Differential Signalling

Hi Hari,

This is partially true, consider a transformer isolated ethernet signal on a twisted pair cable.  In this case the only conductors are the "+" and "-" signal wires so all the return current goes through the other wire.  

Within a system differential signals reside on micorstrip or stripline traces which couple to each other and to a reference plane (usually GND).  This coupling promotes a return current for each "+" and "-" signal in the reference plane itself in addition to the differential current in the "+" and "-" signals.  Considering this current is important for both signal integrity and EMI.

Regards,

Lee

In addition to blogs and application notes on ti.com, there is some really good information on the web about this topic.

I tend to search for Eric Bogatin - I find his books and papers informative and easy to read.

Here is a searchable link. see slides 17-23

hi Hari,

Here is a short paper on the topic.  Just paste the link below into your browser.

electronix.ru/.../index.php

Also there are some very good online presentations that can be searched on Google etc.

Regards,

Lee

Hi Lee,

Thank you very much.Lee one more doubt,Actully what you mean by High speed signal,can we consider 200Mhz as a high speed signal.When will we consider  a signal as High speed signal.

Regards

Hari

Hari,

I prefer to think of all signals as high speed and all interconnects as transmission lines.  This leads to good design decisions and avoids unexpected issues.  

For a 200 MHz signal it is likely the device edge rates will have energy up to ~ 1 GHz. Even a relatively short interconnect will begin to exhibit transmission line effects.  For a differential signal in a point-to-point topology (one Tx -> one Rx), 200 MHz is easy to achieve and takes minimal effort.  Just design for the recommended transmission line impedance, terminate at the receiver, ensure compatibility between the Tx - Rx pair.  On the other hand driving a complex multi-drop topology (one Tx -> many Rx) with single-ended LVCMOS technology at 200 MHz may be a real challenge.  

Regards,

Lee

Common-mode chokes, where two coils are wound on a single core, are useful for prevention of electromagnetic interference (EMI) and radio frequency interference (RFI) from power supply lines and for prevention of malfunctioning of electronic equipment. They pass differential currents (equal but opposite), while blocking common-mode currents. Magnetic fields produced by differential-mode currents in the windings tend to cancel each other out. Thus, the choke presents little inductance or impedance to differential-mode currents. This also means the core will not saturate even for large differential-mode currents and the maximum current rating is instead determined by the heating effect of the winding resistance. Common-mode currents, however, see a high impedance due to the combined inductance of the windings.

Additional searches for Common mode chokes will yield more reference material to read through.

Regards,

Lee