watt metal oxide film resistor can be mounted inside most plastic-bodied jack plugs without difficulty. Figure 2b, below, shows how the capacitor is wired, should you decide to add one. Now, if a crucial signal earth lead is removed, there is no problem, because the 'cold' signal can still make the journey via the screen and the resistor.
This method of tackling ground loops is a compromise, because the induced current isn't eliminated, merely reduced. Nevertheless, it can bring about a dramatic improvement in the level of background hum, and in a system which uses unbalanced cables, it can be very difficult to get rid of hum by any other means.
If you have a desk with balanced line inputs, but have unbalanced outboard gear, you can go one better, as shown in Figure 3, below. A balanced input only 'sees' the difference between the positive and negative input lines, so if both carry identical interference signals, the interference cancels out -- a concept known as 'common mode rejection'. This can be exploited when connecting unbalanced sources to balanced inputs. To prevent significant earth currents flowing in the cable screen (which in extreme conditions could compromise the common mode rejection of the input stage, and allow hum back in), we insert a resistor of around 100 ohms in series with the screen connection. This is more satisfactory than putting a resistor in series with the screen in a completely unbalanced circuit, because we're not relying on the screen to act as a signal return path -- it works purely as a protective screen.
Some mixing consoles use a pseudo-balancing system known as 'ground compensation'. Details of how to connect both balanced and unbalanced signals to these mixers is included in most user manuals, and in most instances, the additional effort involved in making or adapting cables to take advantage of these inputs is very worthwhile.
Not only do we have mixers, tape machines, effects units, MIDI instruments and so on to consider, most systems also include patchbays. Private or semi-pro studios invariably use unbalanced jack patchbays to handle signal connections, and though this doesn't present any great problem, there are one or two points to keep in mind. In the interests of avoiding unnecessary connections between one ground point and another, avoid the type of patchbay where the socket grounds are all linked together along the length of the patchbay -- this is simply asking for trouble. If your patchbay provides the option of removing the ground link between upper and lower socket pairs, then do so wherever a patchbay is being used in a non-normalised application, such as for providing remote console inputs, or for bringing the inputs and outputs of effects and processors to the patchbay.
Normalised patch sockets are usually fed from console insert points, and providing the distance between the console and patchbay is less than around 10 feet, you can get away with using a stereo cable to carry both the insert send and return connections, as shown in Figure 4, at the top of this page. The fact that both signals share a common screen means that there can't be a ground loop between the insert point and the patchbay, even if the upper and lower patchbay socket pairs are ground-linked. However, with very long cable runs, sending both signals down the same cable runs the risk of crosstalk, which may lead to instability.
Usually, the console can be connected directly to the patchbay using conventionally wired cables with no problem; any precautionary measures (such as fitting series resistors) are applied to the cables joining the effects processors, tape machines and instruments to the patchbay. People pontificate over which end of the cable the resistor ought to be connected to, but in practice, I've found it makes little or no difference, so you may as well put it wherever you find most convenient.
Even armed with this knowledge, it's very difficult to track down ground loop problems in a finished system. You might find that you fix one loop and the hum gets louder; this can happen where one ground loop is in antiphase with another! It might sound like a real chore, but the answer is to unplug everything and then start wiring the system from scratch, checking for hum as each new piece of equipment is connected.
The starting points are the mixer and monitor amplifier; if the monitor amp has balanced inputs, use them. Most multitrack mixers have balanced monitor outputs, but even if yours doesn't, you can still use the 'balanced to unbalanced' connection technique outlined earlier in this article. Once you're happy that the system doesn't hum, you can connect the 2-track recorder and try again. Because you have four cables (left and right, in and out) going to the 2-track, you have the conditions for a ground loop, so if hum does rear its ugly head, use the 'resistors in the cables' trick. Even if the 2-track has its own ground lift (see separate side panel on ground lift elsewhere in this article), you'll still need to put resistors in three of the four cables to ensure that there's only one earth signal path to the machine, but try conventional cables first -- you might not need to bother. Of course, some hiss and hum is inevitable if you turn the monitoring system up far enough, but if the hum is at a lower level than the natural background hiss of the circuitry, this is probably as good as you can hope for. At a realistic monitoring level, neither hiss nor hum should be evident, unless you put your ear right against the speaker.
When it comes to connecting the multitrack, the large number of inputs and outputs again increases the risk of multiple ground loops. Very often, you'll get away with using conventional cables, but if problems arise, you'll have to return to using resistors. After you've sorted out the multitrack, it's as good a time as any to connect up the patchbay. The first step is to confirm that all is quiet when no external equipment is plugged into the patchbay. If so, try your outboard units one at a time to see which give trouble. Don't confuse ground loops with the digital noise and hum produced by some budget processors. As a rule, ground loop hum will remain audible, even when the master aux send feeding the outboard equipment is turned down, whereas mix bus noise or other contamination from the console will go up and down as the relevant aux send level or effects input level is adjusted. If you've done your homework, and checked which of your outboard units are ground lifted, you'll have an idea which ones may cause trouble.
Finally come the synths and expanders, and once again the resistor in the cable trick can greatly improve matters. The screen of MIDI cables can also aggravate the ground loop situation, and in extreme cases, you might need to use a DI box to get rid of the hum completely. In my studio, my sampler refuses to play by the book, so I feed it into a spare mic input via a phantom-powered, active DI box. As well as completely curing the hum problem, this also gives the benefit of better level matching at the desk.
Having painted a pretty grim picture of ground loops and the annoying hum they cause, you'll probably find that only a few pieces of equipment give you any real trouble. Providing you test your system as you assemble it, you should have little difficulty in identifying the areas that need attention and the things you can leave alone.
Chasing hum problems isn't nearly as much fun as making music (though on some occasions, you might find it easier), but there's no sense in spending a lot of money on state of the art studio equipment if it isn't going to give its best. Just a couple of days burning your fingers, swearing and sticking your ears in speaker cabs will pay off in the long run -- honest!
If you draw out a wiring diagram for your system, including all the signal cables and mains leads (only those with earths, not mains adaptors), you'll soon see where potential ground loop problems lie. However, problems also arise when a ground signal path is completed by another route -- say the metalwork of a rack system. A well-designed piece of rack equipment should be fitted with an internal ground lift, which will be either fixed or switchable, and this reduces the risk of ground loops when conventional, unbalanced connecting cables are used. Many pieces of semi-pro gear have no ground lift, so how do you tell?
In a ground-lifted unit, there is no direct signal path between the 'cold' or screen side of the audio circuitry and the casework of the box. Instead, the box is earthed, and the 'cold' side of the circuitry connected to the case via a resistor of several hundred ohms. If the manual doesn't make it clear whether a ground lift is fitted or not, simply unplug the unit, plug in a lead, and use a multimeter (set to resistance) to measure the the resistance between the metal case and the barrel of the jack plug, as shown in the diagram on the right. If the resistance measures close to zero, there's no ground lift, but if it measures over 100 ohms, a ground lift is almost certainly fitted.
If no ground lift is evident, then you could run into trouble when fitting the unit into a metal rack; the metal frame creates yet another earth current path between different pieces of equipment. The only solution here is to use nylon mounting bolts and washers, so that the case is isolated from the rack. You may also need to leave extra space to ensure that the device doesn't touch the units above or below it, though a thin cardboard spacer will normally do the trick.
A healthy-sounding studio starts with a good mains supply, so check out part 1 of the 'Studio Wiring' series in the April issue of SOS, and also keep your cable runs as short as possible. Foil-screened cable is best for permanent wiring, as it's reasonably cost-effective, has good screening properties and isn't too thick.
For flexible wiring, cable with a woven copper screen usually performs best, but conductive plastic cables are fine for short patch cables, instrument leads and so on. Though their screening isn't quite so effective as that of woven screen-type cables, their pliability often means they're still working when other leads have fallen apart.
Whatever type of cable you use, try not to run it alongside mains cable for any distance, though crossing it at right angles is no problem. Also be aware that anything containing a large transformer is liable to radiate a strong hum field, so mount power amps and mixer power supplies away from other processors. At the very least, leave a few units of empty rack space between these items and your effects processors.
Don't disconnect the earth leads from pieces of equipment that are designed to be used grounded.
Build up your system a piece at a time, checking for hum at every stage. Cure any ground loop problems before connecting any more equipment. If you experience no hum problems when using standard leads, don't feel you have to fit ground-lifted cables -- move onto the next piece of equipment.
Use balanced wiring where practical.
When working with unbalanced equipment, use ground-lifted leads (see main article) to ensure that each piece of equipment has only one direct earth path, either via a mains earth or via a signal cable screen. In the case of 2-pin mains equipment, or devices that run from mains adaptors, treat as you would for ground-lifted equipment, and ensure that just one of the signal cables provides a true ground. Additional connections should be ground lifted. Again, if you don't find a problem, don't feel you have to provide a cure.
Check individual items of equipment with a meter, to see which have built-in ground lift resistors. Those that are ground lifted should be grounded both via the mains and via one signal cable.
Beware of case-to-case contact causing problems. This is common in metal racks, and can usually be cured by using nylon mounting hardware.