Why Flip-Chip LGA is a Game Changer for RF ESD Protection

How S-parameter simulation results are stored

During the characterization of a device under test (DUT), a network analyzer (NWA) is connected to the DUT and energy waves are applied to the pins of the device. The input and output terminals (ports) of the NWA are where the transmission and reflection of the excitation waves are measured. S-parameter characterization results are stored in a Touchstone file whose format and name depend on the number of pins used. The most common file formats are s1p, s2p and s4p. The s2p file shown in Figure 1(a) has four entries, while the s4p file shown in Figure 1(b) has 16 entries.

Figure 2 shows a standard s2p measurement set-up for an ESD protection diode (characterized as a shunt to ground).

The principal flow paths for the different signal types during an s4p measurement are shown in Figure 3.

Table 1 explains the characterization information stored in each matrix parameter and the relationships between input and output port numbers.

Insertion loss (stored in matrix entries S21 and S12) for a typical device is shown in the graph in Figure 4(a) and illustrates how much of the input excitation signal passed through the DUT over the frequency range of interest, while return loss (S22 and S11), as shown in Figure 4(b), illustrates much of the input is lost due to signal reflections.

A crosstalk (XT) graph (Figure 5) indicates how much signal is likely to be transferred to the wrong port due to factors such as inductive and capacitive coupling. It typically differentiates between far-end crosstalk (FEXT) and near-end crosstalk (NEXT).

Single-ended S-parameters of passive networks under ideal conditions are considered to be symmetric, passive and causal, as explained in Table 2. For small-signal analysis, ESD protection devices are passive.

Measurement setup

Depending on product performance and package, Nexperia recommends measuring S-parameters using any of three principal setups: 

• Direct die contact with air coplanar probes (ACP).
• Contact on adapting PCB without de-embedding.
• Contact on adapting PCB with de-embedding.

Direct die contact with ACP probes 

Figure 6 shows probes, which are connected to the NWA, placed directly in contact with a DUT. This setup is often used for two-pin products with a small bandwidth. The frequency range of this measurement is typically from 300 kHz to 18 GHz, and results are stored in a s2p file.

Contact on adapting PCB without de-embedding

As application data rates increase, higher bandwidth ESD protection devices are required, meaning S-parameter characterization must also be performed at higher frequencies. This requires the use of a small adapting PCB to enable measurements at frequencies ranging from 10 MHz up to 40 GHz.

A methodology called de-embedding can be used to calculate standalone device data, independent of the PCB on which it is mounted. For this approach, a set of calibration structures is required to subtract the PCB behavior from the measurement data. The resulting S-parameters are stored in s2p files describing the device behavior over frequencies ranging from 1 GHz to 40 GHz. This measurement technique is relatively new and not yet applicable to all lower capacitance devices, but it is generally available for devices in SOD962-2 and SOD882 packages. More details on this technique are available in Chapter 6.2 of Nexperia’s ESD Handbook.

Interpreting the Touchstone file

The Touchstone format was created by HP in 1984 and is now the industry standard for storing S-parameter measurement results. These files can be opened using most text editors. Nexperia usually provides *.s2p and *.s4p files for its ESD protection devices, and the meaning of headers typically included at the top are described in Table 3, while an example s4p file for ESD protection device is shown in Figure 7.