Printed Circuit Board Signal Loss Test Technology

Printed circuit board (PCB) signal integrity has been a hot topic in recent years. There have been many research reports in China that analyze the factors affecting PCB signal integrity [1]-[4], but there are relatively few reports on the current status of signal loss testing technology.

The signal loss of PCB transmission lines comes from the conductor loss and dielectric loss of the material, and is also affected by factors such as copper foil resistance, copper foil roughness, radiation loss, impedance mismatch, crosstalk, etc. In the supply chain, the acceptance indicators between copper clad laminate (CCL) manufacturers and PCB express factories use dielectric constant and dielectric loss; while the indicators between PCB express factories and terminals usually use impedance and insertion loss, as shown in Figure 1.

For high-speed PCB design and use, how to quickly and effectively measure the signal loss of PCB transmission lines is of great significance for the setting of PCB design parameters, simulation debugging and production process control.

2 Current Status of PCB Insertion Loss Test Technology

The PCB signal loss test methods currently used in the industry can be classified into two categories based on the instruments used: based on the time domain or based on the frequency domain. The time domain test instrument is the time domain reflectometer (Time Domain Reflectometry, referred to as TDR) or the time domain transmission meter (Time Domain Transmission, referred to as TDT); the frequency domain test instrument is the vector network analyzer (Vector Network Analyzer, referred to as VNA). In the IPC-TM650 test specification, 5 test methods are recommended for testing PCB signal loss: frequency domain method, effective bandwidth method, root pulse energy method, short pulse propagation method, and single-ended TDR differential insertion loss method.

2.1 Frequency domain method

The frequency domain method mainly uses a vector network analyzer to measure the S parameters of the transmission line, directly reads the insertion loss value, and then uses the fitting slope of the average insertion loss in a specific frequency range (such as 1 GHz to 5 GHz) to measure the pass/fail of the board.

The difference in frequency domain measurement accuracy mainly comes from the calibration method. According to the different calibration methods, it can be divided into SLOT (Short-Line-Open-Thru), Multi-Line TRL (Thru-Reflect-Line) and Ecal (Electronic calibration) electronic calibration.

SLOT is generally considered to be the standard calibration method [5]. The calibration model has a total of 12 error parameters. The calibration accuracy of the SLOT method is determined by the calibration parts. High-precision calibration parts are provided by the measurement equipment manufacturer, but the calibration parts are expensive and generally only applicable to coaxial environments. Calibration is time-consuming and increases geometrically with the increase in the number of measurement ends.

The Multi-Line TRL method is mainly used for non-coaxial calibration measurements [6]. The TRL calibration piece is designed and manufactured according to the material of the transmission line used by the user and the test frequency, as shown in Figure 2. Although the Multi-Line TRL method is easier to design and manufacture than the SLOT method, the calibration time of the Multi-Line TRL method also increases exponentially with the increase in the number of measurement terminals.

In order to solve the problem of time-consuming calibration, measurement equipment manufacturers have introduced the Ecal electronic calibration method [7]. Ecal is a transfer standard. The calibration accuracy is mainly determined by the original calibration component. At the same time, the stability of the test cable, the repeatability of the test fixture device and the interpolation algorithm of the test frequency also have an impact on the test accuracy. Generally, the reference surface is calibrated to the end of the test cable using an electronic calibration component, and then the cable length of the fixture is compensated by de-embedding. This is shown in Figure 3.

Taking the insertion loss of a differential transmission line as an example, the comparison of three calibration methods is shown in Table 1.

2.2 Effective Bandwidth Method

The effective bandwidth method (EBW) is strictly speaking a qualitative measurement of the transmission line loss α. It cannot provide a quantitative insertion loss value, but provides a parameter called EBW. The effective bandwidth method uses TDR to transmit a step signal with a specific rise time to the transmission line, and measures the maximum slope of the rise time after the TDR instrument and the device under test are connected. The result is determined as the loss factor in MV/s. More specifically, it determines a relative total loss factor, which can be used to identify the change in the loss of the transmission line between surfaces or layers [8]. Since the maximum slope can be directly measured from the instrument, the effective bandwidth method is often used in mass production testing of printed circuit boards. The schematic diagram of the EBW test is shown in Figure 4.

2.3 Root Pulse Energy Method

The root pulse energy method (RIE) usually uses a TDR instrument to obtain the TDR waveforms of the reference loss line and the test transmission line, and then performs signal processing on the TDR waveforms. The RIE test process is shown in Figure 5:

2.4 Short Pulse Propagation Method

The principle of the short pulse propagation (SPP) test is to use two transmission lines of different lengths, such as 30 mm and 100 mm, to extract the parameter attenuation coefficient and phase constant by measuring the difference between the lengths of the two transmission lines, as shown in Figure 6. This method can minimize the influence of connectors, cables, probes and oscilloscope accuracy. If a high-performance TDR instrument and IFN (Impulse Forming Network) are used, the test frequency can be as high as 40 GHz.

2.5 Single-ended TDR Differential Insertion Loss Method

The single-ended TDR to Differential Insertion Loss method (SET2DIL) is different from the differential insertion loss test using a 4-port VNA. This method uses a two-port TDR instrument to transmit the TDR step response to the differential transmission line, and the end of the differential transmission line is short-circuited, as shown in Figure 7. The typical measurement frequency range of the SET2DIL method is 2 GHz ~ 12 GHz, and the measurement accuracy is mainly affected by the inconsistent delay of the test cable and the impedance mismatch of the device under test. The advantage of the SET2DIL method is that it does not require the use of expensive 4-port VNA and its calibration components. The length of the transmission line of the device under test is only half of that of the VNA method. The calibration component has a simple structure and the calibration time is greatly reduced. It is very suitable for batch testing of PCB manufacturing, as shown in Figure 8.

3 Test equipment and test results

The SET2DIL test board, SPP test board and Multi-Line TRL test board were made using CCL with a dielectric constant of 3.8, a dielectric loss of 0.008 and RTF copper foil. The test equipment was a DSA8300 sampling oscilloscope and an E5071C vector network analyzer. The differential insertion loss test results of each method are shown in Table 2.

4 Conclusion

This article mainly introduces several PCB transmission line signal loss measurement methods currently used in the industry. Due to different test methods, the measured insertion loss values ​​are also different, and the test results cannot be directly compared horizontally. Therefore, the appropriate signal loss test technology should be selected based on the advantages and limitations of various technical methods and combined with one's own needs.