Issue link: https://iconnect007.uberflip.com/i/1190860
40 DESIGN007 MAGAZINE I DECEMBER 2019 nation of electrical performance and how the materials are used in the PCB fabrication pro- cess, the constructions will probably need to be more complex and designed to be specific to the process concerns of the PCB fabricator. The evaluation circuit used to characterize materials electrically should be a simple con- struction with a simple circuit design. It is highly recommended to evaluate the material at lower frequencies and with thick substrates. This will minimize some of the variables that can make the electrical evaluation less accurate. Basically, it is difficult to obtain accurate results at high frequencies. Also, when using thinner materi- als, the influence of etching control has more effect on the electrical results. Typically, when a thinner circuit is used, it will have a narrower conductor to maintain a controlled impedance and the variation of the conductor width due to etching control at the PCB fabricator that will have a greater influence on the electrical per- formance of the circuit as opposed to a thicker substrate with a wider conductor. In general, I am not in favor of ring resona- tors; however, at lower frequencies and with a thicker substrate, the ring resonator can be a good test vehicle. The ring resonator should be based on a double-sided circuit (microstrip), loosely coupled. Lower-order nodes should be tested; test at frequencies below 6 GHz. If the circuit is a double-sided, non-plated through- hole (PTH), bare copper circuit, then this is a pretty good test vehicle, which will have mini- mal electrical variation due to PCB fabrication variables. I have seen many companies use stripline circuits for their test vehicles. Due to the nature of this construction, these circuits are more in- fluenced by PCB fabrication variables than is the simple test vehicle and construction I rec- ommended (microstrip). In the case of strip- line, typically, a core material and prepreg are used to make the stripline construction. These materials are very often dissimilar in electrical properties, which can confuse the electrical re- sults of the evaluation. Additionally, the thick- ness control of the prepreg layer due to circuit fabrication can have a significant impact on the electrical properties of the circuit. The PTH vias, which are used to connect the top and bottom ground planes of the stripline, can also influence the electrical performance of the cir- cuit. If stripline is necessary for the test vehi- cle, again, a thicker test vehicle being used at a lower frequency is advantageous to avoid some inaccuracies in the electrical evaluation. These very simple guidelines I gave for the test vehicle and construction would be good for a materials-only evaluation. This evalua- tion would simply be comparing different ma- terials with the same test vehicle, side-by-side. However, it is common that the material evalu- ations need to be on thinner materials, higher frequencies, complicated multilayer PCBs, etc. When that is the case, the designer needs to consider the effects of PCB fabrication influ- ences on the material properties. A multilayer board using PTH technology means the final conductor thickness will have some normal variation. For a gap-coupled ring resonator, the copper thickness variation can impact the electrical results and cause inaccu- racies in determining the Dk or Df values re- lated to the circuit material. Some test vehi- cles are more sensitive than others to the cop- per thickness variation. As mentioned, the gap- coupled ring resonator can be affected by normal copper plating thickness varia- tion, and so will a grounded coplanar wave- guide (GCPW). However, a microstrip trans- mission line circuit is much less affected by copper thickness variation for its electrical per- formance. Additionally, final plated finishes and their normal thickness variations can also impact the accuracies of the test vehicle. The bottom line is the engineer who defines the test vehicle (circuit design and construc- tion) needs to put a lot of thought into how the normal PCB fabrication variables can impact the test vehicle in a way that influences electri- cal evaluation accuracy. DESIGN007 John Coonrod is technical marketing manager at Rogers Corporation. To read past columns or contact Coonrod, click here.