In smartphone and mediaphone handsets, high-performance electromagnetic interference (EMI) filters with integrated electrostatic discharge (ESD) protection can eliminate the radiated frequencies and preserve the best possible signal integrity on high-speed data lines. These filters can be built from ceramic materials or from silicon-based passives in a monolithic IC.
Ceramic-based and silicon-based EMI filters differ in several important characteristics, including performance, ESD protection, parts count and printed-circuit-board (PCB) area, insertion cost, part-by-part comparison cost, reliability, reuse rate, and total cost of ownership (TCO). Research shows that not only do silicon-based EMI filters perform better on a system basis, they cost less and provide a lower overall TCO than ceramic solutions.
Defining the problem
The increasing demand for advanced digital entertainment applications has resulted in higher signal speeds that require more signal lines and produce more digital noise when traveling across parallel data interfaces. These high-speed signals, traveling either from the camera or to the display, can create noise that interferes with radio functions. At the same time, the number of radios inside a single handset has multiplied, in some cases up to eight, including standard GSM or CDMA, 3G, WiMAX, WiFi, GPS, Bluetooth, and mobile TV, further complicating design efforts to achieve good signal integrity.
EMI occurs when the digital signals of one device distort the digital signals of another, which degrades receiver sensitivity, and can be defined by total isotropic sensitivity (TIS). The sum of all of these factors has led to a much more challenging design environment where EMI issues are a critical concern. Methods for countering EMI in handsets include mechanical measures and high-performance EMI filtering.
Mechanical shielding
Mechanical shielding can accomplish some EMI suppression and is recommended for certain sections of the handset PCB, in particular placing shield cans over sources of EMI radiation such as the RF sections of the Bluetooth and WiFi chips. Other techniques include using EMI-absorbing foam, paint, and tape in select areas of the handset.
However, implementing mechanical measures can be expensive in terms of both time and bill-of-materials (BOM) costs. For example, elaborate tooling is required to productize metal cans, and for fitting EMI-absorbing foam. Painting the interior of plastic housings requires a separate, offline setup and close monitoring of paint composition. The performance of mechanical measures is also inconsistent, both during manufacturing and in the field. Although metal cans are the most robust, foam is especially susceptible to temperature variations and to the wear and tear of daily use.
Finally, purely mechanical measures do not address the root cause of EMI interference because they only enclose EMI, they do not eliminate it. Therefore, mechanical measures may often be necessary but not sufficient solutions for countering EMI.
Low-pass filters
In contrast, high-performance EMI filters with integrated ESD protection can eliminate radiated EMI while preserving the highest possible signal integrity. Low-pass EMI filters allow lower-frequency clock and data signals to pass through, while attenuating the unwanted higher-frequency harmonics, preserving a clean ecosystem for the wireless carrier frequency bands.
Two different types of low-pass filters are commonly employed: those built from ceramic materials or from silicon-based passives in a monolithic IC. Ceramic-based and silicon-based EMI filters differ in several important characteristics, including performance, ESD protection, parts count, PCB area, reliability, and TCO ( See Figure 1).
The type of solution selected matters significantly, not only due to all of these concerns, but also because the solution chosen can directly affect the handset user’s quality of service (QoS). The TIS levels in a wireless phone’s receiver that determine this QoS are a relative measure of the quality of the signal received by the baseband antenna. Low TIS levels result in dropped calls, poor or no service in marginal service areas, and slow or no data capabilities in marginal service areas.
Ceramic-based versus silicon-based low-pass EMI Filters
Discrete ceramic EMI/ESD solutions require many more individual parts, take up by far the most board space, and have a higher placement cost than either quasi-monolithic ceramic filter arrays or silicon-based EMI filters, but they are the least expensive on a part-by-part cost comparison. Filters built exclusively with discrete components experience greater variation in channel-to-channel performance. Long-term reliability also suffers, as time-to-failure grows exponentially with the number of PCB component placements.
Quasi-monolithic ceramic filter arrays, formed by gluing components together, are more commonly used in smartphone and mediaphone handsets. They also have performance issues. Their high level of integration is made possible only by compromising the quality of ESD protection, and the varistor-based ESD solution fails after as few as 10 strikes. In addition, component grounding, one of the most important elements in an EMI filter, suffers because quasi-monolithic assembly dictates placing the grounds at the end of the package. The farther a ground is from a filter channel, the worse its performance becomes. Since ceramic array filters are glued together with weak bonds, they are also susceptible to failure when dropped.
Ceramic solutions filter only one specific band. Filtering a single band at the expense of, and to the detriment of, all of the other bands the handset receives only solves one of multiple receiver sensitivity problems. In general, ceramic low-pass filters may cost less on a part-by-part basis than silicon-based low-pass filters, but the level of performance they offer is significantly lower than that of silicon-based EMI solutions.
Broadband silicon-based low-pass filters permit signals as fast as 400 MHz to pass through with excellent signal integrity; attenuate dangerous harmonics and spikes that occur from 700 MHz through 6 GHz, and provide excellent attenuation through the 2.4-GHz and 5-GHz WiFi bands. They also deliver a better signal-to-noise ratio by lowering the noise floor of the handset ecosystem. Because they are based on zener diodes, they allow for the integration of robust ESD protection.
Silicon-based filters provide higher cutoff frequencies to support high data rates and preserve a high level of signal integrity. With increased data rates, the requirements for higher cutoff frequencies have also increased to maintain signal integrity. However, the requirement to deliver the greatest level of attenuation at critical carrier frequencies (700 MHz to 2.5 GHz) remains unchanged.
Silicon-based filters also have the lowest parts count and the highest reliability due to their fully monolithic, single-chip construction. High reliability results in their superior performance in drop tests compared to ceramic-based filters. Unlike the ceramic solutions that filter only one channel, broadband silicon-based filters can be utilized across multiple phone designs and their high reuse rate improves their utility to the handset manufacturer.
Performance testing and BOM costs
California Micro Devices independently tested handsets from one handset manufacturer and confirmed that silicon-based solutions provide TIS performance of 3 to 6 dBm better than ceramic-based filter array or discrete solutions.
A TIS performance difference of 3 dBm translates into better receiver sensitivity, fewer dropped calls, better audio quality, and faster multimedia access (See Figure 2). In particular, receiver sensitivity directly affects the extremely low-power signals received for GPS applications, as digital noise can interfere with the reception of location-based services. A TIS performance difference as small as 1 or 2 dBm can mean the difference between receiving or not receiving a signal good enough to maintain a GPS connection.
For example, a leading OEM found that using ceramic filters on the handset’s display negatively affected GPS performance, while changing to silicon-based filters improved performance.
Relative BOM costs
Silicon-based solutions, therefore, perform better and provide a greater margin in TIS testing than do ceramic arrays or discretes. To compensate, designers utilizing ceramic filters often employ additional mechanical measures to ensure electromagnetic compliance. Although silicon-based filters may cost more than ceramic filters on a part-to-part comparison, when higher placement costs and the costs of these EMI-reducing mechanical measures are considered, it is clear that solutions employing silicon-based filters have a lower total implementation cost.
California Micro Devices and iSuppli analyzed the bill of materials for two handset models made by the same supplier. One model employed silicon-based EMI filters. The second model was a “cost-reduced” version of the first, in which the manufacturer replaced the silicon-based EMI filters with lower-cost ceramic filters. iSuppli’s analysis showed that the second design costs $0.175 more per handset than the silicon-based solution when insertion costs and additional EMI-reducing measures are accounted for (See Figure 3).
Of course, handsets with ceramic filters do pass the necessary certification testing. But those with silicon filters pass such testing with higher margins and the size of those margins equates to very specific differences in QoS. Compensation strategies, such as combining ceramic filters with mechanical measures, can make handsets with ceramic filters pass certification by a higher, more acceptable degree of margin. The drawback is that, because of their additional cost, using such mechanical measures erodes the perceived benefit of ceramic technology, which is lower cost. That additional cost not only negates the benefits of ceramics but actually makes them more expensive to use than silicon-based solutions.
In conclusion, the EMI filter solution a handset manufacturer chooses makes a difference that has ramifications beyond the immediate considerations of BOM cost. Given the ongoing trend of higher signal speeds, electromechanical compliance (EMC) issues will continue to become more challenging. Although a variety of techniques can achieve EMC, California Micro Devices believes that monolithic silicon-based filters remain the best solution. This is because they are more highly integrated; provide a superior-performing filter; and have the lowest overall total cost of ownership due to their integrated ESD, design reusability, highest reliability, lowest parts count, and smallest PCB footprint.