Simplified audio design based on silicon MEMS technology

This article describes how designers and manufacturers can overcome the many related problems of ECM with next-generation microphones based on CMOS (Complementary Metal Oxide Semiconductor) MEMS (Micro Electro Mechanical Systems) technology.

Evolution of Microphone Technology: From ECM to Silicon Crystal Technology

The traditional ECM is a metal can consisting of a movable permanent charging diaphragm and a rigid backplane parallel to it and a field effect transistor (FET), as shown in Figure 1. The sound wave bends the diaphragm, changing the air gap between the diaphragm and the back plate, so that the capacitance between the diaphragm and the back plate changes. This change is output in the form of a voltage change, which reflects the frequency of entering the sound wave. And magnitude.

Figure 1 shows a typical audio system design in which the source of the FET is grounded and the drain is typically biased by a 2.2k resistor.

Figure 1: A schematic cross-section of an electret condenser microphone (ECM).

It should be noted that the diaphragm of the ECM is connected to the gate of the FET, as shown in Figure 2. The output of the ECM is AC coupled to the preamplifier through a series capacitor. This AC-coupling capacitor provides a single-pole high-pass filter (HPF) that helps filter out unwanted low-frequency components that can further saturate the analog-to-digital converter (ADC). Although the output of the ECM is single-ended, for optimal noise performance, designers typically generate a trace from each unused preamplifier input near the ECM and balance the two traces with a differential input amplifier. The common template level noise source in the two route traces is eliminated.

Figure 2: Typical schematic of an audio system with ECM and integrated FETs.

The challenge of microphone design: reducing noise

The main challenge for frequency system designers is to minimize overall noise in the system design. The noise of the ECM is determined by several sources: electronic noise due to bias voltage fluctuations, FET noise, board-level noise, diaphragm self-noise, and external electromagnetic (EM) fields and radio frequencies coupled to the high impedance input of the FET ( RF) field.

When the system in which the ECM is placed is close to the RF transmitter with power control, the audio component of the RF signal generated by the power control can be demodulated by the microphone and converted into a sound signal audible to the audio path. Low-power portable devices typically use power gating technology to turn off RF when not in use. This threshold appears under audio.

In the ECM, the threshold of the transmit power amplifier (appearing in the audio band) is tuned by the high impedance gate of the FET and the signal is amplified. Once the signal enters the audio band, it is difficult to eliminate. When the audio signal produces audible interference (commonly referred to as breakdown noise), the power threshold of the RF power amplifier is turned on. The most effective way to reduce ECM breakdown noise is to minimize the length of the gate leads and use a capacitor to filter out RF interference in Wi-Fi enabled wireless systems such as cell phones and laptops. This capacitor should be applied to the drain of the FET and is preferably located inside the microphone can. The capacitance value is selected based on the carrier frequency of the interference field and the optimal attenuation frequency of the capacitor. The attenuation frequency of the capacitor can be found in the specification manual provided by the manufacturer.

Another most common source of noise in audio systems is power supply (bias voltage) fluctuations. The ECM is a low-sensitivity microphone that outputs a small analog signal on the order of 10mVrms. Since the ECM does not have any power supply rejection (PSR) capability, small fluctuations in the power supply can cause the user to hear small output signal fluctuations. Therefore, in order to maintain the best signal-to-noise ratio, additional filtering components should be used to keep the microphone bias supply "clean".

The use of ECM in audio systems also presents many mechanical design and manufacturing challenges. First and foremost, although ECM has been shrinking, it has reached its size limit, and further down, it has to pay for the performance degradation of sensitivity, frequency response and noise. Currently, standard sizes of ECMs used in portable electronic devices range from 4 to 6 mm in diameter and from 1.0 to 2.0 mm in height.

Another challenge is that ECM not only detects sound signals, but also detects mechanical vibrations and ultimately converts the vibrations into low-frequency sound signals. When the ECM is placed in a vibrating environment, such as a circuit board mounted near an electric fan or large horn, the primary source of noise in the audio system will be vibration. The only way to reduce vibration at the microphone is to use additional mechanical isolation material when mounting the microphone on the board.

In addition, whether it is the material used to make the ECM diaphragm and backplane, or the permanent diaphragm charge of the ECM, the performance will be significantly reduced at the high temperatures required for surface mounting. Therefore, some form of electronic interconnection (socket or elastic compression connector) must be used between the microphone and the board, so that the already large components are generally taller (too thin with many portable electronic devices today) Compared with the shape). Finally, because ECM can't be surface mounted, it needs to be assembled by hand, so it can be assembled more expensive and reliable than components that can be soldered to the board using an automatic pick and place assembly process. Lower.

Akustica is developing a new generation of single-chip silicon crystal microphones using the latest MEMS technology called CMOS MEMS. Unlike other silicon microphones, at least two silicon chips are required, one for the silicon microphone transducer unit and the other for the integrated circuit (IC). The CMOS MEMS microphone is a monolithic integrated circuit in which the MEMS transducer unit Formed from a metal dielectric structure in a standard CMOS wafer. Because CMOS MEMS microphones are fabricated using industry-standard CMOS processes and equipment currently used to fabricate integrated circuits, the device can be fabricated at any CMOS fab in the world. CMOS MEMS devices have been manufactured in nine different fabs, with 11 different CMOS technologies ranging from 0.6-micron three-layer metal processes to 0.18-micron copper interconnect processes. The result proves that this technology has high yield and repeatability of semiconductor manufacturing and can be mass-produced in extremely high volume.

The monolithic integrated silicon crystal microphone solution developed on the CMOS MEMS platform enables consumer electronics device designers and manufacturers to avoid numerous ECM related issues. Figure 3 is a top plan view and a cross-sectional view of a single chip silicon crystal microphone. This monolithic chip consists of a MEM transducer and an impedance matching circuit. It is also a capacitive sensor with a movable diaphragm and a rigid backplane.

Figure 3: Top view (a) and cross-sectional view (b) of a CMOS MEMS microphone chip.

Figure 4 shows a typical audio system using a CMOS MEMS analog microphone. Since CMOS MEMS microphones are more analogous to analog ICs than ECMs, they also use an IC-like power supply scheme that is directly connected to the power supply. The on-chip isolation between the power supply and the rest of the system adds PSR to the component, making the CMOS MEMS microphone inherently more resistant to power supply noise than the ECM, and eliminating the need for additional filtering circuitry to keep the power line “clean” .

Figure 4: Schematic of a typical audio system with a CMOS MEMS microphone.

When an electronic circuit is fabricated in a micron-scale acoustic structure, the stitch length is short, and the ability to reduce the breakdown noise can be improved. Unlike FETs in ECM, in CMOS MEMS microphones, due to the on-chip amplifier stage, the separation between the diaphragm and the preamplifier is extremely short, and the input and output isolation is better. Because there is better isolation between the power supply and the output signal, and the distance from the diaphragm to the preamplifier is shorter, it is almost impossible to couple the electromagnetic field into the microphone.

CMOS MEMS microphones also address many of the mechanical design and manufacturing challenges encountered with ECM. First, the CMOS MEMS microphone monolithic integrated circuit features a footprint and height that is less than half the size of a conventional ECM. Secondly, the size and quality of the CMOS MEMS microphone diaphragm are small. Compared with the ECM diaphragm with a diameter of 4-6mm, the diameter is less than 0.5mm, which improves the vibration resistance. Third, because CMOS MEMS microphones are fabricated using standard CMOS materials and processes, they are inherently resistant to the high temperatures required for surface mounting. The overall height of such a microphone system is significantly reduced without the need for mechanical interconnection. Finally, CMOS silicon microphones have surface mount and split compatibility, eliminating the need for manual assembly, resulting in lower cost and improved reliability, throughput and yield.

CMOS MEMS microphones can also integrate an analog-to-digital converter on the chip to form a microphone with a robust digital output. Since most portable applications end up converting the analog output of a microphone to a digital signal, the system architecture can be designed to be completely digital, thus removing the analog signal that is prone to noise from the board and Simplified overall design.

The advantages of using a digital CMOS MEMS microphone are most noticeable in applications where a long cable is required between the microphone and the CODEC, such as a laptop platform for optimal sound, a microphone is typically installed in the display, and a CODEC is installed in the main body of the computer. Motherboard. In this case, there are many cable and electronic noise sources that can interfere with small analog sound signals around the laptop display, so shielded cabling and other filter components are needed to minimize interference. However, with digital CMOS MEMS microphones, there is no need to shield wiring or filter components, simplifying design, reducing overall component count and reducing bill of materials (BOM) costs.

Summary of this article

When designing audio systems for today's next-generation portable electronic devices, CMOS MEMS microphones can solve many of the difficulties that cannot be solved with ECM. Table 1 summarizes the differences between ECM microphones and CMOS MEMS microphones, allowing system and mechanical designers and manufacturers to make better use of CMOS MEMS microphones.

Table 1: Comparison of key features of ECM microphones and CMOS MEMS microphones.

Using Akustica's patented CMOS MEMS technology, diaphragms and powerful analog-digital signal processing functions can be integrated into a single chip, enabling next-generation microphones for future portable electronic devices. The design simplicity and productivity offered by CMOS MEMS microphones will enable designers and manufacturers of mobile phones, PCs, PDAs, and countless other consumer electronics products to create more powerful, feature-rich, and lower-cost products. Good service to the market.