scope:

Time domain reflectometry, TDR, is used to measure reflections and time delays of pulses injected into a transmission line. Discontinuities in characteristic impedance, Zo, along the line under test, reflect portions of incident pulses back toward the source where they combine with the injected pulse to produce a voltage-time waveform.

Applicability

If the rate of pulse propagation (distance/time delay) along the line is uniform, then the distance to changes in the line's Zo will be proportional to observable times for voltage changes in the waveform. The observed voltage changes are related in magnitude and direction to the Zo change at the related discontinuity. Thus the TDR method is useful for measuring Zo and locating changes in Zo of a transmission line and is capable of detecting and characterizing defects or the influence of various features along the length of a line.

Conversely, if the locations and types of discontinuity in a line are known, the observed times for corresponding voltage changes in the waveform are useful for determining the time delay of pulses. The time delay is related to the effective permittivity of the dielectric and the transmission structure.

TDR measurements of Zo are used for both engineering development and manufacturing control.

Engineering development requires detailed information on the electrical performance of prototype units to assure the design yields the desired result. Analysis of the effect of variations in design features expected in actual manufacture can be done to assure the proposed article can be manufactured at a useful quality level.

Measurements for manufacturing control are performed to identify and correct process or materials problems occurring during a manufacturing run as well as to assure that a product will perform electrically as designed. Examples of parameter variations detectable by TDR, and that are evidence of process or materials problems, include the following:

Over/under etching (line width problems)

Over/under plating (line width and thickness problems)

Permittivity of the dielectric

Thickness of the dielectric

Residues from process steps including cleaning

Degradation from excessive heating and humidity

Damage from excessive pressure during the multi-layer process.

Increased performance requirements for computers and other electronic products often demand even greater signal fidelity, time precision, and noise immunity than can be easily obtained with a single ended transmission line. That is, a line geometry consisting of a signal trace over ground plane in microstrip or between ground planes in stripline. Differential lines are employed to increase signal fidelity with improved time precision and increased noise immunity. Single ended lines may be called unbalanced transmission lines. Differential lines may be called either balanced or coupled transmission lines. The required TDR method is different for differential lines.

Limitations

Measurements of Zo often vary greatly depending on equipment used and how the tests were performed. Following a specified method helps assure accurate and consistent results. Both "single-ended" and differential line measurements have limitations in common including the following:

The Zo measured units are derived and not directly measured.

These units are for ideal transmission lines where the electrical characteristics do not vary along the length. The measured Zo is only a value that approximates the characteristic impedance of an ideal line that is representative of the line under test.

Lines on a PWB only approximate uniformity. Microstrip lines longer than 6 inches on plated through hole boards often have trace width variations of 40% along the length of the line.

The accuracy of measurements can not be directly traced to a referee standard at NIST (National Institute of Standards and Technology.)

A variety of methods for TDR measurements each have associated accuracies and repeatabilitiies.

If the nominal impedance of the line(s) being measured is significantly different from the nominal impedance of the measurement system (typically 50 ohms), the accuracy and repeatability of the measured numerical valued will be degraded. The greater the difference between the nominal impedance of the line being measured and 50 ohms, the less reliable will be the numerical value of the measured impedance.