Source follows sink magnetic cancellation

[boji]

Quote from Squarewave:

"...there's the return follows "signal" rule so that current going out matches current coming back and the two magnetic fields cancel each other and you're not radiating noise"

What's anyone's take on how to best envision the cancellation field in 3d space at different frequencies when using the cancellation rule, source follows sink*?

*I'm going to guess if you looked at a PCB edge-on, low frequency emission would look like a figure-of-eight, with the most cancellation above and below, and the least cancellation occuring perpendicularly. However since the traces are so thin, emission is also weakest at the edges.

Given cool waveguide pictures like user 37518's avatar picture, going to guess at higher frequencies the emission at trace edges start to matter more, acting like light (The edge of a trace would be brightest if imagined to be translucent and passing photons).

 

[ruffrecords]

Quote from Squarewave:
What's anyone's take on how to best envision the cancellation field in 3d space at different frequencies when using the cancellation rule, source follows sink*?

What causes magnetic radiation is a loop. The bigger the area the bigger the radiation. Using return follows signal means the loop area is no more than the length of the track multiplied by the thickness of the PCB.

Cheers

Ian

 

[JohnRoberts]

Quote from Squarewave:
What's anyone's take on how to best envision the cancellation field in 3d space at different frequencies when using the cancellation rule, source follows sink*?

*I'm going to guess if you looked at a PCB edge-on, low frequency emission would look like a figure-of-eight, with the most cancellation above and below, and the least cancellation occuring perpendicularly. However since the traces are so thin, emission is also weakest at the edges.

Given cool waveguide pictures like user 37518's avatar picture, going to guess at higher frequencies the emission at trace edges start to matter more, acting like light (The edge of a trace would be brightest if imagined to be translucent and passing photons).

I don't know about 3D space but last century I had a senior (not old but very sharp) engineer working for me who used a technique were a single signal trace was bracketed by two adjacent return traces (we were limited to single sided PCB technology. ). As he explained it to me the "effective" center of the two traces were coincident for excellent cancellation of fields.

This may be more about electrostatic than magnetic fields. 

JR

 

[boji]

Thanks for your take guys.

What causes magnetic radiation is a loop.

Hmm. Ok current.  Sounds like there's no try for cancellation, it's just about packing the emissions as tight together as possible?

 

[abbey road d enfer]

Thanks for your take guys.

Hmm. Ok current.  Sounds like there's no try for cancellation, it's just about packing the emissions as tight together as possible?

That's correct. Just imagine; if you had no decoupling caps around a stage, the current delivered to the load would run to the PSU and back. That would be a big loop.
Using decoupling caps of adequate value allowscurrents to be drawn from the caps and returned to their reference point, making a very small loop, on the condition that the return is made to this very point, not to some distant "star point".
Also, reducing the size of the loop shift its radiating efficiency towards high frequencies, which are less of a concern for audio.

 

[Bo Deadly]

I am unclear on how much the proximity of the two conductors matters.

For example, say you have a guitar amp plugged into a mains socket at one end of a room, a space echo plugged into a mains socket at the other end of the room and then you create a ground loop by connecting the space echo ground to amp ground through an instrument cable. AFAIK that is essentially the same scenario. It's a 0V reference that connects to itself but creates a large loop area that acts like the turn of a transformer can converts the noisy magnetic fields present all around us into currents which are then converted to voltage sources at the inputs of non-differential amplifiers.

So if you change the above arrangement so that you run an extension cord from the same mains socket of the guitar amp over to the space echo and power it from that, will that help? Would it make a difference if the extension cord and instrument cable were closer together or would it just make things worse because mains currents would be right next to signal conductors?

 

[abbey road d enfer]

I am unclear on how much the proximity of the two conductors matters.

The closer they are, the smaller the area = the smaller the flux, so the smaller the voltage developed in the loop.

For example, say you have a guitar amp plugged into a mains socket at one end of a room, a space echo plugged into a mains socket at the other end of the room and then you create a ground loop by connecting the space echo ground to amp ground through an instrument cable. AFAIK that is essentially the same scenario. It's a 0V reference that connects to itself but creates a large loop area that acts like the turn of a transformer can converts the noisy magnetic fields present all around us into currents which are then converted to voltage sources at the inputs of non-differential amplifiers.

So if you change the above arrangement so that you run an extension cord from the same mains socket of the guitar amp over to the space echo and power it from that, will that help?

Yes. Because the loop would be considerably minimized. You have to consider the difference of conductance. If the earth connection is much stiffere than teh ground connection of signals (sleeve/shield), the voltage developed in the earth circuit will be almost totally superposed to the signal. That's why usin a cheat plug (eart-lift) reduces hum, at the risk of killing intelligent monkeys. And also why hum reduces when a strong ground conductor is added to teh connection.

Would it make a difference if the extension cord and instrument cable were closer together or would it just make things worse because mains currents would be right next to signal conductors?

Yes it would, but remember that currents in the loop are not only coming from picking up magnetic radiation; there are also those currents that come from capacitive coupling between mains and chassis, either by deliberately adding Cy capacitors or simply by the residual capacitance of the xfmr's windings.