Continue to Site

# Reflection (ringing) issue for transmission line at different frequencies

Status
Not open for further replies.

#### henry kissinger

##### Member level 2
I have a circuit as below.

It consists of a data generator source with source impedance=35 Ohm. Data generated is square wave with certain rise and fall time defined in the schematic.
A transmission line has all the parameters adjusted to have the characteristic impedance of 60 Ohm at frequency=5G Hz (sine wave)
The Load impedance is 60 Ohm.

With the data rate of the source as 1G Hz (square) = 5G Hz (sine) (approximated based on same rise time and fall time), we can see that the voltage at node "V" is good because the load impedance is matched to the characteristic impedance of the line, there is no reflection to distort the signal at "V".

However if I increased the data rate of the source to 5G Hz (square) = 18G Hz (sine), the voltage at node "V" experiences ringing caused by reflection? but why is there a reflection? as the load impedance is matched to characteristic impedance of the line.

To prove that the characteristic impedance remain the same between 1G Hz (square) and 5G Hz (square). I plotted the graph below.
As you can see the S11 is almost the same between 5G Hz(sine) /1G (square), and 18G Hz(sine) /5G(square).
So why is there reflection at data rate of 18G(sine)/5G Hz(square) and causes the node "V" to ring? given that the characteristic impedance of the line is the same based on the fact that S11 stays the same.

There is always a reflection. The line is not settled until
2x "time of flight" when the wavefront returns to the source.
At best; mismatches create -multiple- reflections, manifesting
as either ringing or staircasing, until all the energy has been
scrubbed off by the termination.

You can have intersymbol interference when line length
exceeds symbol rate "length" and when the line is at all
mismatched, this gets pretty chaotic (in my limited
experience with baseband transmission line drivers &

Can you explain what causes the below? because I only know that impedance mismatch causes reflection! what is the mechanism of it? Thank You!

"There is always a reflection. The line is not settled until
2x "time of flight" when the wavefront returns to the source."

There is always a reflection. The line is not settled until
2x "time of flight" when the wavefront returns to the source.
Assuming an ideal transmission line (frequency independent impedance), there's no reflection if the line is terminated with its characteristic impedance. The observed ringing proves that the impedance in the experiment isn't frequency indepedent respectivly exposes higher modes.

Please compare with an ideal line.

Real micro strip lines are hardly ideal in the discussed frequency range. That's why you want a correct termination at both ends which further reduces the residual reflections.

I used to prefer source termination and if you look at the "front end of the line" you will see it "hanging out" at half transition until 2 ToF, while the signal at the far end transitions fully at 1 ToF.

Of course nobody (hardly) cares at the TX end, but this behavior demonstrates that the "bit transition energy" has not left the line entirely until the round trip (hence must be a reflection) has done.

Mismatch makes > 1 reflection as seen in the staircase / up/down activity with each "step" 2X ToF (and perhaps attended by L, C-mismatch HF ringing).

Status
Not open for further replies.