Shadorne and Jea48 -- I went through the Bill Whitlock paper and presentation you linked to -- really super!!! He's absolutely right when he says these issues are not adequately addressed (or addressed at all) in EE curricula, and even EE's often don't know the difference between a ground loop and a fruit loop. :)
Shadorne -- thanks very much for the compliments. As someone with a digital design background, and who builds his own pc's, I agree 1000% that pc's are particularly proficient noise generators, due primarily to the fast edge speeds (rise and fall times) that are present on the innumerable digital signals running around in them, as well as to the large currents that can be change value very quickly as a result of fluctuating demands on the cpu and other devices.
Whitlock very persuasively supports what you said about high impedance shields, in single-ended interconnects, saying that they are typically one of the most major contributors to noise problems. He makes the point that ground offsets and resulting noise are inevitable (although they can be minimized), but it is the resistance of the shields of single-ended interconnects where that noise typically gets introduced into the signal path to the greatest extent. Therefore single-ended interconnects should be as short as possible, and selected for the lowest possible shield resistance.
Blindjim -- Re signal attenuation per unit length of cables, you are probably thinking of video or rf transmission lines, where both the cable impedance and the signal frequencies are well defined and controlled. I don't think that the attenuation is determinable in a meaningful way for noise components that are present at many unknown frequencies and are being conducted through power wiring that has poorly defined and controlled high frequency characteristics. But see page 4 of the reference below (one of those that Shadorne and Jea48 provided) for some rough indications of the impedance of typical power wiring.
For those who don't have time to go through the 140 pages or so of these two references, I'd suggest at least looking at the following pages of this link:
http://www.jensen-transformers.com/an/generic%20seminar.pdf
Page 12: "how the noise gets in"
P. 14: "solutions"
P. 20: "choosing cables"
P. 21: "a checklist" -- especially good!
P. 27: "always ground shield at driver (at least)"
P. 28: "unbalanced to balanced interfaces" (if applicable)
P. 40: "Many of the benefits often ascribed to power treatment schemes are simply due to plugging all system equipment into the same power strip or dedicated branch circuit. For obvious reasons, this is always a good idea"
P. 40: "surge suppressor cautions" (they can do much more harm than good if used improperly)
Thanks again,
-- Al
Shadorne -- thanks very much for the compliments. As someone with a digital design background, and who builds his own pc's, I agree 1000% that pc's are particularly proficient noise generators, due primarily to the fast edge speeds (rise and fall times) that are present on the innumerable digital signals running around in them, as well as to the large currents that can be change value very quickly as a result of fluctuating demands on the cpu and other devices.
Whitlock very persuasively supports what you said about high impedance shields, in single-ended interconnects, saying that they are typically one of the most major contributors to noise problems. He makes the point that ground offsets and resulting noise are inevitable (although they can be minimized), but it is the resistance of the shields of single-ended interconnects where that noise typically gets introduced into the signal path to the greatest extent. Therefore single-ended interconnects should be as short as possible, and selected for the lowest possible shield resistance.
Blindjim -- Re signal attenuation per unit length of cables, you are probably thinking of video or rf transmission lines, where both the cable impedance and the signal frequencies are well defined and controlled. I don't think that the attenuation is determinable in a meaningful way for noise components that are present at many unknown frequencies and are being conducted through power wiring that has poorly defined and controlled high frequency characteristics. But see page 4 of the reference below (one of those that Shadorne and Jea48 provided) for some rough indications of the impedance of typical power wiring.
For those who don't have time to go through the 140 pages or so of these two references, I'd suggest at least looking at the following pages of this link:
http://www.jensen-transformers.com/an/generic%20seminar.pdf
Page 12: "how the noise gets in"
P. 14: "solutions"
P. 20: "choosing cables"
P. 21: "a checklist" -- especially good!
P. 27: "always ground shield at driver (at least)"
P. 28: "unbalanced to balanced interfaces" (if applicable)
P. 40: "Many of the benefits often ascribed to power treatment schemes are simply due to plugging all system equipment into the same power strip or dedicated branch circuit. For obvious reasons, this is always a good idea"
P. 40: "surge suppressor cautions" (they can do much more harm than good if used improperly)
Thanks again,
-- Al