December 15, 2020
Scientists and engineers are conducting considerable research and development in millimeter wave (mmWave) frequencies, especially in the W band (75 to 110 GHz). Some of this effort is fueled by new commercial applications in automotive radar and wireless communications, including 5G. Finding test solutions that produce accurate and repeatable results is critical for the continued use of high frequencies in such advanced designs.
It’s particularly important for many emerging environments. For example, projections are that billions of IoT devices will communicate by 2030. Notable applications include handset communications, apps, Internet streaming, machine learning, and AI.
What will need to change in the test world to address UE verification to ensure such high bandwidth data can effectively transmit through networks? The simple answer is an interface that can propagate a broadband signal with little impact on signal fidelity.
Test ports on most vector network analyzers (VNAs) and similar test equipment are coaxial. Because of this, any connection to a native waveguide interface requires an adapter. Besides the effort of characterizing the adapter and de-embedding the network for accurate measurement analysis, adding an adapter can be costly at frequencies above 100 GHz.
Repeatable measurements are critical when verifying any product but more so for today’s high-frequency designs. Waveguides are well known for their ability to change performance between users. To remove this issue, many design steps have been taken to help increase repeatability.
One approach is to use guiding pins, which align apertures and ensure the interface is matched perfectly. There are also holes in waveguide components that allow interfaces to be fastened to each other using screws (Figure 1). While there are numerous benefits, one drawback is it can be time consuming between measurements.
Limitations of Waveguides
Historically, waveguides are the first component used for measurements up to 100 GHz and higher. The biggest trade-off with waveguides is that they are band limited.
Influence of Power
Power is another factor engineers must consider. RF power transmitted and received through a device is not the only concern. Many active devices require an external direct current or DC (0 Hz) signal input, known as a bias. Often this bias needs to be applied on the same conductors as the RF.
Any high-frequency transceiver used in telecommunications or automotive radar will be accompanied by an amplifier. How does an engineer test those used past 100 GHz? In many RF applications, bias tees are used. No such component exists in the waveguide world.
While large bandwidth coverage provides much more resolution into the impedance characteristics and low pass time domain processing, it also requires a harmonic calibration. It is impossible to achieve an acceptable harmonic calibration using a component that has the lowest frequency at 75 GHz but only spans to 110 GHz. This means that low pass type of measurement is unavailable with waveguide.
Dispersion is a waveguide characteristic that can manifest as a distortion factor on the pulse (Figure 2). Most measurement equipment can handle an analysis of this distortion through the mechanism of dispersion. It can present issues, especially for other time domain techniques.
Benefits of Coaxial Connectors
For all these reasons, coaxial connectors are a better choice for high-frequency designs. Here are four notable advantages:
Broadband frequency coverage/scalability. Coaxial connectors do have frequency limits, but they only exist on the upper frequency. Coaxial connectors do not use apertures to define their frequency coverage. The frequency coverage of a coaxial connector is defined by the cutoff frequency of the next higher order mode dependent on the connector geometry, specifically the radii of the inner and outer conductor.
Coaxial connectors are similar to waveguides in that they typically operate best in mono-mode conditions. The next higher order mode is based on a TE11, similar to waveguides, that occurs near the theoretical limit of the connector. The explanation for this behavior is that it is not feasible for connectors to use support beads as an all air dielectric. From the interface of these support beads to the air dielectric, input signal energy can exchange between modes. Impedance and phase velocity mismatch become resonances or modes.
Frequency Domain. Having broadband frequency coverage for coaxial connectors allows the interface to be used for active device, high-speed digital applications, and corresponding harmonic testing. They still address mmWave frequency requirements, as well. Coaxial connectors use transverse electromagnetic (TEM) propagation, which is not affected by dispersion. Because of the DC start frequency of the coaxial interface, biasing or conducting a DC signal is not an issue either.
Time Domain. Another advantage of the broadband nature of coaxial connectors with a start frequency of DC is all time domain measurements can be made. Low pass time domain and eye diagrams can be tested with even higher resolution, as the frequency moves to 100 GHz and beyond. Engineers can have greater confidence in many high-speed designs because an eye diagram or impedance graph provides insight into signal integrity of the interface, as shown in figure 3.
Measurement Setup. A coaxial interface aids in test system configuration, as well. It allows for a direct connection between interfaces, eliminating the need for alignment screws and guide pins. Issues of repeatability are removed with a coaxial interface. With the use of a proper torque wrench, measurements can be conducted repeatedly with correlating results.
To learn more about the benefits of using coaxial connectors with VNAs in high-frequency designs, download an Anritsu application note.
