0% found this document useful (0 votes)
259 views8 pages

AN-1165 Application Note: Op Amps For MEMS Microphone Preamp Circuits

Circuito preamplificador micrófono MEMS

Uploaded by

Jfgf.20087010
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
259 views8 pages

AN-1165 Application Note: Op Amps For MEMS Microphone Preamp Circuits

Circuito preamplificador micrófono MEMS

Uploaded by

Jfgf.20087010
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

AN-1165

APPLICATION NOTE
One Technology Way • P.O. Box 9106 • Norwood, MA 02062-9106, U.S.A. • Tel: 781.329.4700 • Fax: 781.461.3113 • www.analog.com

Op Amps for MEMS Microphone Preamp Circuits


by Jerad Lewis

INTRODUCTION The voltage noise of an op amp is specified as a noise density


A microphone preamp circuit is used to amplify a microphone’s unit of nV/√Hz. To get the device noise in the circuit’s
output signal to match the input level of the devices following it bandwidth of interest, you need to multiply this noise density
in the signal chain. Matching the peaks of the microphone’s signal by the square root of the bandwidth. Note that this simple
level to the full-scale input voltage of an ADC makes maximum formula only works for a noise spectrum that is uniform in
use of the ADC’s dynamic range and reduces the noise that the frequency range of interest, as shown in Figure 1.
subsequent processing may add to the signal. For a 20 kHz bandwidth, this multiplication factor is 141. Using
A single op amp can be easily used in a circuit as a preamp for a the ADA4075-2 as an example, its noise density of 2.8 nV/√Hz
MEMS microphone output. The MEMS microphone is a single- multiplied by 141 gives a noise level of 0.395 μV, or −128 dBV.
ended output device, so a single op amp stage can be used to add An op amp’s noise density is also typically shown across fre-
gain to the microphone signal or just to buffer the output. quency in a graph in the typical performance characteristics
section of the data sheet. This graph is useful to see below
This application note covers some of the key op amp specifi- what frequency the op amp noise is dominated by 1/f noise.
cations to consider in a preamp design, shows a few basic circuits, For many op amps, this point is below the low end of the audio
and provides a table of Analog Devices, Inc., op amps that may be band (20 Hz), but it is worth looking at the noise density plot
appropriate for a preamp design. The ADMP504 MEMS micro- and not just relying on the noise density specification to
phone is used as an example in this application note to describe completely describe the noise performance. Figure 1 shows an
different design choices. This is an analog microphone with example noise density graph from the ADA4075-2 data sheet.
65 dB SNR. Designs using different microphones may require Notice in Figure 1 that the 1/f corner is at about 10 Hz, well
adjustment from what is described in this application note, below the band of interest for MEMS microphone preamp
depending on the microphone noise, sensitivity, maximum circuits.
acoustic input and other specifications. For more information on
10
Analog Devices MEMS microphones, see www.analog.com/mic. VSY = ±5V

OP AMP SPECIFICATIONS
VOLTAGE NOISE DENSITY (nV/√Hz)

An op amp data sheet has many different specifications and


performance graphs, so it can be overwhelming to try to find
exactly which of these specs matter for your application. For a
microphone preamp design, there are a few specs that matter
more than others; these specs are reviewed here.
Noise
An op amp’s noise spec is given for both voltage noise and
current noise. Typically, you only need to concern yourself with
an op amp’s voltage noise in a preamp design. The current noise 1
07642-038

1 10 100 1k 10k 100k


becomes limiting in the design only when high value (that is,
FREQUENCY (Hz)
noisy) resistors are used. To keep the overall noise of the circuit
Figure 1. ADA4075-2 Voltage Noise Density
low, typically resistors with values less than 10 kΩ are used.

Rev. 0 | Page 1 of 8
AN-1165 Application Note

TABLE OF CONTENTS
Op Amps for MEMS Microphone Preamp Circuits .................... 1 Circuits ................................................................................................4
Introduction ...................................................................................... 1 Op Amp Selection .............................................................................6
Op Amp Specifications .................................................................... 1 Performance Simulations .................................................................6
Revision History ............................................................................... 2 References ...........................................................................................7

REVISION HISTORY
9/12—Revision 0: Initial Version

Rev. 0 | Page 2 of 8
Application Note AN-1165
An ADMP504 analog MEMS microphone has a 65 dB SNR graph of THD (or THD + N) vs. frequency is often included.
(A-weighted) and a −38 dBV sensitivity. This gives it a noise An example of this from the ADA4075-2 data sheet is shown
floor of −103 dBV across a 20 kHz bandwidth. This is in Figure 2.
equivalent to a noise density of about 50 nV/√Hz, which is 1
about the same thermal noise of a 150 kΩ resistor. VSY = ±5V
VIN = 1.5V rms

It is key for the op amp to be significantly quieter than the


microphone so that, with regards to noise, it’s as transparent in 0.1

the preamp circuit as possible. A good rule of thumb is for an

THD + NOISE (%)


op amp’s noise to be at least 10 dB quieter than the microphone
itself to minimize its contribution to the overall noise. To 0.01

achieve this with an ADMP504 preamp, you want the op amp


to have a maximum noise floor of −113 dBV, or 15.9 nV/√Hz. 600Ω
Most of the op amps in Table 1 are well below this threshold 0.001

and the ones that are not are still listed because they have other 2kΩ

specs that may be more important in a particular design, such


as operating current for a low power design. Note that the 0.0001

10958-067
10 100 1k 10k 100k
circuit’s total output noise level will be affected by the gain FREQUENCY (Hz)
applied and the resistors in the circuit, and not just the op amp. Figure 2. ADA4075-2 THD + Noise vs. Frequency
Resistors can usually be chosen to be small enough to have
minimal contribution to the overall circuit noise. Supply Voltage
Slew Rate An op amp supply voltage is usually given as a range, such as
3 V to 30 V, which indicates the minimum and maximum
The slew rate of an op amp describes how quickly its output
differences between the V+ and V− supply pins. This supply
voltage can change (or slew) from one voltage to another. This
could be provided to the op amp as either a single voltage, with
specification is typically given in units of V/μs. The highest slew
V− connected to ground, or a bipolar supply with V+ and V−
rate a preamp circuit will need to support is given by
set to positive and negative voltages of the same value (for
SR = 2 × π × fMAX × VP example, ±15 V).
Here, fMAX is the highest frequency the preamp needs to support The supply voltage needs to be selected so that the op amp’s
(typically 20 kHz for audio) and VP is the peak voltage level at output does not clip given the supply rails. Some op amps are
the output of the op amp. If the peak output voltage is +12 V specified for rail-to-rail operation, which means that the input
(8.5 VRMS), then an op amp should have a slew rate of at least or output voltages (depending on what is being specified) can
1.5 V/μs. go all the way to the rail voltages without clipping. If the op amp
In practice, most audio signals will not reach the full-scale is not rail-to-rail, then the maximum input and output voltages
voltage at higher frequencies, but it is a good idea to design a will be specified in the data sheet; note that the maximum
preamp for that possibility. Slew rate is not a specification for positive and negative voltages will not always be of equal value.
which you typically have to over-design a circuit. You can find The signal’s peak output voltage obviously is directly related to
an op amp with a slew rate that is fast enough to handle the the amount of gain that the preamp circuit provides. The peak
highest frequency of interest for your design, but do not need to output voltage of the ADMP504 is 0.25 Vrms. An ADMP504
include significant margin above this limit. connected to a preamp with a gain of 20 dB (10× gain) will have
THD + N a peak output voltage of 2.5 Vrms, which is 7.0 VP-P. Thus, this
circuit needs a supply voltage of at least 7.0 V, or ±3.5, when
Talking about total harmonic distortion plus noise (THD + N)
using a rail-to-rail output op amp. The supply voltage will need
in an op amp circuit can quickly become a complicated
to be higher if the op amp’s outputs are not rail-to-rail.
discussion. Distortion can come from many different sources,
including slew rate limiting, output loading, and the op amp’s An analog MEMS microphone operates from a 1.5 V to 3.3 V
intrinsic distortion characteristics. THD is always specified as supply. The lowest supply that some of the op amps listed in
a ratio, either as a percentage or as a decibel number. This is the Table 1 support is 2.7 V, so in a low power single-supply circuit,
ratio of the amplitude of the harmonic distortion components the microphone’s VDD supply will need to be set between 2.7 and
of the signal to the fundamental input frequency, so a lower 3.3 V.
number (smaller percentage or more negative dB value) Gain Bandwidth Product
indicates better THD + N performance.
As its name indicates, the gain bandwidth product (GBP) is the
The THD + N specification includes the noise across a specified product of the amplifier’s bandwidth (using the low-pass −3 dB
bandwidth summed with the THD. This is not included in the corner) and gain applied to the input signal. Most preamp
spec table of all op amp data sheets, but even when it is not, a designs for MEMS microphones do not need to add more than
Rev. 0 | Page 3 of 8
AN-1165 Application Note
40 dB of gain, which is a factor of 100. Designing a preamp to Change. You can also find many other more specialized audio
have a bandwidth of at least 50 kHz provides some margin to circuits in the op amp data sheets.
ensure that the higher audio frequencies are not affected by the Noninverting
op amp’s bandwidth limit. An op amp with a 6.5 MHz GBP,
The output of a noninverting preamp circuit has the same
like the ADA4075-2, in a circuit with 40 dB of gain will have a
polarity as its input. This circuit is good to use in applications
pass band up to 65 kHz before the signal begins to roll off.
where it is critical that the signal polarity is not inverted. The
This specification is plotted in the typical performance gain of a noninverting op amp circuit is G = (R1 + R2)/R1, for
characteristics section of the data sheet as gain vs. frequency. a configuration as shown in Figure 4.
This graph from the ADA4075-2 data sheet (see Figure 3) R1 R2
shows that op amp’s open-loop gain vs. frequency.
140 140
VSY = ±15V
120 120 MEMS
MICROPHONE VOUT
100 PHASE 100 C1 VIN

10958-004
80 80
60 60
Figure 4. Noninverting Preamp Circuit
PHASE (Degrees)

40 40
GAIN (dB)

20 GAIN 20 This configuration has a very high input impedance because the
0 0 microphone signal is connected directly to the noninverting
–20 –20 input of the op amp. C1 is a dc-blocking capacitor that is used
–40 –40 because the MEMS microphone’s output is biased at 0.8 V. This
–60 –60 capacitor does not have to be very large in this configuration
–80 –80 because the input impedance of the op amp is very high.
–100 –100
Common-mode rejection is a spec that is of more concern for
10958-015

1k 10k 100k 1M 10M 100M


FREQUENCY (Hz) noninverting circuits than for inverting topologies. In a non-
Figure 3. ADA4075-2 Open-Loop Gain and Phase vs. Frequency inverting circuit, the common-mode voltage can cause
distortion in the output signal. Op amp data sheets typically
Other Specifications specify common-mode rejection ration (CMRR) vs. frequency,
There are some other specifications to consider, depending so you can look at this to determine the performance of a
on how specifically the circuit will be used. For example, if the specific device in the audio frequency band. This is not an issue
preamp needs to be able to drive a low impedance load, like for inverting circuits because there is no dynamic common-
headphones, then you will want to select an op amp with high mode voltage; both inputs are held at ground or virtual ground.
drive capability. Inverting
An op amp’s supply current is typically specified per amplifier Figure 5 shows an inverting op amp circuit. The polarity at the
with no load current. Low power circuit designs are obviously output of this circuit is inverted from the input and the gain is
best done with op amps that have a low supply current. The op G = −R2/R1.
amp noise and supply current are often inversely related, so a C1 R1 R2
tradeoff may need to be made for audio performance vs. power.
VIN
Finally, there are some other specificatons with which you do MEMS
MICROPHONE
not need to be concerned. Offset voltage is often presented as a
VOUT
key op amp spec, but is not critical for preamp applications that
10958-005

are ac-coupled.
CIRCUITS Figure 5. Inverting Preamp Circuit
Basic preamp circuits come in two configurations: inverting and The input impedance of an inverting circuit is equal to R1. This
noninverting. The uses and advantages of each are described in resistor forms a voltage divider with the MEMS microphone’s
this section. output, so the resistor’s value needs to be chosen so that it is
These circuits do not show power supplies or bypass capacitors. high enough not to load the microphone output, but not so
While important for good circuit performance, showing these high that it adds unnecessary noise to the circuit. The analog
is not critical for describing the preamp function. More MEMS microphones typically have an output impedance of
information about decoupling capacitors and proper grounding 200 Ω. If R1 is chosen to be 2.0 kΩ, the resulting voltage divider
for your design can be found in most op amp data sheets and will reduce the level of the signal output from the microphone
the AN-202 Application Note, An IC Amplifier User’s Guide to by 9%
Decoupling, Grounding, and Making Things Go Right for a VOUT = (2.0 kΩ + 200 Ω)/2.0 kΩ × VIN = 0.91 × VIN
Rev. 0 | Page 4 of 8
Application Note AN-1165
DC-blocking capacitor C1 and R1 will form a high-pass filter, the second output has two amplifiers worth of noise and
so C1 should be selected to be large enough so that this filter distortion. A second small issue is that there is a nonzero time
does not interfere with the microphone’s input signal. The delay through each amplifier, so the two legs of the differential
ADMP504 has a low frequency corner at 100 Hz. If R1 is again output are not perfectly aligned. However, this will likely have a
chosen to be 2.0 kΩ, then a capacitor of 2.2 μF results in a high- very small effect on the differential signal performance.
pass filter −3 dB frequency of 40 Hz, which is sufficiently below C1 VIN R1 R2
the microphone’s corner frequency.
A good rule of thumb is to select the cutoff frequency at least MEMS
MICROPHONE
one octave below that of the microphone, unless a specific high-
VOUT–
pass characteristic is sought.
Voltage Follower
A noninverting amplifier can also be implemented as a voltage
follower if no divider circuit is used in the feedback loop. This
circuit is useful for buffering the microphone’s output when it is R3 R4

not able to drive longer traces or cables directly, and may not
need additional gain added to the signal.
VOUT+

10958-008
MEMS
MICROPHONE VOUT
C1 VIN Figure 8. Differential Output Circuit
10958-006

The voltage follower and inverting amplifier circuit shown in


Figure 6. Voltage Follower Figure 7 can also be used to create a differential signal with a
gain of 1. The noninverted output would be taken from the
A voltage follower can also be used as a buffer in front of an
output of the voltage follower amplifier and the inverted output
inverting stage. This configuration may be desired so that lower
would be taken from the output of the inverting amp. In this
value resistors can be used in the inverting circuit. Without the
configuration, the values of R1 and R2 should be equal to
buffer, the input impedance of the inverting stage may need to
achieve unity gain.
be lower than is desired to achieve the target noise performance.
With this buffer in place and the first op amp’s low output A difference amp, like the AD8273, could also be used to create
impedance (compared to the MEMS microphone), the values of this single-ended-to-differential circuit and may have better
resistors R1 and R2 can be made very small so that they do not performance with regards to the issues mentioned above.
contribute extra noise to the circuit. Figure 9 shows the AD8273 configured as a single-ended-
MEMS
to-differential amplifier. Each amplifier is configured for
MICROPHONE R1 R2 G = 2, thus the differential gain is 4×. For more information
C1 VIN
about Analog Devices’ difference amplifiers go to
www.analog.com/difference_amplifiers.
VOUT +VS
10958-007

11
AD8273
Figure 7. Voltage Follower Buffer with Inverting Amplifier 6kΩ 12kΩ
12 2

Differential Output
MEMS 13 +OUT
A MEMS microphone’s single-ended output can be converted MICROPHONE
C1 6kΩ 12kΩ
into a differential signal with two op amps in a simple series 14 3

combination of two inverting circuit stages (see Figure 8). The


ouputs from each stage, which are inverted relative to each 6kΩ 12kΩ
other, then serve as the differential pair. In the circuit shown in 10 6

Figure 8, the signal amplification happens in the first stage and 9 –OUT
is set by R1 and R2. Resistors R3 and R4 should be of equal value to
6kΩ 12kΩ
provide unity gain in the second stage. For best performance, 8 5

1% resistors (or better) should be used to minimize the error


10958-009

4
between the two stages. –VS

One downside of this configuration is that one output has only Figure 9. AD8273 Single-Ended-to-Differential Configuration, G = 4
one amplifier’s contribution to the noise and distortion, while
Rev. 0 | Page 5 of 8
AN-1165 Application Note
OP AMP SELECTION PERFORMANCE SIMULATIONS
Analog Devices has a large selection of different op amps that Analog Devices provides a tool for simulating analog circuits.
are suitable for microphone preamp applications. Table 1 shows NI Multisim™ Component Evaluator Analog Devices Edition
the specs for some of these parts and is sorted by voltage noise. can be used to quickly build a circuit and display its perform-
Whether your application is targeted to be strictly the highest ance specs, including frequency response and noise level. This
performance, or whether you need to design a very cost version of Multisim includes most of the op amps discussed
sensitive circuit, there is an op amp available to meet your here in the library, enabling quick simulation without having to
needs. Interactive selection tools, data sheets, example download and manage SPICE models from different sources.
circuits, and other application notes can be found at Different components, including op amps, can be swapped in
www.analog.com. and out of the circuit to compare performance with these
different devices.

Table 1. Audio Op Amps


Noise GBP Slew Rate THD+N Supply Current/
Op Amp Channels (nV/√Hz) (MHz) (V/μs) (dB, 20 kHz) (V) Amp (mA) Package(s)
AD797 1 0.9 8 20 −120 10 to 36 10.5 PDIP, SOIC
ADA4898 1/2 0.9 65 55 −116 10 to 32 7.9 SOIC
ADA4897 1/2 1 230 120 3 to 10 3 SOIC, SOT-23
AD8597/AD8599 1/2 1.07 10 14 −120 10 to 36 4.8 SOIC, LFCSP
ADA4841 1/2 2.1 80 12 2.7 to 12 1.1 SOIC, SOT-23, MSOP, LFCSP
AD8655/AD8656 1/2 2.7 28 11 −102 2.7 to 5.5 11 MSOP, SOIC
ADA4075-2 2 2.8 6.5 12 −114 9 to 36 1.8 SOIC, LFCSP
OP37 1 3 12 17 8 to 44 5.67 SOIC, DIP
ADA4084-2 2 3.9 13.9 2.6 −83 3 to 30 0.565 MSOP, SOIC, LFCSP
AD8651/AD8652 1/2 4.5 50 41 −104 2.7 to 5.5 8 MSOP, SOIC
ADA4528 1 5.3 4 0.5 −80 2.2 to 5.5 1.4 MSOP, LFCSP
ADA4627 1 6.1 19 40 −127 10 to 30 7 SOIC, LFCSP
AD8605/AD8606/ 1/2/4 6.5 10 5 −86 2.7 to 6 1.2 SOT-23, WLCSP, SOIC, TSSOP
AD8608
AD8646/AD8648 2/4 8 24 11 −80 2.7 to 5.5 2 SOIC, MSOP, TSSOP
OP1177 1/2/4 8 1.3 0.7 5 to 36 0.5 MSOP, SOIC
AD8622/AD8624 2/4 11 0.56 10 −82 5 to 30 0.215 SOIC, MSOP
AD8515 1 22 5 2.7 1.8 to 6 0.5 SOT-23, SC70
AD8541/AD8542/ 1/2/4 40 1 0.75 2.7 to 5.5 0.045 SOT-23, SC70, SOIC
AD8544

Rev. 0 | Page 6 of 8
Application Note AN-1165
REFERENCES
NI Multisim Component Evaluator Analog Devices Edition. www.analog.com/multisim
Jung, Walt. Op Amp Applications Handbook Seminar Notes. www.analog.com/op_amp_applications_handbook
Self, Douglas. Small Signal Audio Design. Focal Press, 2010.
Analog Devices EngineerZone Audio Community. www.analog.com/audio_support_community

Rev. 0 | Page 7 of 8
AN-1165 Application Note

NOTES

©2012 Analog Devices, Inc. All rights reserved. Trademarks and


registered trademarks are the property of their respective owners.
AN10958-0-9/12(0)

Rev. 0 | Page 8 of 8

You might also like