Proper grounding scheme

Status
Not open for further replies.
Jump to Latest
#45
The 1nF capacitors fro the RCA connectors to ground ensure that to RF frequencies the shield of the interconnect cable is connected to the chassis. Any RF impinging on the shield is then shunted to the chassis and away from the input circuits where it could cause problems. All inputs must have this cap fitted. See also THAT 1200/1203/1206 data sheet for examples on how to treat balanced inputs and THAT 1646 line driver data sheet for how to treat outputs.

If you are serious about RF contamination, then fit it. Same on the outputs as well.


If you open up any good commercial equipment, you will see these components. I have a Marantz PM7000 where this is done.
 
#48
scottjoplin said:
I saw that. If you are insulating the sockets from the case it will do no harm but it's not critical. Yes all of them
Click to expand...

It is critical for RF screening... I would suggest Henry Ott and Ralf Morrisons books and articles.

One thing to watch out for is the lower frequencies finding a low resistance path via the thick (and thus low resistance) PE cables between equipment, this will often create a ground loop and thus hum. So ensure that cables used have a good quality low resistance screen.
Return currents follow the path of least impedance but depending on the frequency it can be resistive (low) or inductive (higher).
Just playing about now with some audio grounds (aircraft and missile), one technique I use quite often is diff routing for the signal and its return, usually broad side routed with the return underneath the hot wire and at least 3X the width, preferably 5X.
 
#50
marce said:
It is critical for RF screening... I would suggest Henry Ott and Ralf Morrisons books and articles.

One thing to watch out for is the lower frequencies finding a low resistance path via the thick (and thus low resistance) PE cables between equipment, this will often create a ground loop and thus hum. So ensure that cables used have a good quality low resistance screen.
Return currents follow the path of least impedance but depending on the frequency it can be resistive (low) or inductive (higher).
Just playing about now with some audio grounds (aircraft and missile), one technique I use quite often is diff routing for the signal and its return, usually broad side routed with the return underneath the hot wire and at least 3X the width, preferably 5X.
Click to expand...

Understood. I didn't get in that pdf file that the chassis doesn't appear to be referenced to signal return but then there's an optional connection at MAG. I'm only familiar with the chassis/signal rtn ref. Presumably the capacitors are only required when this is absent. I don't understand when/why it would not be desireable to have sig rtn referenced to the chassis. Can you enlighten me?
 
#51
AndrewT said:
Those measurements and the conclusions drawn from them were commented on.
But the Member would not complete the measurment set to include the other suggested layouts.
He would ONLY publish data for the two extremes, neither of which mimic the way amplifiers are actually assembled.
Click to expand...

Andrew,

The "isolated config" in the last measurements is exactly what you recommend.
None of the measurements and explanations given were seriously discussed.
I have demonstrated that your explanations concerning ground loops are wrong and nobody commented about that.
 
#52
In most design I work on the CHASSIS and on-board signal return (0V, GND whatever you want to call it) are kept as separate as possible and we go to great lengths to make sure small analogue signals are protected. Let me dwell on my reply as I am at work at the moment...
I will also try and dig out previous stuff on this as its a common topic.🙂
 
#54
scottjoplin said:
Yes. But are you saying the signal return does not have to be referenced to an earthed (as in the thing we live on) chassis? Or maybe it's referenced somewhere else, I don't know, I'm asking. Or maybe I'm wrong using the term signal return?
Click to expand...
looks like I missed this question.
has anyone else given an answer?

The enclosure aound the amplifier is not part of the audio. It serves many other purposes but not audio.

Thus it follows that neither the speaker output, nor the signal input, needs to connect to the enclosure for any audio purpose. Imagine the enclosure as an insulating box (plastic, or wood).

The only reason for connecting the MAG to the PE protected enclosure is for Safety.

If you look at a (very simplified) opamp, or chipamp, or power amp you will find a pair of input terminals, an output terminal and two power feed terminals. We have a 5 terminal device.
But in all of these it is actually a 6 terminal device. The missing terminal is the Output Return.
The output needs to be referenced to the same voltage as the input signal.
This is usually done at the local decoupling Zero Volts, where the Signal Return and MF+HF decoupling and Output Return meet up.
The sixth terminal is virtual in many schematics. Often it is referred as ground.

I believe that most of the confusion in Builders minds comes from adopting the word ground. It gets applied to different points/junctions all around our audio gear.

Back to the enclosure.
It protects the internals.
It holds the internals in their correct locations.
It separates the user from the internals.
It blocks interference, depending on how big the holes are.
It gives a connection for the Protective Earth of mains powered equipment.
probably many others.

Does that help?
 
Last edited:
#55
mikolaj612 said:
How about output sockets? Should I do the same?
Click to expand...
Yes, all cables into or out of the chassis can pass interference into the internal circuits.
Yes, all cables can have interference filters where they pass through the enclosure panel.

The simplest interference filter is a capacitor from enclosure to cable at the point of entry.
The lower the inductance in that capacitor and it's connections, the higher the frequency that the filter can be effective at.
We now have many interference sources in our homes that are close to our equipment and generate >1GHz
The 100kW AM transitter sitting in the next county is no longer our biggest interference problem. Our cables, telephones, WiFi, and high efficiency lighting/PSUs have become a problem of our own making.
 
Last edited:
#56
@ AndrewT

Thanks, it helps regards clarity. My next question then would be. what happens to the interference noise that the chassis is screening the equipment from? Is it going to Earth?
"Rule 1. In Morrison [1] (page 39) states, “An electrostatic shield enclosure, to be effective, should be connected to the zero-signal reference potential of any circuitry contained within the shield.”" Quoted from the opening of http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html
 
Last edited:
#57
The enclosure is capacitively coupled to "EARTH", the big ball we all stand/sit on.

The RF interfence passes via the lowest impedance to EARTH.

The internals may have multiple low impedance couplings to the enclosure via all the common mode RF attenuators (those caps from cables to chassis).

Does the low impedance enclosure to audio connection need to be effective down to DC?
Is RF coupling sufficient for the enclosure to be effective?
 
Last edited:
#60
marce said:
Do not think of earth as a current sink, it is there for the PE connection only. !st rule of EMC.
Click to expand...
The cable that links the Protective Earth back to the mains distribution board (Neutral) has too high an impedance to be considered as part of an RF interference solution.

The "Earth" I referred to in post57 is stated as "the big ball we all stand/sit on."
 
Status
Not open for further replies.
Top Bottom