 |
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
| Datasheet File OCR Text: |
| features cmos stereo audio input/output system delta-sigma a/d converters delta-sigma d/a converters input anti-aliasing and output smoothing filters programmable input gain and output attenuation sample frequencies of 4 khz to 50 khz cd quality noise and distortion < 0.01 %thd internal 64x oversampling low power dissipation: 80 ma 1 ma power-down mode general description the cs4216 stereo audio codec is a monolithic cmos device for computer multimedia, automotive, and portable audio applications. it performs a/d and d/a conversion, filtering, and level setting, creating 4 audio inputs and 2 audio outputs for a digital computer system. the digital interfaces of left and right channels are multiplexed into a single serial data bus with word rates up to 50 khz per channel. up to 4 cs4216 de- vices can be attached to a single hardware bus. both the adcs and the dacs use delta-sigma modula- tion with 64x oversampling. the adcs include a digital decimation filter which eliminates the need for external anti-aliasing filters. the dacs include output smoothing filters on-chip. ordering information: cs4216-kl 0 to 70 c 44-pin plcc CS4216-KQ 0 to 70 c 44-pin tqfp cdb4216 evaluation board 16-bit stereo audio codec semiconductor corporation cs4216 ssync lout rout digital filters serial interface control digital filters input mux voltage reference clkin mf7:sfs1/f2 mf8:sfs2/f3 sdin sdout sclk smode1 reset refgnd refbyp refbuf lin1 lin2 rin1 rin2 vd va dgnd agnd a/d a/d d/a d/a do1 mf5:do2/int mf2:do3/f2/cdin mf1:do4/f1/cdout di1 mf6:di2/f1 mf3:di3/f3/cclk output mute output attenuation input gain pdn smode2 smode3 power control mf4:di4/ma/ccs oct 93 ds83f2 1 crystal semiconductor corporation p.o. box 17847, austin, tx 78760 (512) 445-7222 fax: (512) 445-7581 copyright ? crystal semicondutor corporation 1993 (all rights reserved) the cs4216 is an mwave tm audio codec.
analog characteristics ( t a = 25 c; va, vd = +5v; input levels: logic 0 = 0v, logic 1 = vd; 1 khz input sine wave; clkin = 24.576 mhz; sm1; conversion rate = 48 khz; sclk = 12.288 mhz; measurement bandwidth is 10 hz to 20 khz; unless otherwise specified.) parameter * symbol min typ max units analog input characteristics - minimum gain setting (0 db); unless otherwise specified. adc resolution 16 - - bits adc differential nonlinearity (note 1) - - 0.9 lsb instantaneous dynamic range idr 80 85 - db total harmonic distortion thd - - 0.01 % interchannel isolation - 80 - db interchannel gain mismatch - - 0.5 db frequency response (note 1) -0.5 - +0.2 db programmable input gain span 21 22.5 24 db gain step size - 1.5 - db absolute gain step error - - 0.75 db gain drift - 100 - ppm/ c offset error dc coupled inputs - 10 100 lsb ac coupled inputs - 150 400 lsb full scale input voltage 2.5 2.8 3.1 v pp input resistance (notes 1,2) 20 - - k w input capacitance (note 1) - - 15 pf notes: 1. this specification is guaranteed by characterization, not production testing. 2. input resistance is for the input selected. non-selected inputs have a very high (>1m w ) input resistance. * parameter definitions are given at the end of this data sheet. mwave tm is a trademark of the ibm corporation. specifications are subject to change without notice. recommended operating conditions (agnd, dgnd = 0v, all voltages with re- spect to 0v.) parameter symbol min typ max units power supplies: digital vd 4.75 5.0 5.25 v analog va 4.75 5.0 5.25 v operating ambient temperature t a 02570 c cs4216 2 ds83f2 analog characteristics (continued) parameter * symbol min typ max units analog output characteristics - minimum attenuation; unless otherwise specified. dac resolution 16 - - bits dac differential nonlinearity (note 1) - - 0.9 lsb total dynamic range tdr - 93 - db instantaneous dynamic range idr 80 83 - db total harmonic distortion (note 4) thd - - 0.02 % interchannel isolation (note 4) - 80 - db interchannel gain mismatch - - 0.5 db frequency response (note 1) -0.5 - +0.2 db programmable output attenuation span (note 3) -45 -46.5 - db attenuation step size (note 3) - 1.5 - db absolute attenuation step error (note 3) - - 0.75 db gain drift - 100 - ppm/ c refbuf output voltage (note 5) 1.9 2.2 2.5 v maximum output current= 400 m a offset voltage - 10 - mv full scale output voltage (note 4) 2.5 2.8 3.1 v pp deviation from linear phase (note 1) - - 1 degree out of band energy (22 khz to 100 khz) - -60 - db power supply power supply current (note 6) operating - 80 100 ma power down - - 1 ma power supply rejection (1 khz) - 40 - db notes: 3. tested in sm3, slave sub-mode, 128 bpf. 4. 10 k w , 100 pf load. 5. refbuf load current must be dc. to drive dynamic loads, refbuf must be buffered. ac variations in refbuf current may degrade adc and dac performance. 6. typically current: va = 30ma, vd = 50ma. power supply current does not include output loading. * parameter definitions are given at the end of this data sheet. cs4216 ds83f2 3 switching characteristics (t a = 25 c; va, vd = +5v, outputs loaded with 30 pf; input levels: logic 0 = 0v, logic 1 = vd) parameter symbol min typ max units input clock (clkin) frequency sm1: clkin 2.048 24.576 25.6 mhz sm2, sm3, sm4: clkin 1.024 12.288 12.8 mhz clkin low time t ckl 15 - - ns clkin high time t ckh 15 - - ns sample rate (note 1) fs 4 - 50 khz di pins setup time to sclk edge (note 1) t s2 10 - - ns di pins hold time from sclk edge (note 1) t h2 8- -ns do pins delay from sclk edge t pd2 30 - - ns sclk and ssync output delay master mode (note 1) t pd3 - - 50 ns from clkin rising sclk period master mode (note 7) t sckw - 1/(fs*bpf) - s slave mode 75 - - ns sclk high time slave mode t sckh 30 - - ns sclk low time slave mode t sckl 30 - - ns sdin, ssync setup time to sclk edge slave mode t s1 15 - - ns sdin, ssync hold time from sclk edge slave mode t h1 10 - - ns sdout delay from sclk edge t pd1 - - 28 ns output to hi-z state bit 64 (note 1) t hz - - 12 ns output to non-hi-z bit 1 (note 1) t nz 15 - - ns reset pulse width low 500 - - ns ccs low to cclk rising sm4 (note 1) t cslcc 25 - - ns cdin setup to cclk falling sm4 (note 1) t discc 15 - - ns cclk low to cdin invalid (hold time) sm4 (note 1) t ccdih 10 - - ns cclk high time sm4 (note 1) t cclhh 25 - - ns cclk low time sm4 (note 1) t cclhl 25 - - ns cclk period sm4 (note 1) t cclkw 75 - - ns cclk rising to cdout data valid sm4 (note 1) t ccdov - - 30 ns cclk rising to cdout hi-z sm4 (note 1) t ccdot - - 30 ns cclk falling to ccs high sm4 (note 1) t cccsh 0- -ns notes: 7. when the cs4216 is in master mode (ssync and sclk outputs), the sclk duty cycle is 50%. the equation is based on the selected sample frequency (fs) and the number of bits per frame (bpf). cs4216 4 ds83f2 t nz * optional bit 1 bit 2 bit 1 bit 2 frame sync t sckl t sckh t s1 t h1 t sckw t sckh t sckl t h1 t s1 t s1 t h1 ssync [sm1, sm2\ sclk [sm1,sm2\ sclk [sm3,sm4\ ssync [sm3,sm4\ sdin sdout t pd1 t pd1 [sm1,sm2,sm3\ (sm4) [sm1,sm2,sm3\ (sm4) *word sync t hz *word sync bit 63 bit 64 bit 63 bit 64 (bit 32) (bit 31) (bit 31) (bit 32) bit 33 bit 33 (bit 1) (bit 1) t s1 t h1 bit 32 (bit 32) bit 32 (bit 32) serial audio port timing mf3:cclk mf1:cdout mf2:cdin mf4:ccs t ccdov 0msk do1 t ccdih t discc 123 5 89 467 10 adv lcl 11 latt4 latt3 latt2 latt1 latt0 ratt4 ratt3 ratt2 t cclkl t cclkh t cslcc t cclkw mf3:cclk mf1:cdout mf2:cdin mf4:ccs t ccdot 24 28 29 32 31 30 27 26 25 23 22 rgain2 rgain1 0 0 1 rgain0 err1 err0 lcl rcl di1 0 0 0 0 0 00 0 adv t cccsh serial mode 4. control data serial port timing cs4216 ds83f2 5 digital characteristics (t a = 25 c; va, vd = 5v) parameter symbol min typ max units high-level input voltage v ih vd-1.0 - - v low-level input voltage v il --1.0v high-level output voltage at i 0 = -2.0 ma v oh vd-0.3 - - v low-level output voltage at i 0 = +2.0 ma v ol --0.1v input leakage current (digital inputs) - - 10 m a output leakage current (high-z digital outputs) - - 10 m a output capacitance c out - - 15 pf input capacitance c in - - 15 pf sclk* t s2 t h2 t pd2 dix dox * sclk is inverted for sm1 and sm2 di/do timing clkin t ckl sclk sclk & ssync output timing ssync t ckh t pd3 (master mode) (master mode) cs4216 6 ds83f2 a/d decimation filter characteristics parameter symbol min typ max units passband (fs is conversion freq.) 0 - 0.45fs hz frequency response -0.5 - +0.2 db passband ripple - - 0.2 db transition band 0.45fs - 0.55fs hz stop band 3 0.55fs - - hz stop band rejection 80 - - db group delay - 16/fs - s group delay variation vs. frequency - 0.0 m s d/a interpolation filter characteristics parameter symbol min typ max units passband (fs is conversion freq.) 0 - 0.45fs hz frequency response -0.5 - +0.2 db passband ripple - - 0.1 db transition band 0.45fs - 0.55fs hz stop band 3 0.55fs - - hz stop band rejection 74 - - db group delay - 16/fs - s group delay variation vs. frequency - - 0.1/fs m s absolute maximum ratings (agnd, dgnd = 0v, all voltages with respect to 0v.) parameter symbol min typ max units power supplies: digital vd -0.3 - 6.0 v analog va -0.3 - 6.0 v input current (except supply pins) - - 10.0 ma analog input voltage -0.3 - va+0.3 v digital input voltage -0.3 - vd+0.3 v ambient temperature (power applied) -55 - +125 c storage temperature -65 - +150 c warning: operation beyond these limits may result in permanent damage to the device. normal operation is not guaranteed at these extremes. cs4216 ds83f2 7 0.1 m f 2.0 1 m f 0.1 m f + 24 4 va vd 23 5 +5v supply 1 m f + cs4216 agnd dgnd mf3:di3/f3/cclk mf6:di2/f1 mf1:do4/f1/cdout mf2:do3/f2/cdin 36 35 34 40 39 38 mf5:do2/int mf4:di4/ma/ccs p arallel bits sub-mode or settings or control port di1 do1 33 37 rin2 26 line in 2 right ferrite bead note: agnd and dgnd pins must be on the same ground plane + 40 k 40 k rout lout 15 16 refbyp 21 0.1 m f + 10 m f refgnd 22 + > 1.0 m f > 1.0 m f 0.0022 m f npo 600 0.0022 m f npo 600 right audio output left audio output pdn sdout sclk ssync reset sdin 2 13 42 44 1 43 c ontroller clkin 3 mf7:sfs1 mf8:sfs2 smode1 31 30 29 mode setting 32 smode2 41 smode3 lin2 28 line in 2 left rin1 25 line in 1 right refbuf 20 0.47 m f to optional input buffers lin1 27 line in 1 left +5v analog if a separate +5v analog supply is used, remove the 2.0 ohm resistor see analog inputs section for suggested input ciruits. refer to the analog inputs section for terminating unused line inputs. all other unused inputs should be tied to g nd . all n c pins should be left floating. figure 1. typical connection diagram cs4216 8 ds83f2 overview the cs4216 contains two analog-to-digital con- verters, two digital-to-analog converters, adjustable input gain, and adjustable output level control. since the converters contain all the re- quired filters in digital or sampled analog form, the filters frequency responses track the sample rate of the cs4216. only a single-pole rc filter is required on the analog inputs and outputs. the rc filter acts as a charge reserve for the switched-capacitor input and buffers op-amps from a switched-capacitor load. communication with the cs4216 is via a serial port, with sepa- rate pins for data into the device, and data from the device. the filters and converters operate over a sample rate range of 4 khz to 50 khz. functional specifications analog inputs and outputs figure 1 illustrates the suggested connection dia- gram to obtain full performance from the cs4216. the line level inputs, lin1 or lin2 and rin1 or rin2, are selected by an internal input multiplexer. this multiplexer is a source selector and is not designed for switching be- tween inputs at the sample rate. unused analog inputs that are not selected have a very high input impedance, so they may be tied to agnd directly. unused analog inputs that are selected should be tied to agnd through a 0.1 m f capacitor. this prevents any dc current flow. the analog inputs are single-ended and inter- nally biased to the refbuf voltage (nominally 2.2 v). the refbuf output pin can be used to level shift an input signal centered around 0 volts as shown in figure 2. the input buffers shown have a gain of 0.5, yielding a full scale input sensitivity of 2 v rms with the cs4216 pro- grammable gain set to 0. if the source imped- ance is very low, then the inputs can be ac coupled with a series 0.47 m f capacitor, elimi- nating the need for external op-amps (see figure 3). however, the use of ac coupling capacitors will increase dc offset at 0db gain (see analog characteristics table). the analog outputs are also single-ended and centered around the refbuf pin. ac coupling capacitors of >1 m f are recommended. line in right example op-amps are mc34072 or lt1013 line in left op-amps are run from va+5v and agnd rinx (plcc pin 25 or 26) refbuf linx (plcc pin 27 or 28) _ + _ + 56 pf 10 k 20 k 0.47 uf 0.47 uf 0.47 uf 20 k 150 5 k 56 pf 10 k 150 0.01 uf npo 0.01 uf npo figure 2. dc coupled input. line in right line in left 0.01 uf npo 0.01 uf npo 150 150 0.47 uf 0.47 uf rinx (plcc pin 25 or 26) linx (plcc pin 27 or 28) figure 3. ac coupled input cs4216 ds83f2 9 offset calibration both input and output offset voltages are mini- mized by internal calibration. offset calibration occurs after exiting a reset or power down condi- tion. during calibration, which takes 194 frames, output data from the adcs will be all zeros, and will be flagged as invalid. also, the dac outputs will be muted. after power down mode or power up, reset should be held low for a minimum of 50 ms to allow the voltage reference to settle. input gain and output level setting input gain is adjustable from 0 db to +22.5 db in 1.5 db steps. in serial modes sm1 and sm2, the output level attenuation is adjustable from 0 db to -22.5 db. in serial modes sm3 and sm4, the output level attenuation is adjustable from 0 db to -46.5 db. both input and output gain adjustments are internally made on zero- crossings of the analog signal, to minimize "zipper" noise. the gain change automatically takes effect if a zero crossing does not occur within 512 frames. muting and the adc valid counter the mute function allows the output channels to be silenced. it is the controlling processors re- sponsibility to reduce the signal level to a low value before muting, to avoid an audible click. the outputs should be muted before changing the sample frequency. the serial data stream contains a "valid data" indicator for the a/d converters which is false until enough clocks have passed since reset, or low-power (power down mode) operation to have valid a/d data from the filters, i.e., until calibra- tion time plus the full latency of the digital filters has passed. parallel digital input/output pins parallel digital inputs are general purpose pins whose value is reflected in the serial data output stream to the processor. parallel digital outputs provide a way to control external devices using bits in the serial data input stream. all parallel digital pins, with the exception of di1 and do1, are multifunction and are defined by the serial mode selected. serial modes 1 and 2 define all multifunction pins as general purpose digital in- puts and outputs. in serial mode 3 only two digital inputs and two digital outputs are avail- able. in serial mode 4 only one digital input and digital output exists. figure 4 shows when the di pins are latched, and when the do pins are up- dated in sm3 and sm4. reset and power down modes reset places the cs4216 into a known state and must be held low for at least 50 ms after power- up or a hard power down. reset must also occur when the codec is in master mode and a change in sample frequency is desired. in reset, the digi- tal outputs are driven low. reset sets all control data register bits to zero. hard power down mode may be initiated by bringing the pdn pin low. all analog outputs will be driven to the refbuf voltage which will then decay to zero. all digital outputs will be driven low and then will go to a high imped- ance state. minimum power consumption will occur if clkin is held low. after leaving the power down state, reset should be held low for 50 ms to allow the analog voltage reference to settle before calibration is started. ssync sclk di pins latched do pins update start of frame (sm3) figure 4. digital input/output timing cs4216 10 ds83f2 alternatively, soft power down may be initiated, in slave mode, by reducing the sclk frequency below the minimum clkin/12. in soft power down the analog outputs are muted and the serial data from the codec will indicate invalid data and the appropriate error code. the parallel bit i/o is still functional in soft power down mode. this is, in effect, a low power mode with only the parallel bit i/o unit functioning. audio serial interface in serial modes 1, 2, and 3, the audio serial port uses 4 pins: sdout, sdin, sclk and ssync. sdin carries the d/a converters input data and control bits. input data is ignored for frames not allocated to the selected cs4216. sdout car- ries the a/d converters output data and status bits. sdout goes to a high-impedance state during frames not allocated to the selected cs4216. sclk clocks data in to and out of the cs4216. the rising edge of sclk clocks data out on sdout. the falling edge latches data on sdin into the port (sclk polarity is inverted in serial modes 1&2). ssync indicates the start of a frame and/or sub-frame. sclk and ssync must be synchronous to the master clock. serial mode 4 is similar to serial mode 3 with the exception of the control information. in serial mode 4 the control information is entered through a separate asynchronous control port. therefore, the audio serial port only contains audio data which reduces the number of bits on the audio port from 64 to 32 per codec. the serial port protocol is based on frames con- sisting of 1, 2, or 4 sub-frames. the frame rate is the system sample rate. each sub-frame is used by one cs4216 device. up to 4 cs4216s may be attached to the same serial control lines. sfs1 and sfs2 are tied low or high to indicate to each cs4216 which sub-frame is allocated for it to use. serial data format in serial modes 1, 2, and 3, a sub-frame is 64 bits in length and consists of two 16-bit audio values and two 16-bit control fields. in serial mode 4 a sub-frame is 32 bits in length and only contains the two 16-bit audio values; the control data is loaded through a separate port. the audio data is msb first, 2s complement format. the sub-frame bit assignments for serial modes 1, 2, and 3, are numbered 1 through 64 and are shown in figures 5 and 6. control data bits all reset to zero. cs4216 serial interface modes the cs4216 has 4 serial port modes, selected by the smode1, smode2 and smode3 pins. in all modes, clkin, sclk and ssync must be derived from the same clock source. sm1 is an easy interface to asics that use a change in the sclk-to-clkin ratio to determine the sample smode pins serial sclk bit sub-frame bits per sclk & master 3 2 1 mode center width frame (bpf) ssync frequency 0 0 0 sm1 rising 64 bits 256 slave clkin = 512 fs 0 0 1 sm2 rising 64 bits 256 slave sclk = 256 fs 0 1 0 sm3 falling 64 bits 64/128/256 master/slave clkin/sclk = 256 fs 011 factory test mode 1 x x sm4 falling 32 bits ? 32/64/128 ? master/slave clkin = 256 fs ? contains audio data only. control information is entered through a separate serial port. table 1. serial port modes cs4216 ds83f2 11 * ? sub-frame bits 17 to 24 17 18 19 20 21 22 23 24 0000expmute islisr exp expand bit reserved. must be set to zero. mute mute d/a outputs 0 - normal outputs 1 - mute outputs isl select left input mux 0 - select lin1 1 - select lin2 isr select right input mux 0 - select rin1 1 - select rin2 input data bit definitions sub-frame bits 1 to 16 left dac audio data, msb first, 2s comple- ment coded. sub-frame bits 25 to 32 25 26 27 28 29 30 31 32 lg3 lg2 lg1 lg0 rg3 rg2 rg1 rg0 lg3-lg0 sets left input gain. lg3 is the msb. lg0 represents 1.5 db. 0000 = no gain. 1111 = +22.5 db gain rg3-rg0 sets right input gain. rg3 is the msb. rgo represents 1.5 db. 0000 = no gain sub-frame bits 51 to 60 51 52 53 54 55 56 57 58 59 60 la4 la3 la2 la1 la0 ra4 ra3 ra2 ra1 ra0 0 0 la3 la2 la1 la0 ra3 ra2 ra1 ra0 la4-la0 sets left output attenuation ra4-ra0 sets right output attenuation *sm3,4 la4 is the msb. 00000 = no attenuation 11111 = -46.5 db ?sm1, 2 la3 is the msb. 0000 = no attenuation 1111 = -22.5 db la0 represents 1.5 db. ?sm1, 2 ra3 is the msb. 0000 = no attenuation 1111 = -22.5 db *sm3,4 ra4 is the msb. 00000 = no attenuation 11111 = -46.5 db ra0 represents 1.5 db. sub-frame bits 61 to 64 61 62 63 64 do1 do2 do3 do4 do1-do4 set the logic level on the 4 digital output pins. in sm3 do3 and do4 are not available. in sm4 do2, do3, & do4 are not available. sub-frame bits 33 to 48 right dac audio data msb first, 2s comple- ment coded. 64 dac - left word 0000 left a/d gain 30 right a/d gain 30 m in sel. 0000 dac - right word left d/a att. 30 right d/a att. 30 do1 do2 do3 do4 sub-frame word b word a dac - left word 0000 left a/d gain 30 right a/d gain 30 exp m in sel. sm1 and sm2 sm3 00 dac - right word left d/a att. 40 right d/a att. 40 do1 do2 xx msb lsb msb lsb msb lsb msb lsb 01 16 17 21 24 25 28 29 32 22 23 33 52 48 53 56 57 60 61 51 55 exp sub-frame figure 5. serial data input format - sm1, sm2, and sm3. sub-frame bits 49 to 50 must be zero. cs4216 12 ds83f2 sub-frame bits 17 to 24 17 18 19 20 21 22 23 24 reserved 0 adv lcl rcl adv adc valid data bit. 0 - invalid adc data 1 - valid adc data indicates adc has completed initialization after power-up, low power mode, or mute. lcl left adc clipping indicator 0 - normal 1 - clipping rcl right adc clipping indicator 0 - normal 1 - clipping reserved bits can be 0 or 1 sub-frame bits 25 to 32 25 26 27 28 29 30 31 32 er3 er2 er1 er0 ver3 ver2 ver1 ver0 er3-er0 error word 0000 - normal C no errors. 0001 - input sub-frame bit 21 is set. control data will not be loaded 0010 - sync pulse is incorrect. causes the analog output to mute. 0011 - sclk is outside the allowable range. analog output mutes. ver3-ver0 cs4216 version number 0000 = "a" (see appendix a) 0001 = "b", "c", . . . (this data sheet) sub-frame bits 33 to 48 right adc audio data, msb first, 2s comple- ment coded. sub-frame bits 1 to 16 left adc audio data, msb first, 2s comple- ment coded. output data bit definitions 64 adc - left word 30 30 0 adv adc - right word sub-frame word b word a adc - left word 30 30 0 sm1 and sm2 sm3 adc - right word di1 di2 xx msb msb lsb lsb adv lcl rcl error error version version msb msb lsb lsb 00010000 00010000 di1 di2 di3 di4 01 16 17 21 24 25 28 29 32 22 23 33 52 48 53 56 57 60 61 55 lcl rcl . xxxx xxxx xxxx xxxx sub-frame figure 6. serial data output format - sm1, sm2, and sm3. sub-frame bits 49 to 60 these bits are reserved, and can be 0 or 1. sub-frame bits 61 to 64 61 62 63 64 di1 di2 di3 di4 di1-di4 these bits follow the state of the digital input pins. in sm3 di3 and di4 are used and unavailable. in sm4 di2, di3, & di4 are not available as input bits. cs4216 ds83f2 13 frequency. sm2 is similar to sm1 except that clkin is not used and sclk becomes the mas- ter clock and is fixed at 256 fs. sm3 was designed as an easy interface to general purpose dsps and provides extra features such as one more bit of attenuation, a master mode, and vari- able frame sizes. sm4 is similar to sm3 but splits the audio data from the control data thereby reducing the audio serial bus bandwidth by half. the control data is transmitted through a control serial port in sm4. table 1 lists the serial port modes available, along with some of the differences between modes. the first three columns in table 1 select the serial mode. the "sclk bit center" column indicates whether sclk is rising or falling in the center of a bit period. the "sub-frame width" column indicates how many bits are in an individual codecs sub-frame. sm4 differs from all other modes by separating the control data from the audio data. in both sm1 and sm2, there are 256 bits per frame which allows up to four codecs to occupy the same bus. in sm3 and sm4, the number of bits per frame is program- mable. in sm1 and sm2, sclk and ssync must be generated externally; whereas, in sm3 and sm4 the cs4216 can optionally generate those signals. in all modes, sclk and ssync must be synchronous to the master clock. the last column in table 1 lists the master frequency used by the codec. in sm1, the master fre- quency, input on clkin, is 512 times the highest sample frequency available. in sm2, the master frequency is fixed at 256 times the sam- ple frequency and, in this mode, sclk is the master clock. in sm3, the master frequency is 256 times the highest frequency available and is input on clkin or sclk, based on the sub- mode used. in sm4, the master frequency is also 256 times the highest frequency available and is input on clkin. serial mode 1, sm1 serial mode 1 is a slave mode selected by set- ting smode3 = smode2 = smode1 = 0. sclk and sync must be synchronous the mas- ter clock. sm1 uses a two bit wide (minimum) frame sync with an optional word sync. in this mode, ssync low for one sclk period fol- lowed by ssync high for a minimum of two sclk periods indicates the beginning of a frame. the first bit of the frame starts with the rising edge of ssync. an optional word sync, being one sclk period high, may be used to indicate the start of a new 32-bit word. figures 5 and 6 contain the serial data format for sm1. in this serial mode, the ratio of two clocks are used to select sample frequency. these are the master clock clkin and the serial clock sclk. clkin should be set to 512 fs max , where fs max is the maximum required sample rate. sclk must be externally set to a value of clkin/n, such that sclk equals 256 times the desired sample rate. the codec uses the ratio be- tween clkin and sclk to set the internal sample frequency and causes the cs4216 to go into soft power down mode if the sclk fre- quency drops to externally. in master sub-mode, the serial port signal transitions are controlled with respect to the internal analog sampling clock to minimize the amount of digital noise coupled into the ana- log section. since ssync and sclk are externally derived in slave sub-mode, optimum noise management cannot be obtained; therefore, master sub-modes should be used whenever pos- sible. master sub-mode (sm3) master sub-mode is selected by setting mf4:ma = 1, which configures ssync and sclk as outputs from the cs4216. during power down, ssync and sclk are driven high impedance, and during reset they both are driven low. in master sub-mode the number of bits per frame determines how many codecs can occupy the serial bus and is illustrated in figure 8. bits per frame (master sub-mode) mf8:sfs2 selects the number of bits per frame. the two options are mf8:sfs2 = 1 which se- lects 128 bits per frame, and mf8:sfs2 = 0 which selects 64 bits per frame. selecting 128 bits per frame (mf8:sfs2 = 1) al- lows two cs4216s to operate from the same serial bus since each codec requires 64 bit peri- ods. the sub-frame used by an individual codec is selected using mf7:sfs1. mf7:sfs1 = 0 se- lects sub-frame 1 which is the first 64 bits following the ssync pulse. mf7:sfs1 = 1 se- lects sub-frame 2 which is the last 64 bits of the frame. selecting 64 bits per frame (mf8:sfs2 = 0) al- lows only one cs4216 to occupy the serial port. since there is only one sub-frame (which is equal to one frame), mf7:sfs1 is defined differ- ently in this mode. mf7:sfs1 selects the format of ssync. mf7:sfs1 = 0 selects an ssync pulse one sclk period high, directly preceding the data as shown in the center portion of fig- ure 8. this format is used for all other master and slave sub-modes in sm3. if mf7:sfs1 = 1, an alternate ssync format is chosen in which ssync is high during the entire word a (32 bits), which includes the left sample, and low for the entire word b (32 bits), which in- cludes the right sample. this alternate format for ssync is illustrated in the bottom portion of figure 8 and is only available in master sub- mode with 64 bits per frame. a more detailed timing diagram for the 64 bits-per-frame master sub-mode is shown in figure 9. sample frequency selection (master sub-mode) in sm3, master sub-mode, the multifunction pins mf1:f1, mf2:f2, and mf3:f3 are used to select the sample frequency divider. table 3 lists the decoding for the sample frequency select pins where the sample frequency selected is clkin/n. also shown are the sample frequen- cies obtained by using one of two example master clocks: either 12.288 mhz or 11.2896 mhz. the codec must be reset when changing sample frequencies to allow the codec to calibrate to the new sample frequency. slave sub-mode (sm3) in sm3, slave sub-mode is selected by setting mf4:ma = 0 which configures ssync and sclk as inputs to the cs4216. these two sig- nals must be externally derived from clkin. in slave sub-mode, the phase relationship between sclk/ssync and clkin cannot be controlled since sclk and ssync are externally derived. therefore, the noise performance may be slightly worse than when using the master sub-mode. the number of sub-frames on the serial port is selected using mf1:f1 and mf2:f2. in slave sub-mode mf3:f3 works as a general purpose input. figures 10 through 12 illustrate the slave sub-mode formats. cs4216 16 ds83f2 sub-frame 1 sub-frame 2 frame (n+2) word a word b word a word b sub-frame 1 word a word b frame n 128 sclk periods sub-frame 2 word a word b frame (n+3) sub-frame 1 word a word b sub-frame 1 word a word b frame n 64 sclk periods sub-frame 1 word a word b frame (n+1) sub-frame 1 word a word b frame (n+2) sub-frame 1 word a word b frame (n+3) sub-frame 1 word a word b frame (n+4) sub-frame 1 word a word b frame n 64 sclk periods sub-frame 1 word a word b frame (n+1) sub-frame 1 word a word b frame (n+2) sub-frame 1 word a word b frame (n+3) sub-frame 1 word a word b frame (n+4) ssync data ssync data ssync data sfs2 sfs1 frame 1 1 0 1 mf8: mf7: sub- 1 2 sfs2 sfs1 frame 00 mf8: mf7: sub- 1 sfs2 sfs1 frame 01 mf8: mf7: sub- 1 figure 8. sm3, master sub-mode. msb sclk ssync (mf7:sfs1=0) ssync (mf7:sfs1=1) msb 32 clocks 32 clocks lsb word a word b lsb sdin sdout figure 9. detailed master sub-mode, 64 bpf. cs4216 ds83f2 17 bits per frame (slave sub-mode) in slave sub-mode, mf1:f1 and mf2:f2 select the number of bits per frame which determines how many cs4216s can occupy one serial port. table 4 lists the decoding for mf1:f1 and mf2:f2. when set for 64 sclks per frame, one device occupies the entire frame; therefore, a sub-frame is equivalent to a frame. mf7:sfs1 and mf8:sfs2 must be set to zero. see figure 10. when set for 128 sclks per frame, two devices can occupy the serial port, with mf7:sfs1 se- lecting the particular sub-frame. mf8:sfs2 must be set to zero. see figure 11. when set for 256 sclks per frame (mf1:f1, mf2:f2 = 10), four devices can occupy the se- rial port. in this format both mf8:sfs2 and mf7:sfs1 are used to select the particular sub- frame. see figure 12. in all three of the above slave sub-mode for- mats, the frequency of the incoming sclk signal, in relation to the master clock provided on the clkin pin, determines the sample fre- quency. the cs4216 determines the ratio of sclk to clkin and sets the internal operating frequency accordingly. table 5 lists the sclk to clkin frequency ratio used to determine the codecs sample frequency. to obtain a given sample frequency, sclk must equal clkin di- vided by the number in the table, based on the number of bits per frame. as an example, assum- ing 64 bpf (bits per frame) and clkin = 12.288 mhz, if a sample frequency of 24 khz is desired, sclk must equal clkin di- vided by 8 or 1.536 mhz. when mf1:f1 = mf2:f2 = 1, sclk is used as the master clock and is assumed to be 256 times the sample frequency. in this mode, clkin is ignored and the sample frequency is linearly scaled with sclk. (the clkin pin must be tied low.) this mode also fixes sclk at 256 bits per frame with mf7:sfs1 and mf8:sfs2 select- ing the particular sub-frame. fs (khz) mf1: mf2: mf3: n with clkin f1 f2 f3 12.288 11.2896 mhz mhz 0 0 0 256 48.00 44.10 0 0 1 384 32.00 29.40 0 1 0 512 24.00 22.05 0 1 1 640 19.20 17.64 1 0 0 768 16.00 14.70 1 0 1 1024 12.00 11.025 1 1 0 1280 9.60 8.82 1 1 1 1536 8.00 7.35 table 3. sm3-master, fs select mf1: mf2: bits per sample frequency/ f1 f2 frame sclk 0 0 64 ratio to clkin sensed 0 1 128 ratio to clkin sensed 1 0 256 ratio to clkin sensed 1 1 256 fixed ? . = 256 fs ? sclk is master clock. clkin is not used. table 4. sm3-slave, bits per frame. sclk to clkin ratio fs (khz) fs (khz) bpf bpf bpf with clkin with clkin 256 128 64 12.288 mhz 11.2896 mhz 1 2 4 48.00 44.10 1.5 3 6 32.00 29.40 2 4 8 24.00 22.05 2.5 5 10 19.20 17.64 3 6 12 16.00 14.70 4 8 16 12.00 11.025 5 10 20 9.60 8.82 6 12 24 8.00 7.35 table 5. sm3-slave, fs select. cs4216 18 ds83f2 serial mode 4, sm4 serial mode 4 is enabled by setting smode3 = 1. both master and slave sub- modes are available and are selected by setting the smode2 and smode1 pins as shown in table 6. in master sub-mode, the phase relation- ship between sclk/ssync and clkin is controlled to minimize digital noise coupling into the analog section. therefore, master sub- mode may yield slightly better noise performance than slave sub-mode. in slave sub- mode, sclk and ssync must be synchronous to the master clock. in serial mode 4, sm4, the clkin frequency must be 256 times the highest sample frequency needed. also, sm4 has five attenuation bits for each d/a output channel. sm4 differs from sm3 in that sm4 splits the audio data from the con- trol data with the control data input on an independent serial port. this reduces the audio serial bus bandwidth in half, providing an easier interface to low-cost dsps. the audio serial port sub-frame is illustrated in figure 13 for sm4. interrupt pin - mf5: int serial mode 4 also defines the multifunction pin mf5: int as an open-collector interrupt pin. in sm4, this pin requires a pullup resistor and will go low when the adv bit or di1 pin change, or a rising edge on the lcl or rcl bits, or by exiting an sclk out of range condition (er- ror = 3). the interrupt may be masked by setting the msk bit in the control serial data port. sub-frame 1 frame (n+1) word a word b sub-frame 1 word a word b frame n 64 sclk periods frame (n+2) sub-frame 1 word a word b frame (n+3) sub-frame 1 word a word b ssync data sfs2 sfs1 frame 001 mf8: mf7: sub- figure 10. sm3-slave - 64 bpf; mf1:f1, mf2:f2 = 00 sub-frame 1 sub-frame 2 frame (n+1) word a word b word a word b sub-frame 1 word a word b frame n 128 sclk periods sub-frame 2 word a word b frame (n+2) sub-frame 1 word a word b ssync data sfs2 sfs1 frame 0 0 0 1 1 2 mf8: mf7: sub- figure 11. sm3-slave - 128 bpf; mf1:f1, mf2:f2 = 01 sub-frame 1 sub-frame 2 sub-fram e 3 sub-frame 4 frame (n+1) word a word b word a word b w ord a word b word a word b frame n sub-frame 1 word a word b 256 sclk periods ssync data sfs2 sfs1 frame 0 0 1 1 0 1 0 1 1 2 3 4 mf8: mf7: sub- figure 12. sm3-slave - 256 bpf; mf1:f1, mf2:f2 = 10 cs4216 ds83f2 19 mf5: int is reset by reading the control serial port. master sub-mode (sm4) master sub-mode configures ssync and sclk as outputs from the cs4216. during power down, ssync and sclk are driven high im- pedance, and during reset they both are driven low. there are two sm4 master sub-modes. one allows 32 bits per frame and the other allows 64 bits per frame. as shown in table 6, the smode1 and smode2 pins select the particu- lar master sub-mode (as well as the slave sub-mode). when smode1 is set to zero, smode2 selects either master sub-mode with 32-bit frames, or slave sub-mode. smode1,smode2 = 00 selects master sub- mode where a frame = sub-frame = 32 bits. this sub-mode allows only one codec on the audio serial bus, with the first 16 bits being the left channel and the second 16 bits being the right channel. the applications of sm4 section con- tains more information on low-cost implementations of this sub-mode. smode1 = 1 selects master sub-mode with a frame width of 64 bits. this sub-mode allows up to two codecs to occupy the same bus. smode2 is now used to select the particular time slot. if smode2 = 0 the codec selects time slot 1, which is the first 32 bits. if smode2 = 1 the codec selects time slot 2, which is the second 32 bits. in master sub-mode, multifunction pins mf6:f1, mf7:f2, and mf8:f3 select the sample fre- quency as shown in table 7. this table indicates how to obtain standard audio sample frequencies given one of two clkin frequencies: 12.288 mhz or 11.2896 mhz. other clkin frequencies may be used with the corresponding sample frequencies being clkin/n. the codec must be reset when changing sample frequencies to allow a new calibration to occur. slave sub-mode (sm4) in sm4, slave sub-mode is selected by setting smode1,smode2 = 01. this mode configures ssync and sclk as inputs to the cs4216. these two signals must be externally derived from clkin. since the cs4216 has no control over the phase relationship of ssync and smode1 smode2 sm4, sub-mode 0 0 master, 32 bpf 0 1 slave, 128/64/32 bpf 1 0 master, 64 bpf, ts1 1 1 master, 64 bpf, ts2 table 6. sm4 sub-modes. ssync sclk sdin sdout sub-frame 16 17 32 1 1 8 9 24 25 32 14 23 adc - left word msb lsb adc - right word msb lsb adc - left word msb adc - right word lsb dac - left word msb lsb dac - right word msb lsb dac - left word msb dac - right word lsb (slave) (master) figure 13. sm4-audio serial port, 32 bpf cs4216 20 ds83f2 sclk to clkin, the noise performance in slave sub-mode may be slightly worse than when using master sub-mode. the cs4216 inter- nally sets the sample frequency by sensing the ratio of sclk to clkin; therefore, for a given clkin frequency, the sample frequency is se- lected by changing the sclk frequency. sm4-slave allows up to four codecs to occupy the same audio serial port. table 8 lists the pin configurations required to set the serial audio port up for 32, 64, or 128 bits-per-frame (bpf). since each codec requires one sub-frame of 32 bits, 64 bits-per-frame allows up to two codecs to occupy the same audio serial port, and 128 bits-per-frame allows up to four codecs to occupy the same audio serial port. when set up for more than one codec on the bus, other pins are needed to select the particular time slot (ts) associated with each codec. mf8:sfs2 selects the time slot when in 64 bpf mode, and mf8:sfs2 and mf7:sfs1 select one of four time slots when in 128 bits-per-frame mode. ta- ble 8 lists the decoding for time slot selection. in sm4-slave, the frequency of the incoming sclk signal, in relation to clkin, determines the sample frequency on the cs4216. the cs4216 determines the ratio of sclk to clkin and sets the internal sample frequency accord- ingly. table 9 lists the sclk to clkin fre- quency ratio used to determine the codec?s sample frequency. sclk must equal clkin di- vided by the number in the table, based on the selected bits per frame. as an example, assuming 32 bpf and clkin = 11.2896 mhz, if a sample frequency of 11.025 khz is desired, sclk must equal clkin divided by 32 or 352.8 khz. serial control port (sm4) serial mode 4 separates the audio data from the control data. since control data such as gain and attenuation do not change often, this mode re- duces the bandwidth needed to support the audio serial port. the control information is entered through a separate port that can be asynchronous to the audio port and only needs to be updated when changes in the control data are needed. after a reset or power down, the control port must be written once to initialize it if the port will be ac- cessed to read or write control bits. this initial write is considered a "dummy" write since the data is ignored by the codec. a second write is needed to configure the codec as desired. then, the control port only needs to be written to when a change is desired, or to obtain the status infor- mation. the control port does not function if the master clock is not operating. when the control fs (khz) mf6: mf7: mf8: n with clkin f1 f2 f3 12.288 11.2896 mhz mhz 0 0 0 256 48.00 44.10 0 0 1 384 32.00 29.40 0 1 0 512 24.00 22.05 0 1 1 640 19.20 17.64 1 0 0 768 16.00 14.70 1 0 1 1024 12.00 11.025 1 1 0 1280 9.60 8.82 1 1 1 1536 8.00 7.35 table 7. sm4-master, fs select mf6: mf7: mf8: bits per time f1 sfs1 sfs2 frame slot (bpf) (ts) 000 32 1 001 reserved 010 64 1 011 64 2 100 128 1 110 128 2 101 128 3 111 128 4 table 8. sm4-slave, audio port bpf & ts select cs4216 ds83f2 21 port is used asynchronously to the audio port, the noise performance may be slightly degraded due to this asynchronous digital noise. since control data does not need to be accessed each audio frame, an interrupt pin, mf5: int, is included in this mode and will go low when status has changed. the control port serial data format is illustrated in figure 14. the control port uses one of the multifunction pins as a chip select line, mf4: ccs, that must be low for en- tering control data. although only 23 bits contain useful data on mf2:cdin, a minimum of 31 bits must be written. if more than 31 bits are written without toggling mf4: ccs, only the first 31 are recognized. mf1:cdout contains status information that is output on the rising edge of mf3:cclk. status information is re- peated at the end of the frame, bits 25 through 30, to allow a simple 8-bit shift and latch register to store the most important status information using the rising edge of mf4: ccs at the latch control (see figure 17). applications of sm4 figure 15 illustrates one method of using serial mode 4 wherein a dsp controls the audio serial port and a microcontroller controls the control port. each controller is run independently and the micro updates the control information only when needed, or when an interrupt from the cs4216 occurs. figure 16 illustrates the minimum interface to the cs4216. in this application, the dsp sends and receives stereo dac and adc information. the cs4216 is configured for 32 bits per frame, master sub-mode. the control data resets to all zeros, which configures the cs4216 as a simple stereo codec: no gain, no attenuation, line inputs #1, and not muted. figure 17 illustrates how to use all the cs4216 features with a low cost dsp that cannot support the interrupt rate of sm3. using sm4 (32 bits 16 17 32 mf4:ccs mf3:cclk 1 8 9 24 25 10 30 err version msk do1 mf2:cdin mf1:cdout 0 0001 left a/d gain 30 right a/d gain 30 m in sel. left d/a att. 40 right d/a att. 40 adv lcl rcl 10 err adv lcl rcl di1 00000000 di1 0 figure 14. sm4 - control serial port sclk to clkin ratio fs (khz) fs (khz) bpf bpf bpf with clkin with clkin 128 64 32 12.288 mhz 11.2896 mhz 2 4 8 48.00 44.10 3 6 12 32.00 29.40 4 8 16 24.00 22.05 5 10 20 19.20 17.64 6 12 24 16.00 14.70 8 16 32 12.00 11.025 10 20 40 9.60 8.82 12 24 48 8.00 7.35 table 9. sm4-slave, fs select. cs4216 22 ds83f2 per frame, master sub-mode) reduces the dsp interrupts in half since the control data is split from the audio data. this circuit is comprised of three independent sections which may individu- ally be eliminated if not needed. to load control data into the codec, three hc597?s are utilized. these are both latches that store the dsp-sent control data, and shift regis- ters that shift the data into the codec. the codec uses an inverted ssync signal to copy the latches to the shift registers every frame. in this diagram the dsp is assumed to have a data bus bandwidth of at least 24 bits. if the dsp has less than 24_bits, the three hc597s must be split into two addresses. since the hc597 internal latches are copied to the shift registers, the latches con- tinually hold the dsp-sent data; therefore, the dsp only needs to write data to the latches when a change is desired. the second section is comprised of an hc595 shift register and latch that is clocked by an in- verted sclk the data shifted into the hc595 is transferred to the hc595?s latch by the ssync signal. this hc595 captures the 8 bits prior to the ssync signal (which is also mf4: ccs) go- ing high. as shown in figure 14, and assuming the mf4: ccs (ssync) signal rises at bit 32, the 8-bits prior to mf4: ccs rising are a copy of all the important status bits. this allows one shift register to capture all the important information. the interrupt pin cannot reliably be used in this configuration since the interrupt pin is cleared by reading the control port which occurs asynchro- micro- controller serial port general purpose port pins irq dsp sdout sdin ssync sclk cs4216 mf1:cdout mf2:cdin mf4:ccs mf3:cclk reset mf8:f3 mf7:f2 mf6:f1 mf5:int sm4 vd+ 43 42 1 44 40 39 36 35 38 2 34 31 30 figure 15. sm4 - microcontroller interface dsp sdout sdin ssync sclk cs4216 mf1:cdout mf2:cdin mf4:ccs mf3:cclk mf5:int sm4 vd+ reset mf8:f3 mf7:f2 mf6:f1 hard wired or dip switch selectable 32 bpf 43 42 1 44 35 36 38 40 39 2 34 31 30 figure 16. sm4 - minimum dsp interface cs4216 ds83f2 23 dsp sdout sdin ssync sclk cs4216 mf1:cdout mf2:cdin mf4:ccs mf3:cclk reset mf8:f3 mf7:f2 mf6:f1 hc574 hc595 sclk lclk ain a b c d e f g h 0 adv di1 rcl lcl err0 err1 0 hc597 hc597 hc597 dout load din cs_fs oe cs_status cs_control 24+ bit dsp data bus mf5:int sm4 32 bpf vd+ 43 42 1 44 35 36 38 39 40 2 34 31 30 figure 17. sm4 - enhanced dsp interface cs4216 24 ds83f2 nously (every audio frame) with respect to the interrupt occurrence. the third section is only needed if sample fre- quencies need to be changed. this section is comprised of an hc574 octal latch that can be replaced by general purpose port pins if avail- able. this section controls the sample frequency selection bits: mf6:f1, mf7:f2, mf8:f3 and the reset pin. the codec must be reset when changing sample frequencies. power supply and grounding the cs4216, along with associated analog cir- cuitry, should be positioned in an isolated section of the circuit board, and have its own, separate, ground plane. on the cs4216, the analog and digital grounds are internally connected; there- fore, the agnd and dgnd pins must be externally connected with no impedance between them. the best solution is to place the entire chip on a solid ground plane as shown in fig- ure 18. preferably, it should also have its own power plane. the +5v supply must be connected to the cs4216 via a ferrite bead, positioned closer than 1" to the device. the va supply can be derived from vd, as shown in figure 1. al- ternatively, a separate +5v analog supply may be used for va, in which case, the 2.0 w resistor between va and vd should be removed. a sin- gle connection between the cs4216 ground (analog ground) and the board digital ground should be positioned as shown in figure 18. figure 19 illustrates the optimum ground and de- coupling layout for the cs4216 assuming a surface-mount socket and leaded decoupling ca- pacitors. surface-mount sockets are useful since the pad locations are identical to the chip pads; therefore, assuming space for the socket is left on the board, the socket can be optional for pro- duction. figure 19 depicts the top layer, containing signal traces, and assumes the bottom or inter-layer contains a fairly solid ground plane. the important points are that there is solid ground plane under the codec on the same layer as the codec and it connects all ground pins with thick traces providing the absolute lowest imped- ance between ground pins. the decoupling capacitors are placed as close as possible to the device which, in this case, is the socket bound- ary. the lowest value capacitor is placed closest to the codec. vias are placed near the agnd and dgnd pins, under the ic, and should attach to the solid ground plane on another layer. the negative side of the decoupling capacitors should also attach to the same solid ground plane. traces and vias bringing power to the codec should be large, which minimizes the impedance. although not shown in the figures, the trace lay- ers (top layer in the figures) should have ground plane fill in-between the traces to minimize cou- pling into the analog section. see the cdb4216 evaluation board as an example. if using all surface-mount components, the de- coupling capacitors should be placed on the same layer as the codec and in the positions shown in figure 20. the vias shown are assumed to attach to the appropriate power and ground layers. traces and vias bringing power to the codec should be as large as possible to minimize the impedance. if using a through-hole socket, effort should be made to find a socket with minimum height, cs4216 ds83f2 25 which will minimize the socket impedance. when using a through hole socket, the vias un- der the codec in figure 19 are not needed since the pins serve the same function. adc and dac filter response plots figures 21 - 26 shows the overall frequency re- sponse, passband ripple and transition band for the cs4216 adcs and dacs. figure 27 shows the dacs deviation from linear phase. fs is defined as the selected sample frequency and is also the ssync frequency. since the sample frequency is programmable, the filters will adjust to the selected sample frequency. digital ground plane note that the cs4216 is oriented with its digital pins towards the digital end of the board. cpu & digital logic codec digital signals codec analog signals & components > 1/8" cs4216 +5v ferrite bead ground connection analog ground plane figure 18. cs4216 board layout guideline cs4216 26 ds83f2 1 1.0 uf 0.1 uf 10 uf + 1.0 uf 0.1 uf 0.1 uf + + analog supply digital supply figure 19. cs4216 decoupling layout guideline 1 1.0 uf 0.1 uf 10 uf + 1.0 uf 0.1 uf 0.1 uf + + analog supply digital supply figure 20. cs4216 surface mount decoupling layout cs4216 ds83f2 27 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 input frequency (fs) -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 magnitude (db) figure 21. cs4216 adc frequency response 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 input frequency (fs) -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 -0.0 0.2 0.4 0.6 magnitude (db) figure 22. cs4216adc passband ripple 0.40 0.42 0.44 0.46 0.48 0.50 0.52 0.54 0.56 0.58 0.60 input frequency (fs) -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 magnitude (db) figure 23. cs4216 adc transition band 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 input frequency (fs) -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 magnitude (db) figure 24. cs4216 dac frequency response 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 input frequency (fs) -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 -0.0 0.1 0.2 magnitude (db) figure 25. cs4216 dac passband ripple -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 magnitude (db) 0.40 0.42 0.44 0.46 0.48 0.50 0.52 0.54 0.56 0.58 0.60 input frequency (fs) figure 26. cs4216 dac transition band cs4216 28 ds83f2 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 input frequency (fs) -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 phase (degrees) figure 27. cs4216 dac deviation from linear phase cs4216 ds83f2 29 pin descriptions 40 42 34 36 38 6 2 4 8 10 1 3 5 7 9 11 23 25 27 29 31 33 28 24 26 30 32 44 12 14 16 18 20 22 sm mf1 mf2 mf3 mf4 mf5 mf6 mf7 mf8 1 do4 do3 di3 di4 do2 di2 sfs1 sfs2 2 do4 do3 di3 di4 do2 di2 sfs1 sfs2 3 f1f2f3mado2di2sfs1sfs2 4-sl cdout cdin cclk ccs int f1 sfs1 sfs2 4-ma cdout cdin cclk ccs intf1f2f3 ssync reset sclk clkin sdout vd sdin dgnd smode3 nc mf1:do4/f1/cdout nc mf2:do3/f2/cdin nc mf5:do2/int nc do1 nc mf4:di4/ma/ccs nc mf3:di3/f3/cclk nc mf6:di2/f1 pdn di1 nc smode2 rout mf7:sfs1/f2 lout mf8:sfs2/f3 nc smode1 nc lin2 nc lin1 refbuf rin2 refbyp rin1 refgnd va agnd cs4216 44-pin tqfp (q) top view cs4216 30 ds83f2 power supply vd - digital +5v supply, pin 4(l), 42(q). +5v digital supply. va - analog +5v supply, pin 24(l), 18(q). +5v analog supply. dgnd - digital ground, pin 5(l), 43(q). digital ground. must be connected to agnd with zero impedance. cs4216 44-pin plcc (l) top view sm mf1 mf2 mf3 mf4 mf5 mf6 mf7 mf8 1 do4 do3 di3 di4 do2 di2 sfs1 sfs2 2 do4 do3 di3 di4 do2 di2 sfs1 sfs2 3 f1f2f3mado2di2sfs1sfs2 4-sl cdout cdin cclk ccs int f1 sfs1 sfs2 4-ma cdout cdin cclk ccs intf1f2f3 ssync reset sclk clkin sdout vd sdin dgnd smode3 nc mf1:do4/f1/cdout nc mf2:do3/f2/cdin nc mf5:do2/int nc do1 nc mf4:di4/ma/ccs nc mf3:di3/f3/cclk nc mf6:di2/f1 pdn di1 nc smode2 rout mf7:sfs1/f2 lout mf8:sfs2/f3 nc smode1 nc lin2 nc lin1 refbuf rin2 refbyp rin1 refgnd va agnd cs4216 44-pin plcc (l) top view 18 20 22 24 26 28 1 2 4 640 42 44 12 8 10 14 16 7 9 11 13 15 17 29 31 33 35 37 39 34 30 32 36 38 cs4216 ds83f2 31 agnd - analog ground, pin 23(l), 17(q). analog ground. must be connected to dgnd with zero impedance. analog inputs rin1 - right input #1, pin 25(l), 19(q). right analog input #1. full scale input, with no gain, is 1 vrms, centered at refbuf. rin2 - right input #2, pin 26(l), 20(q). right analog input #2. full scale input, with no gain, is 1 vrms, centered at refbuf. lin1 - left input #1, pin 27(l), 21(q). left analog input #1. full scale input, with no gain, is 1 vrms, centered at refbuf. lin2 - left input #2, pin 28(l), 22(q). left analog input #2. full scale input, with no gain, is 1 vrms, centered at refbuf. analog outputs rout - right channel output, pin 15(l), 9(q). right channel analog output. maximum signal is 1 vrms centered at refbuf. lout - left channel output, pin 16(l), 10(q). left channel analog output. maximum signal is 1 vrms centered at refbuf. refbyp - analog reference decoupling, pin 21(l), 15(q). a 10 m f and 0.1 m f capacitor must be attached between refbyp and refgnd. refgnd - analog reference ground connection, pin 22(l), 16(q). connect to agnd. refbuf - buffered reference out, pin 20(l), 14(q). a nominal +2.2 v output for setting the bias level for external analog circuits. serial digital audio interface signals sdin - serial port data in, pin 42(l), 36(q). digital audio data to the dacs and level control information is received by the cs4216 via sdin. sdout - serial port data out, pin 43(l), 37(q). digital audio data from the adcs and status information is output from the cs4216 via sdout. sclk - serial port bit clock, pin 44(l), 38(q). sclk controls the digital audio data on sdout and latches the data on sdin. cs4216 32 ds83f2 ssync - serial port sync signal, pin 1(l), 39(q). indicates the start of a digital audio frame in sm3 and sm4, and also the start of a word in sm1 & sm2. smode1 - serial mode select, pin 29(l), 23(q). one of three pins that select the serial mode and function of the multifunction pins. smode2 - serial mode select, pin 32(l), 26(q). one of three pins that select the serial mode and function of the multifunction pins. smode3 - serial mode select, pin 41(l), 35(q). one of three pins that select the serial mode and function of the multifunction pins. this pin has an internal pull-down making this revision backwards compatible with a previous version (revision a or version3-version0 bits = 0000). multifunction digital pins mf1:do4 - parallel digital bit output #4 in sm1/sm2, pin 40(l), 34(q). in serial modes 1 and 2 this pin reflects the value of the do4 bit in the sub-frame. mf1:f1 - format bit 1 in sm3, pin 40(l), 34(q). in serial mode 3 this pin is a format bit and is used as one of three sample frequency select pins when in master mode, or as one of two bits-per-frame select pins when in slave mode. mf1:cdout - control data output in sm4, pin 40(l), 34(q). in serial mode 4 this pin is the data output for the control port which contains status information. mf2:do3 - parallel digital bit output #3 in sm1/sm2, pin 39(l), 33(q). in serial modes 1 and 2 this pin reflects the value of the do3 bit in the sub-frame. mf2:f2 - format bit 2 in sm3, pin 39(l), 33(q). in serial mode 3 this pin is a format bit and is used as one of three sample frequency select pins when in master mode, or as one of two bits-per-frame select pins when in slave mode. mf2:cdin - control data input in sm4, pin 39(l), 33(q). in serial mode 4 this pin is the control port data input which contains data such as gain and attenuation settings as well as input select, mute, and digital output bits. mf3:di3 - parallel digital bit input #3 in sm1/sm2/sm3 (slave), pin 35(l), 29(q). in serial modes 1 and 2 this pin value is reflected in the di3 bit in the sub-frame. mf3:f3 - format bit 3 in sm3 (master), pin 35(l), 29(q). in serial mode 3 this pin is a format bit and is used as one of three sample frequency select pins when in master mode. in slave mode, the pin reverts to being a general purpose input. mf3:cclk - control data clock in sm4, pin 35(l), 29(q). in serial mode 4 this pin is the control port serial bit clock which latches data from cdin on the falling edge, and outputs data onto cdout on the rising edge. cs4216 ds83f2 33 mf4:di4 - parallel digital bit input #4 in sm1/sm2, pin 36(l), 30(q). in serial modes 1 and 2 this pin value is reflected in the di4 bit in the sub-frame. mf4:ma - master sub-mode in sm3, pin 36(l), 30(q). in serial mode 3 this pin selects either master or slave mode. when mf4:ma = 1, the codec is in master mode and outputs ssync and sclk. when mf4:ma = 0, the codec is in slave mode and receives ssync and sclk from an external source that must be frequency locked to clkin. mf4: ccs - control data chip select in sm4, pin 36(l), 30(q). in serial mode 4 this pin is the control port chip select signal. when low, the control port data is clocked in cdin and status data is output on cdout. when ccs goes high, control data is latched internally. this data remains active until new data is clocked in. the control port may also be asynchronous to the audio data port. mf5:do2 - parallel digital bit output #2 in sm1/sm2/sm3, pin 38(l), 32(q). in serial modes 1, 2, and 3 this pin reflects the value of the do2 bit in the sub-frame. mf5: int - interrupt in sm4, pin 38(l), 32(q). in serial mode 4 this pin is an active low interrupt signal that is maskable using the msk bit in the control port serial data stream. int is an open-collector output and requires and external pull-up resistor. assuming the mask bit is not set, and interrupt is triggered by a change in adv or di1, or a rising edge on lcl or rcl, or a exiting an sclk out of range condition (error = 3) mf6:di2 - parallel digital bit input #2 in sm1/sm2/sm3, pin 34(l), 28(q). in serial modes 1, 2, and 3 this pin value is reflected in the di2 bit of the sub-frame. mf6:f1 - format bit 1 in sm4, pin 34(l), 28(q). in serial mode 4 this pin is a format bit and is used as one of three sample frequency select pins when in master mode. in slave mode, mf6:f1 helps determine the number of sub-frames within a frame. mf7:sfs1 - sub-frame select 1 in sm1/sm2/sm3/sm4-sl, pin 31(l), 25(q). in serial modes 1, 2, and 3, mf7:sfs1 helps select the sub-frame that this particular cs4216 is allocated. in slave sub-mode of serial mode 4, this pin is one of two pins used as a sub-frame select when mf6:f1 = 1 (128-bit frames). when mf6:f1 = 0, this pin is used to select the frame sizes of 32 or 64 bits. mf7:f2 - format bit 2 in sm4-ma, pin 31(l), 25(q). in master sub-mode of serial mode 4, this pin is used as one of three sample frequency select pins. mf8:sfs2 - sub-frame select 2 in sm1/sm2/sm3/sm4-sl, pin 30(l), 24(q). in serial modes 1, 2, 3, and slave sub-mode of 4, mf8:sfs2 helps select the sub-frame that this particular cs4216 is allocated. mf8:f3 - format bit 3 in sm4-ma, pin 30(l), 24(q). in master sub-mode of serial mode 4, this pin is a format bit and is one of three sample frequency select pins. cs4216 34 ds83f2 miscellaneous reset - reset input, pin 2.(l), 40(q). resets the cs4216 into a known state, and must be initiated after power-up or power-down mode. releasing reset caused the cs4216 to initiate a calibration sequence. reset should also be initiated when changing sample frequencies in any master sub-mode. clkin - master clock, pin 3(l), 41(q). clkin is the master clock that operates the internal logic. in serial mode 1, clkin = 512 hfs, where hfs is the highest sample frequency needed. different sample frequencies are obtained by changing the ratio of sclk to clkin. in serial mode 2, clkin is not used and must be tied low. in serial modes 3 and 4, clkin is 256 hfs, where different sample frequencies are obtained by either changing the ratio of sclk to clkin in slave mode, or changing the format pin values (f2-f0) in master mode. pdn - power down, pin 13(l), 7(q). this pin, when low, causes the cs4216 to go into a power down state. reset should be held low for 50 ms when exiting the power down state to allow time for the voltage reference to settle. di1 - parallel digital bit input #1, pin 33(l), 27(q). this pin value is reflected in the di1 bit in the sub-frame. do1 - parallel digital bit output #1, pin 37(l), 31(q). this pin reflects the value of the do1 bit in the sub-frame. nc - no connection, pins 6, 7, 8, 9, 10, 11, 12, 14, 17, 18, 19(l) pins 44, 1, 2, 3, 4, 5, 6, 8, 11, 12, 13(q). these pins should be left floating with no trace attached to allow backwards compatibility with future revisions. they should not be used as a convenient path for signal traces. cs4216 ds83f2 35 parameter definitions resolution the number of bits in the input words to the dacs, and in the output words from the adcs. differential nonlinearity the worst case deviation from the ideal codewidth. units in lsb. total dynamic range tdr is the ratio of the rms value of a full scale signal to the lowest obtainable noise floor. it is measured by comparing a full scale signal to the lowest noise floor possible in the codec (i.e. attenuation bits for the dacs at full attenuation). units in db. instantaneous dynamic range idr is the ratio of a full-scale rms signal to the rms noise available at any instant in time, without changing the input gain or output attenuation settings. it is measured using s/(n+d) with a 1 khz, -60 db input signal, with 60 db added to compensate for the small input signal. use of a small input signal reduces the harmonic distortion components to insignificance when compared to the noise. units in db. total harmonic distortion thd is the ratio of the rms value of a signals first five harmonic components to the rms value of the signals fundamental component. thd is calculated using an input signal which is 3db below typical full-scale, and is referenced to typical full-scale. interchannel isolation the amount of 1 khz signal present on the output of the grounded input channel, with 1 khz 0 db signal present on the other channel. units in db. interchannel gain mismatch for the adcs, the difference in input voltage that generates the full scale code for each channel. for the dacs, the difference in output voltages for each channel with a full scale digital input. units in db. frequency response worst case variation in output signal level versus frequency over the passband. tested over the frequency band of 10 hz to 20 khz, with the sample frequency of 48 khz. units in db. step size typical delta between two adjacent gain or attenuation values. units in db. absolute gain/attenuation step error the deviation of a gain or attenuation step from a straight line passing through the no-gain/attenuation value and the full-gain/attenuation value (i.e. end points). units in db. offset error for the adcs, the deviation of the output code from the mid-scale with the selected input at refbuf. for the dacs, the deviation of the output from refbuf with mid-scale input code. units in lsbs for the adcs and volts for the dacs. cs4216 36 ds83f2 out of band energy the ratio of the rms sum of the energy from 0.46 fs to 2.1 fs compared to the rms full-scale signal value. tested with 48 khz fs giving a out-of-band energy range of 22 khz to 100 khz. cs4216 ds83f2 37 appendix a this data sheet describes version 1 of the cs4216. therefore, this appendix is included to describe the differences between version 0 and version 1. this information is only useful for users that still have version 0 since version 1 devices will supplant the earlier version. the version number is contained in the serial data line, bits 29 - 32 on sdout in sm1-sm3 and, bits 17 - 20 on cdout in sm4. the version number can also be identified by the revision letter stamped on the top of the actual chip. the revision letter immediately precedes the data code on the second line of the package marking (see general information section of the crystal data book). version 0 corresponds to chip revision a, and version 1 corresponds to chip revisions b, and c. the functionally and performance of revisions b and c are identical. likewise, future chip revisions (i.e. d, e, f, . . .) may still be version 1 since the version number only changes if there is a software change to the part. functional differences between version 0 (rev. a) and version 1 (revs. b, c) 1. in version 0, serial mode 4 (sm4) does not exist; the smode3 pin is a no connect. in version 1 the smode3 pin contains an internal pull-down resistor making this version backwards compatible with version 0 sockets. 2. ssync on version 0 must be only one sclk period high in sm3 or 2 sclk periods high in sm1 and sm2 to indicate the start of a frame. also, on version 0 in sm1 or sm2, ssync must be exactly one sclk period high, at the beginning of each word. in version 1, ssync can be high for an arbitrary number of sclks beyond the one in sm3 or two in sm1 and sm2. also in sm1 and sm2, the one-sclk-wide ssyncs at each word are not needed. in version 1, sm3 and sm4, the codec only looks for a low-to-high edge of ssync to start a frame; in sm1 and sm2 a low-to-high edge of ssync, being high for two sclk periods, starts a frame. cs4216 38 ds83f2 features easy dsp hook-up analog-in to analog-out loopback mode correct grounding and layout microphone pre-amplifier line input buffer digital and analog patch areas general description the cdb4216 and cdb4218 evaluation boards allow easy evaluation of the cs4216 and cs4218 audio mul- timedia codecs. analog inputs provided include two bnc line inputs for lin1 and rin1, and two 1 / 4 " micro- phone jacks on the lin2 and rin2 lines. analog outputs are available on two bncs. digital interfacing is facilitated using one to three of the buffered ribbon cable headers. all four serial modes of the cs4216 and all three modes of the cs4218 are supported using a simple dip switch which is decoded to select the proper mode and sub-mode. ordering information: cdb4216 or cdb4218 jun 93 cs83db4 39 crystal semiconductor corporation p.o. box 17847, austin, tx 78760 (512) 445 7222 fax: (512) 445 7581 cs4216 evaluation board semiconductor corporation cdb4216 cs4216 va digital i/o buffers & mux digital patch area agnd vd dgnd clkin audio port control port line inputs microphone jacks line outputs a = 14 db a = - 6 db (+5v) analog patch area lin2 rin2 lin1 rin1 lout rout osc digital i/o port configuration switch (+5v) copyright ? crystal semiconductor corporation 1993 (all rights reserved) general description the cdb4216/8 was designed to provide an easy platform for evaluating the performance of the cs4216 and cs4218 stereo audio codecs. since the evaluation board contains a proper lay- out and is performance tested, the user can concentrate on engineering the rest of the system thereby reducing the development time. the lay- out should also be used as a guideline for obtaining the best possible performance from the cs4216 or cs4218. lastly, the board can be used as a benchmark and debugging tool for user developed pcbs. the evaluation board supports all serial modes and includes decode circuitry to ease the selec- tion of the serial mode and sub-mode of interest. all serial interfaces are buffered for easy connec- tion to the serial port of a dsp or other serial device. the board can also be placed in a loop back mode where the digital data is looped back allowing an analog in to analog out testing vehi- cle without an external processor. a single +5v supply is all that is needed to power the board. analog inputs consist of a pair of line input buff- ers (lin1, rin1) designed to accept a maximum audio signal of 2v rms and bnc-to-phono adapters are included to support various test con- figurations. the second pair of inputs contain a example microphone input buffer supported by two 1 / 4 " mono jacks that are designed to accept standard single-ended dynamic or condenser mi- crophones. the line outputs are supplied via bnc connec- tors with two more bnc-to-phono adapters. the film plots of the evaluation board are in- cluded to provide an example of the optimum layout, grounding, and decoupling arrangement for the cs4216 or cs4218. selecting a serial mode the cs4216 supports four serial modes and many sub-modes, and the cs4218 supports three serial modes and sub-modes. selecting the most appropriate mode for a given application can be time consuming. the cdb4216/8 contains a dip switch that simplifies this selection. since the cs4216/8 contains many multifunction pins, the dip switch lets the user select the configuration and two plds decode the proper multifunction pin values. since these plds are only used to simplify the configuration of the device, they would not be needed in an end application which would hard wire the configuration pins. table 4 describes the multifunction pin values for a given dip switch setting. the pal equations for dip switch decoding are given in figures 8 and 9. all references to sm1 and sm2 apply only to the cs4216. all references to sm3-mm, sm3- ms, and i 2 s apply only to the cs4218. serial port format table 1 lists the dip switches used to select the serial mode. spf2 and spf1 select one the four serial modes of the cs4216 or one of three serial modes for the cs4218. ma selects master (ma = 1) or slave and is only useful in serial modes 3 and 4. the majority of users select spf2 spf1 ma serial mode 00x sm1 slave 01x sm2 slave 100 sm3 slave 1 0 1 sm3 master 110 sm4 slave 1 1 1 sm4 master table 1. dip switch, serial modes cdb4216 40 ds83db4 either serial mode 3, sm3, or serial mode 4, sm4. serial modes 1 and 2, sm1 and sm2, are primarily designed for asics and are less flex- ible. sm1 and sm2 are not available on the cs4218. the cs4218 has additional sm3 sub- modes: multiplier master (sm3-mm) and multiplier slave (sm3-ms). these sub-modes are identical to the sm3 master and slave sub- modes except that the master clock, clkin, must be 16xfs max instead of 256xfs max . the cs4218 also provides a master i 2 s mode. in master sub-modes, the cs4216/8 output ssync and sclk. in slave sub-modes, ssync and sclk must be externally generated and must be synchronous to clkin. bits per frame selection the next decision is selecting the number of bits per frame which defines how many codecs can sit on the same serial bus. each codec occupies a sub-frame and 1 to 4 sub-frames make up a frame. a sub-frame is 64 bits in sm1, sm2, and sm3; and 32 bits in sm4. table 2 lists the possi- ble selections. if the evaluation board serial port is shared with other devices, sdoutub must be used instead of sdout since sdoutub, driven directly from the chip, must only drive the time slot assigned to it. see the audio port header section for more information. time slot selection if the number of bits per frame selected allows for more than one codec sub-frame, then the ac- tual time slot or sub-frame used by the eval board must be selected. this is done with the ts2 and ts1 dip switches. if the number of bits per frame allows only one codec on the serial bus, then ts2 and ts1 are ignored. table 3 list the decoding for ts2 and ts1. time slot 1 is the first sub-frame after ssync goes high, time slot 2 is the next sub-frame, and so on. sample frequency selection - master mode the last decision is selecting the sample fre- quency in master sub-mode. if configured for slave sub-mode, the sample frequency is the ratio of sclk to clkin as described in the cs4216/8 data sheets. in master modes, three pins are used to select the sample frequency di- vide. the dip switches labeled div1, div2, and div3 select the sample frequency and are equivalent to f1, f2, and f3, respectively. the actual f1-f3 pins on the cs4216/8 are different between sm3 and sm4 as shown in table 4 at the end of the data sheets. table 3 and table 9 of the cs4216 data sheet describe the sample fre- quencies obtained using the on-board oscillator of 11.2896 mhz. as an example, if all div switches are off, the sample frequency is 44.1 khz. with only div2 on, the sample fre- quency is 22.05 khz. to obtain a sample frequency of 44.1 khz using the cdb4218, all div switches should be set to zero and a 705.6 khz clock should be connected to the bnc jack (j2). the shunt on j1 should be set to ext. bpf sm1 sm3 sm4 21 sm2 sl ma sl ma 0 0 256 64 64 32 32 0 1 256 128 128 64 64 1 0 256 256 128 128 64 1 1 256 256* 128 128 64 * sclk is master clock. table 2. dip switch, bits per frame ts2 ts1 available sub-frames 421 00 1 1 1 01 2 2 1 10 3 2 1 11 4 2 1 table 3. dip switch, time slots cdb4216 ds83db4 41 dip switch mapping to multi-func- tion pins the two pals on the evaluation board decode the dip switches to configure the codec into a particular mode. these pals are not necessary in a design since only one mode is usually used and can be hard wired. figure 9 and figure 10 list the pal equations used for decoding. table 4 shows the cs4216/8 multi-function pin settings for each possible dip switch configura- tion. refer to the cs4216/8 data sheets to determine pin settings for sample frequencies. once a suitable mode has been chosen using the evaluation board, this table will show the hard wire configuration for each multi-function pin. example mode settings following are two examples of how to set a se- rial mode with the dip switches and then determine the multi-function pin settings for the codec. these modes were chosen for illustration only, not to suggest that they are better than other modes. a commonly used mode is sm3 master (sm3-m on the cs4218), 64 bpf, 44.1khz sampling rate, and bit-long ssync. to configure the codec in this mode, set spf2=ma=1 and all other dip switches to zero. from table 4, the smode3, smode2, and smode1 pins are set to 010, re- spectively. the div1, div2, and div3 dip switches will set the sampling frequency by di- rectly mapping to the mf1, mf2, and mf3 pins as 000. the ma dip switch sets the mf4 pin high for master mode. multi-function pins mf5 and mf6 become the general purpose i/o pins do2 and di2, respectively. in this particular mode, mf7 determines the high time for the ssync signal. the mf7 pin is set low by the ts1 switch to generate the bit-long ssync. in all other applicable cases, the ts1 switch is used for time-slot configuration. 64 bpf is selected by setting the mf8 pin low with the bpf1 switch. sm4 is a powerful mode which reduces data transfer bandwidth to facilitate easier use with low cost dsps. as an example, consider sm4, slave, 64 bpf, time-slot 2, and a 22.05 khz sampling rate. set spf2=spf1=div1=ts1=bpf1=1 and all other dip switches to zero. table 4 shows that the smode3, smode2, and smode1 pins will be set to 110, respectively. the multi-function pins mf1-4 will become the control port interface. mf5 serves as the interrupt pin int. bpf2 and bpf1 will set the codec to 64 bpf by mapping directly to the mf6 and mf7 pins as 01, respec- tively. the second time-slot is chosen with ts1 dip switches spf2 spf1 ma smode smode smode mf1 mf2 mf3 mf4 mf5 mf6 32 1 mf7 mf8 cs4216 multifunction pins do4 00x 0 0 0 do3 di3 di4 do2 di2 ts1 ts2 do4 01x 0 0 1 do3 di3 di4 do2 di2 ts1 ts2 bpf2 010 0 1 0 bpf1 di3 ma=0 do2 di2 ts1 ts2 div1 011 0 1 0 div2 div3 ma=1 do2 di2 ts1 bpf1 cdout 110 1 1 0 cdin cclk ccs int bpf2=0 bpf1 ts1 cdout 110 1 1 0 cdin cclk ccs int bpf2=1 ts1 ts2 cdout 1 1 1 1 bpf>0 ts1 cdin cclk ccs int div1 div2 div3 cdout 1 1 1 1 bpf=0 0 cdin cclk ccs int div1 div2 div3 table 4. cs4216 pin decode cdb4216 42 ds83db4 va vd cs4216 u1 clkin lin2 rin2 lin1 rin1 28 26 27 25 24 4 ferrite bead va (+5v) vd (+5v) r26 2 w l1 va dgnd agnd vd 47 uf + c29 d4 p6ke d3 p6ke 47 uf + c7 0.1 uf c28 1 uf + c3 0.1 uf c5 0.1 uf c30 1 uf + c4 0.1 uf c6 refbyp 21 10 uf + c36 0.1 uf c35 refbuf 20 16 1 uf c32 + c31 2200 pf npo 15 1 uf c34 + c33 2200 pf npo 47.5 k r28 47.5 k r30 lout rout see figure 7, 8 line input buffer see figure 2 microphone input buffer see figure 3 604 r27 604 r29 agnd refgnd dgnd 23 22 5 mf4 mf3 mf6 smode3 smode2 smode1 mf7 mf8 3 32 29 31 30 reset 41 2 36 35 34 see figure 5 see figure 5 see figure 6 lout rout figure 1. cs4216 and power supplies cdb4216 ds83db4 43 setting the mf8 pin high. since the part is in slave mode, the sampling rate must be set by the ratio between clkin and sclk. assuming that clkin has a frequency of 11.2896mhz, this ratio must be eight to give a sampling rate of 22.05khz (refer to the cs4216 data sheet). in all slave modes, ssync and sclk must be synchronous to the master clock. loopback mode the cdb4216/8 may be configured in a simple loop back mode that only requires a power source to operate. no controller of any type is necessary. this mode allows a quick and simple verification of codec operation by sampling the lin1 and rin1 inputs, then looping the digital data back to the lout and rout line outputs. set spf2=spf1=ma=1 and shunt sdout to sdin on stake header j15. this mode uses sm4 with all control settings set to zero, so no gain or attenuation is available. power supply circuitry figure 1 illustrates a portion of the cdb4216/8 schematic and includes the cs4216/8 along with power supply decoupling and circuitry. the evaluation board supports various power supply arrangements. the factory configuration powers the analog portion of the cs4216/8, along with input buffers, from the va binding post, which needs a clean +5 volts. the digital portions of the cs4216/8 are factory configured to obtain power through a 2 w resistor from the va supply. the digital buffers and plds obtain power from the vd binding post, which also needs +5 volts. although binding posts exist for both digital and analog grounds, only one needs to be connected if a single supply is used for both va and vd. note that the cs4216/8 is entirely on the analog ground plane, close to the ground plane split as required by the cs4216/8 data sheets. also note that the two ground planes are connected near the two ground binding posts. space for a ferrite bead, l1, is provided so that the board may be modified to power the codec from the digital supply. selection of l1 will de- pend on the noise characteristics of the digital supply used. analog inputs the analog inputs consist of a pair of line level inputs and a pair of 1 / 4 " mono jacks for two mi- crophones. bnc-to-phono adapters are included to allow testing of the line inputs using coax or standard audio cables. the line-level inputs are connected to the cs4216/8s lin1 and rin1 pins. as shown in figure 2, the line-level inputs go through a buff- er set to a gain of 0.5 which allows input signals of up to 2 v rms . when placed in serial mode 4 with loop back, the lin1 and rin1 inputs are used for analog inputs. the microphone inputs are connected to the cs4216/8s lin2 and rin2 pins. the two mi- crophone inputs are single-ended and are designed to work with both condenser and dy- namic microphones. the microphone input buffer, shown in figure 3, has a gain of 23 db thereby defining a full-scale input voltage to the microphone jacks of 71 mv rms . another 22 db of programmable gain is available on the cs4216/8 to amplify smaller microphone sig- nals. an analog patch area with analog power and ground, included on the cdb4216/8, provides space to develop other input buffer circuits. space for headers are included, j19 and j20, to connect to the lin2 and rin2 inputs. to use these headers, the microphone traces must be cut. analog outputs the cs4216/8 drives the line outputs into an r- c filter and then to a pair of bncs labeled cdb4216 44 ds83db4 150 0.47 uf 26 150 0.47 uf 28 rin2 refbuf lin2 20 cs4216 c19 r24 r25 c20 j19 lmic rmic j20 rin2 1 uf + c21 2 3 va 8 c16 1000 pf npo + r21 13 k 4 r18 47.5 k lin2 (mono) 1 uf + c22 6 5 c15 1000 pf r23 13 k r19 47.5 k npo c26 0.1 uf 1 7 u2 mc33178 + r22 1k r20 1 k c23 10 uf c24 10 uf 1 uf + c25 c37 0.1 uf c17 0.01 uf npo c18 0.01 uf npo figure 2. microphone input buffer refbuf lin1 27 20 cs4216 0.01 uf c13 npo 20 k + _ 10 k 56 pf 1 uf c10 c8 r12 7 6 5 4 8 r14 lin1 npo 0.47 uf c12 va c9 0.1 uf rin1 25 + _ 56 pf c27 3 2 1 rin1 (mono) 20 k r15 npo 10 k r13 1 uf c11 u2 lt1013 0.01 uf npo c14 150 r17 5 k r11 150 r16 0.1 uf c37 + + figure 3. line input buffer cdb4216 ds83db4 45 lout and rout. as with the line inputs, bnc- to-phono adapters are provided for flexibility. the line outputs can drive an impedance of 10 k w or more, which is the typical input imped- ance of most audio gear. audio port header the cdb4216/8 is primarily designed to evalu- ate the cs4216/8 in single chip mode, i.e. only one codec on the serial bus. this is the factory default state of the cdb4216/8. the audio port header j15 provides all buffered signals necessary to connect to the serial port of a dsp or other controller (see figure 4). sdou- tub can provide an unbuffered version of sdout which can be used when connecting multiple codecs on the same bus. the default configuration does not connect sdoutub which may be connected to the sdout of the cs4216/8 through j17 jumper. the eval board supports both master and slave sub-modes. in master sub-modes, ssync and sclk are output (and buffered) from the cs4216/8. in slave sub-modes, ssync and sclk must be provided externally and must be synchronous to the master clock clkin. r32 20 r31 20 a0 a3 a2 a1 oea oeb b3 b2 b1 b0 ssync sclk r39 20 r44 10 r10 10 j17 u6 74hc243 cfsin r40 20 r41 20k r49 100 9 13 8 4 j16 mf5 vd pdn j15 sdoutub sdout sdin sclk ssync r46 47.5k r42 28k vd d2 r38 806 ssync sclk sdout sdin di1 mf1 mf2 mf5 pdn do1 cs4216 1 44 43 42 33 40 39 38 13 37 4 16 8 12 614 317 515 2 18 9 11 713 8 9 10 11 1 13 6 5 4 3 mf1b di1 di4 di3 di2 j13 di4 di3 di2 di1 do4 do3 do2 do1 mf1a mf2 1 19 u7 74hc541 oe1 oe2 j18 u b do1 q1 237k r9 74hc541 u7 figure 4. serial port headers cdb4216 46 ds83db4 19 18 17 16 15 14 13 12 cfsin 1 2 3 4 5 6 7 8 9 smode1 mf8 mf7 smode2 smode3 13 14 16 r49 100 1 2 3 4 5 6 7 8 9 10 11 1 r48 27k di4 di3 di2 di1 2 3 4 5 6 r36 47k r37 47k r33 47k vd 1 r35 47k 2345678910 sw3 13 24 1 12 vd 1 23456 mf1a mf2 mf4 mf3 mf6 mf1b r45 10k vd r53 27k r54 27k r47 27k reset pdn mf5 20 14 13 21 19 18 17 16 reset pdn int ccs cclk cdin cdout u10 vd 1 2 3 4 5 6 7 8 9 r34 10k j14 swx 12 palce22v10z vd 24 c40 0.1uf vd 20 c39 0.1uf 10 lb spf2 spf1 ma bpf2 bpf1 ts2 ts1 div3 div2 div1 swx u9 palce16v8z swy figure 5. dip switch decode + digital header cdb4216 ds83db4 47 control port header the control port header j14 contains the control port pins, available only in sm4, and the pdn and reset pins. serial mode 4, sm4, splits the serial data to the codec into two separate serial ports, the audio port and the control port. the control port pins are available on this header. since cdout is buffered and always driven, it cannot be used on a shared serial port. although the int pin on the codec is open drain, the default factory configu- ration for the eval board is an on-board pull-up resistor and a buffer. therefore, the int header pin cannot share an interrupt pin on a processor since it is buffered and will always be driven. by cutting a trace in the j18 jumper, the unbuffered int signal, labeled u, can be supplied to the header. when using the control port, the lb switch must be off or the control serial port will be blocked. pdn and reset pdn is buffered and controls the pdn pin on the cs4216. pdn contains an on-board pull-up resistor defining the default state as powered. this pin only needs to be controlled when the power down feature is used. reset is also buffered and controls the reset pin on the codec (see figure 6). reset has a pull-up resistor on the board defining the default state as not reset or active. this pin only needs to be controlled when the reset feature on the codec is needed. since the codec requires a reset at power up, a power-up reset circuit is included on the board. a reset switch is also included to allow resetting the device without having to re- move the power supply. the power-up reset plus switch are logically ored with the reset pin on header j14. digital i/o header the digital i/o header, j13 shown in figure 4, contains the four digital inputs, di1-di4, and the four digital outputs, do1-do4. note that all digital i/o except di1 and do1 are multifunc- tion pins and may not be available in a particular mode. since do1 is always a digital output, an led is connected to do1 providing a visual in- dication that software is writing this bit correctly. when the led is on, do1 is high. in sm1 and sm2 all four digital inputs and out- puts are available. in sm3 master sub-modes, only the first two inputs and outputs are func- tional. in sm3 slave sub-modes, three inputs and two outputs are functional. in sm4 only do1 and di1 are functional. see the cs4216/8 data sheet for more details. clocks the cdb4216/8 provides an on-board default clock oscillator of 11.2896 mhz (see figure 7). this allows all 44.1 khz and derivative sample frequencies in sm3 master sub-mode, sm3 slave sub-mode, sm4, and the i 2 s mode. the cs4218 sm3-mm and sm3-ms modes require a master clock of 16xfs max . if using sm1, a master clock with a frequency that is 512xfs max must be supplied. sm2 uses sclk as the master clock ant it must be 256xfs max . a clkin bnc allows the eval board to be driven from an exter- vd vd reset u11 74hc132 r1 47.5k r4 100 r3 100 sw1a r2 47.5k d1 1n4148 cs4216 reset 2 + c1 1 uf 8 9 10 11 12 13 figure 6. reset circuit cdb4216 48 ds83db4 nal source. to select the clkin bnc, the j1 jumper must be placed in the ext position. when the j1 jumper is in the int position, the on-board oscillator is used as the master clock. both clock sources are buffered to guarantee a clean signal and proper clock levels to the codec. if sample frequencies other than the ones pro- vided are needed, the oscillator can be replaced with the proper frequency oscillator. the board accepts crystals and provides the socket y1 (refer to figure 8). when using a crystal, u8 must con- tain an hcu04 unbuffered cmos inverter. the u8 socket is designed to accept either the hcu04 or a crystal oscillator, and can alternate between the two. layout issues figure 11 contains the silk screen, figure 12 contains the component-side copper layer, and figure 13 contains the solder-side copper layer of the cdb4216/8 evaluation board. these plots are included to provide an example of how to correctly layout a pcb for the codec. grounding and power u8 74hcuo4 y1 r55 10m c44 33 pf c43 33 pf clkin-j1 8 c2 0.1uf vd 9 10 11 12 13 14 7123456 figure 8. optional clock circuit u11b 74hc132 vd 4 5 3 1 2 3 clkin 6 clkin u11a cs4216 r8 10 j1 11.2896 mhz oscillator module c2 0.1 uf vd ext int r5 47.5k r7 5k r6 47.5k 8 14 7 figure 7. default clock circuit cdb4216 ds83db4 49 notice in figure 12 and figure 13 how the ground plane split is positioned. the split is next to the part - not under it . the agnd and dgnd pins are connected to the ground plane fill inside the codec pad layout on the component-side layer. this is recommended be- cause agnd and dgnd are connected on the codec die and must have a zero impedance be- tween them. notice how each ground connection has at least four points in thermal relief. the main board grounds at the terminal connections have eight points in thermal relief. this helps minimize the impedance to the main ground terminal from any particular ground pin, reducing the chance of noise coupling. another important design consideration is the ground plane fill between traces on both layers, which minimizes coupling of radiated energy. ground fill on the digital side of the board helps reduce the amount of noisy digital energy radi- ated to the sensitive analog side and to a host system. ground fill on the analog side helps re- duce the amount of radiated digital energy that is coupled into the analog circuitry. all ground plane fills must be connected to their respective grounds - floating ground fill is worse than no fill. all power and ground traces are as thick as the surface mount pads they connect. thick traces minimize impedance, thereby reducing the chance of noise coupling. decoupling notice how the decoupling capacitors are placed as close as possible to the codec. the 0.1 m f ca- pacitors are placed closer than the 10 m f or 1 m f capacitors. this reduces lead inductance at high frequencies and allows the smaller valued ca- pacitors to attenuate unwanted signals more effectively. sockets the cdb4216/8 was designed to accommodate either the 44-pin plcc package or the 44-pin tqfp package. each evaluation board is shipped with a plcc codec loaded into a sur- face mount socket. notice how the socket pads match the footprint of the plcc package. using this socket in a design allows for testing with the socket mounted, and the option to surface mount the codec directly for cost reduction during board production. cdb4216 50 ds83db4 ;palasm design description ;---------------------------------- declaration segment ------------ title cdb4216 pattern 4216s_b revision 4.0b author c. sanchez, m. jordan company crystal semiconductor date 5/28/93 chip _4216s_b palce16v8 ;---------------------------------- pin declarations --------------- pin 1 /spf2 combinatorial ; input pin 2 /spf1 combinatorial ; input pin 3 /ma combinatorial ; input pin 4 /bpf2 combinatorial ; input pin 5 /bpf1 combinatorial ; input pin 6 /ts2 combinatorial ; input pin 7 /ts1 combinatorial ; input pin 8 /div3 combinatorial ; input pin 9 /div2 combinatorial ; input pin 10 gnd pin 11 nc pin 12 /cfsin combinatorial ; output pin 13 nc pin 14 nc pin 15 smode3 combinatorial ; output pin 16 smode2 combinatorial ; output pin 17 mf7 combinatorial ; output pin 18 mf8 combinatorial ; output pin 19 smode1 combinatorial ; output pin 20 vcc ;----------------------------------- boolean equation segment ------ equations /cfsin = spf2 * ma smode3 = spf2 * spf1 smode2 = spf2 * /spf1 + spf2 * spf1 * /ma + spf2 * spf1 * ma * bpf1 * ts1 + spf2 * spf1 * ma * bpf1 * ts2 + spf2 * spf1 * ma * bpf2 * ts1 + spf2 * spf1 * ma * bpf2 * ts2 smode1 = /spf2 * spf1 + spf2 * spf1 * ma * bpf1 + spf2 * spf1 * ma * bpf2 mf8 = /spf2 * ts2 + spf2 * /spf1 * ma * bpf2 + spf2 * /spf1 * ma * bpf1 + spf2 * /spf1 * /ma * bpf2 * ts2 + spf2 * spf1 * /ma * /bpf2 * bpf1 * ts1 figure 9. palce16v8h pal equations. cdb4216 ds83db4 51 + spf2 * spf1 * /ma * /bpf2 * bpf1 * ts2 + spf2 * spf1 * /ma * bpf2 * ts2 + spf2 * spf1 * ma * div3 mf7 = /spf2 * ts1 + spf2 * /spf1 * /ma * /bpf2 * bpf1 * ts1 + spf2 * /spf1 * /ma * /bpf2 * bpf1 * ts2 + spf2 * /spf1 * /ma * bpf2 * ts1 + spf2 * /spf1 * ma * ts1 + spf2 * spf1 * /ma * /bpf2 * bpf1 + spf2 * spf1 * /ma * bpf2 * ts1 + spf2 * spf1 * ma * div2 figure 9. continued. cdb4216 52 ds83db4 ;palasm design description ;---------------------------------- declaration segment ------------ title cdb4216 pattern 4216l_b revision 2.0b author c. sanchez company crystal semiconductor date 4/27/93 chip _4216l_b palce22v10z ;---------------------------------- pin declarations --------------- pin 1 /spf2 combinatorial ; input pin 2 /spf1 combinatorial ; input pin 3 /ma combinatorial ; input pin 4 /bpf2 combinatorial ; input pin 5 /bpf1 combinatorial ; input pin 6 /div3 combinatorial ; input pin 7 /div2 combinatorial ; input pin 8 /div1 combinatorial ; input pin 9 di4 combinatorial ; input pin 10 di3 combinatorial ; input pin 11 di2 combinatorial ; input pin 12 gnd pin 13 cdin combinatorial ; input pin 14 cclk combinatorial ; input pin 15 nc pin 16 mf6 combinatorial ; output pin 17 mf3 combinatorial ; output pin 18 mf4 combinatorial ; output pin 19 mf2 combinatorial ; output pin 20 /ccs combinatorial ; input pin 21 mf1 combinatorial ; output pin 22 nc pin 23 nc pin 24 vcc ;----------------------------------- boolean equation segment ------ equations mf1 = spf2 * /spf1 * ma * div1 + spf2 * /spf1 * /ma * bpf2 mf1.trst = spf2 * /spf1 mf2 = spf2 * /spf1 * ma * div2 + spf2 * /spf1 * /ma * bpf1 + spf2 * spf1 * cdin mf2.trst = spf2 mf3 = /spf2 * di3 + spf2 * /spf1 * /ma * di3 + spf2 * /spf1 * ma * div3 + spf2 * spf1 * cclk mf4 = /spf2 * di4 figure 10. palce22v10z pal equations. cdb4216 ds83db4 53 + spf2 * /spf1 * ma + spf2 * spf1 * /ccs mf6 = /spf2 * di2 + /spf1 * di2 + spf2 * spf1 * ma * div1 + spf2 * spf1 * /ma * bpf2 figure 10. continued. cdb4216 54 ds83db4 e e1 d1 d d2/e2 44 pin plcc no. of terminals d2/e2 max min max min millimeters inches dim a d/e 17.65 17.40 0.685 b e a a1 b e 0.695 16.66 16.51 0.650 0.656 4.57 4.20 0.180 0.165 0.53 0.33 0.021 0.013 2.29 0.090 16.00 14.99 0.590 0.630 1.19 1.35 0.047 0.053 nom 17.53 16.59 4.45 0.41 2.79 15.50 1.27 nom 0.690 0.653 0.175 0.016 0.110 0.610 0.050 3.04 0.120 d1/e1 a1 e millimeters inches dim d/e d1/e1 e b 0.462 0 0 11.75 min max 1.45 min max 0.057 10.10 12.25 3.5 0.398 0.482 b a1 a 1.60 0.05 0.063 0.002 c 0.20 0.008 c 0.45 0.75 0.018 0.030 1.35 0.053 9.90 0.390 a1 a 44 pin tqfp 1 d d1 e1 e a2 a2 l l 0.70 0.45 0.026 0.014 0.036 0.90 0.018 0.30 0.09 0.004 nom 0.15 0.006 nom 3.5 12.0 0.60 1.40 10.0 0.80 0.37 0.145 0.472 0.024 0.055 0.394 0.031 0.016 0.006 ccc 0.10 0.004 7 7 ccc 44 lead tqfp ? notes ? smart analog tm is a trademark of crystal semiconductor corporation |
Price & Availability of CS4216-KQ
 |
|
|
|
|
All Rights Reserved © IC-ON-LINE 2003 - 2022 |
| [Add Bookmark] [Contact Us] [Link exchange] [Privacy policy] |
|
Mirror Sites : [www.datasheet.hk]
[www.maxim4u.com] [www.ic-on-line.cn]
[www.ic-on-line.com] [www.ic-on-line.net]
[www.alldatasheet.com.cn]
[www.gdcy.com]
[www.gdcy.net] |