Shields are your friend, except when... (Part 2)

-August 02, 2013

Last month in Part 1 we looked at one issue of using conductive shields in our high-frequency circuits. We showed that they can and do resonate at frequencies and we can easily calculate the lowest resonant frequency. The simple calculation is useful to determine where your shield will stop acting like a simple box and start acting like a resonant cavity.

These resonances create increased coupling inside the shielding structure itself that can lead to improper operation of our circuits. This month we look at some solutions.

Smoothing the Peaks
What can be done about these resonances?

First off we can make the shielding box smaller in the longest directions. This will increase the resonant frequencies inside the box.

Making the box smaller can be accomplished in two ways. The box itself can shrink or we can add walls and sub divide the box. Sometimes these walls can be made from something as simple as etched or stamped thin brass (shim stock) soldered to the PCB and with fingers making contact with the top of the shield box.

Making the shielding box smaller is not an unlimited option however, as the shield interface on the PCB takes up PCB area where it has to be soldered down. So adding more and more walls to make the shield sections smaller and smaller eventually ends up with a big PCB where all the space is taken up by the shielding tracks with no room for circuitry. This also adds cost to the design as the walls and PCB area for the tracks aren't normally free.

You may already have the circuit partitioned properly into functional blocks and those are broken up nicely into shielding cavities so what else can be done?

Microwave engineers have had this problem for decades, but now the rest of us RF people, and even digital people (Gasp!) also suffer from the dreaded Cavity Resonance syndrome. These hardy microwave souls have devised all sorts of ways to kill the inevitable box resonances. Perhaps the most common solution is to use a Microwave Absorber (Figure 1). These Absorbers are usually flexible silicone based materials that are filled with some low Q radio wave absorbing material [1]. The absorber functions by preventing the reflection of radio waves inside the cavity.

Figure 1: Microwave Absorbers typically come in two types. Left: The Silicone sheet type. Right: The Foam type. Additionally, different thicknesses may be obtained as well. These are samples cut from 12 x 12-inch sheets [1].

As Figure 1 shows, these absorbers normally come in 12 x 12-inch sheets and may be readily cut with scissors, lasers, water jet or die cutting to nearly any shape. To make attachment easy they also come with adhesive backs so they can be stuck in our products, shields or even on troublesome circuits. You can also buy these absorbers in liquid form and pour it in the shield and it hardens as it cures.

For this experiment I cut some samples so that they would fit inside the top of my 1 x 1.5-inch rectangular shield as this is the usual application of these materials [2]. The Silicone Sheets Absorbers come in various thicknesses, so for the first experiment I put 0.010- and 0.050-inch-thick solid material into my test fixture and tested the results (Figure 2).

Figure 2: Comparison of thick and thin Silicone Sheet Absorber material in my shielding test fixture presented last month. The Blue trace is the unmodified shielded enclosure with its resonance peaks. The Green trace is the Emerson BSR-1 material which is 0.010-inch thick, it does not do much to the lower-frequency resonances, but smooths out the responses above 15 GHz better than the thicker absorber. The Orange trace is the Emerson MCS material which is 0.050-inch thick. This thicker material does a good job of removing the resonant peaks.

As expected the thinner material does a better job at high frequencies (greater than 15 GHz) in my test shield than the thicker material. The thicker material does a better job of calming down all the other really high-Q resonant peaks however.

Loading comments...

Write a Comment

To comment please Log In