Here's a good article my dad wrote up about ground loops, you may find this useful, Hartmut!
In the old days we had equipment without the safety ground connection to the chassis, and the signal cables provided the only ground path. This is still the best approach except that it provides no safety if there is a fault in the equipment that would put mains voltages on the chassis.
So we have to think differently and minimize the effect of the ground loop. Just having a ground loop doesn't ensure that you will have hum, the hum is caused by magnetic coupling into the ground loop.
Let's look at where the ground loop is being formed. We have two mains cables going to either a wall outlet or to a plug strip. The safety ground wires are connected together at this point. The mains cables go to separate pieces of equipment. We make the ground loop when we connect a signal cable between the two pieces of equipment, thus creating a continuous connection around a loop. A varying magnetic field can now induce a current into this loop, and create a voltage drop across the wires that form the loop. Even though the two pieces of equipment are grounded thru the signal cable shield and simultaneously grounded to each other thru the safety ground wires in the mains cables, in the presence of a varying magnetic field, the ground potentials at different points in the loop are different.
To understand why you don't hear the hum when no instrument is connected, consider that the center signal wire of the cable between the amplifier input and the power supply doesn't connect to anything at the power supply end. If that wire weren't shielded, then we could induce noise into the center wire by electrostatic (or capacitive) coupling, and we would often describe the resulting noise as buzz since electrostatic coupling more readily couples high frequency components. The shield of the cable (and the metal boxes of both the amplifier and power supply) acts as a Faraday cage and shunts any potential electrostatic noise to the ground (or earth) and prevents this type of interference from being heard. The fact that the system is quiet with the instument disconnected assures us that the electrostatic shielding is performing properly. Even though we still have the magnetic coupling generating voltages in the shields, it has no means of being sampled since the center wire is not terminated.
When the instrument is plugged in we now have a means of listening to the magneticly induced noise (which is stronger at the fundamental frequency of the mains generator) so we percieve it as a hum sound.
So how do we keep our safety grounds and minimize the hum?
It might be easiest to first consider what we would do if we wanted to build a circuit to find a source of magnetic fields. We would take our ?loop? and maximize the area and move the loop until it surrounds the source of the varying magnetic field. Our actual cable to the guitar is of no help to us in maximizing the hum since the signal conductors take exactly the same path as the outer shield and the area enclosed by this portion of the loop is zero. Thus you won't be able to move the guitar and its cable around to make a difference in the amplitude of the hum.
So where is the area of the loop? It's that area between the mains conductors and the signal cable between the amplifier and the power supply, but also includes the metal chassis themselves. So what we want to do is to minimize the area by using a short signal cable, and by dressing the two mains cables, or even twisting them. In some cases this may be sufficient to make the hum acceptably low.
Unfortunately, one of the hard things is that the source of the magnetic hum field is the power transformer inside the amplifier which is also part of the conductor path that makes the loop. So the entire loop is something like: one safety ground connection at one of the mains connectors, thru that power cable to the bolt on the chassis of the amplifier, thru that chassis metal and if the input jack is grounded to the chassis, to the input jack serrated washer that cuts thru paint, then thru the jumper cable to the power supply, whichwhere the jack is grounded to the chassis, then up thru the cable to the instrument, thru the selector switch in standby, and back down thru the cable and to a bolt thru the chassis that connects to the safety ground terminal on the mains connector, thru the mains cable, and to the same safety ground connection we started from. So we can minimize some of the path, but other parts you'd have to smash by driving a truck over the chassis and bend them to minimize them and the equipment wouldn't work after you did that!
So let's go back again to the concept of how to maximize the hum. And realize that if we make a coil out of the jumper wire between the amplifier and the power supply, say about 6 inches in diameter, then we can move the coil around and generate a larger or smaller ground loop current by positioning it closer or farther from the power transformer. Also, if we invert the coil by turning it over (or the same thing if we take the coil apart and wind it in the opposite direction), then the polarity of the signal that is picked up is reversed. So we can by trial an error, determine the winding direction to generate a signal polarity that is opposite that in the loop portion that we can't do anything about. And then position the coil or adjust its area by changing it from a circle to a narrow loop which changes the area which in turn changes the amplitude. If we can set the loop just right we can generate a current that is equal in amplitude to the one we can't do anything about, with the opposite polarity, and the hum is cancelled out.
Quite frequently, the use of a short jumper and dressing the mains cables to follow the same path, and using a power strip so the mains connectors are side-by-side will make the hum sufficiently low. If that isn't effective, then building the loop as described above can work after a period of tedious trials to determine the number of turns and the area to be enclosed as well as the position that will generate the right hum bucking voltage to exactly cancel. Note also that sometimes instead of making the coil with the signal jumper, you can make it with one of the mains cables and position it under the amplifier chassis.
The principle is simple, but since our eyes can't see magnetic fields, then we're blindly trying things. But understanding the fundamental nature of the problem helps us in our blindness.
(Message edited by mica on May 10, 2005)
