Cirrus-logic AN168 Manuel d'utilisateur

Copyright © Cirrus Logic, Inc. 2006(All Rights Reserved)http://www.cirrus.comAN168Application NoteACOUSTIC PATH DESIGN FOR FULL-DUPLEX CELLULAR HANDS-

AN16810 AN168REV2maximum value of this curve is then noted, andgain is added or subtracted at the mic preamp inorder to set this maximum value to -9 d

AN168AN168REV2 11The register settings to accomplish the above are asfollows:reg 0: 47a0 (or c7a0 if internal mic preamp isused)reg 1: 26a2reg 2: 0004

AN16812 AN168REV2Configure the CS6422 from reset with thefollowing:1) Mic set to '1' or '0', depending on whether theinternal mic

AN168AN168REV2 13This parameter can be tested using theHalf-Duplex Alternate Counting Test as de-scribed in the Tests section.9) RSD controls the enab

AN16814 AN168REV2typically employ the sidetone only on the analoginterface, leaving the digital interfacesidetone-free.If it is not possible to disabl

AN168AN168REV2 151.5.1 Acoustic CouplingThe term ‘Acoustic Coupling’ refers to the gainbetween the AO and the APO pins on the CS6422.It includes the s

AN16816 AN168REV2The loop gain measurement procedure is asfollows:1) Configure the CS6422 with its default configu-ration, with the exception of the f

AN168AN168REV2 17the frequency of the howl itself is the frequency atwhich the loop gain, whose frequency response isdominated by the acoustic couplin

AN16818 AN168REV2There are two tests which should be performed, aFrequency Sweep distortion test and a Buzz test.1.5.2.1 Frequency Sweep TestIn this t

AN168AN168REV2 19Similar to harmonic distortion, these ‘buzzing’artifacts cause elevated levels of residual echobecause they result from a non-linear

AN1682 AN168REV2TABLE OF CONTENTS1. DESIGN PROCESS AND CONSIDERATIONS ... 11.1 D

AN16820 AN168REV2discussed earlier in this note.2) Configure the CS6422 from reset with the ex-ception of the following:a) Mic set to '1' or

AN168AN168REV2 21The difference between the two RVol values is theworst-case ERLE of the echo canceler. Here is thedetailed procedure:1) Set up the sp

AN16822 AN168REV2sults.6) Disconnect the signal source from NI (or sim-ply turn it off).7) Close the loop by setting NSdt to ‘-12dB’,RVol to ‘0 dB’, a

AN168AN168REV2 23the two RVol values in dB.1.5.4 Call Testing and Coefficient Optimi-zationThe following tests are useful for optimizing theCS6422 reg

AN16824 AN168REV2reg 2: 0a14reg 3: a046reg 4: 5008reg 5: 018aFor the EC Convergence Test, we also disableHalf-Duplex mode and the Transmit Suppressor

AN168AN168REV2 251.5.4.3 Half-duplex alternate countingThe Half-duplex alternate counting test tests thehalf-duplex behavior of the system. This test

AN16826 AN168REV22) Decoupling and loading capacitors should beplaced as close as possible to the pins they de-couple (AVDD, AGND, DVDD, DGND, MB,CLKI

AN168AN168REV2 271.6.2.3 +12VBATT/BATTGND Components1) Cell-phone battery charger circuitry2) Speaker driver (if powered from +12V source)Most cellula

AN16828 AN168REV21.7.4 Acoustic couplingThe maximum acoustic coupling (acrossfrequency) between the speaker and themicrophone should be limited to -9

AN168AN168REV2 292. APPENDIX - EXAMPLE SPEAKER DRIVER CIRCUITSIn the following pages are 5 example speaker drivercircuits. In each case, the maximum g

AN168AN168REV2 3LIST OF FIGURESFigure 1. Speaker Distortion...

AN16830 AN168REV2Gain = R2/R1C1 = 1/[2 * PI * R1 * (300 Hz)]C2 = 1/[2 * PI * R2 * (4 kHz)]CS6422_A0Filt_OutVbiasVbias+5V+5VU3MC33202+-U1MC33202+-U2MC3

AN168AN168REV2 31+46dB-28.2dB15 Watts into 4 ohmsGain = +17.8dBGain = [R2/(R1+R2)]*[200]CS6422_A0+14V+-+-U1TDA1519A932564781+C310uFC60.1uFC11800pF+C51

AN16832 AN168REV2+40dB-29.3dB3 Watts into 4 ohmsGain = +10.7dBGain = [R2/(R1+R2)]*[1+R4/R5]CS6422_A0+14V+C81000uFR1340KR212.1KR339.2R52.21ohmR61ohmR42

AN168AN168REV2 332.3 Example 3: TDA1905 -- 2.5 Watts into 4ΩThe schematic for the TDA1905 speaker driver isshown in Figure 24. The TDA1905, from ST, i

AN16834 AN168REV22.4 Example 4: LM1877-- 2 Watts into 4 ΩThe schematic for the LM1877 speaker driver isshown in Figure 25. The LM1877, from NationalSe

AN168AN168REV2 35+26dB-17dB2 Watts into 4 ohmsGain = +9dBGain = [R2/(R1+R2)]*[1+R3/R5+R6/R5]CS6422_A0+14VR212.1KR173.2K+C310uFC90.1uFR92.7ohmLS14 Ohm

AN16836 AN168REV21 Watt into 4 ohmsGain = +6dBGain = (2)*(R2/R1)CS6422_A0+5V+-U?LM486164237158LS14 Ohm SPEAKERC40.1uF+C5100uFC23300pFC10.047uFR112.1K+

AN1684 AN168REV21.2 Mechanical DesignThe performance of full-duplex hands-free designsis strongly influenced by the mechanical hardware,far more so th

AN168AN168REV2 51.2.1.3 Speaker Housing RequirementsThe quality of the speaker housing affects theperformance of the system because the speaker canind

AN1686 AN168REV2The Appendix lists five example speaker drivercircuits that are suitable for full-duplex hands-freesystems.1.3.2 Setting the Speaker D

AN168AN168REV2 71.3.3 Volume ControlIn most half-duplex systems, volume control isimplemented by changing the gain of the speakerdriver. In a full-dup

AN1688 AN168REV21.3.4 Acoustic CouplingFigure 5 shows the three most common places fordistortion to be introduced into the acoustic path.These are the

AN168AN168REV2 9Speaker DriverMic PreampSpeakerMicrophoneAirCouplingADCAOAPODAC123CS6422Figure 5. Three Common Sources of Acoustic Path DistortionSpe