While waveguide has been at 110 GHz for a while, the advent of more commercial mmWave applications such as wireless back haul and automotive radar are creating the need for a flexible cable option. Whether used in conjunction with waveguide (e.g., hybrid waveguide-coax-waveguide) or standalone (e.g., a vector network analyzer (VNA) test lead or cable connection between modules), the demand for flexible cable assemblies that operate to 110 GHz is increasing.
This article discusses the technical hurdles and associated decisions to develop a high performance 110 GHz cable assembly, including the 1) cable, 2) connector, 3) test and 4) preparation and resources. These are highly interdependent variables, such that discovering the cause of a problem during development is a combination of science and experience.
CABLE
The flexible cable size chosen for the design is a 0.055 in. (1.4 mm) outside diameter (OD) with jacket, which is a standard size within the industry with an upper frequency above 110 GHz (W-Band). There are other sizes, mostly smaller, that can achieve these frequencies; however, more connector choices are available for the 0.055 in. OD, including 1 mm, SMPS and the proprietary variants of MM4S, G3PO and G4PO.
There are design choices and material challenges for cable construction, depending on which attributes are the primary focus. Is insertion loss more important than ruggedness? Typically, you cannot have both. If loss is important, then a microporous PTFE tape is used, which makes the cable more prone to damage with normal handling. The more rugged choice is an extruded PTFE dielectric, which has an insertion loss penalty. After careful consideration, ruggedness wins, because the target market is test and measurement, where the environment has movement, a fast pace and people are used to a more robust cable. The slightly higher loss is mitigated by the applications that typically use short lengths of cable.
The next design choice is frequency range: should the cable be broadband or one constructed strictly for E-Band (60 to 90 GHz) and W-Band (75 to 110 GHz)? Not knowing how the new E-Band and W-Band applications will use lower frequencies, a broadband cable design is chosen.
Materials and cable constructions that work perfectly well at V-Band and below (i.e., ≤ 70 GHz) sometimes show nonlinear responses above V-Band. When this occurs, there is a diagnostic hunt to identify the problems, with many potential culprits. The factory environment—compressed air and electricity, for example—can introduce intermittent anomalies in the manufacturing process. A fault length resulting in a nonlinear electrical response can be induced either by equipment or materials. Also, a design with tape or wire overlaps can cause periodicity in the electrical performance. Variations in materials from suppliers can be frustrating: one good initial lot of raw material followed by a number of lots of flawed material make the diagnostic hunt more intuitive than scientific.
