PDA

View Full Version : Transmission line matching

likephysics
Apr17-11, 10:16 PM
In high speed digital designs, there is a series terminating resistor to match the output impedance of the driver IC with the transmission line on the PCB. Some think the purpose of this resistor is to decrease the drive of the IC.
How do I convince them its not to decrease the (current) drive but to match the impedance?
Antiphon
Apr17-11, 11:47 PM
Teach them transmission line theory.

Actually it's not there to match the impedance of the output of the driving IC but the impedance of the lines on the board.
sophiecentaur
Apr18-11, 09:39 AM
I guess you could tell them about the effects of echos on pulse shape and inter-symbol interference. Also about the 0.4m per nanosecond speed of signals on transmissions lines.
sophiecentaur
Apr19-11, 07:32 AM
More like 0.2m per nanosecond, actually.
likephysics
Apr20-11, 09:51 AM
I guess you could tell them about the effects of echos on pulse shape and inter-symbol interference. Also about the 0.4m per nanosecond speed of signals on transmissions lines.

This would be too much to digest.
likephysics
Apr20-11, 09:55 AM
The question I am struggling with is - why do reflection only occur when the tx line approaches signal wavelength.
Why doesn't this occur at wave length >> tx line.
I'm not able to come up with a good explanation.
skeptic2
Apr20-11, 12:58 PM
Reflections do occur when transmission lines are shorter than a wavelength it's just that when the time of reflection is less than a cycle, the echo isn't perceived so much as an echo but as a change in impedance of the load. For instance, if a transmission line is 1/4 wavelength and is left unterminated (open) the signal will travel 1/4 wavelength, bounce off the open, and reflect back to the source 180 deg out of phase with the transmitted signal. To the source it appears the load is a short not an open.

In testing RF amplifiers we frequently used stretch lines (or trombone lines) with a short on the end to determine if there were instabilities at any reflection phase angles.
Bob S
Apr20-11, 04:28 PM
The question I am struggling with is - why do reflection only occur when the tx line approaches signal wavelength.
Why doesn't this occur at wave length >> tx line.
I'm not able to come up with a good explanation.
Understand a quarter wave (transmission line) transformer at

http://en.wikipedia.org/wiki/Quarter-wave_impedance_transformer

then you will see that reflections always occur when the termination is mismatched.

Bob S