[Moved]: EMI and Crosstalk Correlation Studies

[engineer-labs]

Is there a direct correlation between Crosstalk and EMI? If so, how does it work?

I'm trying to find papers or source material that discusses this at length; with measurements/graphs but can't find any.

Basically, what I want to find out is what are the physics involved between the two.

 

[erikl]

Changing voltages or currents generate electromagnetic fields, these again generate voltages or currents at nearby conductors by capacitive and/or inductive interaction - called EMI, or crosstalk in audio range.

 

[D.A.(Tony)Stewart]

EMI can be conducted or radiated.

Just like the voltage divider with 2 resistors, if conducted interference can induced by the impedance ratio of the source ( L , C and/or R) times the source voltage or current.

If radiated , in near fields it can also be thought of in the same way.

A simple example is capacitive and inductive coupled lines with a low source impedance and a high impedance at load can be determined by the series and shunt Z ratios times the source levels.

All PCB tracks have distributed inductance and capacitance if there is a nearby ground path that depends on their track width and gap and the dielectric or magnetic coupling .

Since impedance of C drops with rising frequency, crosstalk will also increase with C coupling related noise until transmission line effects at 1/4 wave and 1/2 wave ...3/4 .. etc can invert impedance and create either resonance or nulls.

Noise can be suppressed with differential and common mode capacitance and/or inductance depending on the interference. It can also be suppressed with shielding by diverting these ratios to ground or filtering out unwanted frequencies.


This is just a thumbnal of ideas that fill up an excellent large book by Henry Ott.

For far fields, the distance and effects of frequency and the medium path from the transmitter to the receiver are defined by Friis Loss Equation.

 

[criterion456]

Is there a direct correlation between Crosstalk and EMI? If so, how does it work?

I'm trying to find papers or source material that discusses this at length; with measurements/graphs but can't find any.

Basically, what I want to find out is what are the physics involved between the two.

As erikl said :

Changing voltages or currents generate electromagnetic fields

You can refer to Maxwell equation, or any Electromagnetics text books.



Or we can refer to Fourier Series :



As illustrated above, higher the frequency, more the di(t)/dt, more EMI emission.

Thus, in terms of high speed digital,

there will be transient current (Delta-I noise) during the signal transition.