Connect the shield of an unbalanced signal input to the Chassis and you will mix/add unwanted interference voltages to the wanted input signal, because the amplifier cannot tell the difference between signal and interference.
The amplifier measures the difference in voltage across the unbalanced socket and processes that voltage to send from the output.
An unbalanced interconnect must be kept separate from the inter Chassis Currents that will flow in any system.
In a balanced interconnect the two signals are kept separate from the Chassis. This is the same as an unbalanced interconnect where the Return is taken directly to the input of the amplifier and does NOT connect to Chassis.
The inter Chassis Currents pass into the Chassis via Pin1 of the Balanced interconnect. Pin1 MUST be connected to Chassis to allow these Currents to pass with least impedance.
The amplifier measures the difference in voltage across the unbalanced socket and processes that voltage to send from the output.
An unbalanced interconnect must be kept separate from the inter Chassis Currents that will flow in any system.
In a balanced interconnect the two signals are kept separate from the Chassis. This is the same as an unbalanced interconnect where the Return is taken directly to the input of the amplifier and does NOT connect to Chassis.
The inter Chassis Currents pass into the Chassis via Pin1 of the Balanced interconnect. Pin1 MUST be connected to Chassis to allow these Currents to pass with least impedance.
No wonder people get confused. I seem to remember, from when I had an interest in radio, that coaxial cable screen was directly connected to equipment chassis.
That was getting on for 30 years ago now so I could be mistaken or procedures may have changed since then.
The articles on single ended hifi interconnection I have read (apart from the one speedskater linked to) show the RCA sockets insulated from the chassis.
That was getting on for 30 years ago now so I could be mistaken or procedures may have changed since then.
The articles on single ended hifi interconnection I have read (apart from the one speedskater linked to) show the RCA sockets insulated from the chassis.
Read any competant article on inter-chassis currents and it will be made clear that the return wire of the input pair must NOT carry the inter-chassis current.
Rane and Jensen and Self and Cordell and Jung and Ott all agree.
Rane and Jensen and Self and Cordell and Jung and Ott all agree.
What I've been successful with is connecting the shield from the input RCA jack as the FIRST connection at the MAG point, directly to the chassis at that point. The next to connect is the output ground of the input stage, then the output ground of the output stage, etc. So the noisiest return is always farthest from the input ground chassis connection, and the input ground voltage reference is not bounced by large currents flowing in the rest of the circuits. I use a terminal strip and make the connections as described above, with connecting wires between the pegs of the terminal strip to connect them together and thus to the chassis ground point. If a hum issue does arise, a ground loop breaker (resistor or resistor/diode bridge/cap combo) can easily be installed on the line from input jack ground to its MAG point.
Is this the "correct" solution? I don't know, but it works for me.
Is this the "correct" solution? I don't know, but it works for me.
Last edited:
The above Jim Brown paper was titled "Understanding and Solving RF Interference and Noise Problems". Later he re-edited the paper and re-titled it. Recall that Mr. Brown was Audio Engineering Society past committee chair on EMI/RFI.
Henry W. Ott in Section 2.15.2.1 of his book "Electromagnetic Compatibility" writes:
If the cable shield is connected to the circuit ground, then these RF currents will
enter the equipment and may cause interference.
Therefore, the proper way to terminate the cable shield is to
the equipment’s shielded enclosure, not to the circuit ground. This connection
should have the lowest impedance possible, and the connection should be made
to the outside of the shielded enclosure. In this way, any rf noise current on the
shield will flow harmlessly on the outside surface of the enclosure and through
the enclosure’s parasitic capacitance to ground, bypassing the sensitive electronics
inside the box.
If you think of a cable shield as being an extension of the enclosure’s shield,
then it becomes clear that the shield should be terminated to the enclosure not to
the circuit ground. In the professional audio world, a noise problem that occurs from connecting
a cable shield to circuit ground, instead of to chassis ground, is referred to
as the ‘‘Pin 1 Problem.’’ Neil Muncy coined this term in 1995 in his classic
paper on the subject, Noise Susceptibility in Analog and Digital Signal Processing
Systems. The term Pin 1 refers to the connector pin that is connected to the
cable shield in XLR connectors, which are commonly used in professional
audio systems. For a phone jack, Pin 1 would refer to the sleeve; for a phono
plug or BNC connector, it would refer to the outside shell of the connector.
Shields on coaxial cables, where the shield is the signal return conductor,
must be grounded at both ends, and this ground, for functionality, must be to
the circuit ground. However, for noise considerations, as discussed above, the
shield should first be terminated to the enclosure. This can be accomplished
easily by terminating the cable shield to the enclosure and then connecting the
circuit ground to the enclosure at the same point.
If the cable shield is connected to the circuit ground, then these RF currents will
enter the equipment and may cause interference.
Therefore, the proper way to terminate the cable shield is to
the equipment’s shielded enclosure, not to the circuit ground. This connection
should have the lowest impedance possible, and the connection should be made
to the outside of the shielded enclosure. In this way, any rf noise current on the
shield will flow harmlessly on the outside surface of the enclosure and through
the enclosure’s parasitic capacitance to ground, bypassing the sensitive electronics
inside the box.
If you think of a cable shield as being an extension of the enclosure’s shield,
then it becomes clear that the shield should be terminated to the enclosure not to
the circuit ground. In the professional audio world, a noise problem that occurs from connecting
a cable shield to circuit ground, instead of to chassis ground, is referred to
as the ‘‘Pin 1 Problem.’’ Neil Muncy coined this term in 1995 in his classic
paper on the subject, Noise Susceptibility in Analog and Digital Signal Processing
Systems. The term Pin 1 refers to the connector pin that is connected to the
cable shield in XLR connectors, which are commonly used in professional
audio systems. For a phone jack, Pin 1 would refer to the sleeve; for a phono
plug or BNC connector, it would refer to the outside shell of the connector.
Shields on coaxial cables, where the shield is the signal return conductor,
must be grounded at both ends, and this ground, for functionality, must be to
the circuit ground. However, for noise considerations, as discussed above, the
shield should first be terminated to the enclosure. This can be accomplished
easily by terminating the cable shield to the enclosure and then connecting the
circuit ground to the enclosure at the same point.
A simple test of your circuit design is the John Windt "Hummer", described by Bill Whitlock on page 122 of this paper.
http://centralindianaaes.files.wordpress.com/2012/09/indy-aes-2012-seminar-w-notes-v1-0.pdf
http://centralindianaaes.files.wordpress.com/2012/09/indy-aes-2012-seminar-w-notes-v1-0.pdf
Attachments
Grounding advice is similar to seeking legal advise from many different lawyers, each one in turn giving their own opinion which often contradicts the previous one.
If for example I want the best grounding arrangement for my DCB1.
Should I connect all input and output RCA socket body's together, then take a single short wire link from them to a bolt on the rear panel through which they are all fitted?
OR, should I leave them all insulated at the rear panel and connect the input cable screens together at the volume pot end leaving the output screen separate to take it's own path from the boards common output pin back to the rear panel RCA output sockets. Should I also then take a short wire from the boards 0v at the opposite end of the board from the PSU and connect it to it's own grounding point next to the board which is not directly connected to the mains earth.
Can anyone produce an understandable solution in plain English?
If for example I want the best grounding arrangement for my DCB1.
Should I connect all input and output RCA socket body's together, then take a single short wire link from them to a bolt on the rear panel through which they are all fitted?
OR, should I leave them all insulated at the rear panel and connect the input cable screens together at the volume pot end leaving the output screen separate to take it's own path from the boards common output pin back to the rear panel RCA output sockets. Should I also then take a short wire from the boards 0v at the opposite end of the board from the PSU and connect it to it's own grounding point next to the board which is not directly connected to the mains earth.
Can anyone produce an understandable solution in plain English?
Think of it like you would a Power Amplifier.
Maintain the input (signal) circuits.
Maintain the output circuit.
Take a reference from a centrally located MAG.
The common on the output is probably as good a place as any.
If you use screened two core, then you must connect the screen to chassis.
If you use coax, then you don't connect the screen to chassis. The screen is a signal in this case and must be taken to the amplifier (DCB1).
The DCB1 has a common shared between the channels.
Any connection to the outside world MUST keep the common on the channels VERY close coupled. Otherwise you create a loop that picks up interference.
Maintain the input (signal) circuits.
Maintain the output circuit.
Take a reference from a centrally located MAG.
The common on the output is probably as good a place as any.
If you use screened two core, then you must connect the screen to chassis.
If you use coax, then you don't connect the screen to chassis. The screen is a signal in this case and must be taken to the amplifier (DCB1).
The DCB1 has a common shared between the channels.
Any connection to the outside world MUST keep the common on the channels VERY close coupled. Otherwise you create a loop that picks up interference.
Thanks Andrew. I'm currently trying twin core and screen microphone cable to cut down on wiring, one core L the other R and the screen as common return. I'm not sure if I should connect screen to chassis at any point with this cable.
Maybe I should use just use a separate coax cable for each channel.
Maybe I should use just use a separate coax cable for each channel.
I thought twin core microphone cable would be a reasonable choice to carry both channels in one cable since the DCB1 shares common between both channels.
In this case I should not connect screen to chassis as the screen carries signal return for both channels.
Is it a bad idea to use this cable to carry both channels? My thinking was it would reduce wiring runs inside the chassis.
In this case I should not connect screen to chassis as the screen carries signal return for both channels.
Is it a bad idea to use this cable to carry both channels? My thinking was it would reduce wiring runs inside the chassis.
The problem here though is any RF induced noise on the screen is carried inside the box! The on way would be to connect the screen at the cable entry point with a capacitor to the chassis (this should provide a low impedance path to ground). This is the problem with single ended, the shield and the return path are the same, always a compromise.
Some relevant stuff from Mr Ott.
http://www.hottconsultants.com/pdf_files/Audio Interconnections.pdf
http://www.physics.utah.edu/~kieda/ott.pdf
The premise that you want the induced shield current to return via the earth ground and NOT the local signal ground makes sense to me. The one flaw that seems to call out to me is that most power supplies have the earth ground connected to/at the main audio ground (MAG). Some extra resistance must be introduced to the signal return path that will cause induced noise currents to preferentially flow through the chassis and away via the earth connection while still allowing signal return currents flow through the input circuit and return via the shield to the source. This resistance should be on the order of one or more ohms, and could be located on the signal return between the input jack and the amplifier circuitry and/or between the amplifier circuitry and the main audio ground.
In your last paragraph you state that the circuit ground should be connected to the enclosure where the jacks are located (but no mention of the lifting resistor). I do not see how this would reduce noise since the connection of signal ground would allow noise current to flow through the signal ground just as easily as the chassis.
Can you clarify please?
This is all an expansion of what is called the pin #1 problem. So all external cables should have their shield or ground wire connected to the chassis at that cable's chassis jack. AC power cables should have their Safety Ground (EGC/PE) connected to the chassis where that cable enters the chassis not at the MAG. The MAG should have only one connection to the chassis and that should be near the I/O jacks. But often the Input jacks are on one side of the chassis and the Output jacks on the other. In this case you can connect the MAG near the Output jacks and connect the Input return (it's no longer the shield) to the input circuit common via that 4R7 resistor.
Then you have a classic pin #1 problem.
It took 20 years for the audio industry to buy into the problem for XLR connectors, it may take another 20 years for RCA connectors.
- Status
- Not open for further replies.