This message was received from Joe Gwinn of Raytheon regarding the shielding of Gigabit Ethernet links. These links run at data speeds of 1.25 x 10**9 (yes, 1.25 billion) bits per second, over two-pair, 150-ohm, balanced cabling. We use one pair for the transmit direction, and another pair for the receive direction. The 150-ohms balanced cabling has an overall shield, and here we are discussing whether to ground the shield at one end, the other, or both.
On the matter of how to ground the shields (hardwire to ground, or through a capacitor), and ground currents melting shields, I would like to offer my experience with the care and feeding of ground loops in the shield protecting low- level signals: use a resistor, not a capacitor.
Specifically, the voltage offset between chassis (green wire) grounds rarely exceeds ten volts. If one puts a hundred-ohm one-watt carbon resistor in series with the shield at either end, with the other end directly grounded to the chassis, the ground current will be limited to 0.1 amp, well within the abilities of the shield to carry. The twisted pair within the shield will still be protected from EMI etc, and a suitable differential receiver will have no difficulty handling the power frequency and harmonics 10-volt common-mode voltage.
Actually, I have seen offsets of only a few volts in the laboratory, and have used ten-ohm one-watt carbon series resistors. I have seen several volts in large buildings, and in ships, so I would design for ten volts RMS.
Which end should be hard grounded, and which should have the series resistor? I haven't tried this for communications signals, but my theory would be that the receiver end should be hard grounded, because it's the receiver that handles the lowest-level signal, and a zero-ohm ground is better than a 100-ohm ground. The effect may not be all that large, because shields handle high-impedance noise sources, and 100 ohms isn't much compared to those impedances, except perhaps at very high frequencies.
The 100 ohm resistor could therefore be bypassed with a RF capacitor, which would be protected from ESD puncture by the 100-ohm resistor.
By the way, the ground noise may be at triple the power frequency, if the user system has lots of capacitor-input 5- volt power supplies fed from the three legs of a three-phase prime power system. I have measured 2.4 volts RMS at 240 Hz in an Air Traffic Control automation system, until the green and white grounds were disentangled. The effective source impedance was about one ohm, if I recall. The waveform was pretty close to a sine wave.
When it was able to drive a current through the VMEbus logic ground, the system promptly fell over. I knew I was in trouble when I saw a spark when I touched one ground to another. The tripling comes from the merger of the pulsating currents into the 5-volt power supplies in the common ground impedance.
Thanks for your interest in High-Speed Digital Design.
Joe, I am going to disagree with your suggestion that a shield with a resistor at one end acts as an effective EMI shield. In high-speed digital applications, it doesn't.
In high-speed digital applications, a low impedance connection between the shield and the equipment chassis *at both ends* is required in order for the shield to do its job. The shield connection impedance must be low in the frequency range over which you propose for the shield to operate. The measure of shield connection efficacy for a high-speed connector is called the ground transfer impedance, or shield transfer impedance, of the connector, and it is a crucial parameter. In the example you cite, the ground transfer impedance at one end of the cable would be 100 ohms, rendering the shield useless.
In low-speed applications involving high-impedance circuitry, where most of the near-field energy surrounding the conductors is in the electric field mode (as opposed to the magnetic field mode), shields need only be grounded at one end. In this case the shield acts as a Faraday cage surrounding the conductors, prevent the egress (or ingress) of electric fields.
In high-speed applications involving low-impedance circuitry, most of the near-field energy surrounding the conductors is in the magnetic field mode, and for that problem, only a magnetic shield will work. That’s what the double-grounded shield provides. Grounding both ends of the shield permits high-frequency currents to circulate in the shield, which will counteract the currents flowing in the signal conductors. These counteracting currents create magnetic fields that cancel the magnetic fields emanating from the signal conductors, providing a magnetic shielding effect.
For the magnetic shield to operate properly, we must provide means for current to enter (or exit) at both ends of the cable. As a result, a low-impedance connection to the chassis, operative over the frequency range of our digital signals, is required that *both* ends of our shielded cable. (See Henry Ott, Noise Reduction Techniques in Electronic Systems, 2nd ed., John Wiley & Sons, 1988.)
There are shielding approaches that provide a low ground transfer impedance at high frequencies, while at the same time providing a much higher impedance at 60 Hz. These approaches involve the use of shields that are capacitively- coupled to the chassis. They are used where high-frequency shielding is needed, but where there is a desire to limit the circulation of 60-Hz currents.
For a capacitively-coupled shield to work, the impedance of the capacitor, at the frequency of operation, must be very low. For example, if the signal wires couple to the shield through an impedance of 75 ohms (that’s another way of saying that the common-mode impedance of the cable is 75 ohms), and the shield is tied to ground through an impedance of 0.1 ohm, then we would expect to measure on the shield a voltage equal to (0.1/75) = 0.0013 times the common-mode signal voltage. The shield in this case would be giving us a 57dB shielding effectiveness. These are the specifications that our IEEE 802.3z 1000BASE-CX copper cabling groups feels are necessary to meet FCC/VDE regulations.
For any shield to work in the Gigabit Ethernet application, we will therefore need a ground transfer impedance (that is the impedance between chassis and the shielded of the cable) less than about 0.1 ohms at 625 MHz. If you check the specifications for the BERG MetaGig shielded connector, it beats this specification. It provides a direct metallic connection between chassis and shield that goes all the way around the connector pins, completely enclosing the signal conductors.
To achieve equivalent performance with a capacitively-coupled shield, the effective series inductance of the capacitor would have to be limited to less than about 16 PICO-henries. That small an inductance cannot be implemented in a leaded component, it would have to be a very low-inductance distributed capacitance, possibly implemented as a thin gasket distributed all the way around the connector shell, insulating the connector shell from the chassis. We have seen proposals for this type of connector, but have not seen one work in actual practice.
I do not advocate the use of capacitively-coupled shields for our application because: (1) It would add complexity, (2) It hasn’t been demonstrated to work, and (3) It would not expand the range of our applications. Keep in mind that the short copper link we are discussing (P802.3z clause 39) is intended for use inside a wiring closet. It only goes 25 meters. It will be used between pieces of equipment intentionally tied to the same ground (we call out in the specification that this must be the case). Between such pieces of equipment there will be no large circulating ground currents. For longer connections, we provide other links types which do not require grounding at either end (multimode fiber, singlemode fiber, and category-5 unshielded twisted pairs). Direct grounding of the shield at both ends is the correct choice for our application.
Dr. Howard Johnson