TDR, FDR, and VNA Techniques for Signal Integrity

PCB impedances can adversely affect signal quality and EMI. Make sure those impedance values are what you expect.

Uncontrolled or unexpected impedances on printed circuit board (PCB) traces and power/return planes can negatively impact the performance of electronic devices. Those impedances result from unavoidable but controllable parasitic capacitances and inductances that result from PCB structures and materials. Because you can control them, PCB traces and power/return plane measurements are important aspects of design verification, validation, and production testing. Time-domain reflectometers (TDRs), frequency-domain reflectometers (FDRs), and vector network analyzers (VNAs) let you measure PCB impedance.

A reflectometer is really two instruments: a pulse generator that produces fast-rising pulses and a high-speed sampling oscilloscope that captures and measures the reflected pulses.

As Figure 1 shows, TDR sends a fast-rising pulse down a trace. Because of its fast rise time, the pulse contains a wide range of frequencies. By analyzing the strength and polarity of the reflected signal, you can calculate the impedance of a trace. The test cable and output of the TDR are usually 50 Ω. Impedance matching is important. Mismatches lead to higher capacitance or higher inductance at the load.

Figure 1. A time-domain reflectometer sends a pulse along a trace or wire and measures the response. (Image: HV Technologies)

Instead of using a single fast-rising pulse, an FDR measures the impedance of a trace by sweeping a range of frequencies. The instrument measures reflected energy at multiple frequencies and provides a picture of the impedance across the frequency range.

While TDRs and FDRs measure reflected energy, a VNA can measure the transmission and reflection of a trace to determine its S-parameters. The S-parameters can be used to calculate the impedance of the trace. Transmission measurements can include gain, insertion loss, phase, delay, and so on. Some VNAs are available with software that adds frequency domain measurement capabilities (Figure 2).

Figure 2. Example of a VNA that can make frequency domain measurements using optional software. (Image: Keysight)

Using coupons

Most PCBs include a dedicated area called a test coupon designed for connecting to TDRs, FDRs, and VNAs to measure and validate trace impedance. Test coupons have specific dimensions, such as trace widths, spacings, hole diameters, impedances, and so on.

Test coupons used for measuring power-delivery network (PDN) impedance often use different shapes to identify the power and ground planes, like square pads for the power plane and round pads for the ground plane (Figure 3). In addition, there are often multiple voltage domains that need to be identified and tested individually.

Figure 3. Typical test coupon for PDN impedance measurements. (Image: Altium)

PDN impedance challenges

In addition to dealing with multiple voltage domains, the PDN is usually designed before the final signal bandwidths and impedances of the overall system are known. That means that the PDN impedance needs to be low from DC to very high frequencies, depending on the expectations for the final design. When developing the PDN, it’s important to include all the expected decoupling capacitors and measure the impedance vs. frequency for the whole assembly.

When multiple power supplies are used on the same PCB, the test needs to confirm that the decoupling capacitors are correct for each power plane or supply voltage. It’s usually recommended that you use two coupon structures for each power plane or voltage, one for injecting a signal and the second for measuring the resulting impedance performance.

Compensate for test fixture & probe variations

Like all measurements, impedance measurements are not perfect and can deliver varying levels of accuracy. One challenge is minimizing the effects of the test fixture and probes, which can cause inaccurate measurements. Test systems are available with various tools for compensating for variations in test fixture and probe performance.

For example, you can use mathematical tools such as port extension or “deskew” to extend the calibration reference plane to the device under test (DUT). This technique removes delay from the test but assumes that the test fixture and probe behave as a perfect transmission line. While it’s relatively easy to implement, it only provides good results when using high-quality probes and fixtures.

Explicitly adding loss compensation also extends the calibration plane to the DUT and mathematically eliminates the delay and loss from the test fixture and probe. This solution is more complex to implement but provides improved results.

Fully calibrating the test setup is even more complex but produces the highest accuracy by removing delay, loss, and mismatches from the measurement process. Recently developed electronic calibration (e-cal) tools are a software-based approach that also produces the highest accuracy. Instead of manually calibrating the test setup, an e-cal module is attached to it, and software takes over and provides the calibration.

Summary

Measuring the impedances of PCB traces and power/return planes is an important activity to ensure optimal system performance. In many cases, access points called coupons are included on PCBs to enable efficient impedance testing. In addition to selecting the proper instrument for the situation, several techniques are available for compensating for imperfections in test fixtures and probes and improving the accuracy of test results.