CYP15G0201DXB Dual-channel HOTLink II™ Transceiver Dual-channel HOTLink IITM Transceiver Features ■ Second-generation HOTLink® technology ❐ Compliant to multiple standards • ESCON, DVB-ASI, Fibre Channel and Gigabit Ethernet (IEEE802.3z) • CPRI™ compliant • 8- / 10-B encoded or 10-bit unencoded data ■ Dual-channel transceiver operates from 195 to 1500-MBaud serial data rate ❐ Aggregate throughput of 6-GBits per second ■ Selectable parity check/generate ■ Selectable dual-channel bonding option ❐ One 16-bit channels ■ Skew alignment support for multiple bytes of offset ■ Selectable I/O clocking options ■ MultiFrame™ receive framer ❐ Bit and byte alignment ❐ Comma or full K28.5 detect ❐ Single- or multi-byte framer for byte alignment ❐ Low-latency option ■ Synchronous LVTTL parallel interface ■ Internal phase-locked loops (PLLs) with no external PLL components ■ Optional phase-align buffer in transmit path ■ Optional elasticity buffer in receive path ■ Dual differential positive ECL (PECL) compatible serial inputs per channel ❐ Internal DC-restoration ■ Dual differential PECL-compatible serial outputs per channel ❐ Source matched for 50 W transmission lines ❐ No external bias resistors required ❐ Signaling-rate controlled edge-rates ■ Compatible with ❐ Fiber optic modules ❐ Copper cables ❐ Circuit board traces ■ JTAG boundary scan ■ Built-in self-test (BIST) for at-speed link testing ■ Per-channel link quality indicator ❐ Analog signal detect ❐ Digital signal detect ■ Low power 1.8 W at 3.3 V typical ■ Single 3.3 V supply ■ 196-ball BGA ■ Pb-free package ■ 0.25 m BiCMOS technology Functional Description The CYP15G0201DXB dual-channel HOTLink II™ transceiver is a point-to-point or point-to-multipoint communications building block allowing the transfer of data over high-speed serial links (optical fiber, balanced, and unbalanced copper transmission lines) at signaling speeds ranging from 195- to 1500-MBaud per serial link. 10 10 Serial Links Serial Links 10 10 10 10 System Host 10 CYP15G0201DXB System Host 10 CYP15G0201DXB Figure 1. HOTLink II™ System Connections Backplane or Cabled Connections Cypress Semiconductor Corporation Document Number: 38-02058 Rev. *M • 198 Champion Court • San Jose, CA 95134-1709 • 408-943-2600 Revised March 16, 2015 CYP15G0201DXB The CYP15G0201DXB operates from 195- to 1500-MBaud. The two channels may be combined to allow transport of wide buses across significant distances with minimal concern for offsets in clock phase or link delay. Each transmit channel accepts parallel characters in an input register, encodes each character for transport, and converts it to serial data. Each receive channel accepts serial data and converts it to parallel data, decodes the data into characters, and presents these characters to an Output Register. Figure 1 on page 1 illustrates typical connections between independent host systems and corresponding CYP15G0201DXB parts. As a second-generation HOTLink device, the CYP15G0201DXB extends the HOTLink family with enhanced levels of integration and faster data rates, while maintaining serial-link compatibility (data, command, and BIST) with other HOTLink devices. The transmit (TX) section of the CYP15G0201DXB Dual HOTLink II consists of two byte-wide channels that can be operated independently or bonded to form wider buses. Each channel can accept either 8-bit data characters or pre-encoded 10-bit transmission characters. Data characters are passed from the Transmit Input Register to an embedded 8- /10-B Encoder to improve their serial transmission characteristics. These encoded characters are then serialized and output from dual Positive ECL (PECL) compatible differential transmission-line drivers at a bit-rate of either 10 or 20 times the input reference clock. The receive (RX) section of the CYP15G0201DXB Dual HOTLink II consists of two byte-wide channels that can be operated independently or synchronously bonded for greater bandwidth. Each channel accepts a serial bit-stream from one of two PECL-compatible differential line receivers and, using a completely integrated PLL clock synchronizer, recovers the timing information necessary for data reconstruction. Each recovered bit-stream is deserialized and framed into characters, 8- /10-B decoded, and checked for transmission errors. Recovered decoded characters are then written to an internal Elasticity Buffer, and presented to the destination host system. The integrated 8- /10-B encoder/decoder may be bypassed for systems that present externally encoded or scrambled data at the parallel interface. For those systems using buses wider than a single byte, the two independent receive paths can be bonded together to allow synchronous delivery of data across a two-byte-wide (16-bit) path. The parallel I/O interface may be configured for numerous forms of clocking to provide the highest flexibility in system architecture. In addition to clocking the transmit path interfaces from one of multiple sources, the receive interface may be configured to present data relative to a recovered clock or to a local reference clock. Each transmit and receive channel contains independent BIST pattern generators and checkers. This BIST hardware allows at-speed testing of the high-speed serial data paths in each transmit and receive section, and across the interconnecting links. HOTLink II devices are ideal for a variety of applications where parallel interfaces can be replaced with high-speed, point-to-point serial links. Some applications include interconnecting backplanes on switches, routers, base-stations, servers and video transmission systems. TXDA[7:0] TXCTA[1:0] RXDA[7:0] RXSTA[2:0] TXDB[7:0] TXCTB[1:0] RXDB[7:0] RXSTB[2:0] Transceiver Logic Block Diagram x10 x11 x10 x11 Phase Align Buffer Elasticity Buffer Phase Align Buffer Elasticity Buffer Encoder 8- / 10-B Decoder 8- / 10-B Encoder 8- / 10-B Decoder 8- / 10-B Document Number: 38-02058 Rev. *M Deserializer TX RX TX RX INB1± INB2± Serializer OUTB1± Deserializer OUTB2± Serializer INA1± INA2± Framer OUTA1± OUTA2± Framer Page 2 of 53 CYP15G0201DXB Contents Pin Configuration ............................................................. 4 Pin Configuration ............................................................. 5 Pin Descriptions ............................................................... 6 CYP15G0201DXB HOTLink II Operation ....................... 13 CYP15G0201DXB Transmit Data Path ..................... 13 Transmit Modes ......................................................... 15 Transmit BIST ........................................................... 18 Serial Output Drivers ................................................. 18 Transmit PLL Clock Multiplier .................................... 19 CYP15G0201DXB Receive Data Path ............................ 19 Serial Line Receivers ................................................ 19 Signal Detect/Link Fault ............................................ 20 Clock/Data Recovery ................................................. 21 Deserializer/Framer ................................................... 21 8- / 10-B Decoder Block ............................................ 22 Receive BIST Operation ............................................ 22 Receive Elasticity Buffer ............................................ 23 Receive Modes .......................................................... 23 Power Control ............................................................ 24 Output Bus ................................................................ 25 Parity Generation ...................................................... 25 Receive Synchronization State Machine When Channel Bonding is Enabled ............................................ 26 JTAG Support ............................................................ 31 Maximum Ratings ........................................................... 32 Operating Range ............................................................. 32 CYP15G0201DXB DC Electrical Characteristics .......... 32 Document Number: 38-02058 Rev. *M AC Test Loads and Waveforms ..................................... 33 CYP15G0201DXB AC Characteristics ........................... 34 Switching Waveforms .................................................... 36 Switching Waveforms .................................................... 37 X3.230 Codes and Notation Conventions .................... 41 Notation Conventions ................................................ 41 8- /10-B Transmission Code ...................................... 41 Transmission Order ................................................... 41 Valid and Invalid Transmission Characters ............... 41 Use of the Tables for Generating Transmission Characters .......................... 42 Use of the Tables for Checking the Validity of Received Transmission Characters ........... 42 Ordering Information ...................................................... 48 Ordering Code Definitions ......................................... 48 Package Diagram ............................................................ 49 Acronyms ........................................................................ 50 Document Conventions ................................................. 50 Units of Measure ....................................................... 50 Document History Page ................................................. 51 Sales, Solutions, and Legal Information ...................... 53 Worldwide Sales and Design Support ....................... 53 Products .................................................................... 53 PSoC® Solutions ...................................................... 53 Cypress Developer Community ................................. 53 Technical Support ..................................................... 53 Page 3 of 53 CYP15G0201DXB Pin Configuration Top View [1] A B C D E F G H J K L M N P 1 2 3 4 5 6 7 8 9 10 11 12 13 14 VCC INA2+ OUTA2– VCC INA1+ OUTA1– VCC VCC INB2+ OUTB2– VCC INB1+ OUTB1– VCC TDO INA2– OUTA2+ VCC INA1– OUTA1+ NC NC INB2– OUTB2+ VCC INB1– OUTB1+ BOE[3] NC RFEN VCC LPEN RXLE RXRATE GND GND SPDSEL PARCTL RFMODE VCC SDASEL BOE[2] VCC VCC NC TXRATE RXMODE RXMODE [1] [0] GND GND TCLK TDI INSELB INSELA VCC VCC BISTLE FRAM CHAR TXMODE TXMODE [1] [0] BOE[1] GND GND TXOPB TXPERB TXCKSEL RXCKSEL TRSTZ TMS DECMODE OELE RXCLKC+ RXSTA[2] RXSTA[1] GND GND GND GND TXDB[4] TXDB[3] TXDB[2] TXDB[1] TXDB[0] VCC NC GND GND GND GND GND GND GND GND GND GND NC VCC VCC NC GND GND GND GND GND GND GND GND GND GND NC VCC RXSTA[0] RXOPA RXDA[0] RXDA[1] RXDA[2] GND GND GND GND TXDB[7] TXDB[6] TXDB[5] RXDA[3] RXDA[4] RXDA[5] RXDA[6] TXDA[4] TXCLKA GND GND NC LFIB TXCLKB VCC VCC RXDA[7] LFIA TXDA[3] TXOPA GND GND SCSEL RXDB[7] VCC VCC RXCLKA- TXCTA[1] VCC NC TXDA[2] TXPERA GND GND TXRST RXSTB[0] VCC RXDB[5] RXDB[6] RXCLKA+ TXCTA[0] TXDA[6] VCC TXDA[1] NC NC NC REFCLK- TXCLKO+ VCC RXDB[2] RXDB[3] RXDB[4] TXDA[5] VCC TXDA[0] NC VCC VCC REFCLK+ TXCLKO- VCC RXDB[1] RXDB[0] VCC VCC TXDA[7] BOE[0] TXCTB[0] TXCTB[1] RXOPB RXCLKB+ RXCLKB- RXSTB[2] RXSTB[1] NC Note 1. NC = Do not connect. Document Number: 38-02058 Rev. *M Page 4 of 53 CYP15G0201DXB Pin Configuration Bottom View[1] 14 13 12 11 10 9 8 7 6 5 4 3 2 1 VCC OUTB1– INB1+ VCC OUTB2– INB2+ VCC VCC OUTA1– INA1+ VCC OUTA2– INA2+ VCC BOE[3] OUTB1+ INB1– VCC OUTB2+ INB2– NC NC OUTA1+ INA1– VCC OUTA2+ INA2– TDO BOE[2] SDASEL VCC SPDSEL GND GND RXRATE RXLE LPEN VCC RFEN NC VCC VCC INSELA TCLK GND GND RXMODE[0] RXMODE[1] TXRATE NC VCC VCC TMS TRSTZ TXOPB GND GND BOE[1] TXDB[0] TXDB[1] TXDB[2] TXDB[3] TXDB[4] GND GND GND GND VCC NC GND GND GND GND GND GND GND GND GND VCC NC GND GND GND GND GND GND GND GND TXDB[5] TXDB[6] GND GND GND GND TXCLKB LFIB RXCLKB- RXCLKB+ RXOPB NC GND GND VCC VCC RXDB[7] RXSTB[1] RXSTB[2] SCSEL GND RXDB[6] RXDB[5] VCC RXSTB[0] TXRST RXDB[4] RXDB[3] RXDB[2] VCC VCC RXDB[0] RXDB[1] VCC RFMODE PARCTL INSELB TDI RXCKSEL TXCKSEL TXPERB BOE[0] TXMODE[0] TXMODE[1] FRAMCHAR BISTLE OELE DECMODE GND NC VCC GND GND NC VCC RXDA[2] RXDA[1] RXDA[0] RXOPA RXSTA[0] TXCLKA TXDA[4] RXDA[6] RXDA[5] RXDA[4] RXDA[3] GND TXOPA TXDA[3] LFIA RXDA[7] VCC VCC GND GND TXPERA TXDA[2] NC VCC TXCLKO+ REFCLK- NC NC NC TXDA[1] VCC TXDA[6] TXCTA[0] RXCLKA+ TXCLKO- REFCLK+ VCC VCC NC TXDA[0] VCC TXDA[5] TXDB[7] TXCTB[1] TXCTB[0] NC Document Number: 38-02058 Rev. *M RXSTA[1] RXSTA[2] RXCLKC+ TXCTA[1] RXCLKA- TXDA[7] VCC A B C D E F G H J K L M N P Page 5 of 53 CYP15G0201DXB Pin Descriptions CYP15G0201DXB Dual HOTLink II Transceiver Pin Name I/O Characteristics Signal Description Transmit Path Data Signals TXPERA TXPERB LVTTL Output, changes relative to REFCLK [2] Transmit Path Parity Error. Active HIGH. Asserted (HIGH) if parity checking is enabled and a parity error is detected at the Encoder. This output is HIGH for one transmit character clock period to indicate detection of a parity error in the character presented to the Encoder. If a parity error is detected, the character in error is replaced with a C0.7 character to force a corresponding bad-character detection at the remote end of the link. This replacement takes place regardless of the encoded/non-encoded state of the interface. When BIST is enabled for the specific transmit channel, BIST progress is presented on these outputs. Once every 511 character times (plus a 16-character Word Sync Sequence when the receive channels are clocked by a common clock, that is, RXCKSEL = LOW or HIGH), the associated TXPERx signal pulses HIGH for one transmit-character clock period (if RXCKSEL = MID) or seventeen transmit- character clock periods (if RXCKSEL = LOW or HIGH) to indicate a complete pass through the BIST sequence. For RXCKSEL = LOW or HIGH, if TXMODE[1:0] = LL, then no Word Sync Sequence is sent in BIST, and TXPERx pulses HIGH for one transmit-character clock period. These outputs also provide indication of a transmit Phase-Align Buffer underflow or overflow. When the transmit Phase-Align Buffers are enabled (TXCKSEL LOW, or TXCKSEL = LOW and TXRATE = HIGH), if an underflow or overflow condition is detected, TXPERx for the channel in error is asserted and remains asserted until either an atomic word sync sequence is transmitted or TXRST is sampled LOW to re-center the transmit phase-align buffers. TXCTA[1:0] TXCTB[1:0] LVTTL Input, synchronous, sampled by the selected TXCLKx or REFCLK [2] Transmit Control. These inputs are captured on the rising edge of the transmit interface clock as selected by TXCKSEL, and are passed to the Encoder or Transmit Shifter. They identify how the associated TXDx[7:0] characters are interpreted. When the Encoder is bypassed, these inputs are interpreted as data bits. When the Encoder is enabled, these inputs determine if the TXDx[7:0] character is encoded as Data, a Special Character code, or replaced with other Special Character codes. See Table 1 on page 13 for details. TXDA[7:0] TXDB[7:0] LVTTL Input, synchronous, sampled by the selected TXCLKx or REFCLK [2] Transmit Data Inputs. These inputs are captured on the rising edge of the transmit interface clock (selected by TXCKSEL) and passed to the Encoder or Transmit Shifter. LVTTL Input, asynchronous, internal pull-up, REFCLK [2] Transmit Clock Phase Reset. Transmit Clock Phase Reset. Active LOW. When sampled LOW, the transmit Phase-align Buffers are allowed to adjust their data-transfer timing (relative to the selected input clock) to allow clean transfer of data from the Input Register to the Encoder or Transmit Shifter. When TXRST is sampled HIGH, the internal phase relationship between the associated TXCLKx and the internal character-rate clock is fixed and the device operates normally. TXRST When the Encoder is enabled (TXMODE[1:0] LL), TXDx[7:0] specify the specific data or command character to be sent. When the Encoder is bypassed, these inputs are interpreted as data bits of the 10-bit input character. See Table 1 on page 13 for details. When configured for half-rate REFCLK sampling of the transmit character stream (TXCKSEL = LOW and TXRATE = HIGH), assertion of TXRST is only used to clear Phase-align buffer faults caused by highly asymmetric REFCLK periods or REFCLKs with excessive cycle-to-cycle jitter. During this alignment period, one or more characters may be added to or lost from all the associated transmit paths as the transmit Phase-align Buffers are adjusted. TXRST must be sampled LOW by a minimum of two consecutive rising edges of REFCLK to ensure the reset operation is initiated correctly on all channels. This input is ignored when both TXCKSEL and TXRATE are LOW, since the phase align buffer is bypassed. In all other configurations, TXRST should be asserted during device initialization to ensure proper operation of the Phase-align buffer. TXRST should be asserted after presence of a valid TXCLKx and after allowing enough time for the TXPLL to lock to the reference clock (as specified by parameter tTXLOCK). Note 2. When REFCLK is configured for half-rate operation (TXRATE = HIGH), these inputs are sampled (or the outputs change) relative to both the rising and falling edges of REFCLK. Document Number: 38-02058 Rev. *M Page 6 of 53 CYP15G0201DXB Pin Descriptions (continued) CYP15G0201DXB Dual HOTLink II Transceiver Pin Name I/O Characteristics Signal Description SCSEL LVTTL Input, synchronous, internal pull-down, sampled by TXCLKA or REFCLK [2] Special Character Select. Used in some transmit modes along with TXCTx[1:0] to encode special characters or to initiate a Word Sync Sequence. When the transmit paths are configured for independent inputs clocks (TXCKSEL = MID), SCSEL is captured relative to TXCLKA. TXOPA TXOPB LVTTL Input, Transmit Path Odd Parity. When parity checking is enabled (PARCTL LOW), the parity synchronous, captured at these inputs is XORed with the data on the associated transmit data TXDx bus to internal pull-up, verify the integrity of the captured character. sampled by the respective TXCLKx or REFCLK [2] Transmit Path Clock and Clock Control TXCKSEL 3-Level Select[3] Static Control Input Transmit Clock Select. Selects the clock source, used to write data into the Transmit Input Register, of the transmit channel(s). When LOW, both Input Registers are clocked by REFCLK [2]. When MID, TXCLKx is used as the Input Register clock for TXDx[7:0] and TXCTx[1:0]. When HIGH, TXCLKA is used to clock data into the Input Register of each channel. When TXCKSEL = MID or HIGH (TXCLKx or TXCLKA selected to clock input register), configuring TXRATE = HIGH (Half-rate REFCLK) is an invalid mode of operation. TXRATE LVTTL Input, Transmit PLL Clock Rate Select. When TXRATE = HIGH, the Transmit PLL multiplies Static Control input, REFCLK by 20 to generate the serial symbol-rate clock. When TXRATE = LOW, the transmit internal pull-down PLL multiples REFCLK by 10 to generate the serial symbol-rate clock. See Table 10 on page 19 for a list of operating serial rates. When REFCLK is selected to clock the receive parallel interfaces (RXCKSEL = LOW), the TXRATE input also determines if the clocks on the RXCLKA and RXCLKC outputs are full or half-rate. When TXRATE = HIGH (REFCLK is half-rate), the RXCLKA± and RXCLKC± output clocks are also half-rate clocks and follow the frequency and duty cycle of the REFCLK input. When TXRATE = LOW (REFCLK is full-rate), the RXCLKA± and RXCLKC± output clocks are full-rate clocks and follow the frequency and duty cycle of the REFCLK input. When TXCKSEL = MID or HIGH (TXCLKx or TXCLKA selected to clock input register), configuring TXRATE = HIGH (Half-rate REFCLK) is an invalid mode of operation. TXCLKO± LVTTL Output Transmit Clock Output. This true and complement output clock is synthesized by the transmit PLL and operates synchronous to the internal transmit character clock. It operates at either the same frequency as REFCLK (when TXRATE = LOW), or at twice the frequency of REFCLK (when TXRATE = HIGH). This output clock has no direct phase relationship to REFCLK. TXCLKA TXCLKB LVTTL Clock Input, internal pull-down Transmit Path Input Clocks. These clocks must be frequency-coherent to TXCLKO±, but may be offset in phase. The internal operating phase of each input clock (relative to REFLCK or TXCLKO±) is adjusted when TXRST = LOW and locked when TXRST = HIGH. Transmit Path Mode Control TXMODE[1:0] 3-Level Select[3] Transmit Operating Mode. These inputs are interpreted to select one of nine operating modes Static Control inputs of the transmit path. See Table 3 on page 15 for a list of operating modes. Note 3. 3-Level select inputs are used for static configuration. They are ternary (not binary) inputs that make use of non-standard logic levels of LOW, MID, and HIGH. The LOW level is usually implemented by direct connection to VSS (ground). The HIGH level is usually implemented by direct connection to VCC (power). When not connected or allowed to float, a 3-Level select input self-biases to the MID level. Document Number: 38-02058 Rev. *M Page 7 of 53 CYP15G0201DXB Pin Descriptions (continued) CYP15G0201DXB Dual HOTLink II Transceiver Pin Name I/O Characteristics Signal Description Receive Path Data Signals RXDA[7:0] RXDB[7:0] RXSTA[2:0] RXSTB[2:0] LVTTL Output, synchronous to the selected RXCLKx output or REFCLK [2] input LVTTL Output, synchronous to the selected RXCLKx output or REFCLK [2]input Parallel Data Output. These outputs change following the rising edge of the selected receive interface clock. When the Decoder is enabled (DECMODE = HIGH or MID), these outputs represent either received data or special characters. The status of the received data is represented by the values of RXSTx[2:0]. When the Decoder is bypassed (DECMODE = LOW), RXDx[7:0] become the higher order bits of the 10-bit received character. See Table 16 on page 25 for details. Parallel Status Output. These outputs change following the rising edge of the selected receive interface clock. When the Decoder is bypassed (DECMODE = LOW), RXSTx[1:0] become the two low-order bits of the 10-bit received character, while RXSTx[2] = HIGH indicates the presence of a Comma character in the Output Register. See Table 16 on page 25 for details. When the Decoder is enabled (DECMODE = HIGH or MID), RXSTx[2:0] provide status of the received signal. See Table 18 on page 27, Table 19 and Table 20 on page 30 for a list of Receive Character status. RXOPA RXOPB 3-state, LVTTL Receive Path Odd Parity. When parity generation is enabled (PARCTL LOW), the parity Output, synchronous output at these pins is valid for the data on the associated RXDx bus bits. When parity generto the selected ation is disabled (PARCTL = LOW) these output drivers are disabled (High-Z). RXCLKx output or REFCLK [2] input Receive Path Clock and Clock Control RXRATE LVTTL Input Receive Clock Rate Select. When LOW, the RXCLKx recovered clock outputs are Static Control Input, complementary clocks operating at the recovered character rate. Data for the associated internal pull-down receive channels should be latched on the rising edge of RXCLKx+ or falling edge of RXCLKx–. When HIGH, the RXCLKx recovered clock outputs are complementary clocks operating at half the character rate. Data for the associated receive channels should be latched alternately on the rising edge of RXCLKx+ and RXCLKx–. When REFCLK± is selected to clock the output registers (RXCKSELx = LOW), RXRATEx is not interpreted. The RXCLKA± and RXCLKC± output clocks follow the frequency and duty cycle of REFCLK±. RXCLKA RXCLKB Three-state, LVTTL Output clock or Static control input Receive Character Clock Output or Clock Select Input. When configured such that all output data paths are clocked by the recovered clock (RXCKSEL = MID), these true and complement clocks are the receive interface clocks which are used to control timing of output data (RXDx[7:0], RXSTx[2:0] and RXOPx). These clocks are output continuously at either the dual-character rate (1/20th the serial symbol-rate) or character rate (1/10th the serial symbol-rate) of the data being received, as selected by RXRATE. When configured such that all output data paths are clocked by REFCLK instead of a recovered clock (RXCKSEL = LOW), the RXCLKA and RXCLKC output drivers present a buffered and delayed form of REFCLK. RXCLKA and RXCLKC are buffered forms of REFCLK that are slightly different in phase. This phase difference allows the user to select the optimal setup/hold timing for their specific interface. When RXCKSEL = HIGH and dual-channel bonding is enabled, one of the recovered clocks from channels A or B is selected to present bonded data from channels A and B. RXCLKA output the recovered clock from either receive channel A or receive channel B as selected by RXCLKB+ to clock the bonded output data from channels A and B. See Table 14 on page 24 for details. When RXCKSEL = LOW and dual-channel bonding is enabled, REFCLK is selected to present bonded data from channels A and B. RXCLKA and RXCLKC output drivers present a buffered and delayed form of REFCLK. The master channel for bonding is selected by RXCLKB+ (which acts as an input in this mode) to clock the bonded output data from channels A and B. See Table 14 on page 24 for details. Document Number: 38-02058 Rev. *M Page 8 of 53 CYP15G0201DXB Pin Descriptions (continued) CYP15G0201DXB Dual HOTLink II Transceiver Pin Name RXCKSEL I/O Characteristics 3-Level Select[3] Static Control Input Signal Description Receive Clock Mode. Selects the receive clock-source used to transfer data to the output registers. When LOW, both Output Registers are clocked by REFCLK. RXCLKBoutputs are disabled (High-Z), and RXCLKAand RXCLKC+ present buffered and delayed forms of REFCLK. When MID, each RXCLKxoutput follows the recovered clock for the respective channel, as selected by RXRATE. When the 8- / 10-B Decoder and Elasticity Buffer are bypassed (DECMODE = LOW), RXCKSEL must be MID. When HIGH, and channel bonding is enabled in dual-channel mode (RX modes 2 and 3), RXCLKA outputs the recovered clock from either receive channel A or receive channel B as selected by RXCLKB+. These output clocks may operate at the character-rate or half the character-rate as selected by RXRATE. DECMODE 3-Level Select[3] Static Control Input Decoder Mode Select. This input selects the behavior of the Decoder block. When LOW, the Decoder is bypassed and raw 10-bit characters are passed to the Output Register. When the Decoder is bypassed, RXCKSEL must be MID. When MID, the Decoder is enabled and the Cypress decoder table for Special Code characters is used. When HIGH, the Decoder is enabled and the alternate Decoder table for Special Code characters is used. See Table 25 on page 47 for a list of the Special Codes supported in both encoded modes. RXMODE[1:0] 3-Level Select[3] Receive Operating Mode. These inputs are interpreted to select one of nine operating modes Static Control Inputs of the receive path. See Table 13 on page 23 for details. RFEN LVTTL input, asynchronous, internal pull-down Reframe Enable for All Channels. Active HIGH. When HIGH, the framers in both channels are enabled to frame per the presently enabled framing mode and selected framing character. RFMODE 3-Level Select[3] Static Control Input Reframe Mode Select. Used to control the type of character framing used to adjust the character boundaries (based on detection of one or more framing characters in the received serial bit stream). This signal operates in conjunction with the presently enabled channel bonding mode, and the type of framing character selected. When LOW, the low-latency framer is selected. This frames on each occurrence of the selected framing character(s) in the received data stream. This mode of framing stretches the recovered clock for one or multiple cycles to align that clock with the recovered data. When MID, the Cypress-mode multi-byte parallel framer is selected. This requires a pair of the selected framing character(s), on identical 10-bit boundaries, within a span of 50 bits, before the character boundaries are adjusted. The recovered character clock remains in the same phasing regardless of character offset. When HIGH, the alternate mode multi-byte parallel framer is selected. This requires detection of the selected framing character(s) of the allowed disparities in the received serial bit stream, on identical 10-bit boundaries, on four directly adjacent characters. The recovered character clock remains in the same phasing regardless of character offset. FRAMCHAR 3-Level Select[3] Static Control Input Framing Character Select. Used to control the character or portion of a character used for character framing of the received data streams. When MID, the framer looks for both positive and negative disparity versions of the 8-bit Comma character. When HIGH, the framer looks for both positive and negative disparity versions of the K28.5 character. Configuring FRAMCHAR to LOW is reserved for component test. Document Number: 38-02058 Rev. *M Page 9 of 53 CYP15G0201DXB Pin Descriptions (continued) CYP15G0201DXB Dual HOTLink II Transceiver Pin Name I/O Characteristics Signal Description Device Control Signals PARCTL 3-Level Select[3] Static Control Input Parity Check/Generate Control. Used to control the different parity check and generate functions. When LOW, parity checking is disabled, and the RXOPx outputs are all disabled (High-Z). When MID, and the Encoder/Decoder are enabled (TXMODE[1] LOW, DECMODE LOW), TXDx[7:0] inputs are checked (along with TXOPx) for valid ODD parity, and ODD parity is generated for the RXDx[7:0] outputs and presented on RXOPx. When the Encoder and Decoder are disabled (TXMODE[1] LOW, DECMODE LOW), the TXDx[7:0] and TXCTx[1:0] inputs are checked (along with TXOPx) for valid ODD parity, and ODD parity is generated for the RXDx[7:0] and RXSTx[1:0] outputs and presented on RXOPx. When HIGH, parity checking and generation are enabled. The TXDx[7:0] and TXCTx[1:0] inputs are checked (along with TXOPx) for valid ODD parity, and ODD parity is generated for the RXDx[7:0] and RXSTx[2:0] outputs and presented on RXOPx. REFCLK Differential LVPECL Reference Clock. This clock input is used as the timing reference for the transmit and receive or single-ended PLLs. This input clock may also be selected to clock the transmit and receive parallel interfaces. LVTTL input clock When driven by a single-ended LVCMOS or LVTTL clock source, connect the clock source to either the true or complement REFCLK input, and leave the alternate REFCLK input open (floating). When driven by an LVPECL clock source, the clock must be a differential clock, using both inputs. When TXCKSEL = LOW, REFCLK is also used as the clock for the parallel transmit data (input) interface. When RXCKSEL = LOW, the Elasticity Buffer is enabled and REFCLK is used as the clock for the parallel receive data (output) interface. If the Elasticity Buffer is used, then the framing characters are inserted or deleted to/from the data stream to compensate frequency differences between the reference clock and recovered clock. When addition happens, a K28.5 is appended immediately after a framing character is detected in the elasticity buffer. When deletion happens, a framing character is removed from the datastream when detected in the elasticity buffer. RXCLKC+ 3-state LVTTL Output Delayed REFCLK+ when RXCKSEL=LOW. Delayed form of REFCLK+, used for transfer of recovered data to a host system. This output is only enabled when the receive parallel interface is configured to present data relative to REFCLK (RXCKSEL = LOW). SPDSEL 3-Level Select[3], static control input Serial Rate Select. This input specifies the operating bit-rate range of both transmit and receive PLLs. LOW = 195 to 400 MBaud, MID = 400 to 800 MBaud, HIGH = 800 to 1500 MBaud (800–1540 MBaud for CYW15G0201DXB). When SPDSEL is LOW, setting TXRATE = HIGH (Half-rate Reference Clock) is invalid. TRSTZ LVTTL Input, internal pull-up Device Reset. Active LOW. Initializes all state machines and counters in the device. When sampled LOW by the rising edge of REFLCK, this input resets the internal state machines and sets the Elasticity Buffer pointers to a nominal offset. When the reset is removed (TRSTZ sampled HIGH by REFCLK), the status and data outputs become deterministic in less than 16 REFCLK cycles. The BISTLE, OELE, and RXLE latches are reset by TRSTZ. If the Elasticity Buffer or the Phase Align Buffer are used, TRSTZ should be applied after power-up to initialize the internal pointers into these memory arrays. Analog I/O and Control OUTA1 OUTB1 CML Differential Output Primary Differential Serial Data Outputs. These PECL-compatible CML outputs (+3.3 V referenced) are capable of driving terminated transmission lines or standard fiber-optic transmitter modules. OUTA2 OUTB2 CML Differential Output Secondary Differential Serial Data Outputs. These PECL-compatible CML outputs (+3.3 V referenced) are capable of driving terminated transmission lines or standard fiber-optic transmitter modules. INA1 INB1 LVPECL Differential Primary Differential Serial Data Inputs. These inputs accept the serial data stream for deseriInput alization and decoding. The INx1serial streams are passed to the receiver Clock and Data Recovery (CDR) circuits to extract the data content when INSELx = HIGH. Document Number: 38-02058 Rev. *M Page 10 of 53 CYP15G0201DXB Pin Descriptions (continued) CYP15G0201DXB Dual HOTLink II Transceiver Pin Name I/O Characteristics Signal Description INA2 INB2 LVPECL Differential Secondary Differential Serial Data Inputs. These inputs accept the serial data stream for Input deserialization and decoding. The INx2serial streams are passed to the receiver Clock and Data Recovery (CDR) circuits to extract the data content when INSELx = LOW. INSELA INSELB LVTTL Input, asynchronous Receive Input Selector. Determines which external serial bit stream is passed to the receiver Clock and Data Recovery circuit. When HIGH, the INx1 input is selected. When LOW, the INx2 input is selected. SDASEL 3-Level Select[3], static configuration input Signal Detect Amplitude Level Select. Allows selection of one of three predefined amplitude trip points for a valid signal indication, as listed in Table 11 on page 20. LPEN LVTTL Input, asynchronous, internal pull-down All-Port Loop-Back-Enable. Active HIGH. When asserted (HIGH), the transmit serial data from each channel is internally routed to the associated receiver Clock and Data Recovery (CDR) circuit. All serial drivers are forced to differential logic “1”. All serial data inputs are ignored. OELE LVTTL Input, asynchronous, internal pull-up Serial Driver Output Enable Latch Enable. Active HIGH. When OELE = HIGH, the signals on the BOE[3:0] inputs directly control the OUTxy differential drivers. When the BOE[x] input is HIGH, the associated OUTxy differential driver is enabled. When the BOE[x] input is LOW, the associated OUTxy differential driver is powered down. When OELE returns LOW, the last values present on BOE[3:0] are captured in the internal Output Enable Latch. The specific mapping of BOE[3:0] signals to transmit output enables is listed in Table 9 on page 18. RXLE LVTTL Input, asynchronous, internal pull-up Receive Channel Power-Control Latch Enable. Active HIGH. When RXLE = HIGH, the signals on the BOE[3:0] inputs directly control the power enables for the receive PLLs and analog logic. When the BOE[3:0] input is HIGH, the associated receive channel A and receive channel B PLL and analog logic are active. When the BOE[3:0] input is LOW, the associated receive channel A and receive channel B PLL and analog logic are placed in a non-functional power saving mode. When RXLE returns LOW, the last values present on BOE[3:0] are captured in the internal RX PLL Enable Latch. The specific mapping of BOE[3:0] signals to the associated receive channel enables is listed in Table 9 on page 18. When the device is reset (TRSTZ is sampled LOW), the latch is reset to disable both receive channels. BISTLE LVTTL Input, asynchronous, internal pull-up Transmit and Receive BIST Latch Enable. Active HIGH. When BISTLE = HIGH, the signals on the BOE[3:0] inputs directly control the transmit and receive BIST enables. When the BOE[x] input is LOW, the associated transmit or receive channel is configured to generate or compare the BIST sequence. When the BOE[x] input is HIGH, the associated transmit or receive channel is configured for normal data transmission or reception. When BISTLE returns LOW, the last values present on BOE[3:0] are captured in the internal BIST Enable Latch. The specific mapping of BOE[3:0] signals to transmit and receive BIST enables is listed in Table 9 on page 18. When the latch is closed, if the device is reset (TRSTZ is sampled LOW), the latch is reset to disable BIST on all transmit and receive channels. BOE[3:0] LVTTL Input, asynchronous, internal pull-up BIST, Serial Output, and Receive Channel Enables. These inputs are passed to and through the Output Enable Latch when OELE = HIGH, and captured in this latch when OELE returns LOW. These inputs are passed to and through the BIST Enable Latch when BISTLE = HIGH, and captured in this latch when BISTLE returns LOW. These inputs are passed to and through the Receive Channel Enable Latch when RXLE = HIGH, and captured in this latch when RXLE returns LOW. LFIA LFIB LVTTL Output, Asynchronous Link Fault Indication Output. Active LOW. LFIx is the logical OR of four internal conditions: If the device is reset (TRSTZ is sampled LOW), the latch is reset to disable all outputs. 1. Received serial data frequency outside expected range. 2. Analog amplitude below expected levels. 3. Transition density lower than expected. 4. Receive Channel disabled. Document Number: 38-02058 Rev. *M Page 11 of 53 CYP15G0201DXB Pin Descriptions (continued) CYP15G0201DXB Dual HOTLink II Transceiver Pin Name I/O Characteristics Signal Description JTAG Interface TMS LVTTL Input, internal pull-up Test Mode Select. Used to control access to the JTAG Test Modes. If maintained HIGH for >5 TCLK cycles, the JTAG test controller is reset. The TAP controller is also reset automatically upon application of power to the device. TCLK LVTTL Input, internal pull-down JTAG Test Clock. TDO 3-State LVTTL Output Test Data Out. JTAG data output buffer which is High-Z while JTAG test mode is not selected. TDI LVTTL Input, internal pull-up Test Data In. JTAG data input port. Power VCC +3.3 V power. GND Signal and Power Ground for all internal circuits. Document Number: 38-02058 Rev. *M Page 12 of 53 CYP15G0201DXB CYP15G0201DXB HOTLink II Operation The CYP15G0201DXB is a highly configurable device designed to support reliable transfer of large quantities of data, using high-speed serial links, from one or multiple sources to one or multiple destinations. This device supports two single-byte or single-character channels that may be combined to support transfer of wider buses. CYP15G0201DXB Transmit Data Path Operating Modes The transmit path of the CYP15G0201DXB supports two character-wide data paths. These data paths are used in multiple operating modes as controlled by the TXMODE[1:0] inputs. Input Register The bits in the input register for each channel support different assignments, based on if the character is unencoded, encoded with two control bits, or encoded with three control bits. These assignments are shown in Table 1. Each input register captures a minimum of eight data bits and two control bits on each input clock cycle. When the Encoder is bypassed, the TXCTx[1:0] control bits are part of the pre-encoded 10-bit character. When the encoder is enabled (TXMODE[1] LOW), the TXCTx[1:0] bits are interpreted along with the associated TXDx[7:0] character to generate the specific 10-bit transmission character. When TXMODE[0] HIGH, an additional special character select (SCSEL) input is also captured and interpreted. This SCSEL input is used to modify the encoding of the associated characters. When the transmit input registers are clocked by a common clock (TXCLKA or REFCLK), this SCSEL input can be changed on a clock-by-clock basis and affects both channels. Table 1. Input Register Bit Assignments[4] Encoded Signal Name Unencoded 2-bit Control 3-bit Control TXDx[0] (LSB) DINx[0] TXDx[0] TXDx[0] TXDx[1] DINx[1] TXDx[1] TXDx[1] TXDx[2] DINx[2] TXDx[2] TXDx[2] TXDx[3] DINx[3] TXDx[3] TXDx[3] TXDx[4] DINx[4] TXDx[4] TXDx[4] TXDx[5] DINx[5] TXDx[5] TXDx[5] TXDx[6] DINx[6] TXDx[6] TXDx[6] TXDx[7] DINx[7] TXDx[7] TXDx[7] TXCTx[0] DINx[8] TXCTx[0] TXCTx[0] TXCTx[1] (MSB) DINx[9] TXCTx[1] TXCTx[1] SCSEL N/A N/A SCSEL When operated with a separate input clock on each transmit channel, this SCSEL input is sampled synchronous to TXCLKA. While the value on SCSEL still affects both channels, it is interpreted when the character containing it is read from the transmit phase-align buffer (where both paths are internally clocked synchronously). Phase-Align Buffer Data from the input registers is passed either to the encoder or to the associated phase-align buffer. When the transmit paths are operated synchronous to REFCLK (TXCKSEL = LOW and TXRATE = LOW), the phase-align buffers are bypassed and data is passed directly to the parity check and Encoder blocks to reduce latency. When an Input-Register clock with an uncontrolled phase relationship to REFCLK is selected (TXCKSEL LOW) or if data is captured on both edges of REFCLK (TXRATE = HIGH), the phase-align buffers are enabled. These buffers are used to absorb clock phase differences between the presently selected input clock and the internal character clock. Initialization of these phase-align buffers takes place when the TXRST input is sampled by two consecutive rising edges of REFCLK. When TXRST is returned HIGH, the present input clock phase relative to REFCLK is set. TXRST is an asynchronous input, but is sampled internally to synchronize it to the internal transmit path state machines. After set, the input clocks are allowed to skew in time up to half a character period in either direction relative to REFCLK; that is, 180°. This time shift allows the delay paths of the character clocks (relative to REFLCK) to change due to operating voltage and temperature, while not affecting the design operation. If the phase offset, between the initialized location of the input clock and REFCLK, exceeds the skew handling capabilities of the phase-align buffer, an error is reported on the associated TXPERx output. This output indicates a continuous error until the phase-align buffer is reset. While the error remains active, the transmitter for the associated channel outputs a continuous C0.7 character to indicate to the remote receiver that an error condition is present in the link. In specific transmit modes, it is also possible to reset the phase-align buffers individually and with minimal disruption of the serial data stream. When the transmit interface is configured for generation of atomic word sync sequences (TXMODE[1] = MID) and a phase-align buffer error is present, the transmission of a word sync sequence recenters the phase align buffer and clears the error condition.[5] Notes 4. The TXOPx inputs are also captured in the associated Input Register, but their interpretation is under the separate control of PARCTL. 5. One or more K28.5 characters may be added or lost from the data stream during this reset operation. When used with non-Cypress devices that require a complete 16-character word sync sequence for proper receive elasticity buffer alignment, it is recommend that the sequence be followed by a second word sync sequence to ensure proper operation. Document Number: 38-02058 Rev. *M Page 13 of 53 CYP15G0201DXB Parity Support In addition to the ten data and control bits that are captured at each transmit input register, a TXOPx input is also available on each channel. This allows the CYP15G0201DXB to support ODD parity checking for each channel. This parity checking is available for all operating modes (including encoder bypass). The specific mode of parity checking is controlled by the PARCTL input, and operates per Table 2. Table 2. Input Register Bits Checked for Parity[6] Signal Name TXDx[0] TXDx[1] TXDx[2] TXDx[3] TXDx[4] TXDx[5] TXDx[6] TXDx[7] TXCTx[0] TXCTx[1] TXOPx Transmit Parity Check Mode (PARCTL) MID LOW TXMODE[1] TXMODE[1] HIGH = LOW LOW X[7] X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X When PARCTL is MID (open) and the encoders are enabled (TXMODE[1] L), only the TXDx[7:0] data bits are checked for ODD parity along with the associated TXOPx bit. When PARCTL = HIGH with the Encoder enabled (or MID with the encoder bypassed), the TXDx[7:0] and TXCTx[1:0] inputs are checked for ODD parity along with the associated TXOPx bit. When PARCTL = LOW, parity checking is disabled. When parity checking and the encoder are both enabled (TXMODE[1] LOW), the detection of a parity error causes a C0.7 character of proper disparity to be passed to the transmit shifter. When the encoder is bypassed (TXMODE[1] = LOW), detection of a parity error causes a positive disparity version of a C0.7 transmission character to be passed to the transmit shifter. Encoder The character, received from the input register or phase-align buffer and parity check logic, is then passed to the encoder logic. this block interprets each character and any associated control bits, and outputs a 10-bit transmission character. Depending on the configured operating mode, the generated transmission character may be, ■ The 10-bit pre-encoded character accepted in the input register ■ The 10-bit equivalent of the 8-bit data character accepted in the input register ■ The 10-bit equivalent of the 8-bit special character code accepted in the Input Register ■ The 10-bit equivalent of the C0.7 SVS character if parity checking was enabled and a parity error was detected ■ The 10-bit equivalent of the C0.7 SVS character if a Phase-Align Buffer overflow or underflow error is present ■ A character that is part of the 511-character BIST sequence ■ A K28.5 character generated as an individual character or as part of the 16-character Word Sync Sequence. The selection of the specific characters generated is controlled by the TXMODE[1:0], SCSEL, TXCTx[1:0], and TXDx[7:0] inputs for each character. Data Encoding Raw data, as received directly from the Transmit Input Register, is seldom in a form suitable for transmission across a serial link. The characters must usually be processed or transformed to guarantee ■ A minimum transition density (to allow the serial receive PLL to extract a clock from the data stream) ■ A DC-balance in the signaling (to prevent baseline wander) ■ Run-length limits in the serial data (to limit the bandwidth requirements of the serial link) The remote receiver a way of determining the correct character boundaries (framing) When the encoder is enabled (TXMODE[1] LOW), the characters to be transmitted are converted from data or special character codes to 10-bit transmission characters (as selected by their respective TXCTX[1:0] and SCSEL inputs), using an integrated 8- /10-B encoder. When directed to encode that is, the character as a special character code, it is encoded using the special character encoding rules listed in Table 25 on page 47. When directed to encode the character as a data character, it is encoded using the data character encoding rules in Table 24 on page 43. ■ The 8- /10-B Encoder is standards compliant with ANSI/NCITS ASC X3.230-1994 (Fibre Channel), IEEE 802.3z (Gigabit Ethernet), the IBM® ESCON® and FICON® channels, and Digital Video Broadcast DVB-ASI standards for data transport. Many of the Special Character codes listed in Table 25 on page 47 may be generated by more than one input character. The CYP15G0201DXB is designed to support two independent (but non-overlapping) Special Character code tables. This allows the CYP15G0201DXB to operate in mixed environments with other CYP15G0201DXBs using the enhanced Cypress command code set, and the reduced command sets of other non-Cypress devices. Even when used in an environment that normally uses non-Cypress Special Character codes, the selective use of Cypress command codes can permit operation where running disparity and error handling must be managed. Following conversion of each input character from eight bits to a 10-bit transmission character, it is passed to the Transmit Shifter and is shifted out LSB first, as required by ANSI and IEEE standards for 8- /10-B coded serial data streams. Notes 6. Transmit path parity errors are reported on the associated TXPERx output. 7. Bits marked as X are XORed together. Result must be a logic-1 for parity to be valid. Document Number: 38-02058 Rev. *M Page 14 of 53 CYP15G0201DXB Transmit Modes TX Mode 0—Encoder Bypass The operating mode of the transmit path is set through the TXMODE[1:0] inputs. These 3-level select inputs allow one of nine transmit modes to be selected. The transmit modes are listed in Table 3. When the Encoder is bypassed, the character captured in the TXDx[7:0] and TXCTx[1:0] inputs is passed directly to the Transmit Shifter without modification. If parity checking is enabled (PARCTL LOW) and a parity error is detected, the 10-bit character is replaced with the 1001111000 pattern (+C0.7 character). The encoded modes (TX Modes 3 through 8) support multiple encoding tables. These encoding tables vary by the specific combinations of SCSEL, TXCTx[1], and TXCTx[0] that are used to control the generation of data and control characters. These multiple encoding forms allow maximum flexibility in interfacing to legacy applications, while also supporting numerous extensions in capabilities. Table 4. Encoder Bypass Mode (TXMODE[1:0] = LL) Table 3. Transmit Operating Modes TXMODE [1:0] Mode Number TX Mode With the Encoder bypassed, the TXCTx[1:0] inputs are considered part of the data character and do not perform a control function that would otherwise modify the interpretation of the TXDx[7:0] bits. The bit usage and mapping of these control bits when the Encoder is bypassed is shown in Table 4. Signal Name Operating Mode SCSEL Control TXCTx Function Bus Weight 10B Name 0 2 a TXDx[1] 21 b TXDx[2] 22 c 23 d TXDx[0] (LSB) Word Sync Sequence Support [8] 0 LL None None Encoder Bypass TXDx[3] 1 LM None None Reserved for test TXDx[4] 24 e TXDx[5] 25 i TXDx[6] 26 f TXDx[7] 27 g 2 LH None None Reserved for test 3 ML Atomic Special character Encoder Control 4 MM Atomic Word Sync Encoder Control TXCTx[0] 28 h 5 MH Atomic None Encoder Control TXCTx[1] (MSB) 29 j 6 HL Interruptible Special character Encoder Control 7 HM Interruptible Word sync Encoder Control TX Modes 1 and 2—Factory Test Modes 8 HH Interruptible None Encoder Control In Encoder Bypass the SCSEL input is ignored. All clocking modes interpret the data the same, with no internal linking between channels. These modes enable specific factory test configurations. They are not considered normal operating modes of the device. Entry or configuration into these test modes do not damage the device. Note 8. LSB is shifted out first. Document Number: 38-02058 Rev. *M Page 15 of 53 CYP15G0201DXB TX Mode 3—Atomic Word Sync and SCSEL Control of Special Codes ++––+–+–+–+–+–+– – – + + – + – + – + – + – + – +. When configured in TX Mode 3, the SCSEL input is captured along with the associated TXCTx[1:0] data control inputs. These bits combine to control the interpretation of the TXDx[7:0] bits and the characters generated by them. These bits are interpreted as listed in Table 5. When TXMODE[1] = MID (open, TX modes 3, 4, and 5), the generation of this character sequence is an atomic (non-interruptible) operation. After it has been successfully started, it cannot be stopped until all 16 characters have been generated. The content of the associated Input Register(s) is ignored for the duration of this 16-character sequence. SCSEL TXCTx[1] TXCTx[0] Table 5. TX Modes 3 and 6 Encoding X X 0 Encoded data character 0 0 1 K28.5 fill character 1 0 1 Special character code X 1 1 16-character word sync sequence Characters Generated When TXCKSEL = MID, both transmit channels capture data into their Input Registers using independent TXCLKx clocks. The SCSEL input is sampled only by TXCLKA. When the character (accepted in the Channel-A Input Register) has passed through the Phase-Align Buffer and any selected parity validation, the level captured on SCSEL is passed to the Encoder of Channel-B during this same cycle. To avoid the possible ambiguities that may arise due to the uncontrolled arrival of SCSEL relative to the characters in the alternate channel, SCSEL is often used as a static control input. Word Sync Sequence When TXCTx[1:0] = 11, a 16-character sequence of K28.5 characters, known as a word sync sequence, is generated on the associated channel. This sequence of K28.5 characters may start with either a positive or negative disparity K28.5 (as determined by the current running disparity and the 8- /10-B coding rules). The disparity of the second and third K28.5 characters in this sequence are reversed from what normal 8/10-B coding rules would generate. The remaining K28.5 characters in the sequence follow all 8- /10-B coding rules. The disparity of the generated K28.5 characters in this sequence would follow a pattern of either or At the end of this sequence, if the TXCTx[1:0] = 11 condition is sampled again, the sequence restarts and remains uninterruptible for the following 15 character clocks. If parity checking is enabled, the character used to start the Word Sync Sequence must also have correct ODD parity. After the sequence is started, parity is not checked on the following 15 characters in the Word Sync Sequence. When TXMODE[1] = HIGH (TX modes 6, 7, and 8), the generation of the word sync sequence becomes an interruptible operation. In TX Mode 6, this sequence is started as soon as the TXCTx[1:0] = 11 condition is detected on a channel. In order for the sequence to continue on that channel, the TXCTx[1:0] inputs must be sampled as 00 for the remaining 15 characters of the sequence. If at any time a sample period exists where TXCTx[1:0] 00, the word sync sequence is terminated, and a character representing the associated data and control bits is generated by the Encoder. This resets the word sync sequence state machine such that it starts at the beginning of the sequence at the next occurrence of TXCTx[1:0] = 11. When parity checking is enabled and TXMODE[1] = HIGH, all characters (including those in the middle of a word sync sequence) must have correct parity. The detection of a character with incorrect parity during a word sync sequence (regardless of the state of TXCTx[1:0]) interrupts that sequence and forces the generation of a C0.7 SVS character. Any interruption of the Word Sync Sequence causes the sequence to terminate. When TXCKSEL = LOW, the Input Registers for both transmit channels are clocked by REFCLK[2]. When TXCKSEL = HIGH, the Input Registers for both transmit channels are clocked with TXCLKA. In these clock modes both sets of TXCTx[1:0] inputs operate synchronous to the SCSEL input.[9] Note 9. When operated in any condition where receive channels are bonded together, TXCKSEL must be either LOW or HIGH (not MID) to ensure that associated characters are transmitted in the same character cycle. Document Number: 38-02058 Rev. *M Page 16 of 53 CYP15G0201DXB TXCTx[1] X X 0 Encoded data character 0 0 1 K28.5 fill character 0 1 1 Special character code 1 X 1 16-character Word Sync Sequence Characters Generated When TXCKSEL = MID, both transmit channels operate independently. The SCSEL input is sampled only by TXCLKA. When the character accepted in the Channel-A input register has passed any selected validation and is ready to be passed to the encoder, the level captured on SCSEL is passed to the Encoder of Channel-B during this same cycle. Changing the state of SCSEL changes the relationship of the characters on the alternate channel. SCSEL should either be used as a static configuration input or changed only when the state of TXCTx[1:0] on the alternate channel are such that SCSEL is ignored during the change. TX Mode 4 also supports an atomic word sync sequence. Unlike TX Mode 3, this sequence is started when both SCSEL and TXCTx[0] are sampled HIGH. With the exception of the combination of control bits used to initiate the sequence, the generation and operation of this word sync sequence is the same as that documented for TX Mode 3. TX Mode 5—Atomic Word Sync, No SCSEL Table 7. TX Modes 5 and 8 Encoding, Non-Bonded (RXMODE[1] = LOW) X 0 TXCTx[0] SCSEL TXCTx[0] Table 6. TX Modes 4 and 7 Encoding For non-bonded operation, the TXCTx[1:0] inputs for each channel control the characters generated by that channel. The specific characters generated by these bits are listed in Table 7 on page 17. TXCTx[1] When configured in TX Mode 4, the SCSEL input is captured along with the associated TXCTx[1:0] data control inputs. These bits combine to control the interpretation of the TXDx[7:0] bits and the characters generated by them. These bits are interpreted as listed in Table 6. without atomic word sync sequence generation, two additional encoding mappings are controlled by the channel bonding selection made through the RXMODE[1:0] inputs. SCSEL TX Mode 4—Atomic Word Sync and SCSEL Control of Word Sync Sequence Generation Characters Generated 0 Encoded data character X 0 1 K28.5 fill character X 1 0 Special character code X 1 1 16-character Word Sync Sequence TX Mode 5 also has the capability of generating an atomic Word Sync Sequence. For the sequence to be started, the TXCTx[1:0] inputs must both be sampled HIGH. With the exception of the combination of control bits used to initiate the sequence, the generation and operation of this Word Sync Sequence is the same as that documented for TX Mode 3. Two additional encoding maps are provided for use when receive channel bonding is enabled. When dual-channel bonding is enabled (RXMODE[1] = HIGH), the CYP15G0201DXB is configured such that channels A and B are bonded together to form a two-character-wide path. When operated in this two-channel bonded mode, the TXCTA[0] and TXCTB[0] inputs control the interpretation of the data on both the A and B channels. The characters on each half of these bonded channels are controlled by the associated TXCTx[1] bit. The specific characters generated by these control bit combinations are listed in Table 8. When configured in TX Mode 5, the SCSEL signal is not used. In addition to the standard character encodings, both with and TXCTB[0] TXCTB[1] TXCTA[0] SCSEL TXCTA[1] Table 8. TX Modes 5 and 8, Dual-channel Bonded (RXMODE[1] = HIGH) Characters Generated X 0 0 X 0 Encoded data character on channel A X 0 1 X 0 K28.5 fill character on channel A X 1 0 X 0 Special character code on channel A X 1 1 X 0 16-character word sync on channel A X X 0 0 0 Encoded data character on channel B X X 1 0 0 K28.5 fill character on channel B X X 0 1 0 Special character code on channel B X X 1 1 0 16-character word sync on channel B X X X X 1 16-character word sync on channels A and B Document Number: 38-02058 Rev. *M Page 17 of 53 CYP15G0201DXB Note especially that any time TXCTB[0] is sampled HIGH, both channels A and B start generating an Atomic Word Sync Sequence, regardless of the state of any of the other bits in the A or B Input Registers (with the exception of any enabled parity checking). Transmit BIST Each transmit channel contains an internal pattern generator that can be used to validate both device and link operation. These generators are enabled by the associated BOE[x] signals listed in Table 9 on page 18 (when the BISTLE latch enable input is HIGH). When enabled, a register in the associated transmit channel becomes a signature pattern generator by logically converting to a linear feedback shift register (LFSR). This LFSR generates a 511-character sequence that includes all Data and Special Character codes, including the explicit violation symbols. This provides a predictable yet pseudo-random sequence that can be matched to an identical LFSR in the attached receiver(s). If the receive channels are configured for common clock operation (RXCKSEL MID) and Encoder is enabled (TXMODE[1] LOW) each pass is preceded by a 16-character Word Sync Sequence to allow Elasticity Buffer alignment and management of clock- frequency variations. When the BISTLE signal is HIGH, any BOE[x] input that is LOW enables the BIST generator in the associated transmit channel (or the BIST checker in the associated receive channel). When BISTLE returns LOW, the values of all BOE[x] signals are captured in the BIST Enable Latch. These values remain in the BIST Enable Latch until BISTLE is returned HIGH to open the latch. A device reset, (TRSTZ sampled LOW) presets the BIST Enable Latch to disable BIST on all channels. All data and data-control information present at the associated TXDx[7:0] and TXCTx[1:0] inputs are ignored when BIST is active on that channel. Serial Output Drivers The serial interface output drivers use high-performance differential current mode logic (CML) to provide a source-matched driver for the transmission lines. These drivers accept data from the transmit shifters. these outputs have signal swings equivalent to that of standard PECL drivers, and are capable of driving AC-coupled optical modules or transmission lines. When configured for local loopback (LPEN = HIGH), all enabled Serial Drivers are configured to drive a static differential logic-1. Each serial driver can be enabled or disabled separately through the BOE[3:0] inputs, as controlled by the OELE latch-enable signal. When OELE is HIGH, the signals present on the BOE[3:0] inputs are passed through the Serial Output Enable Latch to control the serial output drivers. The BOE[3:0] input associated with a specific OUTxy driver is listed in Table 9. Table 9. Output Enable, BIST, and Receive Channel Enable Signal Map BOE Input Output Controlled (OELE) BIST Channel Enable (BISTLE) Receive PLL Channel Enable (RXLE) BOE[3] OUTB2 Transmit B X BOE[2] OUTB1 Receive B Receive B BOE[1] OUTA2 Transmit A X BOE[0] OUTA1 Receive A Receive A When OELE is HIGH and BOE[x] is HIGH, the associated serial driver is enabled. When OELE is HIGH and BOE[x] is LOW, the associated driver is disabled and internally powered down. If both outputs for a channel are in this disabled state, the associated internal logic for that channel is also powered down. When OELE returns LOW, the values present on the BOE[3:0] inputs are latched in the Output Enable Latch, and remain there until OELE returns HIGH to enable the latch. A device reset (TRSTZ sampled LOW) clears this latch and disables all output drivers. Note. When all transmit channels are disabled (that is, both outputs disabled in all channels) and a channel is re-enabled, the data on the Serial Drivers may not meet all timing specifications for up to 200 s. Document Number: 38-02058 Rev. *M Page 18 of 53 CYP15G0201DXB Transmit PLL Clock Multiplier The transmit PLL clock multiplier accepts a character-rate or half-character-rate external clock at the REFCLK input, and multiples that clock by 10 or 20 (as selected by TXRATE) to generate a bit-rate clock for use by the Transmit Shifter. It also provides a character-rate clock used by the transmit paths. The clock multiplier PLL can accept a REFCLK input between 10 MHz and 150 MHz, however, this clock range is limited by the operating mode of the CYP15G0201DXB clock multiplier (controlled by TXRATE) and by the level on the SPDSEL input. SPDSEL is a 3-level select[10] (ternary) input that selects one of three operating ranges for the serial data outputs and inputs. The operating serial signaling-rate and allowable range of REFCLK frequencies are listed in Table 10. Table 10. Operating Speed Settings SPDSEL TXRATE REFCLK Frequency (MHz) Signaling Rate (MBaud) LOW 1 Reserved 195-400 0 19.5-40 1 20-40 0 40-80 MID (Open) HIGH 1 40-75 0 80-150 400-800 800-1500 When TXRATE = HIGH (Half-rate REFCLK), TXCKSEL = HIGH or MID (TXCLKx or TXCLKA selected to clock input register) is an invalid mode of operation. The REFCLK input is a differential input with each input internally biased to 1.4 V. If the REFCLK+ input is connected to a TTL, LVTTL, or LVCMOS clock source, the input signal is recognized when it passes through the internally biased reference point. When both the REFCLK+ and REFCLK– inputs are connected, the clock source must be a differential clock. This can be either a differential LVPECL clock that is DC-or AC-coupled, or a differential LVTTL or LVCMOS clock. By connecting the REFCLK– input to an external voltage source or resistive voltage divider, it is possible to adjust the reference point of the REFCLK+ input for alternate logic levels. When doing so it is necessary to ensure that the 0 V differential crossing point remains within the parametric range supported by the input. CYP15G0201DXB Receive Data Path Serial Line Receivers Two differential line receivers, INx1±and INx2±, are available on each channel for accepting serial data streams. The active serial line receiver on a channel is selected using the associated INSELx input. The serial line receiver inputs are differential, and can accommodate wire interconnect and filtering losses or transmission line attenuation greater than 16 dB. For normal operation, these inputs should receive a signal of at least VIDIFF > 100 mV, or 200 mV peak-to-peak differential. Each Line Receiver can be DC- or AC-coupled to +3.3 V powered fiber-optic interface modules (any ECL/PECL logic family, not limited to 100 K PECL) or AC-coupled to +5 V powered optical modules. The common-mode tolerance of these line receivers accommodates a wide range of signal termination voltages. Each receiver provides internal DC-restoration, to the center of the receiver’s common mode range, for AC-coupled signals. The local loopback input (LPEN) allows the serial transmit data outputs to be routed internally back to the clock and data recovery circuit associated with each channel. When configured for local loopback, all transmit serial driver outputs are forced to output a differential logic-1. This prevents local diagnostic patterns from being broadcast to attached remote receivers. Note 10. 3-Level select inputs are used for static configuration. They are ternary (not binary) inputs that make use of non-standard logic levels of LOW, MID, and HIGH. The LOW level is usually implemented by direct connection to VSS (ground). The HIGH level is usually implemented by direct connection to VCC (power). When not connected or allowed to float, a 3-level select input self-biases to the MID level. Document Number: 38-02058 Rev. *M Page 19 of 53 CYP15G0201DXB Signal Detect/Link Fault ■ Analog amplitude above limit specified by SDASEL The clock/data recovery (CDR) circuit includes logic to monitor the frequency of the phase-locked loop (PLL) voltage controlled oscillator (VCO) used to sample the incoming data stream. This logic ensures that the VCO operates at, or near the rate of the incoming data stream for two primary cases: ■ Transition density greater than specified limit ■ ■ Range controller reports the received data stream within normal frequency range (±1500 ppm)[11] Each selected line receiver (that is, that routed to the clock and data recovery PLL) is simultaneously monitored for Receive channel enabled All of these conditions must be valid for the signal detect block to indicate a valid signal is present. This status is presented on the link fault indicator (LFIx) output associated with each receive channel, which changes synchronous to the selected receive interface clock. ■ Table 11. Analog Amplitude Detect Valid Signal Levels[12] SDASEL LOW Typical signal with peak amplitudes above 140 mV p-p differential MID (Open) 280 mV p-p differential HIGH 420 mV p-p differential Analog Amplitude While the majority of these signal monitors are based on fixed constants, the analog amplitude level detection is adjustable to allow operation with highly attenuated signals, or in high-noise environments. This adjustment is made through the SDASEL signal, a 3-level select[13] input, which sets the trip point for the detection of a valid signal at one of three levels, as listed in Table 11. This control input effects the analog monitors for all receive channels. The analog signal detect monitors are active for the line receiver, selected by the associated INSELx input. When configured for local loopback (LPEN = HIGH), no line receivers are selected, and the LFIx output for each channel reports only the receive VCO frequency out-of-range and transition density status of the associated transmit signal. When local loopback is active, the analog signal detect monitors are disabled. Transition Density The transition detection logic checks for the absence of any transitions spanning greater than six transmission characters (60 bits). If no transitions are present in the data received on a channel (within the referenced period), the transition detection logic for that channel asserts LFIx. The LFIx output remains asserted until at least one transition is detected in each of three adjacent received characters. Range Controls when the incoming data stream resumes after a time in which it has been “missing” when the incoming data stream is outside the acceptable frequency range To perform this function, the frequency of the VCO is periodically sampled and compared to the frequency of the REFCLK input. If the VCO is running at a frequency beyond ±1500 ppm[11] as defined by the reference clock frequency, it is periodically forced to the correct frequency (as defined by REFCLK, SPDSEL, and TXRATE) and then released in an attempt to lock to the input data stream. The sampling and relock period of the Range Control is calculated as follows: RANGE CONTROL SAMPLING PERIOD = (REFCLKPERIOD) × (16000). During the time that the range control forces the PLL VCO to run at REFCLK×10 (or REFCLK×20 when TXRATE = HIGH) rate, the LFIx output is asserted LOW. While the PLL is attempting to re-lock to the incoming data stream, LFIx may be either HIGH or LOW (depending on other factors such as transition density and amplitude detection) and the recovered byte clock (RXCLKx) may run at an incorrect rate (depending on the quality or existence of the input serial data stream). After a valid serial data stream is applied, it may take up to one RANGE CONTROL SAMPLING PERIOD before the PLL locks to the input data stream, after which LFIx should be HIGH. ■ Receive Channel Enabled The CYP15G0201DXB contains two receive channels that can be independently enabled and disabled. Each channel can be enabled or disabled separately through the BOE[3:0] inputs, as controlled by the RXLE latch-enable signal. When RXLE is HIGH, the signals present on the BOE[3:0] inputs are passed through the receive channel enable latch to control the PLLs and logic of the associated receive channel. The BOE[3:0] input associated with a specific receive channel is listed in Table 9. When RXLE = HIGH and BOE[x] = HIGH, the associated receive channel is enabled to receive and decode a serial stream. When RXLE = HIGH and BOE[x] = LOW, the associated receive channel is disabled and internally configured for minimum power dissipation. If a single channel of a bonded-pair is disabled, the other receive channels may not bind correctly. If the disabled channel is selected as the master channel for insert/delete functions, or recovered clock select, these functions do not work correctly. Any disabled channel indicates an asserted LFIx output. When RXLE returns LOW, the values present on the BOE[3:0] inputs are latched in the receive channel enable latch, and remain there until RXLE returns HIGH to opened the latch again.[14] Notes 11. REFCLK has no phase or frequency relationship with the recovered clock(s) and only acts as a centering reference to reduce clock synchronization time. REFCLK must be within ±1500 ppm (±0.15%) of the remote transmitter’s PLL reference (REFCLK) frequency. Although transmitting to a HOTLink II receiver necessitates the frequency difference between the transmitter and receiver reference clocks to be within ±1500 ppm, the stability of the crystal needs to be within the limits specified by the appropriate standard when transmitting to a remote receiver that is compliant to that standard. For example, to be IEEE 802.3z Gigabit Ethernet compliant, the frequency stability of the crystal needs to be within ±100 PPM. 12. The peak amplitudes listed in this table are for typical waveforms that have generally 3–4 transitions for every ten bits. In a worse case environment the signals may have a sign-wave appearance (highest transition density with repeating 0101...). Signal peak amplitudes levels within this environment type could increase the values in the table by approximately 100 mV. 13. 3-Level select inputs are used for static configuration. They are ternary (not binary) inputs that make use of non-standard logic levels of LOW, MID, and HIGH. The LOW level is usually implemented by direct connection to VSS (ground). The HIGH level is usually implemented by direct connection to VCC (power). When not connected or allowed to float, a 3-Level select input self-biases to the MID level. 14. When a disabled receive channel is re-enabled, the status of the associated LFIx output and data on the parallel outputs for the associated channel may be indeterminate for up to 10 ms. Document Number: 38-02058 Rev. *M Page 20 of 53 CYP15G0201DXB Clock/Data Recovery The extraction of a bit-rate clock and recovery of bits from each received serial stream is performed by a separate CDR block within each receive channel. The clock extraction function is performed by high-performance embedded PLLs that track the frequency of the transitions in the incoming bit streams and align the phase of their internal bit-rate clocks to the transitions in the selected serial data streams. Each CDR accepts a character-rate (bit-rate 10) or half-character-rate (bit rate 20) reference clock from the REFCLK input. This REFCLK input is used to ■ Ensure that the VCO (within each CDR) is operating at the correct frequency (rather than some harmonic of the bit rate) ■ Improve PLL acquisition time ■ And to limit unlocked frequency excursions of the CDR VCO when no data is present at the selected Serial Line Receiver. Regardless of the type of signal present, the CDR attempts to recover a data stream from it. If the frequency of the recovered data stream is outside the limits set by the range control monitors, the CDR switches to track REFCLK instead of the data stream. After the CDR output (RXCLKx) frequency returns back close to REFCLK frequency, the CDR input is switched back to the input data stream to check its frequency. In case no data is present at the input this switching behavior may result in brief RXCLKx frequency excursions from REFCLK. However, the validity of the input data stream is indicated by the LFIx output. The frequency of REFCLK is required to be within 1500 ppm[11] of the frequency of the clock that drives the REFCLK input of the remote transmitter to ensure a lock to the incoming data stream. For systems using multiple or redundant connections, the LFIx output can be used to select an alternate data stream. When an LFIx indication is detected, external logic can toggle selection of the associated INx1± and INx2± inputs through the associated INSELx input. When a port switch takes place, it is necessary for the receive PLL for that channel to reacquire the new serial stream and frame to the incoming character boundaries. If channel bonding is also enabled, a channel alignment event is also required before the output data may be considered usable. Deserializer/Framer Each CDR circuit extracts bits from the associated serial data stream and clocks these bits into the Shifter/Framer at the bit-clock rate. When enabled, the Framer examines the data stream looking for one or more Comma or K28.5 characters at all possible bit positions. The location of this character in the data stream is used to determine the character boundaries of all following characters. Framing Character The CYP15G0201DXB allows selection of two combinations of framing characters to support requirements of different interfaces. The selection of the framing character is made through the FRAMCHAR input. The specific bit combinations of these framing characters are listed in Table 12. When the specific bit combination of the selected framing character is detected by the framer, the boundaries of the characters present in the received data stream are known. Table 12. Framing Character Selector FRAMCHAR LOW Bits Detected in Framer Character Name Bits Detected Reserved for test MID (Open) Comma+ Comma– HIGH –K28.5 +K28.5 00111110XX [15] or 11000001XX 0011111010 or 1100000101 Framer The framer on each channel operates in one of three different modes, as selected by the RFMODE input. In addition, the framer itself may be enabled or disabled through the RFEN input. When RFEN = LOW, the framers in both receive paths are disabled, and no combination of bits in a received data stream alters the character boundaries. When RFEN = HIGH, the framer selected by RFMODE is enabled on both channels. When RFMODE = LOW, the low-latency framer is selected. This framer operates by stretching the recovered character clock until it aligns with the received character boundaries. In this mode the framer starts its alignment process on the first detection of the selected framing character. To reduce the impact on external circuits that make use of a recovered clock, the clock period is not stretched by more than two bit-periods in any one clock cycle. When operated in with a character-rate output clock (RXRATE = LOW), the output of properly framed characters may be delayed by up to nine character-clock cycles from the detection of the selected framing character. When operated with a half-character-rate output clock (RXRATE = HIGH), the output of properly framed characters may be delayed by up to 14 character-clock cycles from the detection of the selected framing character.[16] When RFMODE is MID (open) the Cypress-mode multi-byte framer is selected. The required detection of multiple framing characters makes the associated link much more robust to incorrect framing due to aliased SYNC characters in the data stream. In this mode, the framer does not adjust the character clock boundary, but instead aligns the character to the already recovered character clock. This ensures that the recovered clock does not contain any significant phase changes or hops during normal operation or framing, and allows the recovered clock to be replicated and distributed to other external circuits or components using PLL-based clock distribution elements. In this framing mode the character boundaries are only adjusted if the selected framing character is detected at least twice within a span of 50 bits, with both instances on identical 10-bit character boundaries. Notes 15. The standard definition of a Comma contains only seven bits. However, since all valid Comma characters within the 8- /10-B character set also have the 8th bit as an inversion of the 7th bit, the compare pattern is extended to a full eight bits to reduce the possibility of a framing error. 16. When Receive BIST is enabled on a channel, the Low-latency Framer must not be enabled. The BIST sequence contains an aliased K28.5 framing character, which causes the Receiver to update its character boundaries incorrectly. Document Number: 38-02058 Rev. *M Page 21 of 53 CYP15G0201DXB When RFMODE = HIGH, the alternate-mode multi-byte framer is enabled. Like the Cypress-mode multi-byte framer, multiple framing characters must be detected before the character boundary is adjusted. In this mode, the data stream must contain a minimum of four of the selected framing characters, received as consecutive characters, on identical 10-bit boundaries, before character framing is adjusted. In this mode, the Framer does not adjust the character clock boundary, but instead aligns the character to the already recovered character clock. Framing for all channels is enabled when RFEN = HIGH. If RFEN = LOW, the framer for each channel is disabled. When the framers are disabled, no changes are made to the recovered character boundaries on any channel, regardless of the presence of framing characters in the data stream. 8- / 10-B Decoder Block The decoder logic block performs three primary functions: ■ Decoding the received transmission characters back into data and special character codes, ■ Comparing generated BIST patterns with received characters to permit at-speed link and device testing, ■ Generation of ODD parity on the decoded characters. 8- / 10-B Decoder The framed parallel output of each deserializer shifter is passed to the 8- / 10-B decoder where, if the decoder is enabled (DECMODE LOW), it is transformed from a 10-bit transmission character back to the original data and special character codes. This block uses the 8- / 10-B Decoder patterns in Table 24 on page 43 and Table 25 on page 47 of this datasheet. Valid data characters are indicated by a 000b bit-combination on the associated RXSTx[2:0] status bits, and special character codes are indicated by a 001b bit-combination on these same status outputs. Framing characters, invalid patterns, disparity errors, and synchronization status are presented as alternate combinations of these status bits. The 8- / 10-B decoder operates in two normal modes, and can also be bypassed. The operating mode for the Decoder is controlled by the DECMODE input. When DECMODE = LOW, the decoder is bypassed and raw 10-bit characters are passed to the output register. In this mode, channel bonding is not possible, the receive Elasticity Buffers are bypassed, and RXCKSEL must be MID. This clock mode generates separate RXCLKx±outputs for each receive channel. When DECMODE = MID (or open), the 10-bit transmission characters are decoded using Table 24 on page 43 and Table 25 on page 47. Received Special Code characters are decoded using the Cypress column of Table 25 on page 47. When DECMODE = HIGH, the 10-bit transmission characters are decoded usingTable 24 on page 43 and Table 25 on page 47. received special code characters are decoded using the Alternate column of Table 25 on page 47. In all settings where the decoder is enabled, the receive paths may be operated as separate channels or bonded to form dual-channel buses. Document Number: 38-02058 Rev. *M Receive BIST Operation The receiver interfaces contain internal pattern generators that can be used to validate both device and link operation. These generators are enabled by the associated BOE[x] signals listed in Table 9 on page 18 (when the BISTLE latch enable input is HIGH). When enabled, a register in the associated receive channel becomes a signature pattern generator and checker by logically converting to a linear feedback shift register (LFSR). This LFSR generates a 511-character sequence that includes all data and special character codes, including the explicit violation symbols. This provides a predictable yet pseudo-random sequence that can be matched to an identical LFSR in the attached transmitter(s). If the receive channels are configured for common clock operation (RXCKSEL MID) each pass is preceded by a 16-character word sync sequence. When synchronized with the received data stream, the associated Receiver checks each character in the decoder with each character generate by the LFSR and indicates compare errors and BIST status at the RXSTx[2:0] bits of the output register. Refer Table 20 on page 30 for details. When the BISTLE signal is HIGH, any BOE[x] input that is LOW enables the BIST generator/checker in the associated receive channel (or the BIST generator in the associated transmit channel). When BISTLE returns LOW, the values of all BOE[x] signals are captured in the BIST enable latch. These values remain in the BIST Enable Latch until BISTLE is returned HIGH. All captured signals in the BIST Enable Latch are set HIGH (that is, BIST is disabled) following a device reset (TRSTZ is sampled LOW). When BIST is first recognized as being enabled in the Receiver, the LFSR is preset to the BIST-loop start-code of D0.0 This code D0.0 is sent only once per BIST loop. The status of the BIST progress and any character mismatches is presented on the RXSTx[2:0] status outputs. Code rule violations or running disparity errors that occur as part of the BIST loop do not cause an error indication. RXSTx[2:0] indicates 010b or 100b for one character period per BIST loop to indicate loop completion. This status can be used to check test pattern progress. These same status values are presented when the decoder is bypassed and BIST is enabled on a receive channel. The specific status reported by the BIST state machine are listed in Table 18 on page 27. These same codes are reported on the receive status outputs regardless of the state of DECMODE. The specific patterns checked by each receiver are described in detail in the Cypress application note entitled “HOTLink built-in self-test.” The sequence compared by the CYP15G0201DXB when RXCKSEL = MID is identical to that in the CY7B933 and CY7C924DX, allowing interoperable systems to be built when used at compatible serial signaling rates. If the number of invalid characters received ever exceeds the number of valid characters by 16, the receive BIST state machine aborts the compare operations and resets the LFSR to the D0.0 state to look for the start of the BIST sequence again. When the receive paths are configured for common clock operation (RXCKSEL MID) each pass must be preceded by a 16-character Word Sync Sequence to allow output buffer alignment and management of clock frequency variations. Page 22 of 53 CYP15G0201DXB This is automatically generated by the transmitter when its local RXCKSEL MID and encoder is enabled. The BIST state machine requires the characters to be correctly framed for it to detect the BIST sequence. If the Low Latency framer is enabled (RFMODE = LOW), the framer misaligns to an aliased SYNC character within the BIST sequence. If the alternate multi-byte framer is enabled (RFMODE = HIGH) and the Receiver outputs are clocked relative to a recovered clock, it is generally necessary to frame the receiver before BIST is enabled. If the receive outputs are clocked relative to REFCLK (RXCKSEL = LOW), the transmitter precedes every 511 character BIST sequence with a 16-character Word Sync Sequence. Receive Elasticity Buffer Each receive channel contains an elasticity buffer that is designed to support multiple clocking modes. These buffers allow data to be read using an elasticity buffer read-clock that is asynchronous in both frequency and phase from the elasticity buffer write clock, or to use a read clock that is frequency coherent but with uncontrolled phase relative to the elasticity buffer write clock. Each elasticity buffer is a minimum of 10 characters deep, and supports a 12-bit-wide data path. It is capable of supporting a decoded character, three status bits, and a parity bit for each character present in the buffer. The write clock for these buffers is always the recovered clock for the associated read channel. The read clock for the elasticity buffers may come from one of three selectable sources. It may be a, ■ Character-rate REFCLK ■ Recovered clock from the same receive channel ■ Recovered clock from an alternate receive channel These elasticity buffers are also used to align the output data streams when both channels are bonded together. More details on how the elasticity buffer is used for independent channel modes and channel bonded modes is discussed in the next section. The elasticity buffers are bypassed whenever the decoders are bypassed (DECMODE = LOW). When the decoders and elasticity buffers are bypassed, RXCKSELx must be set to MID. Document Number: 38-02058 Rev. *M Receive Modes The operating mode of the receive path is set through the RXMODE[1:0] inputs. The ‘Reserved for test’ settings (RXMODE0=M) is not allowed, even if the receiver is not being used. A[1:0] settings are ignored as long as they are not test modes. It stops normal function of the device. When the decoder is disabled, the RX MODE. These modes determine the type (if any) of channel bonding and status reporting. The different receive modes are listed in Table 13. When RXMODE[1] = MID or RXMODE[0] = MID the resulting modes are reserved for test. Table 13. Receive Operating Modes RX Mode Mode Number RXMODE [1:0] Operating Mode Channel Bonding RXSTx Status Reporting 0 LL Independent Status A 1 LH Independent Status B 2 HL Dual Status A 3 HH Dual Status B Independent Channel Modes In independent channel modes (RX Modes 0 and 1, where RXMODE[1] = LOW), both receive paths may be clocked in any clock mode selected by RXCKSEL. When RXCKSEL = LOW, both channels are clocked by REFCLK. RXCLKB± output is disabled (High-Z), and the RXCLKA± and RXCLKC+ outputs presents buffered and delayed forms of REFCLK. In this mode, the receive elasticity buffers are enabled. For REFCLK clocking, the elasticity buffers must be able to insert K28.5 characters and delete framing characters as appropriate. The insertion of a K28.5 or deletion of a framing character can occur at any time on any channel, however, the actual timing on these insertions and deletions is controlled in part by the how the transmitter sends its data. Insertion of a K28.5 character can only occur when the receiver has a framing character in the elasticity buffer. Likewise, to delete a framing character, one must also be in the elasticity buffer. To prevent a receive buffer overflow or underflow on a receive channel, a minimum density of framing characters must be present in the received data streams. When RXCKSEL = MID (or open), each received channel output register is clocked by the recovered clock for that channel. Since no characters may be added or deleted, the receiver elasticity buffer is bypassed. When RXCKSEL = HIGH, all channels are clocked by the selected recovered clock. This selected clock is always output on RXCLKA±. In this mode the receive elasticity buffers are enabled. When data is output using a recovered clock (RXCKSEL = HIGH), receive channels are not allowed to insert and delete characters, except as necessary for elasticity buffer alignment. When the elasticity buffer is used, prior to delivery of valid data, a word sync sequence (or at least four framing characters) must be received to center the elasticity buffers. The elasticity buffer may also be centered by a device reset operation initiated through the TRSTZ input, however, following such an event the CYP15G0201DXB normally requires a framing event before it correctly decodes the characters. When RXCKSEL = HIGH, Page 23 of 53 CYP15G0201DXB since the Elasticity buffer is not allowed to insert or delete framing characters, the transmit clocks on all received channels must all be from a common source. Dual-channel Bonded Modes In dual-channel bonded modes (RX Modes 2 and 3, where RXMODE[1] = HIGH), the associated receive channel pair Output Registers must be clocked by a common clock. This mode does not operate when RXCKSEL = MID. Proper operation in this mode requires that the associated transmit data streams are clocked from a common reference with no long-term character slippage between the bonded channels. In dual-channel mode this means that channels A and B must be clocked from a common reference. Prior to the reception of valid data, a word sync sequence (or that portion of one necessary to align the receive buffers) must be received on the bonded channels (within the allowable inter-channel skew window) to allow the receive elasticity buffers to be centered. While normal characters may be output prior to this alignment event, they are not necessarily aligned within the same word boundaries as when they were transmitted. When RXCKSEL = LOW, both receive channels are clocked by REFCLK. RXCLKB± outputs are disabled (High-Z), and the RXCLKA± and RXCLKC+ outputs present buffered and delayed forms of REFCLK. In this mode, the receive Elasticity Buffers are enabled. For REFCLK clocking, the elasticity buffers must be able to insert K28.5 characters and delete framing characters as appropriate. While these insertions and deletions can take place at any time, they must occur at the same time on both channels that are bonded together. This is necessary to keep the data in the bonded channel-pair properly aligned. This insert and delete process is controlled by the master channel selected by the RXCLKB+ input as listed in Table 14. When RXCKSEL = HIGH, the A and B channels are clocked by the selected recovered clock, as shown in Table 14. The output clock for the channel A/B bonded-pair is output continuously on RXCLKA±. The clock source for this output is selected from the recovered clock for channel A or channel B using the RXCLKB+ input. Table 14. Dual-Channel Bonded Recovered Clock Select and Master Channel Select RXCLKB+ Clock Source RXCLKA± 0 RXCLKA 1 RXCLKB When data is output using a recovered clock (RXCKSEL = HIGH), receive channels are not allowed to insert and delete characters, except as necessary for elasticity buffer alignment. Document Number: 38-02058 Rev. *M Power Control The CYP15G0201DXB supports user control of the powered-up or down state of each transmit and receive channel. The receive channels are controlled by the RXLE signal and values present on the BOE [3:0] bus. The transmit channels are controlled by the OELE signal and the values present on the BOE[3:0] bus. Powering down unused channels saves power and reduces system heat generation. Controlling system power dissipation improves the system performance. Receive Channels When RXLE = HIGH, the signals on the BOE[3:0] inputs directly control the power enables for the receive PLLs and analog circuits. When a BOE[3:0] input is HIGH, the associated receive channel [A and B] PLL and analog logic are active. When a BOE[3:0] input is LOW, the associated receive channel [A and B] PLL and analog logic are powered down. When RXLE returns LOW, the last values present on the BOE[3:0] inputs are captured. The specific BOE[3:0] input signal associated with a receive channel is listed in Table 9 on page 18. Any disabled receive channel indicates a constant LFIx output. When a disable receive channel is re-enabled, the status of the associated LFIx output and data on the parallel outputs for the associated channel may be indeterminate for up to 10 ms. Transmit Channels When OELE is HIGH, the signals on the BOE[3:0] inputs directly control the power enables for the serial drivers. When BOE[3:0] input is HIGH, the associated serial driver is enabled. When BOE[3:0] input is LOW, the associated Serial Driver is disabled and powered down. If both serial drivers of a channel are disabled, the internal logic for that channel is powered down. When OELE returns LOW, the value present on the BOE[3:0] inputs are latched in the output enable latch. Device Reset State When the CYP15G0201DXB is reset by the assertion of TRSTZ, the transmit enable and receive enable latches are both cleared, and the BIST enable latch is preset. In this state, all transmit and receive channels are disabled, and BIST is disabled on all channels. Following a device reset, it is necessary to enable the transmit and receive channels used for normal operation. This can be done by sequencing the appropriate values on the BOE[3:0] inputs while the OELE and RXLE signals are raised and lowered. For systems that do not require dynamic control of power, or want the device to power-up in a fixed configuration, it is also possible to strap the RXLE and OELE control signals HIGH to permanently enable their associated latches. Connection of the associated BOE[3:0] signals HIGH, enables the respective transmit and receive channels as soon as the TRSTZ signal is deasserted. Page 24 of 53 CYP15G0201DXB Output Bus Each receive channel presents a 12-signal output bus consisting of, ■ an 8-bit data bus ■ a 3-bit status bus ■ a parity bit. The signals present on this output bus are modified by the present operating mode of the CYP15G0201DXB as selected by DECMODE. The bits are assigned per Table 15. Table 15. Output Register Bit Assignments[17] Signal Name RXSTx[2] (LSB) RXSTx[1] RXSTx[0] RXDx[0] RXDx[1] RXDx[2] RXDx[3] RXDx[4] RXDx[5] RXDx[6] RXDx[7] (MSB) = MID or DECMODE = LOW DECMODE HIGH COMDETx RXSTx[2] DOUTx[0] RXSTx[1] DOUTx[1] RXSTx[0] DOUTx[2] RXDx[0] DOUTx[3] RXDx[1] DOUTx[4] RXDx[2] DOUTx[5] RXDx[3] DOUTx[6] RXDx[4] DOUTx[7] RXDx[5] DOUTx[8] RXDx[6] DOUTx[9] RXDx[7] When the 8- / 10-B decoder is bypassed (DECMODE = LOW), the framed 10-bit and a single status bit are presented at the receiver output register. The status output indicates if the character in the output register is one of the selected framing characters. The bit usage and mapping of the external signals to the raw 10 B transmission character is shown in Table 16. Table 16. Decoder Bypass Mode (DECMODE = LOW) Signal Name RXSTx[2] (LSB) RXSTx[1] RXSTx[0] RXDx[0] RXDx[1] RXDx[2] RXDx[3] RXDx[4] RXDx[5] RXDx[6] RXDx[7] (MSB) Bus Weight COMDETx 20 21 22 23 24 25 26 27 28 29 10B Name a b c d e i f g h j The COMDETx status outputs operate the same regardless of the bit combination selected for character framing by the FRAMCHAR input. They are HIGH when the character in the Output Register contains the selected framing character at the proper character boundary, and LOW for all other bit combinations. When the low-latency framer and half-rate receive port clocking are also enabled (RFMODE = LOW, RXRATE = HIGH, and RXCKSEL LOW), the framer stretches the recovered clock to the nearest 20-bit boundary such that the rising edge of RXCLKx+ occurs when COMDETx is present on the associated output bus. When the Cypress or Alternate Mode Framer is enabled and half-rate receive port clocking is also enabled (RFMODE LOW and RXRATE = HIGH), the output clock is not modified when framing is detected, but a single pipeline stage may be added or subtracted from the data stream by the framer logic such that the rising edge of RXCLKx+ occurs when COMDETx is present on the associated output bus. This adjustment only occurs when the framer is enabled (RFEN = HIGH). When the framer is disabled, the clock boundaries are not adjusted, and COMDETx may be asserted during the rising edge of RXCLKx– (if an odd number of characters were received following the initial framing). Parity Generation In addition to the eleven data and status bits that are presented by each channel, an RXOPx parity output is also available on each channel. This allows the CYP15G0201DXB to support ODD parity generation for each channel. To handle a wide range of system environments, the CYP15G0201DXB supports multiple different forms of parity generation including no parity. When the decoders are enabled (DECMODE LOW), parity can be generated on ■ the RXDx[7:0] character ■ the RXDx[7:0] character and RXSTx[2:0] status. When the decoders are bypassed (DECMODE = LOW), parity can be generated on ■ the RXDx[7:0] and RXSTx[1:0] bits ■ the RXDx[7:0] and RXSTx[2:0] bits. These modes differ in the number bits which are included in the parity calculation. For all cases, only ODD parity is provided which ensures that at least one bit of the data bus is always a logic-1. Those bits covered by parity generation are listed in Table 17 on page 26. Parity generation is enabled through the 3-level select PARCTL input. When PARCTL = LOW, parity checking is disabled, and the RXOPx outputs are all disabled (High-Z). When PARCTL = MID (open) and the decoders are enabled (DECMODE LOW), ODD parity is generated for the received and decoded character in the RXDx[7:0] signals and is presented on the associated RXOPx output. When PARCTL = MID (open) and the decoders are bypassed (DECMODE = LOW), ODD parity is generated for the received and decoded character in the RXDx[7:0] and RXSTx[1:0] bit positions. When PARCTL = HIGH, ODD parity is generated for the RXDx[7:0] and the associated RXSTx[2:0] status bits. Note 17. The RXOPx outputs are also driven from the associated Output Register, but their interpretation is under the separate control of PARCTL. Document Number: 38-02058 Rev. *M Page 25 of 53 CYP15G0201DXB ■ if The contents of the data bus are valid different types of status reporting. When operated without channel bonding (RXMODE[1] = LOW, RX Modes 0 and 1), these state machines are disabled and characters are decoded directly. In RX Mode 0 the RESYNC (111b) status is never reported. In RX Mode 1, neither the RESYNC (111b) or Channel Lock Detected (010b) status are reported. ■ The type of character present Status Type-A Receive State Machine ■ The state of receive BIST operations (regardless of the state of DECMODE) ■ Character violations This machine has four primary states: NO_SYNC, RESYNC, COULD_NOT_BOND, and IN_SYNC, as shown in Figure 2 on page 28. The IN_SYNC state can respond with multiple status types, while others can respond with only one type. ■ And channel bonding status Receive Status Bits When the 8- / 10-B decoder is enabled (DECMODE LOW), each character presented at the output register includes three associated status bits. These bits are used to identify These conditions normally overlap; that is, a valid data character received with incorrect running disparity is not reported as a valid data character. It is instead reported as a decoder violation of some specific type. This implies a hierarchy or priority level to the various status bit combinations. The hierarchy and value of each status is listed in Table 18 on page 27 when channel bonding enabled and in Table 19 on page 30 when channel bonding is disabled. Table 17. Output Register Parity Generation Receive Parity Generate Mode (PARCTL) Signal Name LOW[18] MID DECMODE = LOW DECMODE LOW HIGH X[19] RXSTx[2] RXSTx[1] X X RXSTx[0] X RXDx[0] X X X X RXDx[1] X X X RXDx[2] X X X RXDx[3] X X X RXDx[4] X X X RXDx[5] X X X RXDx[6] X X X RXDx[7] X X X Receive Synchronization State Machine When Channel Bonding is Enabled Each receive channel contains a Receive Synchronization State Machine. This machine handles loss and recovery of bit, channel, and word framing, and part of the control for channel bonding. This state machine is enabled whenever the receive channels are configured for channel bonding (RXMODE[1] LOW). Separate forms of the state machine exist for the two Status Type-B Receive State Machine This machine has four primary states: NO_SYNC, RESYNC, IN_SYNC, and RESYNC_IN_SYNC, as shown in Figure 3 on page 29. Some of these state can respond with only one status value, while others can respond with multiple status types. BIST Status State Machine When a receive path is enabled to look for and compare the received data stream with the BIST pattern, the RXSTx[2:0] bits identify the present state of the BIST compare operation. Within these status decodes, there are three modes of status reporting. The two normal or data status reporting modes (Type A and Type B) are selectable through the RXMODE[0] input. These status types allow compability with legacy systems, while allowing full reporting in new systems. The third status mode is used for reporting receive BIST status and progress. These status values are generated in part by the Receive Synchronization State Machine, and are listed in Table 18 on page 27. The receive status when the channels are operated independently with channel bonding disabled is shown in Table 19 on page 30. The receive status when Receive BIST is enabled is shown in Table 20 on page 30. The BIST state machine has multiple states, as shown in Figure 4 on page 31 and Table 18 on page 27. When the receive PLL detects an out-of-lock condition, the BIST state is forced to the Start-of-BIST state, regardless of the present state of the BIST state machine. If the number of detected errors ever exceeds the number of valid matches by greater than 16, the state machine is forced to the WAIT_FOR_BIST state where it monitors the interface for the first character of the next BIST sequence (D0.0). Also, if the Elasticity Buffer ever hits and overflow/underflow condition, the status is forced to the BIST_START until the buffer is re centered (approximately nine character periods). To ensure compatibility between the source and destination BIST operating modes, the sending and receiving ends of the link must use the same receive clock setup (RXCKSEL = MID or RXCKSEL MID. Notes 18. Receive path parity output drivers (RXOPx) are disabled (High-Z) when PARCTL = LOW. 19. When the Decoder is bypassed (DECMODE = LOW) and BIST is not enabled (Receive BIST Latch output is HIGH), RXSTx[2] is driven to a logic-0, except when the character in the output buffer is a framing character. Document Number: 38-02058 Rev. *M Page 26 of 53 CYP15G0201DXB Table 18. Receive Character Status Bits when Channel Bonding is Enabled Description RXSTx[2:0] Priority Type-A Status Type-B Status 000 7 Normal Character Received. The valid data character on the output bus meets all the formatting requirements of data characters listed in Table 24 on page 43. 001 7 Special Code Detected. The valid special character on the output bus meets all the formatting requirements of special code characters listed in Table 25 on page 47, but is not the presently selected framing character or a Decoder violation indication. 010 2 Channel Lock Detected. Asserts when the bonded Receive Elasticity Buffer Underrun/Overrun Error. The receive buffer was not able to add/drop channels have detected RESYNC within the allotted window. Presented only on the last cycle before a K28.5 or framing character. aligned data is presented. 011 5 Framing Character Detected. This indicates that a character matching the patterns identified as a framing character (as selected by FRAMCHAR) was detected. The decoded value of this character is present in the associated output bus. 100 4 Codeword Violation. The character on the output bus is a C0.7. This indicates that the received character cannot be decoded into any valid character. 101 1 Loss of Sync. The character on the bus is invalid, due to an event that has caused the receive channels to lose synchronization. When channel bonding is enabled, this indicates that one or more channels have either lost bit synchronization (loss of character framing), or that the bonded channels are no longer in proper character alignment. When the channels are operated independently (with the Decoder enabled), this indicates a PLL Out of Lock condition. 110 6 Running Disparity Error. The character on the output bus is a C4.7, C1.7, or C2.7. 111 3 Resync. The receiver state machine is in the Resynchronization state. In this state the data on the output bus reflects the presently decoded FRAMCHAR. Document Number: 38-02058 Rev. *M Loss of Sync. The character on the bus is invalid, due to an event that has caused the receive channels to lose synchronization. When channel bonding is enabled, this indicates that one or more channels have either lost bit synchronization (loss of character framing), or that the bonded channels are no longer in proper character alignment. When the channels are operated independently (with the Decoder enabled), this indicates a PLL Out of lock condition. Also used to indicate receive Elasticity buffer underflow/ overflow errors. Page 27 of 53 CYP15G0201DXB Figure 2. Status Type-A Receive State Machine Reset NO_SYNC 5 IN_SYNC RXSTx=101 6 4 3 4 COULD_NOT_BOND RXSTx=101 1 RESYNC RXSTx=111 2 # 1 2 3 4 5 6 State Transition Conditions Deskew window expired FRAMCHAR detected (Elasticity buffer under/overrun) OR (RX PLL Loss of Lock) OR (Any Decoder Error) Four consecutive FRAMCHAR detected (Elasticity buffer under/overrun) OR (RX PLL Loss of lock) OR (four consecutive decoder errors) OR (invalid minus valid = 4) Valid character other than a FRAMCHAR Document Number: 38-02058 Rev. *M Page 28 of 53 CYP15G0201DXB Figure 3. Status Type-B Receive State Machine Reset RXSTx = 101 IN_SYNC NO_SYNC 5 RXSTx = 010 6 4 1 RXSTx = 010 4 5 6 7 7 RXSTx = 101 4 3 RXSTx = 111 RESYNC_IN_SYNC RESYNC RXSTx=011 RXSTx=111 2 # 1 2 3 6 2 Conditions (Channels did not bond) AND (deskew window expired) OR (decoder error) FRAMCHAR Detected (elasticity buffer under/overrun) OR (RX PLL loss of lock) OR (any decoder error) OR ((channels did not bond) and (deskew window expired)) Four consecutive FRAMCHAR detected (Elasticity buffer under/overrun) OR (RX PLL loss of lock) OR (four consecutive decoder errors) OR (invalid minus valid = 4) Last FRAMCHAR before a valid character AND bonded (Elasticity buffer under/overrun) OR (RX PLL loss of lock) Document Number: 38-02058 Rev. *M Page 29 of 53 CYP15G0201DXB Table 19. Receive Character Status when Channels are Operated in Independent Mode (RXMODE[1:0] = LL or H) RXSTx[2:0] Priority Type-A Status Type-B Status 000 7 Normal Character Received. The valid data character with the correct running disparity received 001 7 Special Code Detected. Special code other than the selected framing character or decoder violation received 010 2 Receive Elasticity Buffer underrun/overrun INVALID error. The receive elasticity buffer was not able to add/drop a K28.5 or framing character. 011 5 Framing Character Detected. This indicates that a character matching the patterns identified as a framing character was detected. The decoded value of this character is present on the associated output bus. 100 4 Codeword Violation. The character on the output bus is a C0.7. This indicates that the received character cannot be decoded into any valid character. 101 1 PLL Out Of Lock Indication 110 6 Running Disparity Error. The character on the output bus is a C4.7, C1.7 or C2.7 111 3 INVALID Table 20. Receive Character Status when Channels are Operated to Receive BIST Data Receive BIST Status (Receive BIST = Enabled) RXSTx[2:0] Priority 000 7 BIST Data Compare. Character compared correctly 001 7 BIST Command Compare. Character compared correctly 010 2 BIST Last Good. Last Character of BIST sequence detected and valid. 011 5 RESERVED for TEST 100 4 BIST Last Bad. Last Character of BIST sequence detected invalid. 101 1 BIST Start. Receive BIST is enabled on this channel, but character compares have not yet commenced. This also indicates a PLL Out of Lock condition, and Elasticity Buffer overflow/underflow conditions. 110 6 BIST Error. While comparing characters, a mismatch was found in one or more of the decoded character bits. 111 3 BIST Wait. The receiver is comparing characters. but has not yet found the start of BIST character to enable the LFSR. Document Number: 38-02058 Rev. *M Page 30 of 53 CYP15G0201DXB JTAG Support JTAG ID The CYP15G0201DXB contains a JTAG port to allow system level diagnosis of device interconnect. Of the available JTAG modes, only boundary scan is supported. This capability is present only on the LVTTL inputs and outputs and the REFCLK clock input. The high-speed serial inputs and outputs are not part of the JTAG test chain. The JTAG device ID for the CYP15G0201DXB is ‘1C80C069’x. 3-Level Select Inputs Each 3-Level select input reports as two bits in the scan register. These bits report the LOW, MID, and HIGH state of the associated input as 00, 10, and 11, respectively. Figure 4. Receive BIST State Machine Monitor Data Received RXSTx = BIST_WAIT (111) RX PLL Out of Lock RXSTx = BIST_START (101) Elasticity Buffer Error Yes No Receive BIST Detected LOW RXSTx = BIST_START (101) Start of BIST Detected No Yes, RXSTx = BIST_DATA_COMPARE (000)/ BIST_COMMAND_COMPARE(001) Compare Next Character RXSTx = Match BIST_COMMAND_COMPARE (001) Mismatch Yes Command Auto-Abort Condition Data or Command No Data End-of-BIST State End-of-BIST State Yes, RXSTx = BIST_LAST_BAD (100) Yes, RXSTx = BIST_LAST_GOOD (010) RXSTx = BIST_DATA_COMPARE (000) No No, RXSTx = BIST_ERROR (110) Document Number: 38-02058 Rev. *M Page 31 of 53 CYP15G0201DXB Maximum Ratings Static discharge voltage (per MIL-STD-883, Method 3015) .......................... > 2000 V Exceeding maximum ratings may shorten the useful life of the device. User guidelines are not tested. Latch-up current ..................................................... > 200 mA Storage temperature................................. –65 °C to +150 °C Ambient temperature with Power applied........................................... –55 °C to +125 °C Power-up requirements: The CYP15G0201DXB requires one power-supply. The voltage on any input or I/O pin cannot exceed the power pin during power-up. Operating Range Supply voltage to ground potential ...............–0.5 V to +3.8 V Range DC voltage applied to LVTTL outputs in High-Z state ..................................... –0.5 V to VCC + 0.5 V Commercial Output current into LVTTL outputs (LOW) ................... 60 mA Industrial Ambient Temp VCC 0 °C to +70 °C +3.3 V ±5% –40 °C to +85 °C +3.3 V ±5% DC input voltage .................................. –0.5 V to VCC + 0.5 V CYP15G0201DXB DC Electrical Characteristics Over the Operating Range Parameter Description Test Conditions Min Max Unit 2.4 VCC V 0 0.4 V LVTTL-compatible Outputs VOHT Output HIGH voltage IOH = 4 mA, VCC = Min VOLT Output LOW voltage IOL = 4 mA, VCC = Min IOST Output short circuit current VOUT = 0 V[20] IOZL High-Z output leakage current –20 –100 mA –20 20 µA 2.0 VCC + 0.3 V LVTTL-compatible Inputs VIHT Input HIGH voltage VILT Input LOW voltage IIHT Input HIGH current IILT Input LOW current Other Inputs, VIN = 0.0 V – –40 µA IIHPDT Input HIGH current with internal pull-down VIN = VCC – +200 µA IILPUT Input LOW current with internal pull-up VIN = 0.0 V – –200 µA –0.5 0.8 V REFCLK Input, VIN = VCC – 1.5 mA Other Inputs, VIN = VCC – +40 µA REFCLK Input, VIN = 0.0 V – –1.5 mA LVDIFF Inputs: REFCLK± VDIFF[21] Input differential voltage 400 VCC mV VIHHP Highest input HIGH voltage 1.0 VCC V VILLP Lowest input LOW voltage 0.0 VCC/2 V VCOM[22] Common mode range 1.0 VCC – 1.2V V VCC V 3-Level Inputs VIHH 3-level input HIGH voltage Min VCC Max 0.87 × VCC VIMM 3-level input MID voltage Min VCC Max 0.47 × VCC 0.53 × VCC VILL 3-level input LOW voltage Min VCC Max IIHH Input HIGH current VIN = VCC IIMM Input MID current VIN = VCC/2 IILL Input LOW current VIN = GND – V 0.0 0.13 × VCC V – 200 µA –50 50 µA –200 µA Notes 20. Tested one output at a time, output shorted for less than one second, less than 10% duty cycle. 21. This is the minimum difference in voltage between the true and complement inputs required to ensure detection of a logic-1 or logic-0. A logic-1 exists when the true (+) input is more positive than the complement (-) input. A logic-0 exists when the complement (-) input is more positive than true (+) input. 22. The common mode range defines the allowable range of REFCLK+ and REFCLK- when REFCLK+ = REFCLK-. This marks the zero-crossing between the true and complement inputs as the signal switches between a logic-1 and a logic-0. Document Number: 38-02058 Rev. *M Page 32 of 53 CYP15G0201DXB CYP15G0201DXB DC Electrical Characteristics (continued) Over the Operating Range Parameter Description Test Conditions Min Max Unit VCC – 0.5 VCC – 0.2 V Differential CML Serial Outputs: OUTA1±, OUTA2±, OUTB1±, OUTB2± 100 differential load VOHC Output HIGH voltage (VCC referenced) VOLC Output LOW voltage (VCC referenced) VODIF Output differential voltage |(OUT+) (OUT)| 150 differential load VCC – 0.5 VCC – 0.2 V 100 differential load VCC – 1.4 VCC – 0.7 V 150 differential load VCC – 1.4 VCC – 0.7 V 100 differential load 450 900 mV 150 differential load 560 1000 mV 100 1200 mV – VCC V Differential Serial Line Receiver Inputs: INA1±, INA2±, INB1±, INB2± VDIFFS[21] Input differential voltage |(IN+) (IN)| VIHE Highest input HIGH voltage VILE Lowest input LOW voltage IIHE Input HIGH current VIN = VIHE Max IILE Input LOW current VIN = VILE Min VCOM[23, 24] Common mode input range VCC – 2.0 – V – 1350 µA –700 – µA VCC–1.95 VCC – 0.05 Power Supply ICC Power supply current REFCLK = Max Commercial ICC Power supply current REFCLK = 125 MHz Commercial V Typ[23] Max[22] Unit 570 700 mA 710 mA 700 mA 710 mA Industrial 570 Industrial AC Test Loads and Waveforms [27] [27] (a) LVTTL Output Test Load (b) CML Output Test Load 3.3 V R1 R1 = 590 R2 = 435 CL CL 7 pF (Includes fixture and probe capacitance) RL = 100 R2 [28] (c) LVTTL Input Test Waveform (d) CML/LVPECL Input Test Waveform 3.0 V Vth = 1.4 V GND 2.0 V 2.0 V 0.8 V 0.8 V 1 ns RL Vth = 1.4 V VILE 1 ns VIHE VIHE 80% 80% 20% 270 ps 20% VILE 270 ps Notes 23. The common mode range defines the allowable range of INPUT+ and INPUT when INPUT+ = INPUT. This marks the zero-crossing between the true and complement inputs as the signal switches between a logic-1 and a logic-0. 24. Not applicable for AC-coupled interfaces. For AC-coupled interfaces, VDIFFS requirement still needs to be satisfied. 25. Maximum ICC is measured with VCC = MAX, RXCKSEL = LOW, with all TX and RX channels and Serial Line Drivers enabled, sending a continuous alternating 01 pattern to the associated receive channel, and outputs unloaded. 26. Typical ICC is measured under similar conditions except with VCC = 3.3 V, TA = 25 °C, RXCKSEL = LOW, with all TX and RX channels enabled and one Serial Line Driver per transmit channel sending a continuous alternating 01 pattern to the associated receive channel. The redundant outputs on each channel are powered down and the parallel outputs are unloaded. 27. Cypress uses constant current (ATE) load configurations and forcing functions. This figure is for reference only. 5 pF differential load reflects tester capacitance, and is recommended at low data rates only. 28. The LVTTL switching threshold is 1.4 V. All timing references are made relative to the point where the signal edges crosses this threshold voltage. Document Number: 38-02058 Rev. *M Page 33 of 53 CYP15G0201DXB CYP15G0201DXB AC Characteristics Over the Operating Range Parameter Description Min Max Unit Transmitter LVTTL Switching Characteristics fTS TXCLKx Clock Frequency 19.5 150 MHz tTXCLK TXCLKx Period 6.66 51.28 ns tTXCLKH[29] TXCLKx HIGH Time 2.2 – ns tTXCLKL[29] tTXCLKR[29, 30, 31] tTXCLKF[29, 30, 31] TXCLKx LOW Time 2.2 – ns TXCLKx Rise Time 0.2 1.7 ns TXCLKx Fall Time 0.2 1.7 ns tTXDS Transmit Data Setup Time toTXCLKx (TXCKSEL LOW) 1.7 – ns tTXDH Transmit Data Hold Time from TXCLKx(TXCKSEL LOW) 0.8 – ns fTOS TXCLKO Clock Frequency = 1x or 2x REFCLK Frequency 19.5 150 MHz tTXCLKO TXCLKO Period 6.66 51.28 ns tTXCLKOD+ TXCLKO+ Duty Cycle with 60% HIGH time –1.0 +0.5 ns tTXCLKOD– TXCLKO– Duty Cycle with 40% HIGH time –0.5 +1.0 ns Receiver LVTTL Switching Characteristics fRS RXCLKx Clock Output Frequency 9.75 150 MHz tRXCLKP RXCLKx Period 6.66 102.56 ns RXCLKx HIGH Time (RXRATE = LOW) 2.33[29] 26.64 ns RXCLKx HIGH Time (RXRATE = HIGH) 5.66 52.28 ns RXCLKx LOW Time (RXRATE = LOW) 2.33[29] 26.64 ns RXCLKx LOW Time (RXRATE = HIGH) 5.66 52.28 ns tRXCLKD RXCLKx Duty Cycle centered at 50% –1.0 +1.0 ns tRXCLKR[29] RXCLKx Rise Time 0.3 1.2 ns tRXCLKF[29] tRXDV–[32] RXCLKx Fall Time tRXCLKH tRXCLKL tRXDV+[32] 0.3 1.2 ns Status and Data Valid Time to RXCLKx (RXCKSEL = HIGH or MID) 5UI – 1.5 – ns Status and Data Valid Time to RXCLKx(Half Rate Recovered Clock) 5UI – 1.0 – ns Status and Data Valid Time From RXCLKx(RXCKSEL = HIGH or MID) 5UI – 1.8 – ns Status and Data Valid Time From RXCLKx(Half Rate Recovered Clock) 5UI – 2.3 – ns REFCLK Switching Characteristics Over the Operating Range fREF REFCLK Clock Frequency 19.5 150 MHz tREFCLK REFCLK Period 6.66 51.28 ns tREFH REFCLK HIGH Time (TXRATE = HIGH) 5.9 – ns REFCLK HIGH Time (TXRATE = LOW) 2.9[29] – ns REFCLK LOW Time (TXRATE = HIGH) 5.9 – ns REFCLK LOW Time (TXRATE = LOW) 2.9[29] – ns tREFL Notes 29. Tested initially and after any design or process changes that may affect these parameters, but not 100% tested. 30. The ratio of rise time to falling time must not vary by greater than 2:1. 31. For an operating frequency, neither rise or fall specification can be greater than 20% of the clock-cycle period or the datasheet maximum time. 32. Parallel data output specifications are only valid if all inputs or outputs are loaded with similar DC and AC loads. Document Number: 38-02058 Rev. *M Page 34 of 53 CYP15G0201DXB CYP15G0201DXB AC Characteristics (continued) Over the Operating Range Min Max Unit tREFD[33] Parameter REFCLK duty cycle 30 70 % tREFR[29, 30, 31] tREFF[29, 30, 31] REFCLK rise time (20% to 80%) – 2 ns REFCLK fall time (20% to 80%) – 2 ns tTREFDS Transmit data setup time toREFCLK (TXCKSEL LOW) 1.7 – ns tTREFDH Transmit data hold time from REFCLK(TXCKSEL LOW) 0.8 – ns tRREFDA[34] Receive data access time from REFCLK (RXCKSEL LOW) – 9.5 ns tRREFDV Receive data valid time from REFCLK(RXCKSEL LOW) 2.5 – ns tREFADV– Received data valid time to RXCLKA (RXCKSEL = LOW) 10UI – 4.7 – ns tREFADV+ Received data valid time from RXCLKA (RXCKSEL = LOW) 0.5 – ns tREFCDV– Received data valid time to RXCLKC (RXCKSEL = LOW) 10UI – 4.3 – ns tREFCDV+ Received data valid time from RXCLKC (RXCKSEL = LOW) –0.2 – ns REFCLK frequency referenced to extracted received clock frequency –0.02 +0.02 % tREFRX [29, 31] Description Transmit Serial Outputs and TX PLL Characteristics tB Bit Time tRISE[29] CML Output Rise Time 20% to 80% (CML test load) tFALL[29] CML Output Fall Time 80% to 20% (CML test load) 5100 666 ps SPDSEL = HIGH 60 270 ps SPDSEL = MID 100 500 ps SPDSEL = LOW 180 1000 ps SPDSEL = HIGH 60 270 ps SPDSEL = MID 100 500 ps SPDSEL = LOW 180 1000 ps tDJ[29, 35, 37] Deterministic Jitter (peak-peak) IEEE 802.3z – 25 ps tRJ[29, 36, 37] Random jitter () IEEE 802.3z – 11 ps tTXLOCK Transmit PLL lock to REFCLK – 200 us Receive PLL lock to input data stream (cold start) – 376K UI[39] Receive PLL lock to input data stream – 376K UI Receive PLL unlock rate – 46 UI Total jitter tolerance IEEE 802.3z[38] 600 – ps Deterministic jitter tolerance IEEE 802.3z[38] 370 – ps Max Unit Receive Serial Inputs and CDR PLL Characteristics tRXLOCK tRXUNLOCK tJTOL [37] tDJTOL[37] Capacitance [29] Parameter Description Test Conditions CINTTL TTL Input Capacitance TA = 25 °C, f0 = 1 MHz, VCC = 3.3 V 7 pF CINPECL PECL input Capacitance TA = 25 °C, f0 = 1 MHz, VCC = 3.3 V 4 pF Notes 33. The duty cycle specification is a simultaneous condition with the tREFH and tREFL parameters. This means that at faster character rates the REFCLK duty cycle cannot be as large as 30% to 70%. 34. Since this timing parameter is greater than the minimum time period of REFCLK it sets an upper limit to the frequency in which REFCLKx can be used to clock the receive data out of the output register. For predictable timing, users can use this parameter only if REFCLK period is greater than sum of tRREFDA and setup time of the upstream device. When this condition is not true, RXCLKC± or RXCLKA± (a buffered or delayed version of REFCLK when RXCKSELx = LOW) could be used to clock the receive data out of the device. 35. While sending continuous K28.5 s, outputs loaded to a balanced 100 load, measured at the cross point of the differential outputs over the operating range. 36. While sending continuous K28.7 s, after 100,000 samples measured at the cross point of differential outputs, time referenced to REFCLK input, over the operating range. 37. Total jitter is calculated at an assumed BER of 1E12. Hence: Total Jitter (tJ) = (tRJ × 14) + tDJ. 38. Also meets all Jitter Generation and Jitter Tolerance requirements as specified by CPRI, ESCON, FICON, Fibre Channel and DVB-ASI. 39. Receiver UI (Unit Interval) is calculated as 1/(fREF × 20) (when RXRATE = HIGH) or 1/(fREF × 10) (when RXRATE = LOW) if no data is being received, or 1/(fREF × 20) (when RXRATE = HIGH) or 1/(fREF × 10) (when RXRATE = LOW) of the remote transmitter if data is being received. In an operating link this is equivalent to tB. Document Number: 38-02058 Rev. *M Page 35 of 53 CYP15G0201DXB Switching Waveforms For CYP15G0201DXB HOTLink II Transmitter Transmit Interface Write Timing TXCKSEL LOW tTXCLK tTXCLKH tTXCLKL TXCLKx tTXDS TXDx[7:0], TXCTx[1:0], TXOPx, SCSEL tTXDH Transmit Interface Write Timing TXCKSEL = LOW TXRATE = LOW tREFH tREFCLK tREFL REFCLK tTREFDS TXDx[7:0], TXCTx[1:0], TXOPx, SCSEL tTREFDH Transmit Interface Write Timing TXCKSEL = LOW TXRATE = HIGH tREFCLK tREFH tREFL Note 40 Note 40 REFCLK tTREFDS TXDx[7:0], TXCTx[1:0], TXOPx, SCSEL tTREFDH tTREFDS tTREFDH Transmit Interface TXCLKO Timing tREFCLK tREFH TXCKSEL = LOW TXRATE = HIGH tREFL REFCLK Note 41 tTXCLKO tTXCLKOD+ tTXCLKODNote 41 TXCLKO (internal) Notes 40. When REFCLK is configured for half-rate operation (TXRATE = HIGH) and data is captured using REFCLK instead of a TXCLKx clock (TXCKSEL = LOW), data is captured using both the rising and falling edges of REFCLK. 41. The TXCLKO output is at twice the rate of REFCLK when TXRATE = HIGH and same rate as REFCLK when TXRATE = LOW. TXCLKO does not follow the duty cycle of REFCLK. 42. The TXCLKO output remains at the character rate regardless of the state of TXRATE and does not follow the duty cycle of REFCLK. Document Number: 38-02058 Rev. *M Page 36 of 53 CYP15G0201DXB Switching Waveforms (continued) For CYP15G0201DXB HOTLink II Transmitter Transmit Interface TXCLKO Timing tREFCLK tREFH TXCKSEL = LOW TXRATE = LOW tREFL Note 42 REFCLK Note 42 tTXCLKO tTXCLKOD+ tTXCLKOD- TXCLKO Switching Waveforms For CYP15G0201DXB HOTLink II Receiver Receive Interface Read Timing RXCKSEL = LOW TXRATE = LOW tREFCLK tREFH tREFL REFCLK tRREFDA tRREFDV RXDx[7:0], RXSTx[2:0], RXOPx tREFADV+ tREFCDV+ tREFADVtREFCDV- RXCLKA RXCLKC+ Note 43 Receive Interface Read Timing RXCKSEL = LOW TXRATE = HIGH tREFCLK tREFH tREFL REFCLK tRREFDV tRREFDA tRREFDA RXDx[7:0], RXSTx[2:0], RXOPx tREFADV+ tREFCDV+ RXCLKA RXCLKC+ tREFADVt-REFCDVNote 44 Note 43 Notes 43. RXCLKA is delayed in phase from REFCLK, and are different in phase from each other. 44. When operated with a half-rate REFCLK, the setup and hold specifications for data relative to RXCLKA are relative to both rising and falling edges of the respective clock output. Document Number: 38-02058 Rev. *M Page 37 of 53 CYP15G0201DXB Switching Waveforms (continued) For CYP15G0201DXB HOTLink II Receiver Receive Interface Read Timing RXCKSEL = HIGH or MID RXRATE = LOW tRXCLKP tRXCLKH tRXCLKL RXCLKx+ RXCLKx– tRXDVRXDx[7:0], RXSTx[2:0], RXOPx tRXDV+ Receive Interface Read Timing RXCKSEL = HIGH or MID RXRATE = HIGH tRXCLKP tRXCLKH tRXCLKL RXCLKx+ RXCLKx– tRXDVRXDx[7:0], RXSTx[2:0], RXOPx tRXDV+ Document Number: 38-02058 Rev. *M Page 38 of 53 CYP15G0201DXB Table 21. Package Coordinate Signal Allocation Ball ID Signal Name Signal Type Ball ID Signal Name Signal Type Ball ID Signal Name Signal Type A1 VCC POWER C5 RXLE LVTTL IN PU E9 TXOPB LVTTL IN PU A2 INA2+ CML IN C6 RXRATE LVTTL IN PD E10 TXPERB LVTTL OUT A3 OUTA2– CML OUT C7 GND GROUND E11 TXCKSEL 3-LEVEL SEL A4 VCC POWER C8 GND GROUND E12 RXCKSEL 3-LEVEL SEL A5 INA1+ CML IN C9 SPDSEL 3-LEVEL SEL E13 TRSTZ LVTTL IN PU A6 OUTA1– CML OUT C10 PARCTL 3-LEVEL SEL E14 TMS LVTTL IN PU A7 VCC POWER C11 RFMODE A8 VCC POWER C12 VCC CML IN C13 CML OUT C14 A9 INB2+ A10 OUTB2– 3-LEVEL SEL F1 POWER F2 DECMODE 3-LEVEL SEL OELE SDASEL 3-LEVEL SEL F3 RXCLKC+ BOE[2] LVTTL IN PU F4 RXSTA[2] LVTTL OUT LVTTL OUT LVTTL IN PU LVTTL 3-S OUT A11 VCC POWER D1 VCC POWER F5 RXSTA[1] A12 INB1+ CML IN D2 VCC POWER F6 GND A13 OUTB1– CML OUT D3 NC Not Connected F7 GND GROUND A14 VCC POWER D4 TXRATE LVTTL IN PD F8 GND GROUND B1 TDO B2 INA2– B3 OUTA2+ B4 VCC B5 INA1– B6 OUTA1+ LVTTL 3-S OUT CML IN GROUND D5 RXMODE[1] 3-LEVEL SEL F9 GND GROUND D6 RXMODE[0] 3-LEVEL SEL F10 TXDB[4] LVTTL IN CML OUT D7 GND GROUND F11 TXDB[3] LVTTL IN POWER D8 GND GROUND F12 TXDB[2] LVTTL IN CML IN D9 TCLK LVTTL IN PD F13 TXDB[1] LVTTL IN CML OUT D10 TDI LVTTL IN PU F14 TXDB[0] LVTTL IN B7 NC Not Connected D11 INSELB LVTTL IN G1 VCC POWER B8 NC Not Connected D12 INSELA LVTLL IN G2 NC Not Connected B9 INB2– B10 OUTB2+ CML IN D13 VCC POWER G3 GND GROUND CML OUT D14 VCC POWER G4 GND GROUND B11 VCC POWER E1 BISTLE LVTTL IN PU G5 GND GROUND B12 INB1– CML IN E2 FRAMCHAR 3-LEVEL SEL G6 GND GROUND B13 OUTB1+ B14 BOE[3] C1 NC C2 RFEN C3 VCC C4 LPEN CML OUT E3 TXMODE[1] 3-LEVEL SEL G7 GND GROUND LVTTL IN PU E4 TXMODE[0] 3-LEVEL SEL G8 GND GROUND Not Connected E5 BOE[0] LVTTL IN PU G9 GND GROUND LVTTL IN PD E6 BOE[1] LVTTL IN PU G10 GND GROUND POWER E7 GND GROUND G11 GND GROUND LVTTL IN PD E8 GND GROUND G12 GND GROUND G13 NC Not Connected K4 RXDA[6] LVTTL OUT M9 TXRSTn G14 VCC POWER K5 TXDA[4] LVTTL OUT M10 NC H1 VCC POWER K6 TXCLKA LVTTL IN PD M11 RXSTB[0] H2 NC Not Connected K7 GND GROUND M12 VCC H3 GND GROUND K8 GND GROUND M13 RXDB[5] LVTTL OUT H4 GND GROUND K9 NC Not Connected M14 RXDB[6] LVTTL OUT H5 GND GROUND K10 RXOPB LVTTL 3-S OUT N1 RXCLKA+ LVTTL I/O PD H6 GND GROUND K11 RXCLKB+ LVTTL I/O PD N2 TXCTA[0] LVTTL IN H7 GND GROUND K12 RXCLKB– LVTTL I/O PD N3 TXDA[6] LVTTL IN H8 GND GROUND K13 LFIB LVTTL OUT N4 VCC POWER Document Number: 38-02058 Rev. *M LVTTL IN PU Not Connected LVTTL OUT POWER Page 39 of 53 CYP15G0201DXB Table 21. Package Coordinate Signal Allocation (continued) Ball ID Signal Name Ball ID Signal Name H9 GND GROUND K14 TXCLKB H10 H11 GND GROUND L1 GND GROUND L2 H12 GND H13 NC GROUND L3 Not Connected L4 H14 VCC J1 RXSTA[0] POWER L5 LVTTL OUT L6 Signal Type J2 RXOPA J3 RXDA[0] LVTTL OUT LVTTL 3-S OUT J4 RXDA[1] LVTTL OUT J5 RXDA[2] LVTTL OUT Ball ID Signal Name LVTTL IN PD N5 TXDA[1] VCC POWER N6 NC Not Connected VCC POWER N7 NC Not Connected RXDA[7] LVTTL OUT N8 NC Not Connected LFIA LVTTL OUT N9 REFCLK– TXDA[3] LVTTL IN N10 TXCLKO+ TXOPA LVTTL IN N11 VCC Signal Type Signal Type LVTTL IN PECL IN LVTTL OUT POWER L7 GND GROUND N12 RXDB[2] LVTTL OUT L8 GND GROUND N13 RXDB[3] LVTTL OUT L9 SCSEL LVTTL IN N14 RXDB[4] L10 RXSTB[2] LVTTL OUT P1 VCC LVTTL OUT POWER J6 GND GROUND L11 RXSTB[1] LVTTL OUT P2 TXDA[7] LVTTL IN J7 GND GROUND L12 RXDB[7] LVTTL OUT P3 TXDA[5] LVTTL IN J8 GND GROUND L13 VCC POWER P4 VCC POWER J9 GND GROUND L14 VCC POWER P5 TXDA[0] LVTTL IN J10 TXCTB[0] LVTTL IN M1 RXCLKA– LVTTL I/O PD P6 NC Not Connected J11 TXCTB[1] LVTTL IN M2 TXCTA[1] LVTTL IN P7 VCC POWER J12 TXDB[7] LVTTL IN M3 VCC POWER P8 VCC POWER J13 TXDB[6] LVTTL IN M4 NC Not Connected P9 REFCLK+ PECL IN LVTTL OUT J14 TXDB[5] LVTTL IN M5 TXDA[2] LVTTL IN P10 TXCLKO– K1 RXDA[3] LVTTL OUT M6 TXPERA LVTTL OUT P11 VCC K2 RXDA[4] LVTTL OUT M7 GND GROUND P12 RXDB[1] LVTTL OUT K3 RXDA[5] LVTTL OUT M8 GND GROUND P13 RXDB[0] LVTTL OUT P14 VCC POWER POWER Document Number: 38-02058 Rev. *M Page 40 of 53 CYP15G0201DXB X3.230 Codes and Notation Conventions Information to be transmitted over a serial link is encoded eight bits at a time into a 10-bit transmission character and then sent serially, bit by bit. Information received over a serial link is collected ten bits at a time, and those transmission characters that are used for data (data characters) are decoded into the correct eight-bit codes. The 10-bit transmission code supports all 256 8-bit combinations. Some of the remaining transmission characters (special characters) are used for functions other than data transmission. The primary rationale for use of a transmission code is to improve the transmission characteristics of a serial link. The encoding defined by the transmission code ensures that sufficient transitions are present in the serial bit stream to make clock recovery possible at the receiver. Such encoding also greatly increases the likelihood of detecting any single or multiple bit errors that may occur during transmission and reception of information. In addition, some special characters of the transmission code selected by fibre channel standard contain a distinct and easily recognizable bit pattern (the special character COMMA) that assists a receiver in achieving character alignment on the incoming bit stream. Notation Conventions The documentation for the 8- /10-B transmission code uses letter notation for the bits in an 8-bit byte. Fibre Channel Standard notation uses a bit notation of A, B, C, D, E, F, G, H for the 8-bit byte for the raw 8-bit data, and the letters a, b, c, d, e, i, f, g, h, j for encoded 10-bit data. There is a correspondence between bit A and bit a, B and b, C and c, D and d, E and e, F and f, G and g, and H and h. Bits i and j are derived, respectively, from (A,B,C,D,E) and (F,G,H). The bit labeled A in the description of the 8- /10-B transmission code corresponds to bit 0 in the numbering scheme of the FC-2 specification, B corresponds to bit 1, as shown in the following. FC-2 bit designation—76543210 HOTLink D/Q designation—76543210 8B/10B bit designation—HGFEDCBA To clarify this correspondence, the following example shows the conversion from an FC-2 valid data byte to a transmission character (using 8- /10-B transmission code notation) FC-2 45 Bits: 7654 3210 0100 0101 Converted to 8- /10-B notation (note carefully that the order of bits is reversed): Data Byte Name D5.2 Bits:ABCDEFGH 10100 010 Translated to a transmission Character in the 8- /10-B Transmission Code: Bits: abcdeifghj 1010010101 Each valid transmission character of the 8- /10-B transmission code has been given a name using the following convention: cxx.y, where c is used to show whether the transmission character is a data character (c is set to D, and SC/D = LOW) or a special character (c is set to K, and SC/D = HIGH). When c is Document Number: 38-02058 Rev. *M set to D, xx is the decimal value of the binary number composed of the bits E, D, C, B, and A in that order, and the y is the decimal value of the binary number composed of the bits H, G, and F in that order. When c is set to K, xx and y are derived by comparing the encoded bit patterns of the special character to those patterns derived from encoded valid data bytes and selecting the names of the patterns most similar to the encoded bit patterns of the special character. Under the above conventions, the transmission character used for previous examples, is referred to by the name D5.2. The special character K29.7 is so named because the first six bits (abcdei) of this character make up a bit pattern similar to that resulting from the encoding of the unencoded 11101 pattern (29), and because the second four bits (fghj) make up a bit pattern similar to that resulting from the encoding of the unencoded 111 pattern (7).This definition of the 10-bit transmission code is based on the following references. A.X. Widmer and P.A. Franaszek. “A DC-Balanced, Partitioned-Block, 8- /10-B Transmission Code” IBM Journal of Research and Development, 27, No. 5: 440451 (September, 1983). U.S. Patent 4,486,739. Peter A. Franaszek and Albert X. Widmer. “Byte-Oriented DC Balanced (0.4) 8- /10-B Partitioned Block Transmission Code” (December 4, 1984). Fibre Channel Physical and Signaling Interface (ANS X3.2301994 ANSI FCPH Standard). IBM Enterprise Systems Architecture/390 ESCON I/O Interface (document number SA227202). 8- /10-B Transmission Code The following information describes how the tables shall be used for both generating valid transmission characters (encoding) and checking the validity of received transmission characters (decoding). It also specifies the ordering rules to be followed when transmitting the bits within a character and the characters within any higher-level constructs specified by the standard. Transmission Order Within the definition of the 8- /10-B transmission code, the bit positions of the transmission characters are labeled a, b, c, d, e, i, f, g, h, j. Bit “a” is transmitted first followed by bits b, c, d, e, i, f, g, h, and j in that order. Note that bit i is transmitted between bit e and bit f, rather than in alphabetical order. Valid and Invalid Transmission Characters The following tables define the valid data characters and valid special characters (K characters), respectively. The tables are used for both generating valid transmission characters (encoding) and checking the validity of received transmission characters (decoding). In the tables, each valid-data-byte or special-character-code entry has two columns that represent two (not necessarily different) transmission characters. The two columns correspond to the current value of the running disparity (“Current RD” or “Current RD+”). Running disparity is a binary parameter with either a negative () or positive (+) value. After powering on, the transmitter may assume either a positive or negative value for its initial running disparity. Upon transmission of any transmission character, the transmitter selects the proper version of the transmission character based Page 41 of 53 CYP15G0201DXB the transmitter’s current running disparity for the next valid data byte or special character byte to be encoded and transmitted. Table 22 shows naming notations and examples of valid transmission characters. on the current running disparity value, and the transmitter calculates a new value for its running disparity based on the contents of the transmitted character. Special character codes C1.7 and C2.7 can be used to force the transmission of a specific special character with a specific running disparity as required for some special sequences in X3.230. Use of the Tables for Checking the Validity of Received Transmission Characters After powering on, the receiver may assume either a positive or negative value for its initial running disparity. Upon reception of any transmission character, the receiver decides whether the transmission character is valid or invalid according to the following rules and tables and calculates a new value for its running disparity based on the contents of the received character. The column corresponding to the current value of the receiver’s running disparity is searched for the received transmission character. If the received transmission character is found in the proper column, then the transmission character is valid and the associated data byte or special character code is determined (decoded). If the received transmission character is not found in that column, then the transmission character is invalid. This is called a code violation. Independent of the transmission character’s validity, the received transmission character is used to calculate a new value of running disparity. The new value is used as the receiver’s current running disparity for the next received transmission character. The following rules for running disparity are used to calculate the new running-disparity value for transmission characters that have been transmitted and that have been received. Running disparity for a transmission character is calculated from sub-blocks, where the first six bits (abcdei) form one sub-block and the second four bits (fghj) form the other sub-block. Running disparity at the beginning of the 6-bit sub-block is the running disparity at the end of the previous transmission character. Running disparity at the beginning of the 4-bit sub-block is the running disparity at the end of the 6-bit sub-block. Running disparity at the end of the transmission character is the running disparity at the end of the 4-bit sub-block. Table 22. Valid Transmission Characters Data Byte Name Running disparity for the sub-blocks is calculated as follows: 1. Running disparity at the end of any sub-block is positive if the sub-block contains more ones than zeros. It is also positive at the end of the 6-bit sub-block if the 6-bit sub-block is 000111, and it is positive at the end of the 4-bit sub-block if the 4-bit sub-block is 0011 2. Running disparity at the end of any sub-block is negative if the sub-block contains more zeros than ones. It is also negative at the end of the 6-bit sub-block if the 6-bit sub-block is 111000, and it is negative at the end of the 4-bit sub-block if the 4-bit sub-block is 1100 3. Otherwise, running disparity at the end of the sub-block is the same as at the beginning of the sub-block Use of the Tables for Generating Transmission Characters DIN or QOUT Hex Value 765 43210 D0.0 000 00000 00 D1.0 000 00001 01 D2.0 000 00010 02 .g . . . . . . . D5.2 010 000101 45 . . . . . . . . D30.7 111 11110 FE D31.7 111 11111 FF Detection of a code violation does not necessarily show that the transmission character in which the code violation was detected is in error. Code violations may result from a prior error that altered the running disparity of the bit stream which did not result in a detectable error at the Transmission Character in which the error occurred. Table 23 shows an example of this behavior. The appropriate entry in Table 24 on page 43 for the valid data byte or Table 25 on page 47 for the special character byte for which transmission character is to be generated (encoded). The current value of the transmitter’s running disparity is used to select the transmission character from its corresponding column. For each transmission character transmitted, a new value of the running disparity is calculated. This new value shall be used as Table 23. Code Violations Resulting from Prior Errors RD Character RD Character RD Character RD Transmitted data character – D21.1 – D10.2 – D23.5 + Transmitted bit stream – 101010 1001 – 010101 0101 – 111010 1010 + Bit stream after error – 101010 1011 + 010101 0101 + 111010 1010 + Decoded data character – D21.0 + D10.2 + Code Violation + Document Number: 38-02058 Rev. *M Page 42 of 53 CYP15G0201DXB Table 24. Valid Data Characters (TXCTx[0] = 0, RXSTx[2:0] = 000) Data Byte Name Bits Current RD Current RD+ Bits Current RD Current RD+ abcdei fghj Data Byte Name HGF EDCBA abcdei fghj HGF EDCBA abcdei fghj abcdei fghj D0.0 000 00000 100111 0100 011000 1011 D0.1 001 00000 100111 1001 011000 1001 D1.0 000 00001 011101 0100 100010 1011 D1.1 001 00001 011101 1001 100010 1001 D2.0 000 00010 101101 0100 010010 1011 D2.1 001 00010 101101 1001 010010 1001 D3.0 000 00011 110001 1011 110001 0100 D3.1 001 00011 110001 1001 110001 1001 D4.0 000 00100 110101 0100 001010 1011 D4.1 001 00100 110101 1001 001010 1001 D5.0 000 00101 101001 1011 101001 0100 D5.1 001 00101 101001 1001 101001 1001 D6.0 000 00110 011001 1011 011001 0100 D6.1 001 00110 011001 1001 011001 1001 D7.0 000 00111 111000 1011 000111 0100 D7.1 001 00111 111000 1001 000111 1001 D8.0 000 01000 111001 0100 000110 1011 D8.1 001 01000 111001 1001 000110 1001 D9.0 000 01001 100101 1011 100101 0100 D9.1 001 01001 100101 1001 100101 1001 D10.0 000 01010 010101 1011 010101 0100 D10.1 001 01010 010101 1001 010101 1001 D11.0 000 01011 110100 1011 110100 0100 D11.1 001 01011 110100 1001 110100 1001 D12.0 000 01100 001101 1011 001101 0100 D12.1 001 01100 001101 1001 001101 1001 D13.0 000 01101 101100 1011 101100 0100 D13.1 001 01101 101100 1001 101100 1001 D14.0 000 01110 011100 1011 011100 0100 D14.1 001 01110 011100 1001 011100 1001 D15.0 000 01111 010111 0100 101000 1011 D15.1 001 01111 010111 1001 101000 1001 D16.0 000 10000 011011 0100 100100 1011 D16.1 001 10000 011011 1001 100100 1001 D17.0 000 10001 100011 1011 100011 0100 D17.1 001 10001 100011 1001 100011 1001 D18.0 000 10010 010011 1011 010011 0100 D18.1 001 10010 010011 1001 010011 1001 D19.0 000 10011 110010 1011 110010 0100 D19.1 001 10011 110010 1001 110010 1001 D20.0 000 10100 001011 1011 001011 0100 D20.1 001 10100 001011 1001 001011 1001 D21.0 000 10101 101010 1011 101010 0100 D21.1 001 10101 101010 1001 101010 1001 D22.0 000 10110 011010 1011 011010 0100 D22.1 001 10110 011010 1001 011010 1001 D23.0 000 10111 111010 0100 000101 1011 D23.1 001 10111 111010 1001 000101 1001 D24.0 000 11000 110011 0100 001100 1011 D24.1 001 11000 110011 1001 001100 1001 D25.0 000 11001 100110 1011 100110 0100 D25.1 001 11001 100110 1001 100110 1001 D26.0 000 11010 010110 1011 010110 0100 D26.1 001 11010 010110 1001 010110 1001 D27.0 000 11011 110110 0100 001001 1011 D27.1 001 11011 110110 1001 001001 1001 D28.0 000 11100 001110 1011 001110 0100 D28.1 001 11100 001110 1001 001110 1001 D29.0 000 11101 101110 0100 010001 1011 D29.1 001 11101 101110 1001 010001 1001 D30.0 000 11110 011110 0100 100001 1011 D30.1 001 11110 011110 1001 100001 1001 D31.0 000 11111 101011 0100 010100 1011 D31.1 001 11111 101011 1001 010100 1001 Document Number: 38-02058 Rev. *M Page 43 of 53 CYP15G0201DXB Table 24. Valid Data Characters (TXCTx[0] = 0, RXSTx[2:0] = 000) (continued) Data Byte Name Bits Current RD Current RD+ Bits Current RD Current RD+ abcdei fghj Data Byte Name HGF EDCBA abcdei fghj HGF EDCBA abcdei fghj abcdei fghj D0.2 010 00000 100111 0101 011000 0101 D0.3 011 00000 100111 0011 011000 1100 D1.2 010 00001 011101 0101 100010 0101 D1.3 011 00001 011101 0011 100010 1100 D2.2 010 00010 101101 0101 010010 0101 D2.3 011 00010 101101 0011 010010 1100 D3.2 010 00011 110001 0101 110001 0101 D3.3 011 00011 110001 1100 110001 0011 D4.2 010 00100 110101 0101 001010 0101 D4.3 011 00100 110101 0011 001010 1100 D5.2 010 00101 101001 0101 101001 0101 D5.3 011 00101 101001 1100 101001 0011 D6.2 010 00110 011001 0101 011001 0101 D6.3 011 00110 011001 1100 011001 0011 D7.2 010 00111 111000 0101 000111 0101 D7.3 011 00111 111000 1100 000111 0011 D8.2 010 01000 111001 0101 000110 0101 D8.3 011 01000 111001 0011 000110 1100 D9.2 010 01001 100101 0101 100101 0101 D9.3 011 01001 100101 1100 100101 0011 D10.2 010 01010 010101 0101 010101 0101 D10.3 011 01010 010101 1100 010101 0011 D11.2 010 01011 110100 0101 110100 0101 D11.3 011 01011 110100 1100 110100 0011 D12.2 010 01100 001101 0101 001101 0101 D12.3 011 01100 001101 1100 001101 0011 D13.2 010 01101 101100 0101 101100 0101 D13.3 011 01101 101100 1100 101100 0011 D14.2 010 01110 011100 0101 011100 0101 D14.3 011 01110 011100 1100 011100 0011 D15.2 010 01111 010111 0101 101000 0101 D15.3 011 01111 010111 0011 101000 1100 D16.2 010 10000 011011 0101 100100 0101 D16.3 011 10000 011011 0011 100100 1100 D17.2 010 10001 100011 0101 100011 0101 D17.3 011 10001 100011 1100 100011 0011 D18.2 010 10010 010011 0101 010011 0101 D18.3 011 10010 010011 1100 010011 0011 D19.2 010 10011 110010 0101 110010 0101 D19.3 011 10011 110010 1100 110010 0011 D20.2 010 10100 001011 0101 001011 0101 D20.3 011 10100 001011 1100 001011 0011 D21.2 010 10101 101010 0101 101010 0101 D21.3 011 10101 101010 1100 101010 0011 D22.2 010 10110 011010 0101 011010 0101 D22.3 011 10110 011010 1100 011010 0011 D23.2 010 10111 111010 0101 000101 0101 D23.3 011 10111 111010 0011 000101 1100 D24.2 010 11000 110011 0101 001100 0101 D24.3 011 11000 110011 0011 001100 1100 D25.2 010 11001 100110 0101 100110 0101 D25.3 011 11001 100110 1100 100110 0011 D26.2 010 11010 010110 0101 010110 0101 D26.3 011 11010 010110 1100 010110 0011 D27.2 010 11011 110110 0101 001001 0101 D27.3 011 11011 110110 0011 001001 1100 D28.2 010 11100 001110 0101 001110 0101 D28.3 011 11100 001110 1100 001110 0011 D29.2 010 11101 101110 0101 010001 0101 D29.3 011 11101 101110 0011 010001 1100 D30.2 010 11110 011110 0101 100001 0101 D30.3 011 11110 011110 0011 100001 1100 D31.2 010 11111 101011 0101 010100 0101 D31.3 011 11111 101011 0011 010100 1100 Document Number: 38-02058 Rev. *M Page 44 of 53 CYP15G0201DXB Table 24. Valid Data Characters (TXCTx[0] = 0, RXSTx[2:0] = 000) (continued) Data Byte Name Bits Current RD Current RD+ Bits Current RD Current RD+ abcdei fghj Data Byte Name HGF EDCBA abcdei fghj HGF EDCBA abcdei fghj abcdei fghj D0.4 100 00000 100111 0010 011000 1101 D0.5 101 00000 100111 1010 011000 1010 D1.4 100 00001 011101 0010 100010 1101 D1.5 101 00001 011101 1010 100010 1010 D2.4 100 00010 101101 0010 010010 1101 D2.5 101 00010 101101 1010 010010 1010 D3.4 100 00011 110001 1101 110001 0010 D3.5 101 00011 110001 1010 110001 1010 D4.4 100 00100 110101 0010 001010 1101 D4.5 101 00100 110101 1010 001010 1010 D5.4 100 00101 101001 1101 101001 0010 D5.5 101 00101 101001 1010 101001 1010 D6.4 100 00110 011001 1101 011001 0010 D6.5 101 00110 011001 1010 011001 1010 D7.4 100 00111 111000 1101 000111 0010 D7.5 101 00111 111000 1010 000111 1010 D8.4 100 01000 111001 0010 000110 1101 D8.5 101 01000 111001 1010 000110 1010 D9.4 100 01001 100101 1101 100101 0010 D9.5 101 01001 100101 1010 100101 1010 D10.4 100 01010 010101 1101 010101 0010 D10.5 101 01010 010101 1010 010101 1010 D11.4 100 01011 110100 1101 110100 0010 D11.5 101 01011 110100 1010 110100 1010 D12.4 100 01100 001101 1101 001101 0010 D12.5 101 01100 001101 1010 001101 1010 D13.4 100 01101 101100 1101 101100 0010 D13.5 101 01101 101100 1010 101100 1010 D14.4 100 01110 011100 1101 011100 0010 D14.5 101 01110 011100 1010 011100 1010 D15.4 100 01111 010111 0010 101000 1101 D15.5 101 01111 010111 1010 101000 1010 D16.4 100 10000 011011 0010 100100 1101 D16.5 101 10000 011011 1010 100100 1010 D17.4 100 10001 100011 1101 100011 0010 D17.5 101 10001 100011 1010 100011 1010 D18.4 100 10010 010011 1101 010011 0010 D18.5 101 10010 010011 1010 010011 1010 D19.4 100 10011 110010 1101 110010 0010 D19.5 101 10011 110010 1010 110010 1010 D20.4 100 10100 001011 1101 001011 0010 D20.5 101 10100 001011 1010 001011 1010 D21.4 100 10101 101010 1101 101010 0010 D21.5 101 10101 101010 1010 101010 1010 D22.4 100 10110 011010 1101 011010 0010 D22.5 101 10110 011010 1010 011010 1010 D23.4 100 10111 111010 0010 000101 1101 D23.5 101 10111 111010 1010 000101 1010 D24.4 100 11000 110011 0010 001100 1101 D24.5 101 11000 110011 1010 001100 1010 D25.4 100 11001 100110 1101 100110 0010 D25.5 101 11001 100110 1010 100110 1010 D26.4 100 11010 010110 1101 010110 0010 D26.5 101 11010 010110 1010 010110 1010 D27.4 100 11011 110110 0010 001001 1101 D27.5 101 11011 110110 1010 001001 1010 D28.4 100 11100 001110 1101 001110 0010 D28.5 101 11100 001110 1010 001110 1010 D29.4 100 11101 101110 0010 010001 1101 D29.5 101 11101 101110 1010 010001 1010 D30.4 100 11110 011110 0010 100001 1101 D30.5 101 11110 011110 1010 100001 1010 D31.4 100 11111 101011 0010 010100 1101 D31.5 101 11111 101011 1010 010100 1010 Document Number: 38-02058 Rev. *M Page 45 of 53 CYP15G0201DXB Table 24. Valid Data Characters (TXCTx[0] = 0, RXSTx[2:0] = 000) (continued) Data Byte Name Bits Current RD Current RD+ Bits Current RD Current RD+ abcdei fghj Data Byte Name HGF EDCBA abcdei fghj HGF EDCBA abcdei fghj abcdei fghj D0.6 110 00000 100111 0110 011000 0110 D0.7 111 00000 100111 0001 011000 1110 D1.6 110 00001 011101 0110 100010 0110 D1.7 111 00001 011101 0001 100010 1110 D2.6 110 00010 101101 0110 010010 0110 D2.7 111 00010 101101 0001 010010 1110 D3.6 110 00011 110001 0110 110001 0110 D3.7 111 00011 110001 1110 110001 0001 D4.6 110 00100 110101 0110 001010 0110 D4.7 111 00100 110101 0001 001010 1110 D5.6 110 00101 101001 0110 101001 0110 D5.7 111 00101 101001 1110 101001 0001 D6.6 110 00110 011001 0110 011001 0110 D6.7 111 00110 011001 1110 011001 0001 D7.6 110 00111 111000 0110 000111 0110 D7.7 111 00111 111000 1110 000111 0001 D8.6 110 01000 111001 0110 000110 0110 D8.7 111 01000 111001 0001 000110 1110 D9.6 110 01001 100101 0110 100101 0110 D9.7 111 01001 100101 1110 100101 0001 D10.6 110 01010 010101 0110 010101 0110 D10.7 111 01010 010101 1110 010101 0001 D11.6 110 01011 110100 0110 110100 0110 D11.7 111 01011 110100 1110 110100 1000 D12.6 110 01100 001101 0110 001101 0110 D12.7 111 01100 001101 1110 001101 0001 D13.6 110 01101 101100 0110 101100 0110 D13.7 111 01101 101100 1110 101100 1000 D14.6 110 01110 011100 0110 011100 0110 D14.7 111 01110 011100 1110 011100 1000 D15.6 110 01111 010111 0110 101000 0110 D15.7 111 01111 010111 0001 101000 1110 D16.6 110 10000 011011 0110 100100 0110 D16.7 111 10000 011011 0001 100100 1110 D17.6 110 10001 100011 0110 100011 0110 D17.7 111 10001 100011 0111 100011 0001 D18.6 110 10010 010011 0110 010011 0110 D18.7 111 10010 010011 0111 010011 0001 D19.6 110 10011 110010 0110 110010 0110 D19.7 111 10011 110010 1110 110010 0001 D20.6 110 10100 001011 0110 001011 0110 D20.7 111 10100 001011 0111 001011 0001 D21.6 110 10101 101010 0110 101010 0110 D21.7 111 10101 101010 1110 101010 0001 D22.6 110 10110 011010 0110 011010 0110 D22.7 111 10110 011010 1110 011010 0001 D23.6 110 10111 111010 0110 000101 0110 D23.7 111 10111 111010 0001 000101 1110 D24.6 110 11000 110011 0110 001100 0110 D24.7 111 11000 110011 0001 001100 1110 D25.6 110 11001 100110 0110 100110 0110 D25.7 111 11001 100110 1110 100110 0001 D26.6 110 11010 010110 0110 010110 0110 D26.7 111 11010 010110 1110 010110 0001 D27.6 110 11011 110110 0110 001001 0110 D27.7 111 11011 110110 0001 001001 1110 D28.6 110 11100 001110 0110 001110 0110 D28.7 111 11100 001110 1110 001110 0001 D29.6 110 11101 101110 0110 010001 0110 D29.7 111 11101 101110 0001 010001 1110 D30.6 110 11110 011110 0110 100001 0110 D30.7 111 11110 011110 0001 100001 1110 D31.6 110 11111 101011 0110 010100 0110 D31.7 111 11111 101011 0001 010100 1110 Document Number: 38-02058 Rev. *M Page 46 of 53 CYP15G0201DXB Table 25. Valid Special Character Codes and Sequences (TXCTx = special character code or RXSTx[2:0] = 001)[45, 46] S.C. Byte Name S.C. Code Name Cypress S.C. Byte Name[47] Bits HGF EDCBA Alternate S.C. Byte Name[47] Bits HGF EDCBA Current RD abcdei fghj Current RD+ abcdei fghj K28.0 C0.0 (C00) 000 00000 C28.0 (C1C) 000 11100 001111 0100 110000 1011 K28.1[48] C1.0 (C01) 000 00001 C28.1 (C3C) 001 11100 001111 1001 110000 0110 [48] K28.2 C2.0 (C02) 000 00010 C28.2 (C5C) 010 11100 001111 0101 110000 1010 K28.3 C3.0 (C03) 000 00011 C28.3 (C7C) 011 11100 001111 0011 110000 1100 K28.4 C4.0 (C04) 000 00100 C28.4 (C9C) 100 11100 001111 0010 110000 1101 K28.5[48, 49] C5.0 (C05) 000 00101 C28.5 (CBC) 101 11100 001111 1010 110000 0101 K28.6[48] C6.0 (C06) 000 00110 C28.6 (CDC) 110 11100 001111 0110 110000 1001 K28.7[48, 50] C7.0 (C07) 000 00111 C28.7 (CFC) 111 11100 001111 1000 110000 0111 K23.7 C8.0 (C08) 000 01000 C23.7 (CF7) 111 10111 111010 1000 000101 0111 K27.7 C9.0 (C09) 000 01001 C27.7 (CFB) 111 11011 110110 1000 001001 0111 K29.7 C10.0 (C0A) 000 01010 C29.7 (CFD) 111 11101 101110 1000 010001 0111 K30.7 C11.0 (C0B) 000 01011 C30.7 (CFE) 111 11110 011110 1000 100001 0111 001 00010 C2.1 (C22) 001 00010 K28.5,Dn.xxx0 +K28.5,Dn.xxx1 C0.7 (CE0) 111 00000[56] 100111 1000 011000 0111 00001[56] [48] End of Frame Sequence EOFxx[51] C2.1 (C22) Code Rule Violation and SVS Tx Pattern Exception[50, 52] C0.7 (CE0) 111 00000 K28.5[53] C1.7 (CE1) 111 00001 C1.7 (CE1) 111 001111 1010 001111 1010 +K28.5[54] C2.7 (CE2) 111 00010 C2.7 (CE2) 111 00010[56] 110000 0101 110000 0101 C4.7 (CE4) 111 00100[56] 110111 0101 001000 1010 Running Disparity Violation Pattern Exception[55] C4.7 (CE4) 111 00100 Notes 45. All codes not shown are reserved. 46. Notation for Special Character Code Name is consistent with Fibre Channel and ESCON naming conventions. Special Character Code Name is intended to describe binary information present on I/O pins. Common usage for the name can either be in the form used for describing Data patterns (that is, C0.0 through C31.7), or in hex notation (that is, Cnn where nn = the specified value between 00 and FF). 47. Both the Cypress and alternate encodings may be used for data transmission to generate specific Special Character Codes. The decoding process for received characters generates Cypress codes or Alternate codes as selected by the DECMODE configuration input. 48. These characters are used for control of ESCON interfaces. They can be sent as embedded commands or other markers when not operating using ESCON protocols. 49. The K28.5 character is used for framing operations by the receiver. It is also the pad or fill character transmitted to maintain the serial link when no user data is available. 50. Care must be taken when using this Special Character code. When a K28.7(C7.0) or SVS(C0.7) is followed by a D11.x or D20.x,an alias K28.5 sync character is created. These sequences can cause erroneous framing and should be avoided while RFEN = HIGH. 51. C2.1 = Transmit either K28.5+ or +K28.5 as determined by Current RD and modify the Transmission Character that follows, by setting its least significant bit to 1 or 0. If Current RD at the start of the following character is plus (+) the LSB is set to 0, and if Current RD is minus () the LSB becomes 1. This modification allows construction of X3.230 “EOF” frame delimiters wherein the second data byte is determined by the Current RD. For example, to send “EOFdt” the controller could issue the sequence C2.1D21.4 D21.4D21.4, and the HOTLink Transmitter sends either K28.5D21.4D21.4D21.4 or K28.5D21.5 D21.4D21.4 based on Current RD. Likewise to send “EOFdti” the controller could issue the sequence C2.1D10.4D21.4D21.4, and the HOTLink Transmitter sends either K28.5D10.4D21.4 D21.4 or K28.5D10.5D21.4 D21.4 based on Current RD. The receiver, never outputs this Special Character, since K28.5 is decoded as C5.0, C1.7, or C2.7, and the subsequent bytes are decoded as data. 52. C0.7 = Transmit a deliberate code rule violation. The code chosen for this function follows the normal Running Disparity rules. The receiver outputs this Special Character, only if the Transmission Character being decoded is not found in the tables. 53. C1.7 = Transmit Negative K28.5 (K28.5+) disregarding Current RD. The receiver outputs this Special Character, only if K28.5 is received with the wrong running disparity. The receiver outputs C1.7, if K28.5 is received with RD+, otherwise K28.5 is decoded as C5.0 or C2.7. 54. C2.7 = Transmit Positive K28.5 (+K28.5) disregarding Current RD. The receiver outputs this Special Character, only if K28.5 is received with the wrong running disparity. The receiver outputsC2.7, if +K28.5 is received with RD, otherwise K28.5 is decoded as C5.0 or C1.7. 55. C4.7 = Transmit a deliberate code rule violation to indicate a Running Disparity violation. The receiver outputs this Special Character, only if the Transmission Character being decoded is found in the tables, but Running Disparity does not match. This might indicate that an error occurred in a prior byte. 56. Supported only for data transmission. The receive status for these conditions are reported by specific combinations of receive status bits. Document Number: 38-02058 Rev. *M Page 47 of 53 CYP15G0201DXB Ordering Information Speed Ordering Code Package Name Package Type Operating Range Standard CYP15G0201DXB-BBXC BB196A Pb-Free 196-Ball Grid Array Commercial Standard CYP15G0201DXB-BBXI BB196A Pb-Free 196-Ball Grid Array Industrial Ordering Code Definitions CY P 15 02 01 DX B - BB X I/C Temperature Grade : c = Commercial; I = Industrial Pb-free Package Type (196 BGA) Silicon revision Full duplex PHY: 8B/10B Endec, Channel Bonding Number of channel 1.5 = Speed (Gbps) Standard PHY Company Code: CY = Cypress Document Number: 38-02058 Rev. *M Page 48 of 53 CYP15G0201DXB Package Diagram Figure 5. 196-ball FBGA (15 × 15 × 1.5 mm) BB196A Package Outline, 51-85156 51-85156 *D Document Number: 38-02058 Rev. *M Page 49 of 53 CYP15G0201DXB Acronyms Document Conventions Table 26. Acronyms Used in this Document Units of Measure Acronym Description Table 27. Units of Measure AC alternating current BIST built-in self-test °C degree Celsius CDR clock/data recovery Hz hertz CML current mode logic KB 1024 bytes DC direct current Kbit 1024 bits ECL emitter coupled logic kHz kilohertz I/O input/output k kilohm JTAG joint test action group MHz megahertz LFI link fault indicator M megaohm LFSR linear feedback shift register A microampere LFSR linear feedback shift register s microsecond LPEN local loopback input V microvolt PECL positive-ECL or pseudo-ECL W microwatt PLL phase-locked loop mA milliampere TTL transistor transistor logic ms millisecond VCO voltage controlled oscillator mV millivolt nA nanoampere ns nanosecond nV nanovolt ohm ps picosecond Document Number: 38-02058 Rev. *M Symbol Unit of Measure Page 50 of 53 CYP15G0201DXB Document History Page Document Title: CYP15G0201DXB, Dual-channel HOTLink II™ Transceiver Document Number: 38-02058 Revision ECN Origin of Change Submission Date ** 116633 SDR 07/16/02 New data sheet *A 119705 LNM 10/30/02 Revised receive block diagram for RXCLKC+ signal Changed TXPERx description Changed TXCLKO description Corrected RXCLKB- description in REFCLK clocking mode to be disabled Removed reference to ATM support Removed the LOW setting for FRAMCHAR and related references Changed the IOST boundary values Changed VODIF and VOLC for CML output Changed the tTXCLKR and tTXCLKF min. values Changed tTXDS and tTXDH and tTREFDS and tTREFDH Changed tREFADV–, tREFCDV–, and tREFCDV+ Changed the JTAG ID from 0C80C069 to 1C80C069 Added a section for characterization and Standards compliance Changed I/O type of RXCLKC in I/O coordinates table Description of Change *B 122212 RBI 12/28/02 Document Control minor change *C 122547 CGX 12/9/02 Changed Minimum tRISE/tFALL for CML Changed tRXLOCK Changed tDJ, tRJ Changed tJTOL Changed tTXLOCK Changed tRXCLKH, tRXCLKL Changed tTXCLKOD+, tTXCLKODChanged Power specs Changed verbiage...Paragraph: Clock/Data Recovery Changed verbiage...Paragraph: Range Control Added Power-up Requirements *D 124548 LJN 02/13/03 Minor Change: Corrected errors and Power-up notes *E 124995 POT 04/15/03 Changed CYP15G0201DXB to CYP(V)15G0201DXB type corresponding to the Video-compliant parts Reduced the lower limit of the serial signaling rate from 200 Mbaud to 195 Mbaud and changed the associated specifications accordingly *F 128368 PDS 07/28/03 Revised the value of tRREFDV, tREFADV+ and tREFCDV+ *G 131900 PDS 01/30/04 When TXCKSEL = MID or HIGH, TXRATE = HIGH is an invalid mode. Made appropriate changes to reflect this invalid condition. Removed requirement of AC coupling for Serial I/Os for interfacing with LVPECL I/Os. Changed LFIx to Asynchronous output. Expanded the CDR Range Controller’s permissible frequency offset between incoming serial signalling rate and Reference clock from ±200-PPM to ±1500-PPM (changed parameter tREFRX). Added Table for RXSTx[2:0] status for non-bonded (Independent Channel) mode of operation for clarity. Separated the Receive BIST status to a new Table for clarity. *H 338721 SUA See ECN Added CYW15G0201DXB part number for OBSAI RP3 compliance to support operating data rate up to 1540 MBaud. Made changes to reflect OBSAI RP3 and CPR compliance. Added Pb-Free Package option for all parts listed in the datasheet Changed MBd to MBaud in SPDSEL pin description *I 2897032 CGX 03/23/10 Removed inactive parts from Ordering Information. Updated Package Diagram. Document Number: 38-02058 Rev. *M Page 51 of 53 CYP15G0201DXB Document History Page (continued) Document Title: CYP15G0201DXB, Dual-channel HOTLink II™ Transceiver Document Number: 38-02058 Revision ECN Origin of Change Submission Date *J 2905908 CGX 04/06/10 Removed CYV and CYW from the title. Removed CYP15G0201DXB-BBXI, CYV15G0201DXB-BBXC and CYV15G0201DXB-BBXI parts from ordering information. Updated Package Diagram. *K 2955995 CGX 07/30/10 Removed references to CYV15G0201DXB and CYW15G0201DXB. Added Acronyms. Updated to new template. *L 3539174 SAAC 03/01/2012 Updated Ordering Information and added Ordering Code Definitions. Updated Package Diagram. Added Units of Measure. *M 4685875 YLIU 03/16/2015 Updated Package Diagram: spec 51-85156 – Changed revision from *C to *D. Updated to new template. Completing Sunset Review. Document Number: 38-02058 Rev. *M Description of Change Page 52 of 53 CYP15G0201DXB Sales, Solutions, and Legal Information Worldwide Sales and Design Support Cypress maintains a worldwide network of offices, solution centers, manufacturer’s representatives, and distributors. To find the office closest to you, visit us at Cypress Locations. PSoC® Solutions Products Automotive Clocks & Buffers Interface Lighting & Power Control Memory cypress.com/go/automotive cypress.com/go/clocks cypress.com/go/interface cypress.com/go/powerpsoc cypress.com/go/memory PSoC cypress.com/go/psoc Touch Sensing cypress.com/go/touch USB Controllers Wireless/RF psoc.cypress.com/solutions PSoC 1 | PSoC 3 | PSoC 4 | PSoC 5LP Cypress Developer Community Community | Forums | Blogs | Video | Training Technical Support cypress.com/go/support cypress.com/go/USB cypress.com/go/wireless © Cypress Semiconductor Corporation, 2002-2015. The information contained herein is subject to change without notice. Cypress Semiconductor Corporation assumes no responsibility for the use of any circuitry other than circuitry embodied in a Cypress product. Nor does it convey or imply any license under patent or other rights. Cypress products are not warranted nor intended to be used for medical, life support, life saving, critical control or safety applications, unless pursuant to an express written agreement with Cypress. Furthermore, Cypress does not authorize its products for use as critical components in life-support systems where a malfunction or failure may reasonably be expected to result in significant injury to the user. The inclusion of Cypress products in life-support systems application implies that the manufacturer assumes all risk of such use and in doing so indemnifies Cypress against all charges. Any Source Code (software and/or firmware) is owned by Cypress Semiconductor Corporation (Cypress) and is protected by and subject to worldwide patent protection (United States and foreign), United States copyright laws and international treaty provisions. Cypress hereby grants to licensee a personal, non-exclusive, non-transferable license to copy, use, modify, create derivative works of, and compile the Cypress Source Code and derivative works for the sole purpose of creating custom software and or firmware in support of licensee product to be used only in conjunction with a Cypress integrated circuit as specified in the applicable agreement. Any reproduction, modification, translation, compilation, or representation of this Source Code except as specified above is prohibited without the express written permission of Cypress. Disclaimer: CYPRESS MAKES NO WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, WITH REGARD TO THIS MATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. Cypress reserves the right to make changes without further notice to the materials described herein. Cypress does not assume any liability arising out of the application or use of any product or circuit described herein. Cypress does not authorize its products for use as critical components in life-support systems where a malfunction or failure may reasonably be expected to result in significant injury to the user. The inclusion of Cypress’ product in a life-support systems application implies that the manufacturer assumes all risk of such use and in doing so indemnifies Cypress against all charges. Use may be limited by and subject to the applicable Cypress software license agreement. Document Number: 38-02058 Rev. *M Revised March 16, 2015 Page 53 of 53 HOTLink is a registered trademark, and HOTLink II and MultiFrame are trademarks, of Cypress Semiconductor Corporation. CPRI is a trademark of Siemens AG. IBM, ESCON, and FICON are registered trademarks of International Business Machines. All products and company names mentioned in this document may be the trademarks of their respective holders.