• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Let's Talk About Grounding

Status
Not open for further replies.
Jump to Latest
#41
marce said:
Didn't say that you had to go to university, just pointed out that if you are going to try and solve a problem learn something about it first.
I was also pointing out that lifting the screen is not good as it then makes the design susceptible to rf interference.
Click to expand...

OK, maybe that was a bit harsh. I just find it a bit frustrating when people don't directly address the question. I apologise.

Cheers

Ian
 
#43
Don't "break the loop".

Instead decrease the voltage in the part of the loop that is creating the interference.

This can be done by:
A.) reducing the interfering field
and/or
B.) reducing the resistance/impedance of parts of the loop
and/or
C.) increasing the resistance/impedance of parts of the loop.

C.) seems counter intuitive.
But, if the loop is ABCDA and AB is the signal section that has interference imposed on it, then reduce the interference VOLTAGE by reducing the interference current in section AB by increasing the resistance in section BC and/or CD and/or DA.
 
#44
Some 19 years ago Neil Muncy(RIP) wrote a paper that covers most of what we are discussing. My take on the paper is that much of the information was old news in 1995, but he put it together in on paper.

****************************************
"Noise Susceptibility in Analog and Digital Signal Processing Systems"
NElL A. MUNCY, AES Member

A demonstrable cause-and-effect relationship between a popular and widely employed
equipment design practice and electrical noise problems in audio systems of all kinds is
examined. A means of identifying equipment that may exhibit noise problems due to this
design practice is outlined. The relationship between the physical construction of shielded
twisted-pair cable and induced noise in a signal circuit due to cable shield current is explored.
Established "rules" for equipment installation are reexamined. It is shown that noise problems
due to this design practice could be eliminated at the manufacturing level at almost negligible
net cost.
*****************************************
Much later some of this information was incorporated into Audio Engineering Society - Standard 48.
"The Pin 1 Problem"
 
#45
So now these 19 years later, Hi-Fi manufactures still don't understand that "The Pin 1 Problem" also applies to RCA connectors.
*************************
Definition of Pin 1
The term pin 1 will appear repeatedly throughout this
paper. Pin 1 is defined as the terminal or terminals of
any equipment input/output (I/O) connector to which a
cable shield or shields are connected when a mating
cable connector is inserted, regardless of use or connector
type. In the case of XLR connectors, the shield terminal
usually is pin 1. For I/4-in (6.35-mm) connectors,
pin 1 is the sleeve; for RCA connectors, pin 1 refers to
the shell; and so on.
 
#46
DF96 said:
The best option is to ignore recipes, and instead ask yourself where the currents go. Treat every conductor, no matter how thick and short, as a resistor with some inductance (so you will get some voltage drop). Treat every current loop as the primary of a transformer, with every other current loop as a secondary (so you will get some induction).
Click to expand...

From my building practices, and IMO, this is the right answer(For ME). Grounds should be connected from where their signal originates. Otherwise current is being supplied from capacitor and being returned to a different area altogether.
 
#49
AndrewT said:
Many "grounds" are wrongly named.
As you have pointed out they are often the RETURN leg of a signal circuit.
The SIGNAL CIRCUIT cannot work if you omit the RETURN lead.

EVERY signal circuit requires a flow and return.
Click to expand...

Not quite sure what point you are making. Ground loops occur where there is more than one return route which is why uncoupling one route is often recommended (rightly or wrongly).

Cheers

Ian
 
#51
AndrewT said:
There is ONLY ONE Return route to match the Flow route of the circuit.

Multiple grounds occur when folk make willy nilly connections and/or include LOOPS where close coupling would minimise the induced interference.
Click to expand...

The point I am making is that every circuit has a Flow and a Return.
Neither of these are a "ground", even though many here call one of them that.

Get ALL the circuits connected with close coupled Flow and Return.
Then work out which parts of the circuit need a connection to a reference voltage to work properly. This extra "reference" connection is an "extra" wire/trace. It is NOT instead of a Return wire.
 
#52
AndrewT said:
There is ONLY ONE Return route to match the Flow route of the circuit.

Multiple grounds occur when folk make willy nilly connections and/or include LOOPS where close coupling would minimise the induced interference.
Click to expand...

So explain how that works when two pieces of equipment are connected together by a screened cable, connected internally by the manufacturer to the chassis of each piece of equipment, and both have the chassis connected to mains safety earth as required by law.

Cheers

Ian
 
#55
ruffrecords said:
So explain how that works when two pieces of equipment are connected together by a screened cable, connected internally by the manufacturer to the chassis of each piece of equipment, and both have the chassis connected to mains safety earth as required by law.
Click to expand...

For conventional unbalanced interconnection:
At low frequencies, where conduction is limited by Ohmic losses, the "return" currents pass (mostly) through the safety earths. At higher frequencies, where conduction is limited by inductance, (most of) half of each channel's "return" current travels through its channel's shield, half through the other channel's.

"Grounds" are messy indeed.

Thanks,
Chris
 
#59
Renslipevol said:
From my building practices, and IMO, this is the right answer(For ME). Grounds should be connected from where their signal originates. Otherwise current is being supplied from capacitor and being returned to a different area altogether.
Click to expand...
If memory serves... I believe Dave Davenport - in another article of his - refers to this practice as keeping the 'ground' return in the 'domain' of the signal. I try to practice this with all the amps I've built and it seems to work reliably.

Rick
 
Last edited:
#60
An excellent article, my only concern is the recommendation of a long tail for an SPDIF co-ax input back to the star point as shown in Figure 4.2-3, this is splitting the return current for the digital SPDIF input and a so the return current path is through the power supply and round the houses, this is incorrect for a digital signal where the signal and its return should be in intimate contact from source to destination. it is similar to having a very large slot in a ground plane and will affect signal integrity and system noise, the digital return current being superimposed on the whole system ground. With such a digital signal the return path if it is the screen should terminate near the signals termination not have a long tail.... The preferred method if noise is a problem is as Dave says a transformer and this should be used in preference to the long tail on a digital connection.
 
Status
Not open for further replies.
Top Bottom