CMOS MT93L16 Low-Voltage Acoustic Echo Canceller Preliminary Information Features • • • • • • • • • ISSUE3 July 1999 Ordering Information MT93L16AQ 36 Pin QSOP -40 °C to + 85 °C • • AGC on speaker path Handles up to 0 dB acoustic echo return loss and 0dB line ERL Transparent data transfer and mute options 20 MHz master clock operation Low power mode during PCM Bypass Bootloadable for future factory software upgrades 2.7V to 3.6V supply voltage; 5V-tolerant inputs • • • • • Applications • • • • Full duplex speaker-phone for digital telephone Echo cancellation for video conferencing Handsfree in automobile environment Full duplex speaker-phone for PC Limiter Sin µ/A-Law/ + Offset Null Linear - S2 MD1 NBSD Linear/ µ/A-Law ADV NLP + Sout Program RAM S3 S1 Program ROM CONTROL DATA1 Micro Interface DATA2 Howling Adaptive Filter Adaptive Filter Double Talk Detector Controller NBSD R3 R1 SCLK R2 MD2 -24 -> +21dB Linear/ µ/A-Law Rout AGC User Gain CS ADV NLP PORT 1 UNIT Line ECho Path • • ACOUSTIC ECHO PATH • Contains two echo cancellers: 112ms acoustic echo canceller + 16ms line echo canceller Works with low cost voice codec. ITU-T G.711 or signed mag µ/A-Law, or linear 2’s comp Each port may operate in different format Advanced NLP design - full duplex speech with no switched loss on audio paths Fast re-convergence time: tracks changing echo environment quickly Adaptation algorithm converges even during Double-Talk Designed for exceptional performance in high background noise environments Provides protection against narrow-band signal divergence Howling prevention stops uncontrolled oscillation in high loop gain conditions Offset nulling of all PCM channels Serial micro-controller interface ST-BUS, GCI, or variable-rate SSI PCM interfaces User gain control provided for speaker path (-24dB to +21dB in 3dB steps) PORT 2 • DS5068 + Offset Null + µ/A-Law/ Linear Rin Limiter VDD VSS RESET FORMAT ENA2 ENA1 LAW F0i BCLK/C4i MCLK Figure 1 - Functional Block Diagram 1 MT93L16 Preliminary Information ENA1 MD1 ENA2 MD2 Rin Sin IC MCLK IC IC IC LAW FORMAT RESET NC NC SCLK CS 1 36 35 2 34 3 33 4 5 32 31 6 30 7 29 8 9 QSOP 28 27 10 26 11 25 12 24 13 23 14 22 15 21 16 20 17 19 18 IC IC IC MCLK2 NC VSS VDD2 VSS2 IC IC BCLK/C4i F0i Rout Sout VDD NC DATA1 DATA2 Figure 2 - Pin Connections Pin Description Pin # Name Description 1 ENA1 SSI Enable Strobe / ST-BUS & GCI Mode for Rin/Sout (Input). This pin has dual functions depending on whether SSI or ST-BUS/GCI is selected. For SSI, this strobe must be present for frame synchronization. This is an active high channel enable strobe, 8 or 16 data bits wide, enabling serial PCM data transfer for on Rin/Sout pins. Strobe period is 125 microseconds. For ST-BUS or GCI, this pin, in conjunction with the MD1 pin, selects the proper mode for Rin/Sout pins (see ST-BUS and GCI Operation description). 2 MD1 ST-BUS & GCI Mode for Rin/Sout (Input). When in ST-BUS or GCI operation, this pin, in conjunction with the ENA1 pin, will select the proper mode for Rin/Sout pins (see ST-BUS and GCI Operation description). Connect this pin to Vss in SSI mode. 3 ENA2 SSI Enable Strobe / ST-BUS & GCI Mode for Sin/Rout (Input).This pin has dual functions depending on whether SSI or ST-BUS/GCI is selected. For SSI, this is an active high channel enable strobe, 8 or 16 data bits wide, enabling serial PCM data transfer on Sin/Rout pins. Strobe period is 125 microseconds. For ST-BUS/GCI, this pin, in conjunction with the MD2 pin, selects the proper mode for Sin/Rout pins (see ST-BUS and GCI Operation description). 4 MD2 ST-BUS & GCI Mode for Sin/Rout (Input).When in ST-BUS or GCI operation, this pin in conjunction with the ENA2 pin, selects the proper mode for Sin/Rout pins (see ST-BUS and GCI Operation description). Connect this pin to Vss in SSI mode. 5 Rin Receive PCM Signal Input (Input). 128 kbit/s to 4096 kbit/s serial PCM input stream. Data may be in either companded or 2’s complement linear format. This is the Receive Input channel from the line (or network) side. Data bits are clocked in following SSI, GCI or STBUS timing requirements. 6 Sin Send PCM Signal Input (Input). 128 kbit/s to 4096 kbit/s serial PCM input stream. Data may be in either companded or 2’s complement linear format. This is the Send Input channel (from the microphone). Data bits are clocked in following SSI,GCI or ST-BUS timing requirements. 7 IC Internal Connection (Input): Must be tied to Vss. 8 MCLK 9,10,11 IC 12 LAW 13 2 Master Clock (Input): Nominal 20 MHz Master Clock input (may be asynchronous relative to 8KHz frame signal.) Tie together with MCLK2 (pin 33). Internal Connection (Input): Must be tied to Vss. A/µ Law Select (Input). When low, selects µ−Law companded PCM. When high, selects ALaw companded PCM. This control is for both serial pcm ports. FORMAT ITU-T/Sign Mag (Input). When low, selects sign-magnitude PCM code. When high, selects ITU-T (G.711) PCM code. This control is for both serial pcm ports. MT93L16 Preliminary Information Pin Description (continued) Pin # Name 14 RESET 15, 16 NC 17 SCLK 18 CS Serial Port Chip Select (Input). Enables serial microport interface data transfers. Active low. 19 DATA2 Serial Data Receive (Input). In Motorola/National serial microport operation, the DATA2 pin is used for receiving data. In Intel serial microport operation, the DATA2 pin is not used and must be tied to Vss or Vdd. 20 DATA1 Serial Data Port (Bidirectional). In Motorola/National serial microport operation, the DATA1 pin is used for transmitting data. In Intel serial microport operation, the DATA1 pin is used for transmitting and receiving data. 21 NC 22 VDD Positive Power Supply (Input). Nominally 3.3 volts. 23 Sout Send PCM Signal Output (Output). 128 kbit/s to 4096 kbit/s serial PCM output stream. Data may be in either companded or 2’s complement linear PCM format. This is the Send Out signal after acoustic echo cancellation and non-linear processing. Data bits are clocked out following SSI, ST-BUS, or GCI timing requirements. 24 Rout Receive PCM Signal Output (Output). 128 kbit/s to 4096 kbit/s serial PCM output stream. Data may be in either companded or 2’s complement linear PCM format. This is the Receive out signal after line echo cancellation non-linear processing, AGC, and gain control. Data bits are clocked out following SSI, ST-BUS, or GCI timing requirements. 25 F0i Frame Pulse (Input). In ST-BUS (or GCI) operation, this is an active-low (or active-high) frame alignment pulse, respectively. SSI operation is enabled by connecting this pin to Vss. 26 Description Reset / Power-down (Input). An active low resets the device and puts the MT93L16 into a low-power stand-by mode. No Connect (Output). These pins should be left un-connected. Serial Port Synchronous Clock (Input). Data clock for the serial microport interface. No Connect (Output). This pin should be left un-connected. BCLK/C4i Bit Clock/ST-BUS Clock (Input). In SSI operation, BCLK pin is a 128 kHz to 4.096 MHz bit clock. This clock must be synchronous with ENA1, and ENA2 enable strobes. In ST-BUS or GCI operation, C4i pin must be connected to the 4.096MHz (C4) system clock. 27, 28 IC Internal Connection (Input). Tie to Vss. 29 VSS2 Digital Ground (Input): Nominally 0 volts. 30 VDD2 Positive Power Supply (Input): Nominally 3.3 volts (tie together with VDD, pin 22). 31 VSS Digital Ground (Input): Nominally 0 volts (tie together with VSS2, pin 29). 32 NC No Connect (Output). This pin should be left un-connected. 33 MCLK2 34,35,36 IC Master Clock (Input): Nominal 20MHz master clock (tie together with MCLK, pin 8). Internal Connection (Input). Tie to Vss. Notes: 1. All inputs have CMOS compatible, 5V-tolerant logic levels. 2. All outputs have CMOS logic levels. Rout, Sout, and DATA1 are 5V-tolerant when tristated (to withstand other 5V drivers on a shared bus). Glossary Double-Talk Near-end Single-Talk Far-end Single-Talk ADV NLP Howling Narrowband NBSD Noise-Gating Offset Nulling Reverberation time ERL ERLE AGC Simultaneous signals present on Rin and Sin. Signals only present at Sin input. Signals only present at Rin input. Advanced Non-Linear-Processor Oscillation caused by feedback from acoustic and line echo paths Any mono or dual sinusoidal signals Narrow Band Signal Detector Audible switching of background noise Removal of DC component The time duration before an echo level decays to -60dBm Echo Return Loss Echo Return Loss Enhancement Automatic Gain Control 3 MT93L16 Functional Description The MT93L16 device contains two echo cancellers, as well as the many control functions necessary to operate the echo cancellers. One canceller is for acoustic speaker to microphone echo, and one for line echo cancellation. The MT93L16 provides clear signal transmission in both audio path directions to ensure reliable voice communication, even with low level signals. The MT93L16 does not use variable attenuators during double-talk or single-talk periods of speech, as do many other acoustic echo cancellers for speaker-phones. Instead, the MT93L16 provides high performance full-duplex operation similar to network echo cancellers, so that users experience clear speech and un-interrupted background signals during the conversation. This prevents subjective sound quality problems associated with “noise gating” or “noise contrasting”. Preliminary Information • • PCM encoder/decoder compatible with µ/ALaw ITU-T G.711, µ/A-Law Sign-Mag or linear 2’s complement coding. Automatic gain control on the receive speaker path. Adaptation Speed Control The adaptation speed of the acoustic echo canceller is designed to optimize the convergence speed versus divergence caused by interfering near-end signals. Adaptation speed algorithm takes into account many different factors such as relative double-talk condition, far end signal power, echo path change, and noise levels to achieve fast convergence. Advanced Non-Linear Processor (ADV-NLP)2 (2. Patent Pending) The MT93L16 uses an advanced adaptive filter algorithm that is double-talk stable, which means that convergence takes place even while both parties are talking 1. This algorithm allows continual tracking of changes in the echo path, regardless of doubletalk, as long as a reference signal is available for the echo canceller. (1. Patent Pending) The echo tail cancellation capability of the acoustic echo canceller has been sized appropriately (112ms) to cancel echo in an average sized office with a reverberation time of less than 112ms. The 16ms line echo canceller is sufficient to ensure a high ERLE for most line circuits. In addition to the echo cancellers, the following functions are supported: • Control of adaptive filter convergence speed during periods of double-talk, far end singletalk, and near-end echo path changes. • Control of Non-Linear Processor thresholds for suppression of residual non-linear echo. • Howling detector to identify when instability is starting to occur, and to take action to prevent oscillation. • Narrow-Band Detector for preventing adaptive filter divergence caused by narrow-band signals • Offset Nulling filters for removal of DC components in PCM channels. • Limiters that introduce controlled saturation levels. • Serial controller interface compatible with Motorola, National and Intel microcontrollers. 4 After echo cancellation, there is likely to be residual echo which needs to be removed so that it will not be audible. The MT93L16 uses an NLP to remove low level residual echo signals which are not comprised of background noise. The operation of the NLP depends upon a dynamic activation threshold, as well as a double-talk detector which disables the NLP during double-talk periods. The MT93L16 keeps the perceived noise level constant, without the need for any variable attenuators or gain switching that causes audible “noise gating”. The noise level is constant and identical to the original background noise even when the NLP is activated. For each audio path, the NLP can be disabled by setting the NLP- bit to 1 in the LEC or AEC control registers. Narrow Band Signal Detector (NBSD) 3 (3. Patent Pending) Single or multi-frequency tones (e.g. DTMF, or signalling tones) present in the reference input of an echo canceller for a prolonged period of time may cause the adaptive filter to diverge. The Narrow Band Signal Detector (NBSD) is designed to prevent this divergence by detecting single or multi-tones of arbitrary frequency, phase, and amplitude. When narrow band signals are detected, the filter adaptation process is stopped but the echo canceller continues to cancel echo. The NBSD can be disabled by setting the NB- bit to 1 in the MC control registers. MT93L16 Preliminary Information Howling Detector (HWLD) 4 (4. Patent Pending) The Howling detector is part of an Anti-Howling control, designed to prevent oscillation as a result of positive feedback in the audio paths. The HWLD can be disabled by setting the AH- bit to 1 in the (MC) control register. The AGC can be disabled by setting the AGC- bit to 1 in MC control register. Mute Function A pcm mute function is provided for independent control of the Receive and Send audio paths. Setting the MUTE_R or MUTE_S bit in the MC register, causes quiet code to be transmitted on the Rout or Sout paths respectively. Offset Null Filter To ensure robust performance of the adaptive filters at all times, any DC offset that may be present on either the Rin signal or the Sin signal, is removed by highpass filters. These filters have a corner frequency placed at 40Hz. The offset null filters can be disabled by setting the HPF- bit to 1 in the LEC or AEC control registers. Limiters To prevent clipping in the echo paths, two limiters with variable thresholds are provided at the outputs. The Rout limiter threshold is in Rout Limiter Register 1 and 2. The Sout limiter threshold is in Sout Limiter Register. Both output limiters are always enabled. User Gain The user gain function provides the ability for users to adjust the audio gain in the receive path (speaker path). This gain is adjustable from -24dB to +21dB in 3dB steps. It is important to use ONLY this user gain function to adjust the speaker volume. The user gain function in the MT93L16 is optimally placed between the two echo cancellers such that no reconvergence is necessary after gain changes. The gain can be accessed through Receive Gain Control Register. AGC The AGC function is provided to limit the volume in the speaker path. The gain of the speaker path is automatically reduced during the following conditions: • When clipping of the receive signal occurs. • When initial convergence of the acoustic echo canceller detects unusually large echo return. • When howling is detected. Quiet code is defined according to the following table. LINEAR SIGN/ 16 bits MAGNITUDE 2’s µ-Law complement A-Law +Zero (quiet code) 0000h 80h CCITT (G.711) µ-Law A-Law FFh D5h Table 1 - Quiet PCM Code Assignment Bypass Control A PCM bypass function is provided to allow transparent transmission of pcm data through the MT93L16. When the bypass function is active, pcm data passes transparently from Rin to Rout and from Sin to Sout, with bit-wise integrity preserved. When the Bypass function is selected, most internal functions are powered down to provide low power consumption. The BYPASS control bit is located in the main control MC register. Adaptation Enable/Disable Adaptation control bits are located in the AEC and LEC control registers. When the ADAPT- bit is set to 1, the adaptive filter is frozen at the current state. In this state, the device continues to cancel echo with the current echo model. When the ADAPT- bit is set to 0, the adaptive filter is continually updated. This allows the echo canceller to adapt and track changes in the echo path. This is the normal operating state. MT93L16 Throughput Delay In all modes, voice channels always have 2 frames of delay. In ST-BUS/GCI operation, the D and C channels have a delay of one frame. 5 MT93L16 Preliminary Information Power Down / Reset Holding the RESET pin at logic low will keep the MT93L16 device in a power-down state. In this state all internal clocks are halted, and the DATA1, Sout and Rout pins are tristated. high) frame pulse is applied to the F0i pin, the device will assume GCI operation. If F0i is tied continuously to Vss, the device will assume SSI operation. Figures 11 to 13 show timing diagrams of these 3 PCM-interface operation conventions. ST-BUS and GCI Operation The user should hold the RESET pin low for at least 200 msec following power-up. This will insure that the device powers up in a proper state. Following any return of RESET to logic high, the user must wait for 8 complete 8 KHz frames prior to writing to the device registers. During this time, the initialization routines will execute and set the MT93L16 to default operation (program execution from ROM using default register values). PCM Data I/O The PCM data transfer for the MT93L16 is provided through two PCM ports. One port consists of Rin and Sout pins while the second port consists of Sin and Rout pins. The data are transferred through these ports according to either ST-BUS, GCI, or SSI conventions, and the device automatically detects the correct convention. The device determines the convention by monitoring the signal applied to the F0i pin. When a valid ST-BUS (active low) frame pulse is applied to the F0i pin, the MT93L16 will assume ST-BUS operation. When a valid GCI (active The ST-BUS PCM interface conforms to Mitel’s STBUS standard, with an active-low frame pulse. Input data is clocked in by the rising edge of the bit clock (C4i) three-quarters of the way into the bitcell, and output data bit boundaries (Rout, Sout) occur every second falling edge of the bit clock (see Figure 11.) The GCI PCM interface corresponds to the GCI standard commonly used in Europe, with an activehigh frame pulse. Input data is clocked in by the falling edge of the bit clock (C4i) three-quarters of the way into the bitcell, and output data bit boundaries (Rout, Sout) occur every second rising edge of the bit clock (see Figure 12.) Either of these interfaces (STBUS or GCI) can be used to transport 8 bit companded PCM data (using one timeslot) or 16 bit 2’s complement linear PCM data (using two timeslots). The MD1/ENA1 pins select the timeslot on the Rin/Sout port while the MD2/ENA2 pin selects the timeslot on the Sin/Rout port, as in Table 2. Figures 3 to 6 illustrate the timeslot allocation for each of these four modes. C4i start of frame (stbus & GCI) F0i (ST-BUS) 0 1 2 3 4 B F0i (GCI) PORT1 Rin 7 6 5 4 3 2 1 0 EC Sout 7 6 5 4 3 2 1 0 PORT2 Sin 7 6 5 4 3 2 1 0 EC Rout 7 6 5 4 3 2 1 0 outputs = High impedance inputs = don’t care In ST-BUS/GCI Mode 1, echo canceller I/O channels are assigned to ST-BUS/GCI timeslot 0. Note that the user can configure PORT1 and PORT2 into different modes. Figure 3 - ST-BUS and GCI 8-Bit Companded PCM I/O on Timeslot 0 (Mode 1) 6 MT93L16 Preliminary Information C4i start of frame (stbus & GCI) F0i (ST-BUS) 0 1 2 3 4 B F0i (GCI) PORT1 Rin 7 6 5 4 3 2 1 0 EC 7 6 5 4 3 2 1 0 Sout PORT2 7 6 5 4 3 2 1 0 Sin EC 7 6 5 4 3 2 1 0 Rout outputs = High impedance inputs = don’t care In ST-BUS/GCI Mode 2, echo canceller I/O channels are assigned to ST-BUS/GCI timeslot 2. Note that the user can configure PORT1 and PORT2 into different modes. Figure 4 - ST-BUS and GCI 8-Bit Companded PCM I/O on Timeslot 2 (Mode 2) C4i start of frame (stbus & GCI) F0i (ST-BUS) 0 1 2 D C B 3 4 F0i (GCI) PORT1 Rin 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 EC Sout 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 PORT2 Sin 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 EC Rout 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 outputs = High impedance inputs = don’t care indicates that an input channel is bypassed to an output channel ST-BUS/GCI Mode 3 supports connection to 2B+D devices where timeslots 0 and 1 transport D and C channels and echo canceller (EC) I/O channels are assigned to ST-BUS timeslot 2 (B). Both PORT1 and PORT2 must be configured in Mode 3. Figure 5 - ST-BUS and GCI 8-Bit Companded PCM I/O with D and C channels (Mode 3) 7 MT93L16 Preliminary Information C4i start of frame (stbus & GCI) F0i (stbus) F0i (GCI) Rin S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 PORT1 EC S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 Sout S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 Sin PORT2 EC S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 Rout outputs = High impedance inputs = don’t care ST-BUS/GCI Mode 4 allows 16 bit 2’s complement linear data to be transferred using ST-BUS/GCI I/O timing. Note that PORT1 and PORT2 need not necessarily both be in mode 4. Figure 6 - ST-BUS and GCI 16-Bit 2’s complement linear PCM I/O (Mode 4) PORT1 Rin/Sout ST-BUS/GCI Mode Selection PORT2 Sin/Rout Enable Pins Enable Pins MD1 ENA1 MD2 ENA2 0 0 Mode 1. 8 bit companded PCM I/O on timeslot 0 0 0 0 1 Mode 2. 8 bit companded PCM I/O on timeslot 2. 0 1 1 0 Mode 3. 8 bit companded PCM I/O on timeslot 2. Includes D & C channel bypass in timeslots 0 & 1. 1 0 1 1 Mode 4. 16 bit 2’s complement linear PCM I/O on timeslots 0 & 1. 1 1 Table 2 - ST-BUS & GCI Mode Select In SSI operation, the frame boundary is determined by the rising edge of the ENA1 enable strobe (see Figure 7). The other enable strobe (ENA2) is used for parsing input/output data and it must pulse within 125 microseconds of the rising edge of ENA1. In SSI operation, the enable strobes may be a mixed combination of 8 or 16 BCLK cycles allowing the flexibility to mix 2’s complement linear data on one port (e.g., Rin/Sout) with companded data on the other port (e.g., Sin/Rout). Enable Strobe Pin Designated PCM I/O Port ENA1 Line Side Echo Path (PORT 1) ENA2 Acoustic Side Echo Path (PORT 2) Table 3 - SSI Enable Strobe Pins SSI Operation PCM Law and Format Control (LAW, FORMAT) The SSI PCM interface consists of data input pins (Rin, Sin), data output pins (Sout, Rout), a variable rate bit clock (BCLK), and two enable pins (ENA1, ENA2) to provide strobes for data transfers. The active high enable may be either 8 or 16 BCLK cycles in duration. Automatic detection of the data type (8 bit companded or 16 bit 2’s complement linear) is accomplished internally. The data type cannot change dynamically from one frame to the next. 8 The PCM companding/coding law used by the MT93L16 is controlled through the LAW and FORMAT pins. ITU-T G.711 companding curves for µ-Law and A-Law are selected by the LAW pin. PCM coding ITU-T G.711 and Sign-Magnitude are selected by the FORMAT pin. See Table 4. MT93L16 Preliminary Information BCLK start of frame (SSI) PORT1 ENA1 8 or 16 bits Rin EC Sout 8 or 16 bits PORT2 ENA2 8 or 16 bits Sin EC 8 or 16 bits Rout outputs = High impedance inputs = don’t care Note that the two ports are independent so that, for example, PORT1 can operate with 8-bit enable strobes and PORT2 can operate with 16-bit enable strobes. Figure 7 - SSI Operation Sign-Magnitude ITU-T (G.711) FORMAT=0 FORMAT=1 PCM Code µ/A-LAW µ-LAW A-LAW LAW = 0 or 1 LAW = 0 LAW =1 + Full Scale 1111 1111 1000 0000 1010 1010 + Zero 1000 0000 1111 1111 1101 0101 - Zero 0000 0000 0111 1111 0101 0101 - Full Scale 0111 1111 0000 0000 0010 1010 Table 4 - Companded PCM Linear PCM The 16-bit 2’s complement PCM linear coding permits a dynamic range beyond that which is specified in ITU-T G.711 for companded PCM. The echo-cancellation algorithm will accept 16 bits 2’s complement linear code which gives a maximum signal level of +15dBm0. Bit Clock (BCLK/C4i ) The BCLK/C4i pin is used to clock the PCM data for GCI and ST-BUS (C4i) interfaces, as well as for the SSI (BCLK) interface. In SSI operation, the bit rate is determined by the BCLK frequency. This input must contain either eight or sixteen clock cycles within the valid enable strobe window. BCLK may be any rate between 128 KHz to 4.096 MHz and can be discontinuous outside of the enable strobe windows defined by ENA1, ENA2 pins. Incoming PCM data (Rin, Sin) are sampled on the falling edge of BCLK while outgoing PCM data (Sout, Rout) are clocked out on the rising edge of BCLK. See Figure 13. In ST-BUS and GCI operation, connect the system C4 (4.096MHz) clock to the C4i pin. Master Clock (MCLK) A nominal 20MHz, continuously-running master clock (MCLK) is required. MCLK may be asynchronous with the 8KHz frame. 9 MT93L16 Microport The serial microport provides access to all MT93L16 internal read and write registers, plus write-only access to the bootloadable program RAM (see next section for bootload description.) This microport is compatible with Intel MCS-51 (mode 0), Motorola SPI (CPOL=0, CPHA=0), and National Semiconductor Microwire specifications. The microport consists of a transmit/receive data pin (DATA1), a receive data pin (DATA2), a chip select pin (CS) and a synchronous data clock pin (SCLK). The MT93L16 automatically adjusts its internal timing and pin configuration to conform to Intel or Motorola/National requirements. The microport dynamically senses the state of the SCLK pin each time CS pin becomes active (i.e. high to low transition). If SCLK pin is high during CS activation, then Intel mode 0 timing is assumed. In this case DATA1 pin is defined as a bi-directional (transmit/ receive) serial port and DATA2 is internally disconnected. If SCLK is low during CS activation, then Motorola/National timing is assumed and DATA1 is defined as the data transmit pin while DATA2 becomes the data receive pin. The MT93L16 supports Motorola half-duplex processor mode (CPOL=0 and CPHA=0). This means that during a write to the MT93L16, by the Motorola processor, output data from the DATA1 pin must be ignored. This also means that input data on the DATA2 pin is ignored by the MT93L16 during a valid read by the Motorola processor. All data transfers through the microport are two bytes long. This requires the transmission of a Command/ Address byte followed by the data byte to be written to or read from the addressed register. CS must remain low for the duration of this two-byte transfer. As shown in Figures 8 and 9, the falling edge of CS indicates to the MT93L16 that a microport transfer is about to begin. The first 8 clock cycles of SCLK after the falling edge of CS are always used to receive the Command/Address byte from the microcontroller. The Command/Address byte contains information detailing whether the second byte transfer will be a read or a write operation and at what address. The next 8 clock cycles are used to transfer the data byte between the MT93L16 and the microcontroller. At the end of the two-byte transfer, CS is brought high again to terminate the session. The rising edge of CS will tri-state the DATA1 pin. The DATA1 pin will remain tristated as long as CS is high. Intel processors utilize Least Significant Bit (LSB) first transmission while Motorola/National processors use Most Significant Bit (MSB) first transmission. 10 Preliminary Information The MT93L16 microport automatically accommodates these two schemes for normal data bytes. However, to ensure timely decoding of the R/W and address information, the Command/ Address byte is defined differently for Intel and Motorola/National operations. Refer to the relative timing diagrams of Figure 8 and Figure 9. Receive data bits are sampled on the rising edge of SCLK while transmit data is clocked out on the falling edge of SCLK. Detailed microport timing is shown in Figure 14 and Figure 15. Bootload Process and Execution from RAM A bootloadable program RAM (BRAM) is available on the MT93L16 to support factory-issued software upgrades to the built-in algorithm. To make use of this bootload feature, users must include 4096 X 8bits of memory in their microcontroller system (i.e. external to the MT93L16), from which the MT93L16 can be bootloaded. Registers and program data are loaded into the MT93L16 in the same fashion via the serial microport. Both employ the same command / address / data byte specification described in the previous section on serial microport. Either intel or motorola mode may be transparently used for bootloading. There are also two registers relevant to bootloading (BRC=control and SIG=signature, see Register Summary). The effect of these register values on device operation is summarized in Table 5. Bootload mode is entered and exited by writing to the bootload bit in the Bootload RAM Control (BRC) register at address 3fh (see Register Summary). During bootload mode, any serial microport "write" (R/W command bit =0) to an address other than that of the BRC register will contribute to filling the program BRAM. Call these transactions "BRAM-fill" writes. Although a command/address byte must still precede each data byte (as described for the serial microport), the values of the address fields for these "BRAM-fill" writes are ignored (except for the value 3fh, which designates the BRC register.) Instead, addresses are internally generated by the MT93L16 for each "BRAM-fill" write. Address generation for "BRAM-fill" writes resumes where it left off following any read transaction while bootload mode is enabled. The first 4096 such "BRAM-fill" writes while bootload is enabled will load the memory, but further ones after that are ignored. Following the write of the first 4096 bytes, the program BRAM will be filled. Before bootload mode is disabled, it is recommended that users then read back the value from the signature register (SIG) and compare it to the one supplied by the factory along with the code. Equality verifies that the correct data has been loaded. The signature calculation uses an 8-bit MISR which only incorporates input from "BRAM-fill" MT93L16 Preliminary Information FUNCTIONAL DESCRIPTION FOR USING THE BOOTABLE RAM BOOTLOAD MODE - Microport Access is to bootload RAM (BRAM) R/W Address Data W 3fh (= 1 1 1 1 1 1 b) Writes "data" to BRC reg. - Bootload frozen; BRAM contents are NOT affected. C3C2C1C0 W other than 3fh Writes "data" to next byte in BRAM (bootloading.) X 1 0 0 R 1x xxxxb Reads back "data" = BRC reg value. - Bootload frozen; BRAM contents are NOT affected. R 0x xxxxb Reads back "data" = SIG reg value. - Bootload frozen; BRAM contents are NOT affected. BRC Register Bits NON-BOOTLOAD MODE - Microport Access is to device registers (DREGs) BRC Register Bits C3C2C1C0 R/W Address Data W any (= a5 a4 a3 a2 a1 a0 b) Writes "data" to corresponding DREG. R any (= a5 a4 a3 a2 a1 a0 b) Reads back "data" = corresponding DREG value. X 0 0 0 PROGRAM EXECUTION MODES C3C2C1C0 0 0 0 0 Execute program in ROM, bootload mode disabled. - BRAM address counter reset to initial (ready) state. - SIG reg reseeded to initial (ready) state C3C2C1C0 Execute program in ROM, while bootloading the RAM. - BRAM address counter increments on microport writes (except to 3fh) - SIG reg recalculates signature on microport writes (except to 3fh) C3C2C1C0 1 0 0 0 Execute program in RAM, bootload mode disabled. - BRAM address counter reset to initial (ready) state. - SIG reg reseeded to initial (ready) state C3C2C1C0 - NOT RECOMMENDED (Execute program in RAM, while bootloading the RAM) 0 1 0 0 1 1 0 0 Table 5 - Bootload RAM Control (BRC) Register States Note: bits C1 C0 are reserved, and must be set to zero. writes. Resetting the bootload bit (C2) in the BRC register to 0 (see Register Summary) exits bootload mode, resetting the signature (SIG) register and internal address generator for the next bootload. A hardware reset (RESET=0) similarly returns the MT93L16 to the ready state for the start of a bootload. Once the program has been loaded, to begin execution from RAM, bootload mode must be disabled (BOOT bit, C2=0) and execution from RAM enabled (RAM_ROMb bit, C3=1) by setting the appropriate bits in the BRC register. During the bootload process, however, ROM program execution (RAM_ROMb bit, C3=0) should be selected. See Table 5 for the effect of the BRC register settings on Microport accesses and on program execution. Following program loading and enabling of execution from RAM, it is recommended that users set the software reset bit in the Main Control (MC) register, to ensure that the device updates the default register values to those of the new program in RAM. Note: it is important to use a software reset rather than a hardware (RESET=0) reset, as the latter will return the device to its default settings (which includes execution from program ROM instead of RAM.) To verify which code revision is currently running, users can access the Firmware Revision Code (FRC) register (see Register Summary). This register reflects the identity code (revision number) of the last program to run register initialization (which follows a software or hardware reset.) 11 MT93L16 Preliminary Information COMMAND/ADDRESS ➄ DATA 1 R/W A0 A1 A2 A3 A4 A5 DATA INPUT/OUTPUT D0 D1 D2 D3 D4 D5 D6 D7 X ➀ SCLK ➁ ➃ CS ➂ ➀ ➁ This delay is due to internal processor timing and is equal to Tsch time. The delay is transparent to MT93L16. The MT93L16: latches receive data on the rising edge of SCLK outputs transmit data on the falling edge of SCLK ➂ The falling edge of CS indicates that a COMMAND/ADDRESS byte will be transmitted from the microprocessor. The subsequent byte is always data followed by CS returning high. ➃ A new COMMAND/ADDRESS byte may be loaded only by CS cycling high then low again. ➄ The COMMAND/ADDRESS byte contains: 1 bit - Read/Write 6 bits - Addressing Data 1 bit - Unused Figure 8 - Serial Microport Timing for Intel Mode 0 COMMAND/ADDRESS ➄ DATA 2 Receive R/W A5 A4 A3 A2 A1 A0 DATA INPUT D7 D6 D5 D4 D3 D2 D1 D0 X DATA OUTPUT DATA 1 Transmit D7 D6 D5 D4 D3 D2 D1 D0 High Impedance ➀ SCLK ➁ ➃ CS ➂ ➀ This delay is due to internal processor timing and is equal to Tsch time. The delay is transparent to MT93L16. ➁ The MT93L16: latches receive data on the rising edge of SCLK outputs transmit data on the falling edge of SCLK ➂ The falling edge of CS indicates that a COMMAND/ADDRESS byte will be transmitted from the microprocessor. The subsequent byte is always data followed by CS returning high. ➃ A new COMMAND/ADDRESS byte may be loaded only by CS cycling high then low again. ➄ The COMMAND/ADDRESS byte contains: 1 bit - Read/Write 6 bits - Addressing Data 1 bit - Unused Figure 9 - Serial Microport Timing for Motorola Mode 00 or National Microwire 12 MT93L16 Preliminary Information Absolute Maximum Ratings* Parameter Symbol Min Max Units VDD-VSS -0.5 5.0 V 1 Supply Voltage 2 Input Voltage Vi VSS-0.3 5.5 V 3 Output Voltage Swing Vo VSS-0.3 5.5 V 4 Continuous Current on any digital pin Ii/o ±20 mA 5 Storage Temperature TST 150 °C -65 6 Package Power Dissipation PD 90 (typ) * Exceeding these values may cause permanent damage. Functional operation under these conditions is not implied. mW . Recommended Operating Conditions - Voltages are with respect to ground (VSS) unless otherwise stated Characteristics Sym Min Typ Max Units VDD 2.7 3.3 3.6 V 1 Supply Voltage 2 Input High Voltage 1.4 VDD V 3 Input Low Voltage VSS 0.4 V 4 Operating Temperature TA -40 +85 °C Min Typ Test Conditions Echo Return Limits Characteristics Max Units Test Conditions 1 Acoustic Echo Return 0 dB Measured from Rout -> Sin 2 Line Echo Return 0 dB Measured from Sout -> Rin DC Electrical Characteristics*- Voltages are with respect to ground (VSS) unless otherwise stated. Characteristics Sym Min Typ‡ Max Units 70 µA RESET = 0 mA RESET = 1, clocks active Standby Supply Current: ICC 3 Operating Supply Current: IDD 20 2 Input HIGH voltage VIH 3 Input LOW voltage VIL 4 Input leakage current IIH/IIL 5 High level output voltage VOH 6 Low level output voltage VOL 7 High impedance leakage IOZ 1 8 Output capacitance Co 10 1 0.7VDD Conditions/Notes V 0.1 0.3VDD V 10 µA VIN=VSS to VDD V IOH=2.5mA 0.4VDD V IOL=5.0mA 10 µA VIN=VSS to VDD 0.8VDD pF 9 Input capacitance Ci 8 pF ‡ Typical figures are at 25°C and are for design aid only: not guaranteed and not subject to production testing. *DC Electrical Characteristics are over recommended temperature and supply voltage. 13 MT93L16 Preliminary Information AC Electrical Characteristics† - Serial Data Interfaces - Voltages are with respect to ground (VSS) unless otherwise stated Characteristics Sym Min Typ Max Units 20.5 MHz Test Notes 1 MCLK Frequency fCLK 19.15 2 BCLK/C4i Clock High tBCH, tC4H 90 ns 3 BCLK/C4i Clock Low tBLL, tC4L 90 ns 4 BCLK/C4i Period tBCP 240 5 SSI Enable Strobe to Data Delay (first bit) tSD 80 ns CL=150pF 6 SSI Data Output Delay (excluding first bit) tDD 80 ns CL=150pF 7 SSI Output Active to High Impedance tAHZ 80 ns CL=150pF 8 SSI Enable Strobe Signal Setup tSSS 10 tBCP -15 ns 9 SSI Enable Strobe Signal Hold tSSH 15 tBCP -10 ns 10 SSI Data Input Setup tDIS 10 ns 11 SSI Data Input Hold tDIH 15 ns 12 ST-BUS/GCI F0i Setup tF0iS 20 150 ns 13 ST-BUS/GCI F0i Hold tF0iH 20 150 ns 14 ST-BUS/GCI Data Output delay tDSD 80 ns CL=150pF 15 ST-BUS/GCI Output Active to High Impedance tASHZ 80 ns CL=150pF 16 ST-BUS/GCI Data Input Hold time tDSH 20 ns 17 ST-BUS/GCI Data Input Setup time tDSS 20 † Timing is over recommended temperature and power supply voltages. ns 14 7900 ns MT93L16 Preliminary Information AC Electrical Characteristics† - Microport Timing Characteristics Sym Min Typ Max Units 1 Input Data Setup tIDS 30 ns 2 Input Data Hold tIDH 30 ns 3 Output Data Delay tODD 100 ns 4 Serial Clock Period tSCP 500 ns 5 SCLK Pulse Width High tSCH 250 ns 6 SCLK Pulse Width Low tSCL 250 ns 7 CS Setup-Intel tCSSI 200 ns 8 CS Setup-Motorola tCSSM 100 ns 9 CS Hold tCSH 100 ns 10 CS to Output High Impedance tOHZ 100 ns Test Notes CL=150pF CL=150pF † Timing is over recommended temperature range and recommended power supply voltages. Characteristic Symbol CMOS Level Units CMOS reference level VCT 0.5*VDD V Input HIGH level VH 0.9*VDD V Input LOW level VL 0.1*VDD V Rise/Fall HIGH measurement point VHM 0.7*VDD V Rise/Fall LOW measurement point VLM 0.3*VDD V Table 8 - Reference Level Definition for Timing Measurements T=1/fCLK MCLK (I) VH VCT VL Notes: O. CMOS output I. CMOS input (5V tolerant) (see Table 8 for symbol definitions) Figure 10 - Master Clock - MCLK 15 MT93L16 Preliminary Information Bit 7 Bit 6 Sout/Rout (O) VCT tDSD C4i (I) tASHZ tC4H VH VCT input sampled VL tF0iS tF0iH F0i (I) VH tC4L VCT VL tDSS tDSH start of frame Rin/Sin (I) VH VCT VL Bit 6 Bit 7 Figure 11 -GCI Data Port Timing ) Bit 7 Bit 6 Sout/Rout (O) VCT tDSD C4i (I) tC4H VH VCT input sampled VL tF0iS tF0iH F0i (I) tASHZ VH tC4L VCT VL tDSS tDSH start of frame Rin/Sin (I) VH VCT VL Bit 6 Bit 7 Figure 12 - ST-BUS Data Port Timing Bit 7 Bit 6 Bit 5 VCT Sout/Rout (O) tSD VH VCT VL tSSS ENA1 (I) or ENA2 (I) tAHZ tBCH tBCP VH input sampled BCLK (I) tDD tBCL tSSH VCT VL tDIS tDIH start of frame VH Rin/Sin (1) VCT VL Bit 7 Notes: O. CMOS output Bit 6 I. CMOS input (5V tolerant) Bit 5 (see Table 8 for symbol definitions) Figure 13 - SSI Data Port Timing 16 MT93L16 Preliminary Information DATA OUTPUT DATA INPUT DATA1 (I,O) VCT tIDS tIDH SCLK (I) tSCH tODD tOHZ VH VCT VL tCSSI CS (I) tSCL tSCP tCSH VH VCT VL Notes: O. CMOS output I. CMOS input (5V tolerant) (see Table 8 for symbol definitions) Figure 14 - INTEL Serial Microport Timing DATA2 (I) (Input) VH VCT VL tIDS tIDH SCLK (I) tSCH tSCP VH VCT VL tCSSM CS (I) tSCL tCSH VH VCT VL tODD DATA1 (O) (Output) tOHZ VCT Notes: O. CMOS output I. CMOS input (5V tolerant) (see Table 8 for symbol definitions) Figure 15 - Motorola Serial Microport Timing 17 MT93L16 Preliminary Information Register Summary Address: 00h R/W Power Up Reset 00h RESET AHAGCNBBYPASS Main Control Register (MC) 7 MSB LIMIT 6 MUTE_R 5 MUTE_S NB- 2 AGC- 1 AH- When high, AGC is disabled and when low AGC is enabled. When high, Narrowband signal detectors in Rin and Sin paths are disabled and when low the signal detectors are enabled When high, the Send and Receive paths are transparently by-passed from input to output and when low the Send and Receive paths are not bypassed When high, the Sin path is muted to quite code (after the NLP) and when low the Sin path is not muted When high, the Rin path is muted to quite code (after the NLP) and when low the Rin path is not muted When high, the 2-bit shift mode is enabled in conjunction with bit 7 of LEC register and when low 2-bit shift mode is disabled Address: 21h R/W ECBY ADAPT- Acoustic Echo Canceller Control Register (AEC) 7 MSB P- 6 ASC- 5 NLP- 4 INJ- 3 HPF- 2 HCLR 1 ADAPT- 0 ECBY When high, the Echo canceller adaptation is disabled and when low the adaptation is enabled When high, Adaptive filter coefficients are cleared and when low the filter coefficients are not cleared HPF- When high, Offset nulling filter is bypassed in the Sin/Sout path and when low the Offset nulling filter in not bypassed INJ- When high, the Noise filtering process is disabled in the NLP and when low the Noise filtering process is enabled NLP- When high, the Non Linear Processor is disabled in the Sin/Sout path and when low the NLP is enabled ASC- When high, the Internal Adaptation speed control is disabled and when low the Adaptation speed is enabled When high, the Exponential weighting function for the adaptive filter is disabled and when low the weighting function is enabled Address: 01h R/W Power Up Reset 00h ECBY ADAPT- LSB When high, the Echo estimate from the filter is not subtracted from the input (Sin), when low the estimate is subtracted HCLR P- LSB When high, the Howling detector is disabled and when low the Howling detector is enabled. MUTE_S Power Up Reset 00h 0 RESET When high, the power initialization routine is executed presetting all registers to default values. This bit automatically clears itself to’0’ when reset is complete. MUTE_R LIMIT 18 4 BYPASS 3 Line Echo Canceller Control Register (LEC) MSB 7 SHFT 6 ASC- 5 NLP- 4 INJ- 3 HPF- 2 HCLR 1 ADAPT- 0 ECBY LSB When high, the Echo estimate from the filter is not substracted from the input (Rin), when low the estimate is substracted When high, the Echo canceller adaptation is disabled and when low the adaptation is enabled HCLR When high, Adaptive filter coefficients are cleared and when low the filter coefficients are not cleared HPF- When high, Offset nulling filter is bypassed in the Rin/Rout path and when low the Offset nulling filter in not bypassed INJ- When high, the Noise filtering process is disabled in the NLP and when low the Noise filtering process is enabled NLP- When high, the Non Linear Processor is disabled in the Rin/Rout path and when low the NLP is enabled ASC- When high, the Internal Adaptation speed control is disabled and when low the Adaptation speed is enabled SHFT when high the 16-bit linear mode, inputs Sin, Rin, are shift right by 2 and outputs Sout, Rout are shift left by 2. This bit is ignored when 16-bit linear mode is not selected in both ports. This bit is also ignored if bit 7 of MC register is set to zero MT93L16 Preliminary Information Address: 22h Read Power Up Reset 00h NBS Acoustic Echo Canceller Status Register (ASR) (* Do not write to this register) 7 6 ACMUND 5 HWLNG - MSB 4 3 - NLPDC 2 LOGICAL OR of the status bit NBS + NBR from LSR Register DT When high the Double Talk is detected and when low, the Double talk is not detected HWLNG ACMUND - RESERVED. When high, Howling is occurring in the loop and when low, no Howling is detected When high, No active signal in the Rin/Rout path RESERVED. Line Echo Canceller Status Register (LSR) (* Do not write to this register) Power Up Reset 00h 6 - 5 - 4 - 3 - NLPC 2 DT 1 0 NB NBR LSB When high, a narrowband signal has been detected in the Receive (Rin) path. When low no narrowband signal is not detected in the Rin path NB This bit indicates a LOGICAL-OR of Status bits NBR + NBS (from ASR Register) DT When high, double-talk is detected and when low double-talk is not detected NLPC LSB When high, the NLP is activated and when low the NLP is not activated Address: 02h Read NBR 0 NBS NB When high, the Narrowband signal has been detected in the Sin/Sout path and when low, the Narrowband signal has not been detected in the Sin/Sout path NB NLPDC 1 DT When high, NLP is activated and when low NLP is not activated - RESERVED. . -- Address: 20h R/W Power Up Reset 6Dh Receive Gain Control Register (RGC) 7 MSB - 6 - 5 - 4 - 3 G3 2 G2 1 0 G1 GO LSB G0 G1 G2 User Gain Control on the Rin/Rout path (Tolerance of gains: +/- 0.15 dB). The hexadecimal number represents G3 to G0 value in the table below. G3 - RESERVED - Gain Values for Receive Gain Control Register Bit G3 to G0 (RGC) 0h -24dB 4h -12dB 8h 0 dB Ch +12 dB 1h -21dB 5h -9 dB 9h + 3 dB Dh + 15 dB 2h -18dB 6h -6 dB Ah + 6 dB Eh + 18 dB 3h -15dB 7h -3 dB Bh +9 dB Fh + 21 dB 19 MT93L16 Preliminary Information Address: 16h Read Power Up Reset 00h Receive (Rin) Peak Detect Register 1 MSB 7 RIPD 7 6 RIPD 6 5 RIPD5 4 RIPD4 3 RIPD3 2 (RIPD1) RIPD2 1 RIPD1 0 RIPD0 LSB RIPD0 RIPD1 RIPD2 RIPD3 These peak detector registers allow the user to monitor the receive in signal (Rin) peak level at reference point R1 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. RIPD4 RIPD5 RIPD6 RIPD7 Address: 17h Read Receive (Rin) Peak Detect Register 2 Power Up Reset 00h MSB RIPD8 7 RIPD9 MSB 7 RIPD 15 6 RIPD 5 14 RIPD13 4 RIPD12 3 RIPD11 2 (RIPD2) RIPD10 1 RIPD9 0 RIPD8 LSB RIPD10 RIPD11 See Above Description RIPD12 RIPD13 RIPD14 RIPD15 Address: 18h Read Power Up Reset 00h Receive (Rin) ERROR Peak Detect Register 1 MSB 7 REPD 7 6 REPD 6 5 REPD 5 4 REPD 4 3 REPD 3 2 REPD 2 (REPD1) 1 REPD1 0 REPD 0 LSB REPD0 REPD1 REPD2 REPD3 These peak detector registers allow the user to monitor the error signal peak level at reference point R2 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. REPD4 REPD5 REPD6 REPD7 Receive (Rin) ERROR Peak Detect Register 2 Address: 19h Read Power Up Reset 00h MSB 7 REPD 6 REPD14 15 REPD8 REPD9 REPD10 REPD11 REPD12 REPD13 REPD14 REPD15 20 See above description 5 REPD13 4 REPD 3 REPD11 12 2 REPD 10 (REPD2) 1 REPD 9 0 REPD8 LSB MT93L16 Preliminary Information Address: 3Ah Read Power Up Reset 00h Receive (Rout) Peak Detect Register 1 MSB 7 ROPD 7 6 ROPD 6 5 ROPD 5 4 ROPD4 3 ROPD3 2 (ROPD1) ROPD2 1 ROPD1 0 ROPD 0 LSB ROPD0 ROPD1 ROPD2 ROPD3 These peak detector registers allow the user to monitor the receive out signal (Rout) peak level at reference point R3 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. ROPD4 ROPD5 ROPD6 ROPD7 Address: 3Bh Read Power Up Reset 00h Receive (Rout) Peak Detect Register 2 MSB 7 ROPD 6 ROPD 5 ROPD13 15 14 (ROPD2) 4 ROPD 3 ROPD 2 ROPD 11 12 10 1 ROPD 9 0 ROPD8 LSB ROPD8 ROPD9 ROPD10 ROPD11 ROPD12 See Above description ROPD13 ROPD14 ROPD15 Address: 36h Read Power Up Reset 00h Send (Sin) Peak Detect Register 1 MSB 7 SIPD 7 6 SIPD 6 5 SIPD5 4 SIPD4 3 SIPD3 2 (SIPD1) SIPD2 1 SIPD1 0 SIPD0 LSB SIPD0 SIPD1 SIPD2 SIPD3 These peak detector registers allow the user to monitor the receive in signal (Sin) peak level at reference point S1 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. SIPD4 SIPD5 SIPD6 SIPD7 Send (Sin) Peak Detect Register 2 Address: 37h Read Power Up Reset 00h MSB 7 SIPD 15 6 SIPD 14 5 SIPD13 4 SIPD12 3 SIPD11 2 (SIPD2) SIPD10 1 SIPD9 0 SIPD8 LSB SIPD8 SIPD9 SIPD10 SIPD11 See above description SIPD12 SIPD13 SIPD14 SIPD15 21 MT93L16 Preliminary Information Address: 38h Read Power Up Reset 00h Send ERROR Peak Detect Register 1 MSB 7 SEPD 7 6 SEPD 6 5 SEPD5 4 SEPD4 3 SEPD3 2 (SEPD1) SEPD2 1 SEPD1 0 SEPD0 LSB SEPD0 SEPD1 SEPD2 SEPD3 These peak detector registers allow the user to monitor the error signal peak level in the send path at reference point S2 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. SEPD4 SEPD5 SEPD6 SEPD7 Address: 39h Read Power Up Reset 00h Send ERROR Peak Detect Register 2 7 SEPD 6 SEPD 15 14 MSB 5 SEPD13 4 SEPD 3 SEPD11 12 (SEPD2) 2 SEPD 10 1 SEPD 9 0 SEPD8 LSB SEPD8 SEPD9 SEPD10 SEPD11 SEPD12 See Above description SEPD13 SEPD14 SEPD15 Address: 1Ah Read Power Up Reset 00h Send (Sout) Peak Detect Register 1 MSB 7 SOPD 7 6 SOPD 6 5 SOPD 5 4 SOPD 4 3 SOPD 3 2 (SOPD1) SOPD2 1 SOPD1 0 SOPD 0 LSB SOPD0 SOPD1 SOPD2 SOPD3 These peak detector registers allow the user to monitor the Send out signal (Sout) peak level at reference point S3 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. SOPD4 SOPD5 SOPD6 SOPD7 Address: 1Bh Read Power Up Reset 00h Send (Sout) Peak Detect Register 2 MSB 7 SOPD 6 SOPD 15 14 SOPD8 SOPD9 SOPD10 SOPD11 SOPD12 SOPD13 SOPD14 SOPD15 22 See Above description 5 SOPD13 4 SOPD 3 SOPD11 12 (SOPD2) 2 SOPD 10 1 SOPD 9 0 SOPD8 LSB MT93L16 Preliminary Information Address: 3Ch R/W Power Up Reset 00h A_AS0 A_AS1 A_AS2 A_AS3 Acoustic Echo Canceller Adaptation Speed Register 1 MSB 7 A_AS 7 6 A_AS 6 5 A_AS 5 4 A_AS 4 3 A_AS 3 2 A_AS2 1 (A_AS1) A_AS1 0 A_AS 0 LSB This register allows the user to program control the adaptation speed of the Acoustic Echo Canceller. This register value changes dynamically when the ’ASC-’ bit in the Acoustic Echo Canceller Control Register is low. The ’ASC-’ bit must be 1 when this register is under user control. The valid range is from 0000h to 7FFFh. The high byte is in Register 2 and the low byte is in Register 1. Smaller values correspond to slower adaptation speed. A_AS4 A_AS5 A_AS6 A_AS7 Address: 3Dh R/W Power Up Reset 10h Acoustic Echo Canceller Adaptation Speed Register 2 MSB 7 A_AS 6 A_AS 15 14 5 A_AS13 4 A_AS 12 3 A_AS11 2 A_AS 10 (A_AS2) 1 A_AS 9 0 A_AS8 LSB A_AS8 A_AS9 A_AS10 A_AS11 See Above description A_AS12 A_AS13 A_AS14 A_AS15 Address: 1Ch R/W Power Up Reset 00h L_AS0 L_AS1 L_AS2 L_AS3 Line Echo Canceller Adaptation Speed Register 1 MSB 7 L_AS 7 6 L_AS 6 5 L_AS 5 4 L_AS 4 3 L_AS 3 2 L_AS2 (L_AS1) 1 L_AS1 0 L_AS 0 LSB This register allows the user to program control the adaptation speed of the Line Echo Canceller. This register value changes dynamically when the ’ASC-’ bit in the Acoustic Echo Canceller Control Register is low. The ’ASC-’ bit must be 1 when this register is under user control. The valid range is from 0000h to 7FFFh. The high byte is in Register 2 and the low byte is in Register 1. Smaller values correspond to slower adaptation speed. L_AS4 L_AS5 L_AS6 L_AS7 Address: 1Dh Read Power Up Reset 08h Line Echo Canceller Adaptation Speed Register 2 MSB 7 L_AS 6 L_AS 15 14 5 L_AS13 4 L_AS 12 3 L_AS11 2 L_AS 10 (L_AS2) 1 L_AS 9 0 L_AS8 LSB L_AS8 L_AS9 L_AS10 L_AS11 See Above description L_AS12 L_AS13 L_AS14 L_AS15 23 MT93L16 Preliminary Information Address: 24h R/W Power Up Reset 80h Rout Limiter Register 1 (RL1) 7 MSB L0 6 - 5 - 4 - 3 - 2 - 1 - 0 LSB - RESERVED L0 This bit is used in conjunction with Rout Limiter Register 2. (See description below.) Address: 25h R/W Power Up Reset 3Eh Rout Limiter Register 2 (RL2) 7 MSB L8 6 L7 5 L6 4 L5 3 L4 2 L3 1 L2 0 L1 LSB L1 L2 L3 L4 L5 L6 In conjunction with bit 7 (L0) of the above (RL1) register, this register (RL2) allows the user to program the output Limiter threshold value in the Rout path. Default value is (1f40)h which is equal to 3.14dBmo Maximum value is (7FC0 )h = 15 dBmo Minimum value is (0040)h = -38 dBmo L7 L8 Address: 26h R/W Power Up Reset 3Dh Sout Limiter Register (SL) 7 MSB L4 6 L3 5 L2 4 L1 3 L0 2 - 1 - RESERVED L0 L1 L2 L3 L4 24 This register allows the user to program the output Limiter threshold value in the Rout path Default value is (1f40)h which is equal to 3.14dBmo Maximum value is (7F40 )h - 0 LSB MT93L16 Preliminary Information Address: 03h Read Power Up Reset 00h Firmware Revision Code Register (FRC) 7 MSB FRC2 6 FRC1 5 FRC0 4 - 3 - 2 - 1 - 0 - 0 - LSB - RESERVED - FRC0 FRC1 Revision code of the firmware program currently being run (default=rom=00). FRC2 Address: 3fh R / W Power Up Reset 00h Bootload RAM Control Register (BRC) 7 MSB - 6 - 5 - 4 - 3 RAM_ROMb 2 BOOT 1 C0 RESERVED. Must be set to zero. C1 RESERVED. Must be set to zero. C2 BOOT bit. When high, puts device in bootload mode. When low, bootload is disabled. C3 RAM_ROMb bit. When high, device executes from RAM. When low, device executes from ROM. - LSB - RESERVED - Address: 07h Read Power Up Reset FFh Bootload RAM Signature Register (SIG) MSB 7 SIG 7 6 SIG 6 5 SIG5 4 SIG 4 3 SIG3 2 SIG 2 1 SIG 1 0 SIG0 LSB SIG7 SIG6 SIG5 SIG4 This register provides the signature of the bootloaded data to verify error-free delivery into the device. Note: this register is only accessible if BOOT bit is high (bootload mode enabled) in the above BRC register. While bootload is disabled, the register value is held constant at its reset seed value of FFh. SIG3 SIG2 SIG1 SIG0 25 Package Outlines D e ZD R E H A A1 Pin #1 B 70 ±0.20 0.51 x 45° ±0.10 7° 0.63 ±.004 (.025) (.014) GAGE PLANE 0.335 (.020) ±.008 C L DETAIL - A Q Notes: A 1. Lead Coplanarity should be 0 to 0.10mm (.004") max 2. Package surface finishing (2.1) Top Matte: (Charmilles #18-30) (2.2) All Sides: (Charmilles #18-30) (2.3) Bottom Matte: (Charmilles #18-30) 3. All dimensions excluding mold flashes 4. Max. deviation of center of package and center of leadrame to be 0.10mm (.004") 5. Max. misalignment between top and bottom center of package to 0.10mm (.004") 6. End flash from the package body shall not exceed 0.152 (.006") per side (D) 7. Dimension B shall not include dambar protrusion/intrusion and solder coverage. 8. Not to scale 9. Dimension in inches 10.Dimensions in (millimeters) QSOP - Quad Shrink Outline Package 36-Pin Dim 36-Pin Dim Min Max Min Max A .096 (2.44) .104 (2.64) e .0315 inches (ref) 0.80mm A1 .004 (0.10) .012 (0.30) H .398 (10.11) .414 (10.51) B .011 (0.28) .020 (0.51) L 0.16 (0.40) .050 (1.27) C .0091 (0.23) .0125 (0.32) Q 0° 8° D .598 (15.20) .606 (15.40) R .025 (0.63) .035 (0.89) E .291 (7.40) .299 (7.60) ZD .0335 inches (ref) 0.85 http://www.mitelsemi.com World Headquarters - Canada Tel: +1 (613) 592 2122 Fax: +1 (613) 592 6909 North America Tel: +1 (770) 486 0194 Fax: +1 (770) 631 8213 Asia/Pacific Tel: +65 333 6193 Fax: +65 333 6192 Europe, Middle East, and Africa (EMEA) Tel: +44 (0) 1793 518528 Fax: +44 (0) 1793 518581 Information relating to products and services furnished herein by Mitel Corporation or its subsidiaries (collectively “Mitel”) is believed to be reliable. However, Mitel assumes no liability for errors that may appear in this publication, or for liability otherwise arising from the application or use of any such information, product or service or for any infringement of patents or other intellectual property rights owned by third parties which may result from such application or use. Neither the supply of such information or purchase of product or service conveys any license, either express or implied, under patents or other intellectual property rights owned by Mitel or licensed from third parties by Mitel, whatsoever. Purchasers of products are also hereby notified that the use of product in certain ways or in combination with Mitel, or non-Mitel furnished goods or services may infringe patents or other intellectual property rights owned by Mitel. This publication is issued to provide information only and (unless agreed by Mitel in writing) may not be used, applied or reproduced for any purpose nor form part of any order or contract nor to be regarded as a representation relating to the products or services concerned. The products, their specifications, services and other information appearing in this publication are subject to change by Mitel without notice. No warranty or guarantee express or implied is made regarding the capability, performance or suitability of any product or service. Information concerning possible methods of use is provided as a guide only and does not constitute any guarantee that such methods of use will be satisfactory in a specific piece of equipment. It is the user’s responsibility to fully determine the performance and suitability of any equipment using such information and to ensure that any publication or data used is up to date and has not been superseded. Manufacturing does not necessarily include testing of all functions or parameters. These products are not suitable for use in any medical products whose failure to perform may result in significant injury or death to the user. All products and materials are sold and services provided subject to Mitel’s conditions of sale which are available on request. M Mitel (design) and ST-BUS are registered trademarks of MITEL Corporation Mitel Semiconductor is an ISO 9001 Registered Company Copyright 1999 MITEL Corporation All Rights Reserved Printed in CANADA TECHNICAL DOCUMENTATION - NOT FOR RESALE