another one of RFMD’s TxM products, the RF3282, was used as the test vehicle. Figure 7 shows the radiated power from the RF3282 TxM. The red graph represents the TxM with no shield and the blue graph represents the TxM using MicroShield. Note that the blue graph has been shifted slightly to the right to better illustrate the difference between the two devices under test. As shown in the graph, the MicroShield integrated RF shielding reduces the radiated power sig- nicantly. Only one caveat was observed at 10.5 GHz. It appears that either another mode exists (cavity mode) or perhaps the results could be related to ground current owing on the top surface of the shield. Nevertheless, the average attenuation in radiated power amounts to 15 dB or better. We have discussed the benets of Mi- croShield as it relates to EMI and RFI emis- sions, improving the ability to meet specica- tions. Additionally, MicroShield integrated RF shielding minimizes exposure to external EMI/ RFI as well, resulting in lower susceptibility to performance shifts seen during handset design. A component’s sensitivity to board placement is a critical factor as handset designers and manufacturers increasingly rely upon handset platforms to satisfy their time and cost require- ments. When these platforms are applied to individual handset designs, performance can suffer, with EMI and RFI emissions often prime contributors to performance inconsistencies. With MicroShield, handset manufactur- ers are able to place highly complex RF modules as they would any component that is insensitive to EMI/RFI, providing a true “plug-and- play” solution that is robust to board design and layout changes. By eliminating sensitivity to board placement, MicroShield eliminates the risk of circuit retuning, thereby accelerating time-to-market and reducing the cost of RF implementation. RFD Figure 6. The device under test is placed on top of a non-absorbing and non-reflective material. -20 -25 -30 -35 -40 -45 -50 -55 -60 -65 -70 -75 -80 2750 4000 5000 6000 7000 8000 9000 10000 11000 12000 12750 X = Frequency in MHz Y = Power in dBm Figure 7. Comparing Tx module with no shield (red graph) to Tx module using conformal shielding (blue graph). mechanisms and impact on higher-order harmonics is still valid. A second area of concern is the presence of eld lines around a microstrip line, which are strongest near the ground plane. As long as the distance between the shield and ground plane is noticeably greater than the distance between the microstrip and ground plane, the effect of adding a shield should be minimal. Bond wires and surface-mount inductors have weaker direct coupling to the ground plane, and it is expected that the eld lines will change when the shield is added. A 3-D simulated E-eld distribution is shown in Figure 3. As expected, the eld lines around the surface-mount inductor and bond wires are less conned and, therefore, more prone to alterations if a shield is added on top of the overmold. Acknowledgments The authors would like to thank in particular Mick Zhou in the modeling team at RFMD and also Scott Morris and Milind Shah at RFMD Corporate R&D packaging. Figure 4 is a plot of the output match without a shield, which illustrates the electrical eld contours in volts per meter. Crimson colors denote strong eld lines whereas dark blue colors denote that an electrical eld is virtually non-existent. The next step is to plot and inspect the two-port S parameter simulation for any changes to higher- order harmonics with and without the shield present. References 1. David M. Pozar, Microwave engineering, ISBN 0-201- 50418-9
The 3-D EM simulation of the output match reveals a change in resonance at a higher frequency. In a TxM, the complexity of circuitry is much higher than just a simple output match. Furthermore, high Q tank circuits implemented to eliminate higher-order harmonics will be affected to a level clearly outpacing the simple change in one resonance, as observed in the simulation.
The nal task is to conduct radiated measurements on a TxM without the shield and compare the results to those of a TxM with MicroShield integrated RF shielding technology applied. For ac- curate measurements, leaking RF power from connectors and other onboard circuitry on the test board must be prevented; therefore, the test board designed for these measurements contains separate shield cans as shown in Figure 5.
All radiated measurements were performed at Delta Technologies in Copenhagen, Denmark. The device under test is placed on top of a non-absorbing and non-reective material (Figure 6). In this example
ABOUT THE AUTHORS
Ulrik Riis Madsen received his B.Sc with honours in Electrical Engineering fromAarhus University, Denmark, in June 1996. Until October 1999, he was employed with Dancall Telecom in Denmark. He joined RFMD in 1999 and has designed several key power amplier modules and also conceptualized the PowerStar concept. Madsen is currently a staff engineer in the corporate research and development group with RF Micro Devices.
Carsten Hinrichsen received his M.Sc.EE. from AAU, Denmark in 1993. During his career in the RF industry he has worked in DanPhone (PMR), Dancall/Bosh telecom and Texas Instruments. For the past three years he has worked as a GSM PA designer at RF Micro Devices. Hinrichsen holds several patents in various aspects of RF design.