Fig 1. A traditional analog JFET-ECM includes a pair of connection rings, a metal can, the JFET, the electret layer, the printed-circuit board, and other components.
Fig 2. The placement of the digital mic IC relative to the ECM mic diaphragm is key to the device’s performance.
Fig 3. Mobile phones are inherently noisy, and components with inadequate linearity performance don’t work well in such environments. Positioning an ADC within the mic enclosure alleviates this problem and contributes to better system power efficiency.
Fig 4. In today’s MEMS mics, the mic diaphragm and the amp IC or amp/ADC IC are wire-bonded together and housed in a single package.
This year, more than 2 billion mics will be used in cell phones and other mobile communication systems. The increasing rate of mic usage has driven investment in new mic technologies to increase performance, enhance manufacturability, reduce size, and reduce costs.
The reason is simple. With the adoption of highly featured smart phones, more than one mic per phone is needed for audio processing such as echo cancelling, noise cancellation, wind noise filtering, beam steering, 3D sound, and other interesting effects. Such sound processing will become pervasive as phone makers seek to further differentiate their products via more sophisticated audio features.
For years, manufacturers have been striving to improve the performance of electret condenser microphones (ECMs), including sensitivity, signal-to-noise ratio (SNR), and reflow soldering. Analogue-to-digital converter (ADCs) ICs, particularly those designed for microphones based on microelectromechanical systems (MEMS), are now significantly contributing to improved microphone performance.
The MEMS microphone market developed in successive waves. The first began in 2002 with an iconic (at that time) cell phone that adopted a MEMS mic. The second wave was not huge as additional MEMS mics emerged from 2005 to 2008 to challenge the incumbent’s hegemony, but with limited success.
The third wave has the potential to change everything. Mainstream semiconductor makers may supplant traditional mic suppliers as mics morph into pure semiconductor products made entirely in semiconductor production facilities (i.e., wafer fab, package, and test).
Cell-phone makers are driving this paradigm shift, as they want to source a mic as a semiconductor component that can be soldered onto a board like a semiconductor component, priced like a semiconductor, and sourced from an approved semiconductor supplier.
Severe Price Competition
Getting mics from a chip company reduces the number of vendors that cell-phone makers have to deal with, which they like. The severe price competition endemic to every level of the cell-phone supply chain is driving disintermediation (i.e., cutting the middle man—in this case the mic maker—out of the loop).
This process narrows the field to fewer vendors, giving cell-phone makers more pricing power. It also eliminates one level of margin stacking, which lowers the price even further. The sheer purchasing power of high-volume cell-phone companies provides the juice to make such a major shift in the supply chain possible.
The end customer’s desire for increased performance also is driving the MEMS mic market. Digital technology improves performance because of its inherent RF and electromagnetic interference (EMI) immunity, which helps enable the implementation of features such as beam steering, 3D spatialization, and noise and echo cancellation.
These types of advanced features are surfacing on smart-phone platforms. Now that digital mics are part of the feature enhancement formula, smart-phone growth translates directly into digital mic growth, especially with the multiplier effect of more than one mic per smart phone.
As smart phones gain more popularity, the features once reserved for the most expensive phones will move down-market to less expensive phones. Digital mics, increasingly MEMS mics, will spread beyond high-end smart phones.
ECM mics have dominated the market for 50 years. They now use junction field effect transistor (JFET) amplifiers integrated right into the mic enclosure itself in mobile applications. A traditional analog JFET-ECM includes a pair of connection rings, a metal can, the JFET, the electret layer, the printed-circuit board, and other components (Fig. 1).
JFET amplifiers present low power consumption and simplicity. They have poor linearity and do not work well in noisy environments like cell phones, though. Analogue mics are vulnerable to the injection of interference prior to the analogue-to-digital conversion, which happens later in the signal path.
Therefore, bringing an ADC right into the mic enclosure to create a “digital” mic makes good sense. Right now, semiconductor suppliers like Fairchild Semiconductor are integrating ADCs onto the same IC substrate as the amplifier for ECM mics.
The market for digital ECM mics is being driven by the need for higher performance to support high-end features in smart phones, tablets, mobile PCs, and headphones. Mic ICs using fourth-order sigma-delta modulators on their ADCs to reach better than 60.5-dB SNR are now available with further improvements under development.
The block diagram and mic drawings in Fig. 2 and Fig. 3 illustrate a digital mic IC’s architecture. They also show where the chip fits in the physical ECM enclosure relative to the ECM mic diaphragm.
An ECM mic diaphragm comprises “electret” material (usually a type of polymer) that has an electrical charge. Being charged, the diaphragm’s movement inherently changes the electrical field to transduce sound into an electrical signal. A MEMS mic, in contrast, has a silicon diaphragm, which is charge-neutral and therefore must be biased externally (using a charge pump) to create the transduction field.
So, a MEMS mic IC can be viewed as a superset of an ECM mic IC with the addition of the charge pump. The final packages of MEMS and ECM mics are currently shaped very differently due to the differing diaphragm and IC assemblies. MEMS mics typically have rectangular packages and ECM mics are mostly cylindrical, though not always.
The MEMS mic market has undergone an evolution of its own. It started with a two-chip analog MEMS microphone comprising one chip for the mechanical microphone diaphragm and one for the amplifier (in the case of the analogue mic) or the amp plus ADC (in the case of the digital mic). The mic diaphragm and the amp IC or amp/ADC IC are wire-bonded together and housed in a single package (Fig. 4).
Taking A Different Approach
After the two-chip MEMS mic established its foothold in the cell-phone market, other suppliers announced different approaches including an innovative one-chip configuration. A few years ago, the one-chip digital MEMS microphone managed to penetrate laptop PCs for voice over IP (VoIP) usage by enabling the placement of two beam-steering digital mics in the bezel of a PC.
The bezel is the ideal mic location in a PC since it is closer to the level of the speaker’s mouth and directly faces the speaker. This placement was made possible because of the natural RF and EMI immunity endemic to digital mics, which allowed the signals to be routed over a distance, through the hinge, and near the noisy processor and wireless ICs without fear of noise injection.
Such meandering routing would not be feasible with analogue mics without expensive and clumsy cable shielding. The one-chip approach has not made it into cell phones as of yet, but it will.
In single-chip MEMS mic architectures, the integration of additional functions on the same die as the mic element can be less costly by using standard semiconductor processing (CMOS), packaging, and testing while providing a higher level of reliability because wire bonds are eliminated.
Conversely, with a one-chip architecture, adapting to inevitable design changes becomes more limited since any signal-processing circuit modifications require a redesign that includes or is contiguous with the MEMS diaphragm.
A design flexibility versus cost tradeoff between the single-chip and dual-chip approaches may likely lead to both styles co-existing in the future as the market sorts out the winning approach.
Although late to the mic-making party, semiconductor IC companies are now emerging as credible MEMS mic providers (in addition, of course, to already being pre-amp and ECM mic IC suppliers). To get into the game, IC companies have been acquiring mic IC and MEMS technology and entering into supply and development partnerships with MEMS providers.
Logos may be changing, but the excellent technologies are living on. Mic makers recognize that a shakeout could be forthcoming, so they have started to strategically partner with semiconductor companies to participate in the conversion to MEMS mics.
To summarize, the mic market is migrating from analogue to digital, from ECM to MEMS, and increasingly from mic makers to silicon suppliers. The timing of these shifts is modulated by the price, performance, reliability, and alternate sourcing as smart phones drive the need for differentiated audio features enabled by multiple-mic architectures.