Mike, Understood your question, but the answer is basically "It depends".
The discussion of skin depth, etc... is only one concern. Ott does an entire chapter devoted to shielding. His concern is inductive coupling of the current paths defined by the complete circuit topology and a mathematical treatment of shields in general. He discusses the Shield Cutoff Frequency Fc ending with the observation that Aluminum foil shield has a much higher Fc due to its higher resistance Rs. Starting with an unshielded wire as a 0dB reference, he builds his case through experimental data based upon various combinations and circuit topologies of - twisted pair / untwisted, shielded / unshielded, differential mode / common mode circuits, no shield ground / one end grounded / both ends grounded. Starting with a single wire as 0dB reference, twisted pair gave him 13dB attenuation, versus shielded twisted pair of a common mode circuit with the shield grounded at BOTH ends giving him 28dB attenuation. But, I would again say to be careful with these numbers due to the unknown specifics and topology of the circuit with which he was dealing. This should also be kept in mind with what you and your colleagues are dealing. So, just for the sake of this discussion, off the cuff, I'd say that with this in mind, generally you're correct that the configuration you're discussing, i.e. the added attenuation of the shield, is capable of far more attenuation than merely 6dB. By how much better than 6dB? I don't know. How's that for a non-statement? ;) My immediate concern would be any real-world current through or along the shield due to any imbalance. I would direct your colleagues to Henry Ott's book "Noise Reduction Techniques in Electronic Systems" for a better explanation than I can do here. So, I guess I didn't answer your question, but that's about the best I could do. Regards, Doug ---------- > From: [email protected] > To: [email protected] > Cc: [email protected] > Subject: Re: SE of cable outer shield 2 > Date: Tuesday, November 04, 1997 4:55 AM > > This is a followup clarification (or at least a mind tickling comment) on the > original question I raised as to the ADDED shielding effectiveness of an > outer cable shield over a group of (inner) twisted, shielded pairs carrying > high frequency digital signals (balanced ECL drivers and receivers). > > The ability to reliably model circuit and cable performance to 22 GHz is more > difficult to prove than to verify by test; however, knowing "ballpark" > numbers and gaining the insight that models offer can lead to innovative > designs and potentially save many man-hours of cost. In this particular > case, the models are most useful in predicting the spectral content of the > potentially offensive radiated emissions. > > As a qualitative comparison, I see that a small shield current (for a given > ECL-driven, balanced wire pair) will flow on the INTERIOR of its associated > shield. By consideration of skin depth, some (but less) current will flow on > the OUTER surface of that same shield. This latter surface current can be > likened to the surface current on a wire inside another shield (i.e., the > outer, added cable assembly shield). Hence, I see the added cable shield > like the outer shield of a coaxial cable, except the "center" wire is > randomly displaced from the center, and mixed with other wires. From > practical experience or otherwise, do you have an estimate for the ADDED > attenuation (or SE)? > > I am anticipating 30 to 40 dB added reduction of radiated emissions from > addition of this outer cable shield, assuming 360 degree termination at both > ends of the cable (two to ten feet in length). > > Mike Conn > Owner/Principal Consultant > Mikon Consulting
