Originally Posted by SAM64
Once again , this only works if there's no connection between chassis and signal ground at either end of the cable, not true 99% of the time.
Once again, I totally disagree, partially because I've done the circuit analysis and partially because I've used the method incessantly with good measured and audible results.
What really happens is that the better the chassis are tied together, the less hum voltage there is for the differential input to cancel out.
Here is the circuit with the chassis floating:
The chassis ground voltage at the ends of the balanced line are indeterminate because nothing ties them together except the balanced line and any casual return path though input resistors, etc.. It is possible to exceed the common mode dynamic range of the balanced input stage and still get hum. Been there, done that.
Here is what happens if there is a connection between the two chassis (inter-chassis bonding)
The first question that pops into my mind is "What's wrong with that?"
The chassis ground voltages at each end of the balanced line are now determinate and known to be roughly equal. Common mode cancellation is not interfered with.