Happy New Year to all!
More interesting opinions about shield grounding.
------------- Answers to shield grounding question ----------------
In my last newsletter, I asked for opinions about which end of the shield on a signal cable should be grounded: the source end, or the load end. Eighteen engineers from 12 different countries offered interesting ideas.
* Several engineers argued that the cable shield should be grounded at both ends - especially engineers with experience at electric power substations - for the following reasons: high frequency radiated interference from arcs, cell phones, etc. are not attenuated by single-ended grounding; grounding both ends effectively introduces a common-mode choke around the signal conductors. (Marc points out that this function used to be performed by ferrous conduit, but that modern plastic conduit has made the problem worse.)
* On the other hand, there are engineers who argued that cable shields should only be grounded at one point, especially at electrical substations (!) - otherwise the shield itself will carry earth currents between locations that are at different earth potentials. During earth faults, when a major power conductor accidentally touches the earth, these currents can rise to thousands of amps, and will quickly destroy any cable that happens to be connected to earth at both ends. One of the engineers suggested solving this problem by breaking shields on long cables into shorter segments of shield, with each segment locally earthed.
* Yet another argument: engineers in the telecommunications industry apparently ground both ends of shields, and they've been doing it this way for more than a hundred years, so who can argue with them? On the other hand, engineers who construct petrochemical plants always ground just one end of the shield, and their equipment works pretty well, too. And I'm told that the elevator industry, which runs signal cables long linear distances near power cables, follows the single point ground rule, too.
* An engineer with great practical experience wrote that initially he followed the "always ground the transmitter end" cook-book instruction until he constructed a steel mill in Australia in the 1970's, and found that his feedback signals to DC drives were swamped by noise. He then switched to the philosophy of grounding the end of the shield that is closest to the sensitive electronics, which fixed the problem, and he's stuck with that solution ever since.
* Bill Moncrief's answer was so clear, as usual, that I will reproduce it here: "To protect the interior wiring from having a current induced, you connect both ends and let current flow. That makes the inside of the shielded cable a net-zero space; an electromagnetic shield. To protect the inside wires from collecting a voltage from a transmitter, you ground one end and let the voltage appear on the shield. The inside wires are a net-zero space again from the electrostatic shield. And we used tri-ax cable in EHV substations to carry one-watt radio signals. Double ground the inside, single ground the outside, but I think that the order was just convenient."
* A different idea was suggested by a couple of design engineers, who work on signal grounding inside equipment: you must ground the source end, because for any signal there is only one source (but there might be many receivers). If you're going to ground just one end, you don't have any choice.
The information above came from a news letter from Power Standards Labs.
http://www.powerstandards.com/
More interesting opinions about shield grounding.
------------- Answers to shield grounding question ----------------
In my last newsletter, I asked for opinions about which end of the shield on a signal cable should be grounded: the source end, or the load end. Eighteen engineers from 12 different countries offered interesting ideas.
* Several engineers argued that the cable shield should be grounded at both ends - especially engineers with experience at electric power substations - for the following reasons: high frequency radiated interference from arcs, cell phones, etc. are not attenuated by single-ended grounding; grounding both ends effectively introduces a common-mode choke around the signal conductors. (Marc points out that this function used to be performed by ferrous conduit, but that modern plastic conduit has made the problem worse.)
* On the other hand, there are engineers who argued that cable shields should only be grounded at one point, especially at electrical substations (!) - otherwise the shield itself will carry earth currents between locations that are at different earth potentials. During earth faults, when a major power conductor accidentally touches the earth, these currents can rise to thousands of amps, and will quickly destroy any cable that happens to be connected to earth at both ends. One of the engineers suggested solving this problem by breaking shields on long cables into shorter segments of shield, with each segment locally earthed.
* Yet another argument: engineers in the telecommunications industry apparently ground both ends of shields, and they've been doing it this way for more than a hundred years, so who can argue with them? On the other hand, engineers who construct petrochemical plants always ground just one end of the shield, and their equipment works pretty well, too. And I'm told that the elevator industry, which runs signal cables long linear distances near power cables, follows the single point ground rule, too.
* An engineer with great practical experience wrote that initially he followed the "always ground the transmitter end" cook-book instruction until he constructed a steel mill in Australia in the 1970's, and found that his feedback signals to DC drives were swamped by noise. He then switched to the philosophy of grounding the end of the shield that is closest to the sensitive electronics, which fixed the problem, and he's stuck with that solution ever since.
* Bill Moncrief's answer was so clear, as usual, that I will reproduce it here: "To protect the interior wiring from having a current induced, you connect both ends and let current flow. That makes the inside of the shielded cable a net-zero space; an electromagnetic shield. To protect the inside wires from collecting a voltage from a transmitter, you ground one end and let the voltage appear on the shield. The inside wires are a net-zero space again from the electrostatic shield. And we used tri-ax cable in EHV substations to carry one-watt radio signals. Double ground the inside, single ground the outside, but I think that the order was just convenient."
* A different idea was suggested by a couple of design engineers, who work on signal grounding inside equipment: you must ground the source end, because for any signal there is only one source (but there might be many receivers). If you're going to ground just one end, you don't have any choice.
The information above came from a news letter from Power Standards Labs.
http://www.powerstandards.com/
