Analog Angle Blog

3D printing challenges multi-GHz component constraints

It’s not news that operating frequencies for wireless and even wired circuits are rapidly moving up the spectrum. The reality is that while it wasn’t long ago that operating at just a gigahertz or two was considered a testbench accomplishment, we now have mass-market consumer products designed for the multi-GHz range and moving up fast with 5G. The physical implications are well known; as frequencies increase and wavelengths shrink, so do the relevant dimensions and allowable tolerances of components, board tracks, and well, everything.

At these tiny dimensions, making and using even basic components such as connectors are a major challenge. Take, for instance, coaxial cables that have diameters on the order of a millimeter or two. Connectors and waveguides always had tight dimensional tolerances and the need for some ruggedness. However, the problem is now intensified as previously “lesser” issues such as surface finish and smoothness now can significantly affect performance of connectors, circuit-board laminates, and more. Fabricating surfaces and connectors at these dimensions is, in many ways, a non-silicon, all-metal version of MEMS.

Researchers and commercial vendors are now investigating how to possibly solve the problem with the help of a more-precise version of another leading-edge technology that has gone mainstream. Stereolithography, known by the acronym SLA, is also called 3D printing or additive manufacturing (AM). It’s being used to fabricate tiny connectors and other components where conventional precision machining or etching is running into barriers. This is especially the case in low-to-moderate volumes where the set-up and tooling effort and cost are relatively high. Yet the volume can’t justify costly dies, molds, fixtures, and whatever else it takes to make these minute devices with the dimensional precision and finish needed.

But why stop with passive devices? One very interesting example of an active device with integral waveguides—those component interconnects are a big part of the challenge—has originated from a team at the University of Birmingham in UK. They designed and fabricated a 62.5 GHz to 125 GHz Schottky diode frequency doubler—yes, that’s 125, not 12.5—with a split-block waveguide structure using a high-precision SLA printing process. See their published paper titled “125 GHz Frequency Doubler using a Waveguide Cavity Produced by Stereolithography”.

The waveguide cavity and its waveguide flanges were printed using a system from BMF Boston Micro Fabrication, which uses projection micro stereolithography (PμSL) technology, as shown in Figure 1 and Figure 2. You can see more about how this system works in the short BMF video.

Figure 1 Configuration of the 125-GHz frequency doubler shows (a) layout of one split-block and (b) close-up picture of the Schottky diode MMIC. Source: University of Birmingham

Figure 2 It’s the photo of the as-fabricated polymer waveguides created by using SLA process (left) and an optical microscope image of the area where the MMIC sits (right). Source: University of Birmingham

The printed polymer waveguide parts were plated with copper and a thin protective layer of gold. They characterized the surface roughness of the printed waveguide parts and measured the critical dimensions. The data showed good printing quality as well as a dimensional accuracy that meets the tight tolerance requirements for such a sub-terahertz active device (Figure 3).

Figure 3 Pictures of the fabricated frequency doubler show (a) the fabricated MMIC placed in the 3D-printed waveguide split-block and (b) the assembled doubler. Source: University of Birmingham

The doubler, which is claimed to be the first ever produced using SLA, consists of a 20-μm thick GaAs Schottky-diode monolithic microwave integrated circuit (MMIC) fabricated in the waveguide. It has a maximum output power of 33 mW at 126 GHz and an input power of 100 mW. Peak conversion efficiency, an important figure of merit, is about 32% with input power ranging from 80 mW to 110 mW.

If you are unfamiliar with use of a Schottky diode as a frequency multiplier, the approach uses the common technique of employing a non-linear element—here, a diode—to create harmonics when driven by a fundamental frequency waveform (Figure 4).

Figure 4: The block diagram shows frequency doubler using a non-linear element (top) and the core schematic of that frequency doubler (bottom). Source:

Of course, for GHz-range work, the simple schematic can only be a hint at what it really takes to build a doubler in practice, since those nice, discrete, lumped elements have a very different manifestation in the GHz reality than what’s indicated by those simple symbols in the line drawing.

Using precision SLA and the materials it supports opens up a new path for creating unique high-GHz passive and active components for custom, low volume, and perhaps even higher-volume applications.

Do you see a near-term use for precision SLAs for fabricating these components using designs and arrangements which would be difficult or simply not possible with conventional precision techniques? Is there something you’d like to try to create, and which would make your RF design much easier if it were feasible?

Related Content

0 comments on “3D printing challenges multi-GHz component constraints

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.