Overcoming PCB Insertion Loss in mmWave Interconnects with 3D Printing

Ziv Cohen

Application Manager, Nano Dimension

Whether you work on high-speed digital or high-frequency analog devices, losses and signal distortion are important performance metrics in these systems. In planar PCBs on rigid substrates, the industry is moving away from FR4 PCBs, and a variety of high-speed laminate materials are available with low losses and nearly flat dispersion. Despite the availability of these materials, reducing PCB insertion loss, which contains indicators about the discontinuities in a channel and reflections in mmWave interconnects, is still a real challenge.

Designing for reduced insertion loss in mmWave analog channels, and in digital channels operating with mmWave-equivalent bandwidths, requires carefully engineering signal traces and layer transitions on interconnects. Parasitics in interconnects can be difficult to control in traditional plated through-hole via structures, and these structures are likely to remain with us in the future. However, using an additive manufacturing system allows you to engineer unique structures for layer transitions with lower insertion loss, which prevents signal degradation at high frequencies.

Digital channels operating at mmWave bandwidths can experience significant PCB insertion loss, jitter, and ringing if not designed properly.

What Causes PCB Insertion Loss in mmWave Devices?

If you’ve never worked with RF systems, mmWave might sound like an esoteric buzz word. However, the principles involved in mmWave design build on those used in lower-frequency designs. Just as is the case with any controlled impedance design, PCB insertion loss arises due to inconsistent impedance throughout an interconnect. This quantity represents the amount of signal that is transmitted at an impedance discontinuity along an interconnect. 

As an mmWave signal travels between a source and load, there is the potential for reflection off of an impedance discontinuity, such as a change in trace dimensions or via. Fiber weave effects also create impedance discontinuities and resulting signal integrity problems above approximately 40 GHz.

In mmWave devices, by far the most prominent source of impedance discontinuities in any impedance controlled interconnect is vias. Layer transitions require some kind of via, whether it is a blind, buried, or through-hole via. Plated through-hole vias or stacked blind/buried vias are typical in traditional multilayer PCB designs. These structures have some parasitics with them, and these parasitics are responsible for producing PCB insertion loss at higher frequencies.

If you look at a time-domain reflectometry trace for an mmWave interconnect on a test coupon and calculate the impedance from the interconnect return loss, you’ll likely see a capacitive impedance drop at plated through-hole vias that are used for layer transitions. These impedance discontinuities are undesirable for two primary reasons. First, these capacitive impedance discontinuities at vias cause reflections, leading to the formation of standing waves and strong radiated EMI. Second, this reduces the signal level that reaches the downstream load (i.e., insertion loss).

As a designer, your goal is to engineer these layer transitions to have nearly the same impedance as the transmission lines connected to them and to prevent any via resonance and noise coupling. At a minimum, any through-hole vias should be backdrilled to eliminate resonances. As parasitics can be the major determinant of PCB insertion loss at mmWave frequencies, you need to carefully manage your via size to reduce these parasitics, particularly parasitic capacitance at the via anti-pad (see below).

Parasitic capacitance at a via anti-pad.

With the typical via structure, this requires careful management of the return path and reducing the parasitic capacitance at the via anti-pad. In the example above, the solution is to increase the spacing between the via barrel and the ground planes. To maintain a consistent return path, you would then need to place a second via between the ground planes (to the left of the through-hole via) with sufficiently low parasitic capacitance and parasitic inductance. The third option, which eliminates these extra steps, is to design a completely new layer transition structure. This is where a new fabrication process is needed and where 3D printing becomes indispensable.

How 3D Printing Helps with Engineering Low PCB Insertion Loss

Via structures in PCBs on standard rigid laminates are constrained by the traditional manufacturing process. These structures are barrel-shaped simply because they must be formed with a drilling process (either mechanical or laser drilling). Alternative structures, such as vertical conducting structures (VeCS), can be formed in a typical planar PCB, but this incurs higher cost and fabrication time for each layer transition.

The layer-by-layer deposition process in 3D printing eliminates these design constraints on any structures used for layer transitions. You can easily place a structure like VeCS or any other unique layer transition structure you can imagine. This allows you to precisely engineer the impedance of your interconnects to prevent PCB insertion loss and ensure signal integrity. This is vital in mmWave interconnects, including in 5G systems, car/UAV radar, and cellular-enabled IoT devices.

There are many other benefits of using 3D printing for mmWave devices and high-speed digital devices. The fabrication time and cost structure involved in additive manufacturing are independent of the complexity of the product being produced and instead only depend on the weight of materials being deposited. Inkjet systems that co-deposit the insulating substrate and conductive tracks have a very fine resolution, making them ideal for advanced HDI designs. This allows extremely complex, high-density digital or mmWave designs to be produced in a matter of hours with any geometry and interconnect architecture. These devices can then be immediately tested and redesigned as necessary.

As more insulating polymers, semiconducting polymers, and other advanced materials become adapted for use in a variety of 3D printing systems, designers will be able to print a broader range of fully-functional devices. The expansion into a broader range of insulating polymer materials is interesting as their material properties can be tuned through doping and functionalization. This allows complex devices to be precisely designed with ultra-low losses at mmWave and higher frequencies, which is simply not possible with standard planar PCB laminate materials.

Designing and modeling mmWave devices is a complicated affair, but manufacturing unique mmWave devices with low PCB insertion loss is easy when you work with the right additive manufacturing system. The DragonFly LDM system from Nano Dimension is ideal for high-mix, low-volume production of complex additively manufactured electronics (AMEs) in-house and at scale. This advanced system is ideal for producing planar or nonplanar high-speed/high-frequency devices with unique interconnect and routing architecture. Read a case study or contact us today to learn more about the DragonFly LDM system.

Ziv Cohen

Application Manager, Nano Dimension

Ziv Cohen has both an MBA and a bachelor’s degree in physics and engineering from Ben Gurion University, as well as more than 20 years of experience in increasingly responsible roles within R&D. In his latest position, he was part of Mantis Vision team—offering advanced 3D Content Capture and Sharing technologies for 3D platforms. The experience that he brings with him is extensive and varied in fields such as satellites, 3D, electronic engineering, and cellular communications. As our Application Manager, he’ll be ensuring the objectives of our customers and creating new technology to prototype and manufacture your PCBs.