In describing what happens when an electromagnetic field or wave is traveling on a transmission line, we often talk about the various things that can disrupt that wave. Among the most common of these things are reflections. Part 1 of this article will describe what reflections are and how they are caused. The second part will discuss what they do to a signal. Then, a separate article will address how reflections are controlled by terminations.
Note: In numerous articles provided by Speeding Edge, we have touched on the topic of reflections but this article is the first “ground-up” examination of terminations. Some of those previous articles include:
- PCB Design For Test Structures And Tests Performed, Part 2
- It’s Your Loss: Determining And Controlling Loss In The PCB Design Process
- The Evolution Of Simulation Technology and Crosstalk
- High-Speed PCB Design: Just How Fast Is Fast?
Other articles on the Altium blog discuss various other problems with transmission lines, such as why their length matters.
Some Basic Elements
When an electromagnetic wave or field is launched down a transmission line, if there is no absorber or terminator at the load end of the line all of the energy, which is contained within the EM field, is reflected back toward the source or the driver. If this reflected energy is unwanted there has to be a way to absorb it at the load end. This is done through the use of a parallel termination. If it’s not possible to use this type of termination, the energy will reflect back down the transmission line and reflect off the source or driver unless that source or driver is of such an impedance that it can absorb it.
This type of absorber is called a series termination. Allowing the energy to reflect in this manner is often referred to as reflected wave switching.
What Happens When Energy on a Transmission Line Is Reflected?
When some of the energy in the EM field is reflected back toward the source, some of the energy that is in that field and continues on is diminished by the amount of energy that has been reflected. The amount of reflection and its polarity can be calculated using the following equation.
The Reflection Equation
This equation predicts the percentage of the incident EM field that will be reflected back to the source based on the two impedances on each side of a change where Zl is the downstream impedance and ZO is the upstream impedance. The equation reflects the voltage amplitude of the reflection.
As can be seen, when the load or downstream impedance at a change is higher than the upstream impedance, the polarity of the reflection is positive. The reflected voltage adds to the incident voltage. If the downstream impedance is infinity or the transmission line ends in an open circuit, all of the incident voltage is reflected. In this scenario, the entire voltage amplitude and all of the energy in the EM field is reflected back to the source and adds to the incident voltage level. This results in double the voltage at the open circuit end of the line.
On the other hand, when the impedance at a change is lower than the line or the upstream impedance, the polarity of the reflection is negative. The reflected voltage subtracts from the incident voltage. In the most extreme circumstance, if the downstream impedance is a short circuit or zero ohms, all of the energy in the EM field reflects back toward the source. None of the energy is absorbed. Here, the reflected voltage is inverted and subtracts from the incident voltage yielding zero volts. The figure below shows a 50-ohm transmission line driven by a logic signal. It is terminated in an impedance that has three different values.
A 50-ohm Parallel Terminated Transmission Line
The waveforms appearing at the output of the driver are plotted in the graph below.
Output Voltage Waveforms for the circuit shown above
First, a zero to one logic transition takes place with a terminating resistor of 50 ohms that has the same value as the line impedance. Here, the resulting signal, which has no reflection, is the center trace. The termination absorbs all of the energy in the EM field.
Next, the value of the termination is set to 70 ohms which yields a downstream impedance that is higher than the upstream impedance. Some of the energy in the EM field is reflected back to the source or the driver. The polarity of this reflection is positive and adds to the incident voltage level. The name given to this “positive” or “adds to” reflection is overshoot. Normally, this kind of reflection does not degrade the logic signal levels, so it is not usually harmful. However, it is harmful if the sum of the incident voltage level and the reflected voltage exceeds the voltage rating of a logic circuit. This doubling that engineers see at a load can be mystifying and often leads to the inappropriate addition of AC
When the value of the termination is set to 30 ohms it creates a downstream impedance that is lower than the upstream impedance. Again, some of the energy in the EM field is reflected back to the source or the driver. The name given to this “negative” or “takes away” reflection is undershoot. Sometimes, this type of reflection is designated as “ring back” but this is an unnecessary nametag. Undershoot is the proper designation. What’s important to remember is that undershoot degrades the logic level and is always harmful. During the design process, when a product developer has any control over the way in which impedance mismatches occur, it is desirable to ensure that downstream impedances are always equal to or higher than upstream ones. One method for accomplishing this is through the use of termination resistors.
In the graph above, there is a rising edge or a logic transition from a 0 to a 1. If the switching edge is falling or the logic transition is from a 1 to a 0 (producing overshoot) and there is a termination resistor value that is higher than the line impedance, the results is a negative-going reflection. This reflection still adds to the incident voltage level as in the example cited earlier. Even though the reflection is negative going, it still results in overshoot because the incident logic level has been “added to”. When the termination resistor is smaller in value than the line impedance, the reflection still subtracts from the incident logic signal. Hence, although the reflection is positive-going, it still results in undershoot. With the concept of reflections and what causes them in mind, it’s now possible to illustrate the behavior of reflections and how they impact signals. This will be addressed in Part 2 of this article.
About the AuthorMore Content by Kella Knack