DP83950B RIC TM Repeater Interface Controller Y General Description Y The DP83950B Repeater Interface Controller ‘‘RIC’’ may be used to implement an IEEE 802.3 multiport repeater unit. It fully satisfies the IEEE 802.3 repeater specification including the functions defined by the repeater, segment partition and jabber lockup protection state machines. The RIC has an on-chip phase-locked-loop (PLL) for Manchester data decoding, a Manchester encoder and an Elasticity Buffer for preamble regeneration. Each RIC can connect to 13 cable segments via its network interface ports. One port is fully AUI compatible and is able to connect to an external MAU using the maximum length of AUI cable. The other 12 ports have integrated 10BASE-T transceivers. These transceiver functions may be bypassed so that the RIC may be used with external transceivers, for example DP8392 coaxial transceivers. In addition, large repeater units, containing several hundred ports may be constructed by cascading RICs together over an Inter-RIC bus. The RIC is configurable for specific applications. It provides port status information for LED array displays and a simple interface for system processors. The RIC posseses multifunction counter and status flag arrays to facilitate network statistics gathering. A serial interface, known as the Management Interface is available for the collection of data in Managed Hub applications. Features Y Y Y Y Y Y Y Y Y Y Y Y Separate partition state machines for each port Provides port status information for LED displays including: receive, collision, partition and link status Power-up configuration options: Repeater and Partition Specifications, Transceiver Interface, Status Display, Processor Operations Simple processor interface for repeater management and port disable On-chip Event Counters and Event Flag Arrays Serial Management Interface to combine packet and repeater status information together CMOS process for low power dissipation Single 5V supply Table of Contents 1.0 SYSTEM DIAGRAM 2.0 CONNECTION DIAGRAM 3.0 PIN DESCRIPTIONS 4.0 BLOCK DIAGRAM 5.0 FUNCTIONAL DESCRIPTION 6.0 HUB MANAGEMENT SUPPORT 7.0 PORT LOGIC FUNCTIONS Compliant with the IEEE 802.3 Repeater Specification 13 network connections (ports) per chip Selectable on-chip twisted-pair transceivers Cascadable for large hub applications Compatible with AUI compliant transceivers On-chip Elasticity Buffer, Manchester encoder and decoder 8.0 RIC REGISTER DESCRIPTIONS 9.0 AC AND DC SPECIFICATIONS 10.0 AC TIMING TEST CONDITIONS 11.0 PHYSICAL DIMENSIONS 1.0 System Diagram Simple RIC Hub TL/F/11096 – 1 TRI-STATEÉ is a registered trademark of National Semiconductor Corporation. RICTM and SONICTM are trademarks of National Semiconductor Corporation. C1995 National Semiconductor Corporation TL/F/11096 RRD-B30M16/Printed in U. S. A. DP83950B RIC Repeater Interface Controller October 1995 2.0 Connection DiagramÐ160 Pin PQFP Package Pin Table (12 T.P. Ports a 1 AUI Bottom View) Pin Name Pin No. TXO12Pb 40 NC Pin Name Pin No. 160 VCC Pin Name 120 VCC 80 TXO12 a 39 RXI7b 159 GND 119 GND 79 TXO12b 38 RXI7 a 158 TXO2Pb 118 IRC 78 TXO12P a 37 TXO7P a 157 TXO2 a 117 IRE 77 RXI12b 36 TXO7b 156 TXO2b 116 IRD 76 RXI12 a 35 TXO7 a 155 TXO2P a 115 COLN 75 VCC 34 TXO7Pb 154 RXI2b 114 VCC 74 GND 33 VCC 153 RXI2 a 113 GND 73 RXI11b 32 GND 152 VCC 112 PKEN 72 RXI11 a 31 TXO6Pb 151 GND 111 RXMPLL 71 TXO11P a 30 TXO6 a 150 RX1b 110 BUFEN 70 TXO11b 29 TXO6b 149 RX1 a 109 RDY 69 TXO11 a 28 TXO6P a 148 CD1b 108 ELI 68 TXO11Pb 27 RXI6b 147 CD1 a 107 RTI 67 VCC 26 RXI6 a 146 TX1b 106 STR1 66 GND 25 VCC 145 TX1 a 105 VCC 65 TXO10Pb 24 GND 144 VCC 104 GND 64 TXO10 a 23 RXI5b 143 GND 103 STR0 63 TXO10b 22 RXI5 a 142 VCC 102 ACTND 62 TXO10P a 21 TXO5P a 141 GND 101 ANYXND 61 RXI10b 20 TXO5b 140 CLKIN 100 ACKO 60 RXI10 a 19 TXO5 a 139 RA4 99 MRXC 59 VCC 18 TXO5Pb 138 RA3 98 MEN 58 GND 17 VCC 137 RA2 97 MRXD 57 RXI9b 16 GND 136 RA1 96 MCRS 56 RXI9 a 15 TXO4Pb 135 RA0 95 VCC 55 TXO9P a 14 TXO4 a 134 VCC 94 GND 54 TXO9b 13 TXO4b 133 GND 93 ACKI 53 TXO9 a 12 TXO4P a 132 MLOAD 92 ACTNS 52 TXO9Pb 11 RXI4b 131 CDEC 91 ANYXNS 51 VCC 10 RXI4 a 130 WR 90 PCOMP 50 GND 9 VCC 129 RD 89 NC 49 TXO8Pb 8 GND 128 D7 88 RXI13b 48 TXO8 a 7 RXI3b 127 D6 87 RXI13 a 47 TXO8b 6 RXI3 a 126 D5 86 TXO13P a 46 TXO8P a 5 TXO3P a 125 D4 85 TXO13b 45 RXI8b 4 TXO3b 124 D3 84 TXO13 a 44 RXI8 a 3 TXO3 a 123 D2 83 TXO13Pb 43 VCC 2 TXO3Pb 122 D1 82 VCC 42 GND 1 NC 121 D0 81 GND 41 Note: NC e No Connect 2 Pin No. Pin Name Pin No. 2.0 Connection DiagramÐ160 Pin PQFP Package (Continued) TL/F/11096 – 42 Ports 2 – 13 TP Port 1 AUI Order Number DP83950BVQB See NS Package Number VUL160A 3 2.0 Connection DiagramÐ160 Pin PQFP Package (Continued) Pin Table (1–5 AUI a 6 – 13 T.P. Ports) Pin Name Pin No. TXO12Pb 40 NC Pin Name Pin No. 160 VCC Pin Name 120 VCC 80 TXO12 a 39 RXI7b 159 GND 119 GND 79 TXO12b 38 RXI7 a 158 TX2b 118 IRC 78 TXO12P a 37 TXO7P a 157 TX2 a 117 IRE 77 RXI12b 36 TXO7b 156 CD2b 116 IRD 76 RXI12 a 35 TXO7 a 155 CD2 a 115 COLN 75 VCC 34 TXO7Pb 154 RX2 a 114 VCC 74 GND 33 VCC 153 RX2b 113 GND 73 RXI11b 32 GND 152 VCC 112 PKEN 72 RXI11 a 31 TXO6Pb 151 GND 111 RXMPLL 71 TXO11P a 30 TXO6 a 150 RX1b 110 BUFEN 70 TXO11b 29 TXO6b 149 RX1 a 109 RDY 69 TXO11 a 28 TXO6P a 148 CD1b 108 ELI 68 TXO11Pb 27 RXI6b 147 CD1 a 107 RTI 67 VCC 26 RXI6 a 146 TX1b 106 STR1 66 GND 25 VCC 145 TX1 a 105 VCC 65 TXO10Pb 24 GND 144 VCC 104 GND 64 TXO10 a 23 RX5 a 143 GND 103 STR0 63 TXO10b 22 RX5b 142 VCC 102 ACTND 62 TXO10P a 21 CD5 a 141 GND 101 ANYXND 61 RXI10b 20 CD5b 140 CLKIN 100 ACKO 60 RXI10 a 19 TX5 a 139 RA4 99 MRXC 59 VCC 18 TX5b 138 RA3 98 MEN 58 GND 17 VCC 137 RA2 97 MRXD 57 RXI9b 16 GND 136 RA1 96 MCRS 56 RXI9 a 15 TX4b 135 RA0 95 VCC 55 TXO9P a 14 TX4 a 134 VCC 94 GND 54 TXO9b 13 CD4b 133 GND 93 ACKI 53 TXO9 a 12 CD4 a 132 MLOAD 92 ACTNS 52 TXO9Pb 11 RX4 a 131 CDEC 91 ANYXNS 51 VCC 10 RX4b 130 WR 90 PCOMP 50 GND 9 VCC 129 RD 89 NC 49 TXO8Pb 8 GND 128 D7 88 RXI13b 48 TXO8 a 7 RX3 a 127 D6 87 RXI13 a 47 TXO8b 6 RX3b 126 D5 86 TXO13P a 46 TXO8P a 5 CD3 a 125 D4 85 TXO13b 45 RXI8b 4 CD3b 124 D3 84 TXO13 a 44 RXI8 a 3 TX3 a 123 D2 83 TXO13Pb 43 VCC 2 TX3b 122 D1 82 VCC 42 GND 1 NC 121 D0 81 GND 41 Note: NC e No Connect 4 Pin No. Pin Name Pin No. 2.0 Connection DiagramÐ160 Pin PQFP Package (Continued) TL/F/11096 – 43 Ports 6 – 13 TP Ports 1 – 5 AUI Order Number DP83950BVQB See NS Package Number VUL160A 5 2.0 Connection DiagramÐ160 Pin PQFP Package (Continued) Pin Table (1–7 AUI a 8 – 13 T.P. Ports) Pin Name Pin No. TXO12Pb 40 NC Pin Name Pin No. 160 VCC Pin Name 120 VCC 80 TXO12 a 39 RX7 a 159 GND 119 GND 79 TXO12b 38 RX7b 158 TX2b 118 IRC 78 TXO12P a 37 CD7 a 157 TX2 a 117 IRE 77 RXI12b 36 CD7b 156 CD2b 116 IRD 76 RXI12 a 35 TX7 a 155 CD2 a 115 COLN 75 VCC 34 TX7b 154 RX2 a 114 VCC 74 GND 33 VCC 153 RX2b 113 GND 73 RXI11b 32 GND 152 VCC 112 PKEN 72 RXI11 a 31 TX6b 151 GND 111 RXMPLL 71 TXO11P a 30 TX6 a 150 RX1b 110 BUFEN 70 TXO11b 29 CD6b 149 RX1 a 109 RDY 69 TXO11 a 28 CD6 a 148 CD1b 108 ELI 68 TXO11Pb 27 RX6 a 147 CD1 a 107 RTI 67 VCC 26 RX6b 146 TX1b 106 STR1 66 GND 25 VCC 145 TX1 a 105 VCC 65 TXO10Pb 24 GND 144 VCC 104 GND 64 TXO10 a 23 RX5 a 143 GND 103 STR0 63 TXO10b 22 RX5b 142 VCC 102 ACTND 62 TXO10P a 21 CD5 a 141 GND 101 ANYXND 61 RXI10b 20 CD5b 140 CLKIN 100 ACKO 60 RXI10 a 19 TX5 a 139 RA4 99 MRXC 59 VCC 18 TX5b 138 RA3 98 MEN 58 GND 17 VCC 137 RA2 97 MRXD 57 RXI9b 16 GND 136 RA1 96 MCRS 56 RXI9 a 15 TX4b 135 RA0 95 VCC 55 TXO9P a 14 TX4 a 134 VCC 94 GND 54 TXO9b 13 CD4b 133 GND 93 ACKI 53 TXO9 a 12 CD4 a 132 MLOAD 92 ACTNS 52 TXO9Pb 11 RX4 a 131 CDEC 91 ANYXNS 51 VCC 10 RX4b 130 WR 90 PCOMP 50 GND 9 VCC 129 RD 89 NC 49 TXO8Pb 8 GND 128 D7 88 RXI13b 48 TXO8 a 7 RX3 a 127 D6 87 RXI13 a 47 TXO8b 6 RX3b 126 D5 86 TXO13P a 46 TXO8P a 5 CD3 a 125 D4 85 TXO13b 45 RXI8b 4 CD3b 124 D3 84 TXO13 a 44 RXI8 a 3 TX3 a 123 D2 83 TXO13Pb 43 VCC 2 TX3b 122 D1 82 VCC 42 GND 1 NC 121 D0 81 GND 41 Note: NC e No Connect 6 Pin No. Pin Name Pin No. 2.0 Connection DiagramÐ160 Pin PQFP Package (Continued) TL/F/11096 – 44 Ports 8 – 13 TP Ports 1 – 7 AUI Order Number DP83950BVQB See NS Package Number VUL160A 7 2.0 Connection DiagramÐ160 Pin PQFP Package (Continued) Pin Table (All AUI Ports) Pin Name Pin No. Pin Name Pin No. Pin Name Pin No. Pin Name Pin No. TX12b 40 NC 160 VCC 120 VCC 80 TX12 a 39 RX7 a 159 GND 119 GND 79 CD12b 38 RX7b 158 TX2b 118 IRC 78 CD12 a 37 CD7 a 157 TX2 a 117 IRE 77 RX12 a 36 CD7b 156 CD2b 116 IRD 76 RX12b 35 TX7 a 155 CD2 a 115 COLN 75 VCC 34 TX7b 154 RX2 a 114 VCC 74 GND 33 VCC 153 RX2b 113 GND 73 RX11 a 32 GND 152 VCC 112 PKEN 72 RX11b 31 TX6b 151 GND 111 RXMPLL 71 CD11 a 30 TX6 a 150 RX1b 110 BUFEN 70 CD11b 29 CD6b 149 RX1 a 109 RDY 69 TX11 a 28 CD6 a 148 CD1b 108 ELI 68 TX11b 27 RX6 a 147 CD1 a 107 RTI 67 VCC 26 RX6b 146 TX1b 106 STR1 66 GND 25 VCC 145 TX1 a 105 VCC 65 TX10b 24 GND 144 VCC 104 GND 64 TX10 a 23 RX5 a 143 GND 103 STR0 63 CD10b 22 RX5b 142 VCC 102 ACTND 62 CD10 a 21 CD5 a 141 GND 101 ANYXND 61 RX10 a 20 CD5b 140 CLKIN 100 ACKO 60 RX10b 19 TX5 a 139 RA4 99 MRXC 59 VCC 18 TX5b 138 RA3 98 MEN 58 GND 17 VCC 137 RA2 97 MRXD 57 RX9 a 16 GND 136 RA1 96 MCRS 56 RX9b 15 TX4b 135 RA0 95 VCC 55 CD9 a 14 TX4 a 134 VCC 94 GND 54 CD9b 13 CD4b 133 GND 93 ACKI 53 TX9 a 12 CD4 a 132 MLOAD 92 ACTNS 52 TX9b 11 RX4 a 131 CDEC 91 ANYXNS 51 VCC 10 RX4b 130 WR 90 PCOMP 50 GND 9 VCC 129 RD 89 NC 49 TX8b 8 GND 128 D7 88 RX13 a 48 TX8 a 7 RX3 a 127 D6 87 RX13b 47 CD8b 6 RX3b 126 D5 86 CD13 a 46 CD8 a 5 CD3 a 125 D4 85 CD13b 45 RX8 a 4 CD3b 124 D3 84 TX13 a 44 RX8b 3 TX3 a 123 D2 83 TX13b 43 VCC 2 TX3b 122 D1 82 VCC 42 GND 1 NC 121 D0 81 GND 41 Note: NC e No Connect 8 2.0 Connection DiagramÐ160 Pin PQFP Package (Continued) TL/F/11096 – 45 All AUI Ports Order Number DP83950BVQB See NS Package Number VUL160A 9 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) Pin Table (12 T.P. Ports a 1 AUI Bottom View) Pin Name Pin No. TXO12Pb A15 RXI7b Pin Name Pin No. C2 VCC Pin Name Pin No. S1 VCC TXO12 a A14 RXI7 a A1 GND P4 GND P15 TXO12b B14 TXO7P a B1 TXO2Pb S2 IRC N14 TXO12P a C13 TXO7b D2 TXO2 a S3 IRE P16 RXI12b B13 TXO7 a E3 TXO2b R4 IRD N15 RXI12 a A13 TXO7Pb F3 TXO2P a P5 COLN N16 VCC C12 VCC C1 RXI2b R5 VCC M15 GND C11 GND D1 RXI2 a S4 GND M14 RXI11b B12 TXO6Pb E2 VCC S5 PKEN L14 RXI11 a B11 TXO6 a G3 GND S6 RXM L15 TXO11P a A12 TXO6b F2 RX1b P6 BUFEN M16 TXO11b A11 TXO6P a E1 RX1 a R6 TXO11 a C10 RXI6b G2 CD1b S7 TXO11Pb A10 RXI6 a H3 CD1 a R7 VCC B10 NC F1 TX1b P7 GND B9 NC G1 TX1 a P8 TXO10Pb C9 VCC H2 VCC R8 TXO10 a C8 GND J3 GND S8 TXO10b A9 RXI5b J2 VCC S9 TXO10P a A8 RXI5 a H1 GND R9 RXI10b B8 TXO5P a J1 CLKIN P9 RXI10 a B7 TXO5b K1 RA4 S10 VCC C7 TXO5 a K3 RA3 R10 GND A7 TXO5Pb K2 RA2 S11 RXI9b RXI9 a A6 B6 VCC L1 GND L2 RA1 RA0 P10 R11 TXO9P a C6 TXO4Pb M1 VCC S12 TXO9b C5 TXO4 a L3 GND R12 Pin Name Pin No. N13 RDY L16 ELI K16 RTI K14 STR1 K15 VCC J16 GND J15 STR0 J14 ACTND H16 ANYXND H15 ACKO H14 MRXC G14 MEN G15 MRXD G16 MCRS F16 VCC F14 GND F15 E15 E14 TXO9 a B5 TXO4b M2 MLOAD P11 ACKI TXO9Pb A5 TXO4P a N1 CDEC S13 ACTNS R13 ANYXNS E16 D16 VCC A4 RXI4b N2 WR GND B4 RXI4 a M3 RD S14 PCOMP TXO8Pb C4 VCC P1 D7 P12 RXI13b D15 R14 RXI13 a D14 S15 TXO13P a C16 P13 TXO13b C15 P14 TXO13 a B16 B15 TXO8 a TXO8b TXO8P a RXI8b A3 C3 GND RXI3b D4 RXI3 a R1 D6 P2 D5 N3 D4 B3 TXO3P a RXI8 a B2 TXO3b P3 D3 R2 D2 R15 TXO13Pb VCC A2 TXO3 a N4 D1 S16 VCC D13 GND D3 TXO3Pb R3 D0 R16 GND C14 Note: NC e No Connect 10 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) TL/F/11096 – 2 Bottom View 1 AUI a 2 – 13 T.P. Ports Order Number DP83950BNU See NS Package Number UP159A 11 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) Pin Table (1–5 AUI a 6 – 13 T.P. Ports) Pin Name Pin No. TXO12Pb A15 RXI7b Pin Name Pin No. C2 VCC Pin Name Pin No. S1 VCC TXO12 a A14 RXI7 a A1 GND P4 GND P15 TXO12b B14 TXO7P a B1 TX2b S2 IRC N14 TXO12P a C13 TXO7b D2 TX2 a S3 IRE P16 RXI12b B13 TXO7 a E3 CD2b R4 IRD N15 RXI12 a A13 TXO7Pb F3 CD2 a P5 COLN N16 VCC C12 VCC C1 RX2 a R5 VCC M15 GND C11 GND D1 RX2b S4 GND M14 RXI11b B12 TXO6Pb E2 VCC S5 PKEN L14 RXI11 a B11 TXO6 a G3 GND S6 RXM L15 TXO11P a A12 TXO6b F2 RX1b P6 BUFEN M16 TXO11b A11 TXO6P a E1 RX1 a R6 TXO11 a C10 RXI6b G2 CD1b S7 TXO11Pb A10 RXI6 a H3 CD1 a R7 VCC B10 NC F1 TX1b P7 GND B9 NC G1 TX1 a P8 TXO10Pb C9 VCC H2 VCC R8 TXO10 a C8 GND J3 GND S8 TXO10b A9 RX5 a J2 VCC S9 TXO10P a A8 RX5b H1 GND R9 RXI10b B8 CD5 a J1 CLKIN P9 RXI10 a B7 CD5b K1 RA4 S10 VCC C7 TX5 a K3 RA3 R10 GND A7 TX5b K2 RA2 S11 RXI9b RXI9 a A6 B6 VCC L1 GND L2 RA1 RA0 P10 R11 TXO9P a C6 TX4b M1 VCC S12 TXO9b C5 TX4 a L3 GND R12 Pin Name Pin No. N13 RDY L16 ELI K16 RTI K14 STR1 K15 VCC J16 GND J15 STR0 J14 ACTND H16 ANYXND H15 ACKO H14 MRXC G14 MEN G15 MRXD G16 MCRS F16 VCC F14 GND F15 E15 E14 TXO9 a B5 CD4b M2 MLOAD P11 ACKI TXO9Pb A5 CD4 a N1 CDEC S13 ACTNS R13 ANYXNS E16 D16 VCC A4 RX4 a N2 WR GND B4 RX4b M3 RD S14 PCOMP TXO8Pb C4 VCC P1 D7 P12 RXI13b D15 R14 RXI13 a D14 S15 TXO13P a C16 P13 TXO13b C15 P14 TXO13 a B16 B15 TXO8 a TXO8b TXO8P a RXI8b A3 C3 GND RX3 a D4 RX3b R1 D6 P2 D5 N3 D4 B3 CD3 a RXI8 a B2 CD3b P3 D3 R2 D2 R15 TXO13Pb VCC A2 TX3 a N4 D1 R16 VCC D13 GND D3 TX3b R3 D0 R16 GND C14 Note: NC e No Connect 12 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) TL/F/11096 – 3 Bottom View 1–5 AUI a 6 – 13 T.P. Ports Order Number DP83950BNU See NS Package Number UP159A 13 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) Pin Table (1–7 AUI a 8 – 13 T.P. Ports) Pin Name Pin No. TXO12Pb A15 RX7 a Pin Name Pin No. C2 VCC Pin Name Pin No. S1 VCC TXO12 a A14 RX7b A1 GND P4 GND P15 TXO12b B14 CD7 a B1 TX2b S2 IRC N14 TXO12P a C13 CD7b D2 TX2 a S3 IRE P16 RXI12b B13 TX7 a E3 CD2b R4 IRD N15 RXI12 a A13 TX7b F3 CD2 a P5 COLN N16 VCC C12 VCC C1 RX2 a R5 VCC M15 GND C11 GND D1 RX2b S4 GND M14 RXI11b B12 TX6b E2 VCC S5 PKEN L14 RXI11 a B11 TX6 a G3 GND S6 RXM L15 TXO11P a A12 CD6b F2 RX1b P6 BUFEN M16 TXO11b A11 CD6 a E1 RX1 a R6 TXO11 a C10 RX6 a G2 CD1b S7 TXO11Pb A10 RX6b H3 CD1 a R7 VCC B10 NC F1 TX1b P7 GND B9 NC G1 TX1 a P8 TXO10Pb C9 VCC H2 VCC R8 TXO10 a C8 GND J3 GND S8 TXO10b A9 RX5 a J2 VCC S9 TXO10P a A8 RX5b H1 GND R9 RXI10b B8 CD5 a J1 CLKIN P9 RXI10 a B7 CD5b K1 RA4 S10 VCC C7 TX5 a K3 RA3 R10 GND A7 TX5b K2 RA2 S11 RXI9b RXI9 a A6 B6 VCC L1 GND L2 RA1 RA0 P10 R11 TXO9P a C6 TX4b M1 VCC S12 TXO9b C5 TX4 a L3 GND R12 Pin Name Pin No. N13 RDY L16 ELI K16 RTI K14 STR1 K15 VCC J16 GND J15 STR0 J14 ACTND H16 ANYXND H15 ACKO H14 MRXC G14 MEN G15 MRXD G16 MCRS F16 VCC F14 GND F15 E15 E14 TXO9 a B5 CD4b M2 MLOAD P11 ACKI TXO9Pb A5 CD4 a N1 CDEC S13 ACTNS R13 ANYXNS E16 D16 VCC A4 RX4 a N2 WR GND B4 RX4b M3 RD S14 PCOMP TXO8Pb C4 VCC P1 D7 P12 RXI13b D15 R14 RXI13 a D14 S15 TXO13P a C16 P13 TXO13b C15 P14 TXO13 a B16 B15 TXO8 a TXO8b TXO8P a RXI8b A3 C3 GND RX3 a D4 RX3b R1 D6 P2 D5 N3 D4 B3 CD3 a RXI8 a B2 CD3b P3 D3 R2 D2 R15 TXO13Pb VCC A2 TX3 a N4 D1 S16 VCC D13 GND D3 TX3b R3 D0 R16 GND C14 Note: NC e No Connect 14 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) TL/F/11096 – 4 Bottom View 1–7 AUI a 8 – 13 T.P. Ports Order Number DP83950BNU See NS Package Number UP159A 15 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) Pin Table ( All AUI Ports) Pin Name Pin No. Pin Name Pin No. Pin Name Pin No. Pin Name Pin No. TX12b A15 RX7 a C2 VCC S1 VCC TX12 a A14 RX7b A1 GND P4 GND P15 CD12b B14 CD7 a B1 TX2b S2 IRC N14 CD12 a C13 CD7b D2 TX2 a S3 IRE P16 RX12 a B13 TX7 a E3 CD2b R4 IRD N15 RX12b A13 TX7b F3 CD2 a P5 COLN N16 VCC C12 VCC C1 RX2 a R5 VCC M15 GND C11 GND D1 RX2b S4 GND M14 RX11 a B12 TX6b E2 VCC S5 PKEN L14 RX11b B11 TX6 a G3 GND S6 RXM L15 CD11 a A12 CD6b F2 RX1b P6 BUFEN M16 CD11b A11 CD6 a E1 RX1 a R6 TX11 a C10 RX6 a G2 CD1b S7 TX11b A10 RX6b H3 CD1 a R7 VCC B10 NC F1 TX1b P7 GND B9 NC G1 TX1 a P8 TX10b C9 VCC H2 VCC R8 TX10 a C8 GND J3 GND S8 CD10b A9 RX5 a J2 VCC S9 CD10 a A8 RX5b H1 GND R9 RX10 a B8 CD5 a J1 CLKIN P9 RX10b B7 CD5b K1 RA4 S10 VCC C7 TX5 a K3 RA3 R10 GND A7 TX5b K2 RA2 S11 RX9 a RX9b A6 B6 VCC L1 GND L2 RA1 RA0 P10 R11 CD9 a C6 TX4b M1 VCC S12 CD9b C5 TX4 a L3 GND R12 N13 RDY L16 ELI K16 RTI K14 STR1 K15 VCC J16 GND J15 STR0 J14 ACTND H16 ANYXND H15 ACKO H14 MRXC G14 MEN G15 MRXD G16 MCRS F16 VCC F14 GND F15 E15 E14 TX9 a B5 CD4b M2 MLOAD P11 ACKI TX9b A5 CD4 a N1 CDEC S13 ACTNS R13 ANYXNS E16 D16 VCC A4 RX4 a N2 WR GND B4 RX4b M3 RD S14 PCOMP TX8b C4 VCC P1 D7 P12 RX13 a D15 R14 RX13b D14 S15 CD13 a C16 P13 CD13b C15 P14 TX13 a B16 B15 TX8 a CD8b CD8 a RX8 a A3 C3 GND RX3 a D4 RX3b R1 D6 P2 D5 N3 D4 B3 CD3 a RX8b B2 CD3b P3 D3 R2 D2 R15 TX13b VCC A2 TX3 a N4 D1 S16 VCC D13 GND D3 TX3b R3 D0 R16 GND C14 Note: NC e No Connect 16 2.0 Connection DiagramÐ160 Pin PGA Package (Continued) TL/F/11096 – 5 Bottom View All AUI Ports Order Number DP83950BNU See NS Package Number UP159A 17 3.0 Pin Descriptions Pin No. Pin Name Driver Type I/O Description NETWORK INTERFACE PINS (On-Chip Transceiver Mode) RXI2b to RXI13b TP RXI2 a to RXI13 a I Twisted Pair Receive Input Negative TXOP2b to TXOP13b TP I Twisted Pair Receive Input Positive TT O Twisted Pair Pre-emphasis Transmit Output Negative TXO2b to TXO13b TT O Twisted Pair Transmit Output Negative TXO2 a to TXO13 a TT O Twisted Pair Transmit Output Positive TXOP2 a to TXOP13 a TT O Twisted Pair Pre-emphasis Transmit Output Positive CD1 a AL I AUI Collision Detect Input Positive CD1b AL I AUI Collision Detect Input Negative RX1 a AL I AUI Receive Input Positive RX1b AL I AUI Receive Input Negative TX1 a AD O AUI Transmit Output Positive TX1b AD O AUI Transmit Output Negative NETWORK INTERFACE PINS (External Transceiver Mode AUI Signal Level Compatibility Selected) TX2 a to TX13 a AL O Transmit Output Positive TX2b to TX13b AL O Transmit Output Negative CD2 a to CD13 a AL I Collision Input Positive CD2b to CD13b AL I Collision Input Negative RX2 a to RX13 a AL I Receive Input Positive RX2b to RX13b AL I Receive Input Negative CD1 a AL I AUI Collision Detect Input Positive CD1b AL I AUI Collision Detect Input Negative RX1 a AL I AUI Receive Input Positive RX1b AL I AUI Receive Input Negative TX1 a AD O AUI Transmit Output Positive TX1b AD O AUI Transmit Output Negative Note: AD e AUI level and Drive compatible, TP e Twisted Pair interface compatible, AL e AUI Level compatible, TT e TTL compatible, I e Input, O e Output. 18 3.0 Pin Descriptions (Continued) Pin No. Pin Name Driver Type I/O Description PROCESSOR BUS PINS RA0–RA4 TT I REGISTER ADDRESS INPUTS: These five pins are used to select a register to be read or written. The state of these inputs are ignored when the read, write and mode load input strobes are high. (Even under these conditions these inputs must not be allowed to float at an undefined logic state). STR0 C O DISPLAY UPDATE STROBE 0 Maximum Display Mode: This signal controls the latching of display data for network ports 1 to 7 into the off chip display latches. Minimum Display Mode: This signal controls the latching of display data for the RIC into the off chip display latch. During processor access cycles (read or write is asserted) this signal is inactive (high). STR1 C O DISPLAY UPDATE STROBE 1 Maximum Display Mode: This signal controls the latching of display data for network ports 8 to 13 into the off chip display latches. Minimum Display Mode: No operation During processor access cycles (read or write is asserted) this signal is inactive (high). D0– D7 TT B, Z DATA BUS Display Update Cycles: These pins become outputs providing display data and port address information. Address information only available in Maximum Display mode. Processor Access Cycles: Data input or output is performed via these pins. The read, write and mode load inputs control the direction of the signals. Note: The data pins remain in their display update function, i.e., asserted as outputs unless either the read or write strobe is asserted. BUFEN C O BUFFER ENABLE: This output controls the TRI-STATEÉ operation of the bus transceiver which provides the interface between the RIC’s data pins and the processor’s data bus. Note: The buffer enable output indicates the function of the data pins. When it is high they are performing display update cycles, when it is low a processor access or mode load cycle is occurring. RDY C O DATA READY STROBE: The falling edge of this signal during a read cycle indicates that data is stable and valid for sampling. In write cycles the falling edge of RDY denotes that the write data has been latched by the RIC. Therefore data must have been available and stable for this operation to be successful. ELI C O EVENT LOGGING INTERRUPT: A low level on the ELI output indicates the RIC’s hub management logic requires CPU attention. The interrupt is cleared by accessing the Port Event Recording register or Event Counter that produced it. All interrupt sources may be masked. RTI C O REAL TIME INTERRUPT: A low level on the RTI output indicates the RIC’s real time (packet specific) interrupt logic requires CPU attention. The interrupt is cleared by reading the Real Time Interrupt Status register. All interrupt sources may be masked. CDEC TT I COUNTER DECREMENT: A low level on the CDEC input strobe decrements all of the RIC’s Port Event Counters by one. This input is internally synchronized and if necessary the operation of the signal is delayed if there is a simultaneous internally generated counting operation. WR TT I WRITE STROBE: Strobe from the CPU used to write an internal register defined by the RA0–RA4 inputs. RD TT I READ STROBE: Strobe from the CPU used to read an internal register defined by the RA0 – RA4 inputs. MLOAD TT I DEVICE RESET AND MODE LOAD: When this input is low all of the RIC’s state machines, counters and network ports are reset and held inactive. On the rising edge of MLOAD the logic levels present on the D0 – 7 pins and RA0 – RA4 inputs are latched into the RIC’s configuration registers. The rising edge of MLOAD also signals the beginning of the display test operation. 19 3.0 Pin Descriptions (Continued) Pin No. Pin Name Driver Type I/O Description INTER-RIC BUS PINS ACKI TT I ACKNOWLEDGE INPUT: Input to the network ports’ arbitration chain. ACKO TT O IRD TT B, Z INTER-RIC DATA: When asserted as an output this signal provides a serial data stream in NRZ format. The signal is asserted by a RIC when it is receiving data from one of its network segments. The default condition of this signal is to be an input. In this state it may be driven by other devices on the Inter-RIC bus. ACKNOWLEDGE OUTPUT: Output from the network ports’ arbitration chain. IRE TT B, Z INTER-RIC ENABLE: When asserted as an output this signal provides an activity framing enable for the serial data stream. The signal is asserted by a RIC when it is receiving data from one of its network segments. The default condition of this signal is to be an input. In this state it may be driven by other devices on the Inter-RIC bus. IRC TT B, Z INTER-RIC CLOCK: When asserted as an output this signal provides a clock signal for the serial data stream. Data (IRD) is changed on the falling edge of the clock. The signal is asserted by a RIC when it is receiving data from one of its network segments. The default condition of this signal is to be an input. When an input IRD is sampled on the rising edge of the clock. In this state it may be driven by other devices on the Inter-RIC bus. COLN TT B, Z COLLISION ON PORT N: This denotes that a collision is occurring on the port receiving the data packet. The default condition of this signal is to be an input. In this state it may be driven by other devices on the Inter-RIC bus. PKEN C O PACKET ENABLE: This output acts as an active high enable for an external bus transceiver (if required) for the IRE, IRC IRD and COLN signals. When high the bus transceiver should be transmitting on to the bus, i.e., this RIC is driving the IRD, IRE, IRC and COLN bus lines. When low the bus transceiver should receive from the bus. CLKIN TT I 40 MHz CLOCK INPUT: This input is used to generate the RIC’s timing reference for the state machines, and phase lock loop decoder. ACTND OD O ACTIVITY ON PORT N DRIVE: This output is active when the RIC is receiving data or collision information from one of its network segments. ACTNS TT I ACTIVITY ON PORT N SENSE: This input senses when this or another RIC in a multi-RIC system is receiving data or collision information. ANYXND OD O ACTIVITY ON ANY PORT EXCLUDING PORT N DRIVE: This output is active when a RIC is experiencing a transmit collision or multiple ports have active collisions on their network segments. ANYXNS TT I ACTIVITY ON ANY PORT EXCLUDING PORT N SENSE: This input senses when this RIC or other RICs in a multi-RIC system are experiencing transmit collisions or multiple ports have active collisions on their network segments. 20 3.0 Pin Descriptions (Continued) Pin No. Pin Name Driver Type I/O Description MANAGEMENT BUS PINS MRXC TT O, Z MANAGEMENT RECEIVE CLOCK: When asserted this signal provides a clock signal for the MRXD serial data stream. The MRXD signal is changed on the falling edge of this clock. The signal is asserted when a RIC is receiving data from one of its network segements. Otherwise the signal is inactive. MCRS TT B, Z MANAGEMENT CARRIER SENSE: When asserted this signal provides an activity framing enable for the serial output data stream (MRXD). The signal is asserted when a RIC is receiving data from one of its network segments. Otherwise the signal is an input. MRXD TT O, Z MANAGEMENT RECEIVE DATA: When asserted this signal provides a serial data stream in NRZ format. The data stream is made up of the data packet and RIC status information. The signal is asserted when a RIC is receiving data from one of its network segments. Otherwise the signal is inactive. MEN C O MANAGEMENT BUS OUTPUT ENABLE: This output acts as an active high enable for an external bus transceiver (if required) for the MRXC, MCRS and MRXD signals. When high the bus transceiver should be transmitting on to the bus. PCOMP TT I PACKET COMPRESS: This input is used to activate the RIC’s packet compress logic. A low level on this signal when MCRS is active will cause that packet to be compressed. If PCOMP is tied low all packets are compressed, if PCOMP is tied high packet compression is inhibited. POWER AND GROUND PINS VCC Positive Supply GND Negative Supply EXTERNAL DECODER PINS RXM TT O RECEIVE DATA MANCHESTER FORMAT: This output makes the data, in Manchester format, received by port N available for test purposes. If not used for testing this pin should be left open. Note: TT e TTL compatible, B e Bi-directional, C e CMOS compatible, OD e Open Drain, I e Input, O e Output, Z e TRI-STATE 21 FIGURE 5.1 TL/F/11096 – 6 4.0 Block Diagram 22 5.0 Functional Description data and collision status is sent to the main state machine via the receive multiplexor and collision activity status signals. This enables the main state machine to determine the source of the data to be repeated and the type of data to be transmitted. The transmit data may be either the received packet’s data field or a preamble/jam pattern consisting of a 1010 . . . bit pattern. Associated with the main state machine are a series of timers. These ensure various IEEE specification times (referred to as the TW1 to TW6 times) are fulfilled. A repeater unit is required to meet the same signal jitter performance as any receiving node attached to a network segment. Consequently, a phase locked loop Manchester decoder is required so that the packet may be decoded, and the jitter accumulated over the receiving segment recovered. The decode logic outputs data in NRZ format with an associated clock and enable. In this form the packet is in a convenient format for transfer to other devices, such as network controllers and other RICs, via the Inter-RIC bus (described later). The data may then be re-encoded into Manchester data and transmitted. Reception and transmission via physical layer transceiver units causes a loss of bits in the preamble field of a data packet. The repeater specification requires this loss to be compensated for. To accomplish this an elasticity buffer is employed to temporarily store bits in the data field of the packet. The sequence of operation is as follows: Soon after the network segment receiving the data packet has been identified, the RIC begins to transmit the packet preamble pattern (1010 . . . ) onto the other network segments. While the preamble is being transmitted the Elasticity Buffer monitors the decoded received clock and data signals (this is done via the Inter-RIC bus as described later). When the start of frame delimiter ‘‘SFD’’ is detected the received data stream is written into the elasticity buffer. Removal of data from the buffer for retransmission is not allowed until a valid length preamble pattern has been transmitted. The I.E.E.E. repeater specification details a number of functions a repeater system must perform. These requirements allied with a need for the implementation to be multiport strongly favors the choice of a modular design style. In such a design, functionality is split between those tasks common to all data channels and those exclusive to each individual channel. The RIC follows this approach, certain functional blocks are replicated for each network attachment, (also known as a repeater port), and others are shared. The following section briefly describes the functional blocks in the RIC. 5.1 OVERVIEW OF RIC FUNCTIONS Segment Specific Block: Network Port As shown in the Block Diagram, the segment specific blocks consist of: 1. One or more physical layer interfaces. 2. A logic block required for performing repeater operations upon that particular segment. This is known as the ‘‘port’’ logic since it is the access ‘‘port’’ the segment has to the rest of the network. This function is repeated 13 times in the RIC (one for each port) and is shown on the right side of the Block Diagram, Figure 5.1 . The physical layer interfaces provided depends upon the port under examination. Port 1 has an AUI compliant interface for use with AUI compatible transceiver boxes and cable. Ports 2 to 13 may be configured for use with one of two interfaces: twisted pair or an external transceiver. The former utilizes the RIC’s on-chip 10BASE-T transceivers, the latter allows connection to external transceivers. When using the external transceiver mode the interface is AUI compatible. Although AUI compatible transceivers are supported the interface is not designed for use with an interface cable, thus the transceivers are necessarily internal to the repeater equipment. Inside the port logic there are 3 distinct functions: 1. The port state machine ‘‘PSM’’ is required to perform data and collision repetition as described by the repeater specification, for example, it determines whether this port should be receiving from or transmitting to its network segment. 2. The port partition logic implements the segment partitioning algorithm. This algorithm is defined by the IEEE specification and is used to protect the network from malfunctioning segements. 3. The port status register reflects the current status of the port. It may be accessed by a system processor to obtain this status or to perform certain port configuration operations, such as port disable. Inter-RIC Bus Interface Using the RIC in a repeater system allows the design to be constructed with many more network attachments than can be supported by a single chip. The split of functions already described allows data packets and collision status to be transferred between multiple RICs, and at the same time the multiple RICs still behave as a single logical repeater. Since all RICs in a repeater system are identical and capable of performing any of the repetition operations, the failure of one RIC will not cause the failure of the entire system. This is an important issue in large multiport repeaters. RICs communicate via a specialized interface known as the Inter-RIC bus. This allows the data packet to be transferred from the receiving RIC to the other RICs in the system. These RICs then transmit the data stream to their segments. Just as important as data transfer is the notification of collisions occurring across the network. The Inter-RIC bus has a set of status lines capable of conveying collision information between RICs to ensure their main state machines operate in the appropriate manner. Shared Functional Blocks: Repeater Core Logic The shared functional blocks consist of the Repeater Main State Machine (MSM) and Timers, a 32 bit Elasticity Buffer, PLL Decoder, and Receive and Transmit Multiplexors. These blocks perform the majority of the operations needed to fulfill the requirements of the IEEE repeater specification. When a packet is received by a port it is sent via the Receive Multiplexor to the PLL Decoder. Notification of the 23 5.0 Functional Description (Continued) ensures bus contention is avoided during simultaneous display update cycles and processor accesses of other devices on the system bus. When the processor accesses a RIC register, the RIC enables the data buffer and selects the operation, either input or output, of the data pins. LED Interface and Hub Management Function Repeater systems usually possess optical displays indicating network activity and the status of specific repeater operations. The RIC’s display update block provides the system designer with a wide variety of indicators. The display updates are completely autonomous and merely require SSI logic devices to drive the display devices, usually made up of light emitting diodes, LEDs. The status display is very flexible allowing the user to choose those indicators appropriate for the specification of the equipment. The RIC has been designed with special awareness for system designers implementing large repeaters possessing hub management capabilities. Hub management uses the unique position of repeaters in a network to gather statistics about the network segments they are attached to. The RIC provides hub management statistical data in 3 steps. Important events are gathered by the management block from logic blocks throughout the chip. These events may then be stored in on-chip latches or counted in on-chip counters according to user supplied latching and counting masks. The fundamental task of a hub management system implementation is to associate the current packet and any management status information with the network segment, i.e., repeater port where the packet was received. The ideal system would place this combined data packet and status field in system memory for examination by hub management software. The ultimate function of the RIC’s hub management support logic is to provide this function. To accomplish this the RIC utilizes a dedicated hub management interface. This is similar to the Inter-RIC bus since it allows the data packet to be recovered from the receiving RIC. Unlike the Inter-RIC bus the intended recipient is not another RIC but National Semiconductor’s DP83932 ‘‘SONICTM ’’ Network controller. The use of a dedicated bus allows a management status field to be appended at the end of the data packet. This can be done without affecting the operation of the repeater system. 5.2 DESCRIPTION OF REPEATER OPERATIONS In order to implement a multi-chip repeater system which behaves as though it were a single logical repeater, special consideration must be paid to the data path used in packet repetition. For example, where in the path are specific operations such as Manchester decoding and elasticity buffering performed. Also the system’s state machines which utilize available network activity signals, must be able to accommodate the various packet repetition and collision scenarios detailed in the repeater specification. The RIC contains two types of inter-acting state machines. These are: 1. Port State Machines (PSMs). Every network attachment has its own PSM. 2. Main State Machine (MSM). This state machine controls the shared functional blocks as shown in the block diagram Figure 5.1. Repeater Port and Main State Machines These two state machines are described in the following sections. Reference is made to expressions used in the IEEE Repeater specification. For the precise definition of these terms please refer to the specification. To avoid confusion with the RIC’s implementation, where references are made to repeater states or terms as described in the IEEE specification, these items are written in italics. The IEEE state diagram is shown in Figure 5-3 , the Inter-RIC bus state diagram is shown in Figure 5-2 . Processor Interface The RIC’s processor interface allows connection to a system processor. Data transfer occurs via an octal bi-directional data bus. The RIC has a number of on-chip registers indicating the status of the hub management functions, chip configuration and port status. These may be accessed by providing the chosen address at the Register Address (RA4 – RA0) input pins. Display update cycles and processor accesses occur utilizing the same data bus. An on-chip arbiter in the processor/ display block schedules and controls the accesses and ensures the correct information is written into the display latches. During the display update cycles the RIC behaves as a master of its data bus. This is the default state of the data bus. Consequently, a TRI-STATE buffer must be placed between the RIC and the system processor’s data bus. This TL/F/11096 – 7 FIGURE 5.2. Inter-RIC Bus State Diagram 24 5.0 Functional Description (Continued) TL/F/11096 – 8 FIGURE 5.3. IEEE Repeater Main State Diagram 25 5.0 Functional Description (Continued) The interaction of the main and port state machines is visible, in part, by observing the Inter-RIC bus. Port State Machine (PSM) There are two primary functions for the PSM as follows: 1. Control the transmission of repeated data and jam signals over the attached segment. 2. Decide whether a port will be the source of data or collision information which will be repeated over the network. This repeater port is known as PORT N . An arbitration process is required to enable the repeater to transition from the IDLE state to the SEND PREAMBLE PATTERN or RECEIVE COLLISION states, see Figure 5.3 . This process is used to locate the port which will be PORT N for that particular packet. The data received from this port is directed to the PLL decoder and transmitted over the Inter-RIC bus. If the repeater enters the TRANSMIT COLLISION state a further arbitration operation is performed to determine which port is PORT M . PORT M is differentiated from the repeater’s other ports if the repeater enters the ONE PORT LEFT state. In this state PORT M does not transmit to its segment; where as all other ports are still required to transmit to their segments. Inter-RIC Bus Operation Overview The Inter-RIC Bus consists of eight signals. These signals implement a protocol which may be used to connect multiple RICs together. In this configuration, the logical function of a single repeater is maintained. The resulting multi-RIC system is compliant to the IEEE 802.3 repeater specification and may connect several hundred network segments. An example of a multi-RIC system is shown in Figure 5.4 . The Inter-RIC Bus connects multiple RICs to realize the following operations: Port N Identification (which port the repeater receives data from) Port M Identification (which port is the last one experiencing a collision) Data Transfer RECEIVE COLLISION identification TRANSMIT COLLISION identification DISABLE OUTPUT (jabber protection) The following tables briefly describes the operation of each bus signal, the conditions required for a RIC to assert a signal and which RICs (in a multi-RIC system) would monitor a signal: Main State Machine (MSM) The MSM controls the operation of the shared functional blocks in each RIC as shown in the block diagram, Figure 5.1 , and it performs the majority of the data and collision propagation operations as defined by the IEEE specification, these include: Function Preamble Regeneration ACKI Action Function Restore the length of the preamble pattern to the defined size. Fragment Extension Extend received data or collision fragments to meet the minimum fragment length of 96 bits. Elasticity Buffer Control A portion of the received packet may require storage in an Elasticity Buffer to accommodate preamble regeneration. Jam/ Preamble Pattern Generation In cases of receive or transmit collisions a RIC is required to transmit a jam pattern (1010 . . . ). Note: A RIC which contains PORT N or PORT M may be identified by its ACKO signal being low when its ACKI input is high. Note: This pattern is the same as that used for preamble regeneration. Transmit Collision Enforcement Once the TRANSMIT COLLISION state is entered a repeater is required to stay in this state for at least 96 network bit times. Data Encoding Control NRZ format data from the elasticity buffer must be encoded into Manchester format data prior to retransmission. Tw1 Enforcement Enforce the Transmit Recovery Time specification. Tw2 Enforcement Enforce Carrier Recovery Time specification on all ports with active collisions. Input signal to the PSM arbitration chain. This chain is employed to identify PORT N and PORT M . Conditions required for a RIC to drive this signal Not applicable RIC Receiving the signal This is dependent upon the method used to cascade RICs, described in a following section. ACKO Function 26 Output signal from the PSM arbitration chain. Conditions required for a RIC to drive this signal This is dependent upon the method used to cascade RICs, described in a following section. RIC Receiving the Signal Not applicable 5.0 Functional Description (Continued) ACTN Function IRD This signal denotes there is activity on PORT N or PORT M . Function Decoded serial data, in NRZ format, received from the network segment attached to PORT N . Conditions required for a RIC to drive this signal A RIC must contain PORT N or PORT M . Conditions required for a RIC to drive this signal A RIC must contain PORT N . Note: Although this signal normally has only one source asserting the signal active it is used in a wired-or configuration. RIC Receiving the Signal The signal is monitored by all RICs in the repeater system. RIC Receiving the Signal The signal is monitored by all other RICs in the repeater system. ANYXN Function This signal denotes that a repeater port that is not PORT N or PORT M is experiencing a collision. Conditions required for a RIC to drive this signal Any RIC which satisfies the above condition. Note: This bus line is used in a wired-or configuration. RIC Receiving the Signal The signal is monitored by all RICs in the repeater system. IRC Function Denotes PORT N or PORT M is experiencing a collision. Conditions required for a RIC to drive this signal A RIC must contain PORT N or PORT M . (Note 1) RIC Receiving the Signal The Signal is monitored by all other RICs in the repeater system. Function This signal acts as an activity framing signal for the IRC and IRD signals. IRE Conditions required for a RIC to drive this signal A RIC must contain PORT N . RIC Receiving the Signal The Signal is monitored by all other RICs in the repeater system. Conditions required for a RIC to drive this signal A RIC must contain PORT N . RIC Receiving the Signal The signal is monitored by all other RICs in the repeater system. Methods of RIC Cascading In order to build multi-RIC repeaters PORT N and PORT M identification must be performed across all the RICs in the system. Inside each RIC the PSMs are arranged in a logical arbitration chain where port 1 is the highest and port 13 the lowest. The top of the chain, the input to port 1 is accessible to the user via the RIC’s ACKI input pin. The output from the bottom of the chain becomes the ACKO output pin. In a single RIC system PORT N is defined as the highest port in the arbitration chain with receive or collision activity. Port N identification is performed when the repeater is in the IDLE state. PORT M is defined as the highest port in the chain with a collision when the repeater leaves the TRANSMIT COLLISION state. In order for the arbitration chain to function, all that needs to be done is to tie the ACKI signal to a logic high state. In multi-RIC systems there are two methods to propagate the arbitration chain between RICs: The first and most straight forward is to extend the arbitration chain by daisy chaining the ACKI ACKO signals between RICs. In this approach one RIC is placed at the top of the chain (its ACKI input is tied high), then the ACKO signal from this RIC is sent to the ACKI input of the next RIC and so on. This arrangement is simple to implement but it places some topological restrictions upon the repeater system. In particular, if the repeater is constructed using a backplane with removable printed circuit boards. (These boards contain the RICs and their associated components). If one of the boards is removed then the ACKI ACKO chain will be broken and the repeater will not operate correctly. COLN Function Clock signal associated with IRD and IRE. Note 1: Refer to note on page 25 for the transmit collision case. 27 5.0 Functional Description (Continued) The second method of PORT N or M identification avoids this problem. This second technique relies on an external parallel arbiter which monitors all of the RIC’s ACKO signals and responds to the RIC with the highest priority. In this scheme each RIC is assigned with a priority level. One method of doing this is to assign a priority number which reflects the position of a RIC board on the repeater backplane, i.e., its slot number. When a RIC experiences receive activity and the repeater system is in the IDLE state, the RIC board will assert ACKO. External arbitration logic drives the identification number onto an arbitration bus and the RIC containing PORT N will be identified. An identical procedure is used in the TRANSMIT COLLISION state to identify PORT M . This parallel means of arbitration is not subject to the problems caused by missing boards, i.e., empty slots in the backplane. The logic associated with asserting this arbitration vector in the various packet repetition scenarios could be implemented in programmable logic type devices. To perform PORT N or M arbitration both of the above methods employ the same signals: ACKI, ACKO and ACTN. The Inter-RIC bus allows multi-RIC operations to be performed in exactly the same manner as if there is only a single RIC in the system. The simplest way to describe the operation of Inter-RIC bus is to see how it is used in a number of common packet repetition scenarios. Throughout this description the RICs are presumed to be operating in external transceiver mode. This is advantageous for the explanation since the receive, transmit and collision signals from each network segment are observable. In internal transceiver mode this is not the case, since the collision signal for the non-AUI ports is derived by the transceivers inside the RIC. PLL decoder. Once phase lock has been achieved, the decoded data, in NRZ format, with its associated clock and enable signals are asserted onto the IRD IRE and IRC InterRIC bus lines. This serial data stream is received from the bus by all RICs in the repeater and directed to their Elasticity Buffers. Logic circuits monitor the data stream and look for the Start of Frame Delimiter (SFD). When this has been detected data is loaded into the elasticity buffer for later transmission. This will occur when sufficient preamble has been transmitted and certain internal state machine operations have been fulfilled. Figure 5.4 shows two RICs A and B, daisy chained together with RIC A positioned at the top of the chain. A packet is received at port B1 of RIC B and is then repeated by the other ports in the system. Figure 5.5 shows the functional timing diagram for this packet repetition represented by the signals shown in Figure 5.4 . In this example only two ports in the system are shown, obviously the other ports also repeat the packet. It also indicates the operation of the RICs’ state machines in so far as can be seen by observing the Inter-RIC bus. For reference, the repeater’s state transitions are shown in terms of the states defined by the IEEE specification. The location, i.e., which port it is, of PORT N is also shown. The following section describes the repeater and Inter-RIC bus transitions shown in Figure 5.5. The repeater is stimulated into activity by the data signal received by port B1. The RICs in the system are alerted to forthcoming repeater operation by the falling edges on the ACKI ACKO daisy chain and the ACTN bus signal. Following a defined start up delay the repeater moves to the SEND PREAMBLE state. The RIC system utilizes the start up delay to perform port arbitration. When packet transmission begins the RIC system enter the REPEAT state. The expected, for normal packet repetition, sequence of repeater states, SEND PREAMBLE , SEND SFD and SEND DATA is followed but is not visible upon the Inter-RIC bus. They are merged together into a single REPEAT state. This is also true for the WAIT and IDLE states, they appear as a combined Inter-RIC bus IDLE state. Once a repeat operation has begun, i.e., the repeater leaves the IDLE state. It is required to transmit at least 96 bits of data or jam/preamble onto its network segments. If the duration of the received signal from PORT N is smaller than 96 bits, the repeater transitions to the RECEIVE COLLISION state (described later). This behavior is known as fragment extension. After the packet data has been repeated, including the emptying of the RICs’ elasticity buffers, the RIC performs the Tw1 transmit recovery operation. This is performed during the WAIT state shown in the repeater state diagram. 5.3 EXAMPLES OF PACKET REPETITION SCENARIOS Data Repetition The simplest packet operation performed over the Inter-RIC Bus is data repetition. In this operation a data packet is received at one port and transmitted to all other segments. The first task to be performed is PORT N identification. This is an arbitration process performed by the Port State Machines in the system. In situations where two or more ports simultaneously receive packets the Inter-RIC bus operates by choosing one of the active ports and forcing the others to transmit data. This is done to faithfully follow the IEEE specification’s allowed exit paths from the IDLE state, i.e., to the SEND PREAMBLE PATTERN or RECEIVE COLLISION states. The packet begins with a preamble pattern derived from the RIC’s on chip jam/preamble generator. The data received at PORT N is directed through the receive multiplexor to the 28 5.0 Functional Description (Continued) TL/F/11096 – 9 Note: In this example the Inter-RIC bus is configured to use active low signals. FIGURE 5.4. RIC System Topology 29 5.0 Functional Description (Continued) *Note 1: The activity shown in RXA1 represents the transmitted signal on TXA1 after being looped back by the attached transceiver. Note: In this example the Inter-RIC bus is configured to use active low signals. FIGURE 5.5. Data Repetition 30 TL/F/11096 – 10 5.0 Functional Description (Continued) TL/F/11096 – 11 *Note 1: SEND PREAMBLE, SEND SFD, SEND DATA . Note: In this example the Inter-RIC bus is configured to use active low signals. FIGURE 5.6. Receive Collision 31 5.0 Functional Description (Continued) TRANSMIT COLLISION state when the port which has been PORT N starts to transmit a Manchester encoded 1 on to its network segment. Whilst in the TRANSMIT COLLISION state all ports of the repeater must transmit the 1010 . . . jam pattern and PORT M arbitration is performed. Each RIC is obliged, by the IEEE specification, to ensure all of its ports transmit for at least 96 bits once the TRANSMIT COLLISION state has been entered. This transmit activity is enforced by the ANYXN bus signal. Whilst ANYXN is active all RIC ports will transmit jam. To ensure this situation lasts for at least 96 bits, the MSMs inside the RICs assert the ANYXN signal throughout this period. After this period has elapsed, ANYXN will only be asserted if there are multiple ports with active collisions on their network segments. There are two possible ways for a repeater to leave the TRANSMIT COLLISION state. The most straight forward is when network activity, i.e., collisions and their Tw2 extensions, end before the 96 bit enforced period expires. Under these conditions the repeater system may move directly to the WAIT state when 96 bits have been transmitted to all ports. If the MSM enforced period ends and there is still one port experiencing a collision the ONE PORT LEFT state is entered. This may be seen on the Inter-RIC bus when ANYXN is deasserted and PORT M stops transmitting to its network segment. In this circumstance the Inter-RIC bus transitions to the RECEIVE COLLISION state. The repeater will remain in this state whilst PORT M’s collision, Tw2 collision extension and any receive signals are present. When these conditions are not true, packet repetition finishes and the repeater enters the WAIT state. Figure 5.7 shows a multi-RIC system operating under transmit collision conditions. There are many different scenarios which may occur during a transmit collision, this figure illustrates one of these. The diagram begins with packet reception by port A1. Port B1 experiences a collision, since it is not PORT N it asserts ANYXN. This alerts the main state machines in the system to switch from data to jam pattern transmission. Port A1 is also monitoring the ANYXN bus line. Its assertion forces A1 to relinquish its PORT N status, start transmitting, stop asserting ACTN and release its hold on the PSM arbitration signals (ACKO A and ACKI B). The first bit it transmit will be a Manchester encoded ‘‘1’’ in the jam pattern. Since port B1 is the only port with a collision it attains PORT M status and stops asserting ANYXN. It does however assert ACTN, and exert its presence upon the PSM arbitration chain (forcesACKO B low). The MSMs ensure that ANYXN stays active and thus force all of the ports, including PORT M , to transmit to their segments. After some time port A1 experiences a collision. This arises from the presence of the packet being received from port A1’s segment and the jam signal the repeater is now transmitting onto this segment. Two packets on one segment results in a collision. PORT M now moves from B1 to A1. Port A1 fulfills the same criteria as B1, i.e., it has an active collision on its segment, but in addition it is higher in the arbitration chain. This priority yields no benefits for port A1 since the ANYXN signal is still active. There are now two sources driving ANYXN, the MSMs and the collision on port B1. Eventually the collision on port B1 ends and the ANYXN extension by the MSMs expires. There is only one collision Receive Collisions A receive collision is a collision which occurs on the network segment attached to PORT N , i.e., the collision is ‘‘received’’ in a similar manner as a data packet is received and then repeated to the other network segments. Not surprisingly receive collision propagation follows a similar sequence of operations as is found with data repetition: An arbitration process is performed to find PORT N and a preamble/jam pattern is transmitted by the repeater’s other ports. When PORT N detects a collision on its segment the COLN Inter-RIC bus signal is asserted. This forces all the RICs in the system to transmit a preamble/jam pattern to their segments. This is important since they may be already transmitting data from their elasticity buffers. The repeater moves to the RECEIVE COLLISION state when the RICs begin to transmit the jam pattern. The repeater remains in this state until both the following conditions have been fulfilled: 1. At least 96 bits have been transmitted onto the network, 2. The activity has ended. Under close examination the repeater specification reveals that the actual end of activity has its own permutations of conditions: 1. Collision and receive data signals may end simultaneously, 2. Receive data may appear to end before collision signals, 3. Receive data may continue for some time after the end of the collision signal. Network segments using coaxial media may experience spurious gaps in segment activity when the collision signal goes inactive. This arises from the inter-action between the receive and collision signal squelch circuits, implemented in coaxial transceivers, and the properties of the coaxial cable itself. The repeater specification avoids propagation of these activity gaps by extending collision activity by the Tw2 wait time. Jam pattern transmission must be sustained throughout this period. After this, the repeater will move to the WAIT state unless there is a data signal being received by PORT N . The functional timing diagram, Figure 5.6 , shows the operation of a repeater system during a receive collision. The system configuration is the same as earlier described and is shown in Figure 5.4 . The RICs perform the same PORT N arbitration and data repetition operations as previously described. The system is notified of the receive collision on port B1 by the COLN bus signal going active. This is the signal which informs the main state machines to output the jam pattern rather than the data held in the elasticity buffers. Once a collision has occurred the IRC, IRD AND IRE bus signals may become undefined. When the collision has ended and the Tw2 operation performed, the repeater moves to the WAIT state. Transmit Collisions A transmit collision is a collision that is detected upon a segment to which the repeater system is transmitting. The port state machine monitoring the colliding segment asserts the ANYXN bus signal. The assertion of ANYXN causes PORT M arbitration to begin. The repeater moves to the 32 5.0 Functional Description (Continued) mon with the operation of the RECEIVE COLLISION state, the repeater remains in this condition until the collision and receive activity on PORT M sibsides. The packet repetition operation completes when the Tw1 recovery time in the WAIT state has been performed. on the network (this may be deduced since ANYXN is inactive) so the repeater will move to the ONE PORT LEFT state. The RIC system treats this state in a similar manner to a receive collision with PORT M fulfilling the role of the receiving port. The difference from a true receive collision is that the switch from packet data to the jam pattern has already been made (controlled by ANYXN). Thus the state of COLN has no effect upon repeater operations. In com- Note: In transmit collision conditions COLN will only go active if the RIC which contained PORT N at the start of packet repetition contains PORT M during the TRANSMIT COLLISION and ONE PORT LEFT states. TL/F/11096 – 12 Note: In this example the Inter-RIC bus is configured to use active low signals. FIGURE 5.7. Transmit Collision 33 5.0 Functional Description (Continued) Figure 5.8 shows the effect of a jabber length packet upon a RIC based repeater system. The JABBER PROTECT state is entered from the SEND DATA state. While the Tw4 period is observed the Inter-RIC bus displays the IDLE state. This is misleading since new packet activity or continuous activity (as shown in the diagram) does not result in packet repetition. This may only occur when the Tw4 requirement has been satisified. Jabber Protection A repeater is required to disable transmit activity if the length of its current transmission reaches the jabber protect limit. This is defined by the specification’s Tw3 time. The repeater disables output for a time period defined by the Tw4 specification, after this period normal operation may resume. TL/F/11096 – 13 *Note 1: The IEEE Specification does not have a jabber protect state defined in its main state diagram, this behaviour is defined in an additional MAU Jabber Lockup Protection state diagram. Note: In this example the Inter-RIC bus is configured to use active low signals. FIGURE 5.8. Jabber Protect 34 5.0 Functional Description (Continued) Note: DE e Bus Drive Enable Active High, RE e Bus Receive Enable active low. TL/F/11096 – 14 Note: In this example the Inter-RIC bus is shown as using active low signals. FIGURE 5.9. External Bus Transceiver Connection Diagram TL/F/11096 – 15 FIGURE 5.10. Mode Load Operation 35 5.0 Functional Description (Continued) 5.4 DESCRIPTION OF HARDWARE CONNECTION FOR INTER-RIC BUS be configured to invert the active states of the ACTN, ANYXN, COLN and IRE signals. Instead of being active low they are active high. When considering the hardware interface the Inter-RIC bus may be viewed as consisting of three groups of signals: 1. Port Arbitration chain, namely: ACKI and ACKO. Thus they become active low once more when passed through an inverting bus driver. This is particularly important for the ACTN and ANYXN bus lines, since these signals must be used in a wired-or configuration. Incorrect signal polarity would make the bus unusable. 2. Simultaneous drive and sense signals, i.e., ACTN and ANYXN. (Potentially these signals may be driven by multiple devices). 3. Drive or sense signals, i.e., IRE, IRD, IRC and COLN. (Only one device asserts these signals at any instance in time.) The first set of signals are either used as point to point links or with external arbitration logic. In both cases the load on these signals will not be large so that the on-chip drivers are adequate. This may not be true for signal classes (2) and (3). The Inter-RIC bus has been designed to connect RICs together directly or via external bus transceivers. The latter is advantageous in large repeaters. In the second application the backplane is often heavily loaded and is beyond the drive capability of the on-chip bus drivers. The need for simultaneous sense and drive capabilities on the ACTN and ANYXN signals and the desire to allow operation with external bus transceivers makes it necessary for these bus signals to each have a pair of pins on the RIC. One driving the bus the other sensing the bus signal. When external bus transceivers are used they must be open collector/open drain to allow wire-ORing of the signals. Additionally, the drive and sense enables of the bus transceiver should be tied in the active state. When the RIC is used in a stand alone configuration, it is required to tie ACTND to ACTNS and ANYXND to ANYXNS. The uni-directional nature of information transfer on the IRE, IRD, IRC and COLN signals, means a RIC is either driving these signals or receivng them from the bus but not both at the same time. Thus a single bi-directional input/output pin is adequate for each of these signals. In an external bus transceiver is used with these signals the Packet Enable ‘‘PKEN’’ RIC output pin performs the function of a drive enable and sense disable. 5.5 PROCESSOR AND DISPLAY INTERFACE The processor interface pins, which include the data bus, address bus and control signals, actually perform three operations which are multiplexed on these pins. These operations are: 1. The Mode Load Operation, which performs a power up initialization cycle upon the RIC. 2. Display Update Cycles, which are refresh operations for updating the display LEDs. 3. Processor Access Cycles, which allows mP’s to communicate with the RIC’s registers. These three operations are described below. Mode Load Operation The Mode Load Operation is a hardware initialization procedure performed at power on. It loads vital device configuration information into on-chip configuration registers. In addition to its configuration function the MLOAD pin is the RIC’s reset input. When MLOAD is low all of the RIC’s repeater timers, state machines, segment partition logic and hub management logic are reset. The Mode Load Operation may be accomplished by attaching the appropriate set of pull up and pull down resistors to the data and register address pins to assert logic high or low signals onto these pins, and the providing a rising edge on the MLOAD pin as is shown in Figure 5.10 . The mapping of chip functions to the configuration inputs is shown in Table 5.1. Such an arrangement may be performed using a simple resistor, capacitor, diode network. Performing the Mode Load Operation in this way enables the configuration of a RIC that is in a simple repeater system (one without a processor). Alternatively in a complex repeater system, the Mode Load Operation may be performed using a processor write cycle. This would require the MLOAD pin be connected to the CPU’s write strobe via some decoding logic, and included in the processor’s memory map. Figure 5.9 shows the RIC connected to the Inter-RIC bus via external bus transceivers, such as National’s DS3893A bus transceivers. Some bus transceivers are of the inverting type. To allow the Inter-RIC bus to utilize these transceivers the RIC may 36 5.0 Functional Description (Continued) TABLE 5.1. Pin Definitions for Options in the Mode Load Operation Pin Name Programming Function Effect When Bit is 0 Effect When Bit is 1 D0 resv Not Permitted Required To ensure correct device operation, this bit must be written with a logic one during the mode load operation. D1 tw2 5 bits 3 bits This allows the user to select one of two values for the repeater specification tw2 time. The lower limit (3 bits) meets the IEEE specification. The upper limit (5 bits) is not specification compliant but may provide users with higher network throughput by avoiding spurious network activity gaps when using coaxial (10BASE2, 10BASE5) network segments. D2 CCLIM 63 31 The partition specification requires a port to be partitioned after a certain number of consecutive collisions. The RIC has two values available to allow users to customize the partitioning algorithm to their environment. Please refer to the Partition State Machine, in data sheet Section 7.3. D3 LPPART Selected Not Selected The RIC may be configured to partition a port if the segment transceiver does not loopback data to the port when the port is transmitting to it, as described in the Partition State Machine. D4 OWCE Selected Not Selected This configuration bit allows the on-chip partition algorithm to include out of window collisions into the collisions it monitors, as described in the Partition State Machine. D5 TXONLY Selected Not Selected This configuration bit allows the on-chip partition algorithm to restrict segment reconnection, as described in the Partition State Machine. D6 DPART Selected Not Selected The Partition state machines for all ports may be disabled by writing a logic zero to this bit during the mode load operation. D7 MIN/MAX Minimum Mode Maximum Mode Function The operation of the display update block is controlled by the value of this configuration bit, as described in the Display Update Cycles section. 37 5.0 Functional Description (Continued) TABLE 5.1 Pin Definitions for Options in the Mode Load Operation (Continued) Pin Name Programming Function Effect When Bit is 0 Effect When Bit is 1 RA0 BYPAS1 These configuration bits select which of the repeater ports (numbers 2 to 13) are configured to use the on-chip internal 10BASE-T transceivers or the external transceiver interface. The external transceiver interface operates using AUI compatible signal levels. RA1 BYPAS2 BYPAS2 BYPAS1 0 0 All ports (2 to 13) use the external Transceiver Interface. 0 1 Ports 2 to 7 use the external interface, 8 to 13 use the internal 10BASE-T transceivers. 1 0 Ports 2 to 5 use the external interface, 6 to 13 use the internal 10BASE-T transceivers. 1 1 All ports (2 to 13) use the internal 10BASE-T transceivers. Function Information RA2 BINV Active High Signals Active Low Signals This selection determines whether the Inter-RIC signals: IRE, ACTN, ANYXN, COLN and Management bus signal MCRS are active high or low. RA3 EXPLL External PLL Internal PLL If desired, the RIC may be used with an external decoder, this configuration bit performs the selection. RA4 resv Not Permitted Required To ensure correct device operation, this bit must be written with a logic one during the mode load operation. 38 5.0 Functional Description (Continued) The signals provided and their timing relationships have been designed to interface directly with 74LS259 type addressable latches. The number of latches used being dependant upon the complexity of the display. Since the latches are octal, a pair of latches is needed to display each type of segment specific data (13 ports means 13 latch bits). The accompanying tables (5.1 and 5.2) show the function of the interface pins in minimum and maximum modes. Figure 5.12 shows the location of each port’s status information when maximum mode is selected. This may be compared with the connection diagram Figure 5.11 . Immediately following the Mode Load Operation (when the MLOAD pin transitions to a high logic state), the display logic performs an LED test operation. This operation lasts one second and while it is in effect all of the utilized LEDs will blink on. Thus an installation engineer is able to test the operation of the display by forcing the RIC into a reset cycle (MLOAD forced low). The rising edge on the MLOAD pin starts the LED test cycle. During the LED test cycle the RIC does not perform packet repetition operations. The status display possesses a capability to lengthen the time an LED is active. At the end of the repetition of a packet, the display is frozen showing the current activity. This freezing lasts for 30 milliseconds or until a subsequent packet is repeated. Thus at low levels of packet activity the display stretches activity information to make it discernable to the human eye. At high traffic rates the relative brightness of the LEDs indicates those segments with high or low activity. It should be mentioned that when the Real Time Interrupt (RTI) occurs, the display update cycle will stop and after RTI is serviced, the display update cycle will resume activity. 5.6 DESCRIPTION OF HARDWARE CONNECTION FOR PROCESSOR AND DISPLAY INTERFACE Display Update Cycles The RIC possesses control logic and interface pins which may be used to provide status information concerning activity on the attached network segments and the current status of repeater functions. These status cycles are completely autonomous and require only simple support circuitry to produce the data in a form suitable for a light emitting diode ‘‘LED’’ display. The display may be used in one of two modes: 1. Minimum Mode: General Repeater Status LEDs 2. Maximum Mode: Individual Port Status LEDs Minimum mode, intended for simple LED displays, makes available four status indicators. The first LED denotes whether the RIC has been forced to activate its jabber protect functions. The remaining 3 LEDs indicate if any of the RIC’s network segments are: (1) experiencing a collision, (2) receiving data, (3) currently partitioned. When minimum display mode is selected the only external components required are a 74LS374 type latch, the LEDs and their current limiting resistors. Maximum mode differs from minimum mode by providing display information specific to individual network segments. This information denotes the collision activity, packet reception and partition status of each segment. In the case of 10BASE-T segments the link integrity status and polarity of the received data are also made available. The wide variety of information available in maximum mode may be used in its entirety or in part. Thus allowing the system designer to choose the appropriate complexity of status display commensurate with the specification of the end equipment. TABLE 5.2. Status Display Pin Functions in Minimum Mode Signal Pin Name Function in MINIMUM MODE D0 No operation D1 Provides status information indicating if there is a collision occurring on one of the segments attached to this RIC. D2 Provides status information indicating if one of this RIC’s ports is receiving a data or collision packet from a segment attached to this RIC. D3 Provides status information indicating that the RIC has experienced a jabber protect condition. D4 Provides Status information indicating if one of the RIC’s segments is partitioned. D(7:5) No operation STR0 This signal is the latch enable for the 374 type latch. STR1 This signal is held at a logic one. 39 5.0 Functional Description (Continued) Table 5.3 Status Display Pin Functions in MAXIMUM MODE Signal Pin Name Function in Maximum Mode D0 Provides status information concerning the Link Integrity status of 10BASE-T segments. This signal should be connected to the data inputs of the chosen pair of 74LS259 latches. D1 Provides status information indicating if there is a collision occurring on one of the segments attached to this RIC. This signal should be connected to the data inputs of the chosen pair of 74LS259 latches. D2 Provides status information indicating if one of this RIC’s ports is receiving a data or a collision packet from its segment. This signal should be connected to the data inputs of the chosen pair of 74LS259 latches. D3 Provides Status information indicating that the RIC has experienced a jabber protect condition. Additionally it denotes which of its ports are partitioned. This signal should be connected to the data inputs of the chosen pair of 74LS259 latches. D4 Provides status information indicating if one of this RIC’s ports is receiving data of inverse polarity. This status output is only valid if the port is configured to use its internal 10BASE-T transceiver. The signal should be connected to the data inputs of the chosen pair of 74LS259 latches. D(7:5) These signals provide the repeater port address corresponding to the data available on D(4:0). STR0 This signal is the latch enable for the lower byte latches, that is the 74LS259s which display information concerning ports 1 to 7. STR1 This signal is the latch enable for the upper byte latches, that is the 74LS259s which display information concerning ports 8 to 13. Maximum Mode LED Definitions 74LS259 Latch Inputs e STR0 259 Output Q0 259 Addr S2 – 0 000 RIC Port Number Q1 Q2 Q3 Q4 Q5 Q6 Q7 001 010 011 100 101 110 111 1 (AUI) 2 3 4 5 6 7 LINK LINK LINK LINK LINK LINK COL COL COL COL COL COL RIC D0 259 Ý1 RIC D1 259 Ý2 ACOL RIC D2 259 Ý3 COL AREC REC REC REC REC REC REC REC RIC D3 259 Ý4 JAB PART PART PART PART PART PART PART BDPOL BDPOL BDPOL BDPOL BDPOL BDPOL RIC D4 259 Ý5 74LS259 (or Equiv.) Latch Inputs e STR1 259 Output Q0 Q1 Q2 Q3 Q4 Q5 Q6 Q7 259 Addr S2 – 0 000 001 010 011 100 101 110 111 RIC Port Number RIC D0 259 Ý6 8 9 10 11 12 13 LINK LINK LINK LINK LINK LINK RIC D1 259 Ý7 COL COL COL COL COL COL RIC D2 259 Ý8 REC REC REC REC REC REC RIC D3 259 Ý9 PART PART PART PART PART PART RIC D4 259 Ý10 BDPOL BDPOL BDPOL BDPOL BDPOL BDPOL This shows the LED Output Functions for the LED Drivers when 74LS259s are used. The top table refers to the bank of 4 74LS259s latched with STR0, and the lower table refers to the bank of 4 74LS259s latched with STR1. For example the RIC’s D0 data signal goes to 259 Ý1 and Ý5. These two 74LS259s then drive the LINK LEDs). Note: ACOL e Any Port Collision, AREC e Any Port Reception, JAB e Any Port Jabbering, LINK e Port Link, COL e Port Collision, REC e Port Reception, PART e Port Partitioned, BDPOL e Bad (inverse) Polarity or received data. FIGURE 5.12 40 FIGURE 5.11. Maximum Mode LED Display (All Available Status Bits Used) TL/F/11096 – 16 5.0 Functional Description (Continued) 41 5.0 Functional Description (Continued) TL/F/11096 – 17 FIGURE 5.13. Processor Connection Diagram 42 6.1 EVENT COUNTING FUNCTION 5.0 Functional Description (Continued) The counters may increment upon the occurrence of one of the categories of event as described below. Processor Access Cycles Access to the RIC’s on-chip registers is made via its processor interface. This utilizes conventional non-multiplexed address (five bit) and data (eight bit) busses. The data bus is also used to provide data and address information to off chip display latches during display update cycles. While performing these cycles the RIC behaves as a master of its data bus. Consequently a TRI-STATE bi-directional bus transceiver, e.g., 74LS245 must be placed between the RIC and any processor bus. The processor requests a register access by asserting the read ‘‘RD’’ or write ‘‘WR’’ input strobes. The RIC responds by finishing any current display update cycle and asserts the tri-state buffer enable signal ‘‘BUFFEN’’. If the processor cycle is a write cycle then the RIC’s data buffers are disabled to prevent contention. In order to interface to the RIC in a processor controlled system it is likely a PAL device will be used to perform the following operations: 1. Locate the RIC in the processor’s memory map (address decode), 2. Generate the RIC’s read and write strobes, 3. Control the direction signal for the 74LS245. An example of the processor and display interfaces is shown in Figure 5.13 . Potential sources for Counter increment: Jabber Protection (JAB): The port counter increments if the length of a received packet from its associated port, causes the repeater state machine to enter the jabber protect state. Elasticity Buffer Error (ELBER): The port counter increments if a Elasticity Buffer underflow or overflow occurs during packet reception. The flag is held inactive if a collision occurs during packet reception or if a phase lock error, described below, has already occurred during the repetition of the packet. Phase Lock Error (PLER): A phase lock error is caused if the phase lock loop decoder looses lock during packet reception. Phase lock onto the received data stream may or may not be recovered later in the packet and data errors may have occurred. This flag is held inactive if a collision occurs. Non SFD Packet (NSFD): If a packet is received and the start of frame delimiter is not found, the port counter will increment. Counting is inhibited if the packet suffers a collision. Out of Window Collision (OWC): The out of window collision flag for a port goes active when a collision is experienced outside of the network slot time. Transmit Collision (TXCOL): The transmit collision flag for a port is enabled when a transmit collision is experienced by the repeater. Each port experiencing a collision under these conditions is said to have suffered a transmit collision. Receive Collision (RXCOL): The receive collision flag for a port goes active when the port is the receive source of network activity and suffers a collision, provided no other network segments experience collision then the receive collision flag for the receiving port will be set. Partition (PART): The port counter increments when a port becomes partitioned. Bad Link (BDLNK): The port counter increments when a port is configured for 10BASE-T operation has entered the link lost state. Short Event reception (SE): The port counter increments if the received packet is less than 74 bits long and no collision occurs during reception. Packet Reception (REC): When a packet is received the port counter increments. In order to utilize the counters the user must choose, from the above list, the desired statistic for counting. This counter mask information must be written to the appropriate, Event Count Mask Register. There are two of these registers, the Upper and Lower, Event Count Mask registers. For the exact bit patterns of these registers please see Section 8 of the data sheet. For example if the counters are configured to count network collisions and the appropriate masks have been set, then whenever a collision occurs on a segment, this information is latched by the hub management support logic. At the end of repetition of the packet the collision status, respective to each port, is loaded into that port’s counter. This operation is completely autonomous and requires no processor intervention. 6.0 Hub Management Support The RIC provides information regarding the status of its ports and the packets it is repeating. This data is available in three forms: 1. Counted EventsÐNetwork events accumulated into the RIC’s 16-bit Event Counter Registers. 2. Recorded EventsÐNetwork events that set bits in the Event Record Registers. 3. Hub Management Status PacketsÐThis is information sent over the Management Bus in a serial function to be decoded by an Ethernet Controller board. The counted and recorded event information is available through the processor interface. This data is port specific and may be used to generate interrupts via the Event Logging Interrupt ‘‘ELI’’ pin. Since the information is specific to each port, each repeater port has its own event record register and event counter. The counters and event record registers have user definable masks which enable them to be configured to count and record a variety of events. The counters and record registers are designed to be used together so that detailed information, i.e., a count value can be held on-chip for a specific network condition, and more general information, i.e., certain types of events have occurred, may be retained in on-chip latches. Thus the user may configure the counters to increment upon a rapidly occurring event (most likely to be used to count collisions), and the record registers may log the occurrence of less frequent error conditions such as jabber protect packets. 43 6.0 Hub Management Support (Continued) the counter array. Other NON COUNTER registers may be read in between these read cycles and also write cycles may be performed. If another counter is read or the same byte of the original counter is read, then the holding register is updated from the counter array and the unread byte is lost. If the reset on read option is employed then the counter is reset after the transfer to the holding register is performed. Processor read and write cycles are scheduled in such a manner that they do not conflict with count up or count down operations. That is to say, in the case of a processor read the count value is stable when it is loaded into the holding register. In the case of a processor write, the newly written value is stable so it maybe incremented or decrement by any subsequent count operation. During the period the MLOAD pin is low, (power on reset) all counters are reset to zero and all count masks are forced into the disabled state. Section 8 of the data sheet details the address location of the port event counters. Each counter is 16 bits long and may be directly read by the processor. Additionally each counter has a number of decodes to indicate the current value of the count. There are three decodes: Low Count (a value of 00FF Hex and under), High Count (a value of C000 Hex and above), Full Count (a value of FFFF Hex). The decodes from each counter are logically ‘‘ORed’’ together and may be used as interrupt sources for the ELI interrupt pin. Additionally the status of these bits may be observed by reading the Page Select Register (PSR), (see Section 8 for register details). In order to enable any of these threshold interrupts, the appropriate interrupt mask bit must be written to the Management and Interrupt Configuration Register; see Section 8 for register details. In addition to their event masking functions the Upper Event Counting Mask Register (UECMR) possesses two bits which control the operation of the counters. When written to a logic one, the reset on read bit ‘‘ROR’’ resets the counter after a processor read cycle is performed. If this operation is not selected then in order to zero the counters they must either be written with zeros by the processor or allowed to roll over to all zeros. The freeze when full bit ‘‘FWF’’ prevents counter roll over by inhibiting count up cycles (these happen when chosen events occur), thus freezing the particular counter at FFFF Hex. The port event counters may also be controlled by the Counter Decrement (CDEC) pin. As its name suggests a logic low state on this pin will decrement all the counters by a single value. The pulses on CDEC are internally synchronized and scheduled so as not to conflict with any ‘‘up counting’’ activity. If an up count and a down count occur simultaneously then the down count is delayed until the up count has completed. This combination of up and down counting capability enables the RIC’s on-chip counters to provide a simple rolling average or be used as extensions of larger off chip counters. 6.2 EVENT RECORD FUNCTION As previously stated each repeater port has its own Event Recording Register. This is an 8-bit status register each bit is dedicated to logging the occurrence of a particular event (see Section 8 for detailed description). The logging of these events is controlled by the Event Recording Mask Register, for an event to be recorded the particular mask bit must be set, (see Section 8 description of this register). Similar to the scheme employed for the event counters, the recorded events are latched during the repetition of a packet and then automatically loaded into the recording registers at the end of transmission of a packet. When one of the unmasked events occurs, the particular port register bit is set. This status is visible to the user. All of the register bits for all of the ports are logically ‘‘ORed’’ together to produce a Flag Found ‘‘FF’’ signal. This indicator may be found by reading the Page Select Register. Additionally an interrupt may be generated if the appropriate mask bit is enabled in the Management and Interrupt Configuration Register. A processor read cycle to a Event Record Register resets any of the bits set in that register. Read operations are scheduled to guarantee non changing data during a read cycle. Any internal bit setting event which immediately follows a processor read will be successful. The events which may be recorded are described below: Jabber Protection (JAB): This flag goes active if the length of a received packet from the relevant port, causes the repeater state machine to enter the Jabber Protect state. Elasticity Buffer Error (ELBER): This condition occurs if an Elasticity Buffer full or overflow occurs during packet reception. The flag is held inactive if a collision occurs during packet reception or if a phase lock error has already occurred during the repetition of the packet. Phase Lock Error (PLER): A phase lock error is caused if the phase lock loop decoder loses lock during packet reception. Phase lock onto the received data stream may or may not be recovered later in the packet and data errors may have occurred. This flag is held inactive if a collision occurs. Non SFD Packet (NSFD): If a packet is received and the start of frame delimiter is not found, the flag will go active. The flag is held inactive if a collision occurs in during packet repetition. Note: If the FWF option is enabled then the count down operation is disabled from those registers which have reached FFFF Hex and consequently have been frozen. Thus, if FWF is set and CDEC has been employed to provide a rate indication. A frozen counter indicates that a rate has been detected which has gone out of bounds, i.e., too fast increment or too slow increment. If the low count and high count decodes are employed as either interrupt sources or in a polling cycle, the direction of the rate excursion may be determined. Reading the Event Counters The RIC’s external data bus is eight bits wide, since the event counters are 16 bits long two processor read cycles are required to yield the counter value. In order to ensure that the read value is correct and to allow simultaneous event counts with processor accesses, a temporary holding register is employed. A read cycle to either the lower or upper byte of a counter, causes both bytes to be latched into the holding register. Thus when the other byte of the counter is obtained the holding register is accessed and not the actual counter register. This ensures that the upper and lower bytes contain the value sampled at the same instance in time, i.e., when the first read cycle to that counter occurred. There is no restriction concerning whether the upper or lower byte is read first. However to ensure the ‘‘same instance value’’ is obtained, the reads of the upper then lower byte (or vice versa) should be performed as consecutive reads of 44 6.0 Hub Management Support (Continued) memory required to buffer packets when they are received and are waiting to be processed by hub management software. In this kind of application an address decoder, which forms part of the packet compress logic, would monitor the address fields as they are received over the management bus. If the destination address is not the address of the management node inside the hub, then packet compression could be employed. In this manner only the portion of the packet meaningful for hub management interrogation, i.e., the address fields, is transferred to the SONIC and is buffered in memory. If the repeated packet ends before PCOMP is asserted or before the required number of bytes have been transferred, then the hub management status field is directly appended to the received data at a byte boundary. If the repeated packet is significantly longer than the value in the Decode Register requires and PCOMP is asserted the status fields will be delayed until the end of packet repetition. During this delay period MRXC clocks are inhibited but the MCRS signal remains asserted. Out of Window Collision (OWC): The out of window collision flag for a port goes active when a collision is experienced outside of the network slot time. Partition (PART): This flag goes active when a port becomes partitioned. Bad Link (BDLNK): The flag goes active when a port is configured for 10BASE-T operation has entered the link lost state. Short Event reception (SE): This flag goes active if the received packet is less than 74 bits long and no collision occurs during reception. 6.3 MANAGEMENT INTERFACE OPERATION The HUB Management interface provides a mechanism to combine repeater status information with packet information to form a hub management status packet. The interface, a serial bus consisting of carrier sense, received clock and received data, is designed to connect one or multiple RIC’s over a backplane bus to a DP83932 ‘‘SONIC’’ network controller. The SONIC and the RICs form a powerful entity for network statistics gathering. The interface consists of four pins: MRXC Management Receive ClockÐ10 MHz NRZ Clock output. MCRS MRXD Note: If PCOMP is asserted late in the packet, i.e., after the number of bytes defined by the packet compression register, then packet compression will not occur. The Management Interface may be fine tuned to meet the timing consideration of the SONIC and the access time of its associated packet memory. This refinement may be performed in two ways: 1. The default mode of operation of the Management interface is to only transfer packets over the bus which have a start of frame delimiter. Thus ‘‘packets’’ that are only preamble/jam and do not convey any source or destination address information are inhibited. This filtering may be disabled by writing a logic zero to the Management Interface Configuration or ‘‘MIFCON’’ bit in the Management and Interrupt Configuration Register. See Section 8 for details. 2. The Management bus has been designed to accommodate situations of maximum network utilization, for example when collision generated fragments occur; (these collision fragments may violate the IEEE802.3 IFG specification). The IFG required by the SONIC is a function of the time taken to release space in the receive FIFO and to perform end of packet processing (write status information into memory). These functions are primarily memory operations and consequently depend upon the bus latency and the memory access time of the system. In order to allow the system designer some discretion in choosing the speed of this memory, the RIC may be configured to protect the SONIC from a potential FIFO overflow. This is performed by utilizing the Inter Frame Gap Threshold Select Register. The value held in this register, plus one, defines, in network bit times, the minimum allowed gap between frames on the management bus. If the gap is smaller than this number then MCRS is asserted but MRXC clocks are inhibited. Consequently no data transfer is performed. Thus the system designer may make the decision whether to gather statistics on all packets even if they occur with very small IFGs or to monitor a subset. The status field, shown in Figure 6.1 , contains information which may be conveniently analyzed by considering it as Management Carrier SenseÐInput/Output indicating of valid data stream. Management Receive DataÐNRZ Data output synchronous to MRXC. PCOMP Packet CompressÐInput to truncate the packet’s data field. The first three signals mimic the interface between an Ethernet controller and a phase locked loop decoder (specifically the DP83932 SONIC and DP83910 SNI), these signals are driven by the RIC receiving the packet. MRXC and MRXD compose an NRZ serial data stream compatible with the DP83932. The PCOMP signal is driven by logic on the processor board. The actual data stream transferred over MRXD is derived from data transferred over the IRD InterRIC bus line. These two data streams differ in two important characteristics: 1. At the end of packet repetition a hub management status field is appended to the data stream. This status field, consisting of 7 bytes is shown in Figure 6.1 and 6.2 . The information field is obtained from a number of packet status registers described below. In common with the 802.3 protocol the least significant bit of a byte is transmitted first. 2. While the data field of the repeated packet is being transferred over the management bus, received clock signals on the MRXC pin may be inhibited. This operation is under the control of the Packet Compress pin PCOMP. If PCOMP is asserted during repetition of the packet then MRXC signals are inhibited when the number of bytes (after SFD) transferred over the management bus equals the number indicated in the Packet Compress Decode Register. This register provides a means to delay the effect of the PCOMP signal, which may be generated early in the packet’s repetition, until the desired moment. Packet compression may be used to reduce the amount of 45 6.0 Hub Management Support (Continued) counts the number of received bytes and transfers this over the management bus [PSR(4), (5)]. providing information of six different types. They are held in seven Packet Status Registers ‘‘PSRs’’: 1. The RIC and port address fields [PSR(0) and (1)] can uniquely identify the repeater port receiving the packet out of a potential maximum of 832 ports sharing the same management bus (64 RICs each with 13 ports). Thus all of the other status fields can be correctly attributed to the relevant port. 2. The status flags the RIC produces for the event counters or recording latches are supplied with each packet [PSR(2)]. Additionally the clean receive CLN status is supplied to allow the user to determine the reliability of the address fields in the packet. The CLN status bit [PSR(1)] is set if no collisions are experienced during the repetition of the address fields. 3. The RIC has an on-chip timer to indicate when, relative to the start of packet repetition, a collision, if any, occurred [PSR(3)]. There is also a timer which indicates how many bit times of IFG was seen on the network between repetition of this packet and the preceding one. This is provided by [PSR(6)]. 5. Appending a status field to a data packet will obviously result in a CRC error being flagged by the SONIC. For this reason the RIC monitors the repeated data stream to check for CRC and FAE errors. In the case of FAE errors the RIC provides additional dummy data bits, so that the status fields are always byte aligned. 6. As a final check upon the effectiveness of the management interface, the RIC transfers a bus specific status bit to the SONIC. This flag Packet Compress Done PCOMPD [PSR(0)], may be monitored by hub management software to check if the packet compression operation is enabled. Figure 6.2 shows an example of a packet being transmitted over the management bus. The first section of the diagram (moving from left to right) shows a short preamble and SFD pattern. The second region contains the packet’s address and the start of the data fields. During this time logic on the processor/SONIC card would determine if packet compression should be used on this packet. The PCOMP signal is asserted and packet transfer stops when the number of bytes transmitted equals the value defined in the decode register. Hence the MRXC signal is idle for the remainder of the packet’s data and CRC fields. The final region shows the transfer of the RIC’s seven bytes of packet status. The following pages describe these Hub Management registers which constitute the management status field. 4. If packet compression is employed, the receive byte count contained in the SONIC’s packet descriptor will indicate the number of bytes transferred over the management bus rather than the number of bytes in the packet. For this reason the RIC which receives the packet, Packet Status Register PSR D7 D6 D5 D4 D3 D2 D1 D0 TXCOL PSR(0) A5 A4 A3 A2 A1 A0 PCOMPD PSR(1) CRCER FAE COL CLN PA3 PA2 PA1 PA0 PSR(2) SE OWC NSFD PLER ELBER JAB CBT9 CBT8 PSR(3) Collision Bit Timer CBT7 CBT6 CBT5 CBT4 CBT3 CBT2 CBT1 CBT0 PSR(4) Lower Repeat Byte Count RBY7 RBY6 RBY5 RBY4 RBY3 RBY2 RBY1 RBY0 PSR(5) Upper Repeat Byte Count RBY15 RBY14 RBY13 RBY12 RBY11 RBY10 RBY9 RBY8 PSR(6) Inter Frame Gap Bit Timer IBT7 IBT6 IBT5 IBT4 IBT3 IBT2 IBT1 IBT0 Note: These registers may only be reliably accessed via the management interface. Due to the nature of these registers they may not be accessed (read or write cycles) via the processor interface. FIGURE 6.1. Hub Management Status Field 46 47 FIGURE 6.2. Operation of the Management Bus Note: In this example the Management Bus is configured to use active low signals. TL/F/11096 – 18 6.0 Hub Management Support (Continued) 6.0 Hub Management Support (Continued) Packet Status Register 0 D7 D6 D5 D4 D3 D2 D1 D0 A5 A4 A3 A2 A1 A0 PCOMPD resv Bit Symbol D0 resv D1 PCOMPD D(7:2) Description RESERVED FOR FUTURE USE: This bit is currently undefined, management software should not examine the state of this bit. A(5:0) PACKET COMPRESSION DONE: If packet compression is utilized, this bit informs the user that compression was performed, i.e., the packet was long enough to require compression. RIC ADDRESS (5:0): This address is defined by the user and is supplied when writing to the RIC Address Register. It is used by hub management software to distinguish between RICs in a multi-RIC system. Packet Status Register 1 D7 D6 D5 D4 D3 D2 D1 D0 CRCER FAE COL CLN PA3 PA2 PA1 PA0 Bit Symbol D(3:0) PA(3:0) Description PORT ADDRESS: This field defines the port which is receiving the packet. D4 CLN CLEAN RECEIVE: This bit is asserted from the start of reception, and is deasserted if a collision occurs within a window from the start of reception to the end of the 13th byte after SFD detection. If no SFD is detected the window is extended to the end of reception. D5 COL COLLISION: If a receive or transmit collision occurs during packet repetition the collision bit is asserted. D6 FAE FRAME ALIGNMENT ERROR: This bit is asserted if a Frame Alignment Error occurred in the repeated packet. D7 CRCER CRC ERROR: This bit is asserted if a CRC Error occurred in the repeated packet. This status flag should not be tested if the COL bit is asserted since the error may be simply due to the collision. 48 6.0 Hub Management Support (Continued) Packet Status Register 2 D7 D6 D5 D4 D3 D2 D1 D0 SE OWC NSFD PLER ELBER JAB CBT9 CBT8 Bit Symbol D(1:0) CT(9:8) Description COLLISION TIMER BITS 9 AND 8: These two bits are the upper bits of the collision bit timer. D2 JAB JABBER EVENT: This bit indicates that the receive packet was so long the repeater was forced to go into a jabber protect condition. D3 ELBER D4 PLER PHASE LOCK LOOP ERROR: The packet suffered sufficient jitter/noise corruption to cause the phase lock loop decoder to lose lock. D5 NSFD NON SFD: The repeated packet did not contain a Start of Frame Delimiter. When this bit is set the Repeat Byte Counter counts the length of the entire packet. When this bit is not set the byte counter only counts post SFD bytes. ELASTICITY BUFFER ERROR: During the packet an Elasticity Buffer under/overflow occurred. Note: The operation of this bit is not inhibited by the occurrence of a collision during packet repetition (see description of the Repeat Byte Counter below). D6 OWC D7 SE OUT OF WINDOW COLLISION: The packet suffered an out of window collision. SHORT EVENT: The receive activity was so small it met the criteria to be classed as a short event. The other registers comprise the remainder of the collision timer register [PSR(3)], the Repeat Byte Count registers [PSR(4), (5)], and the Inter Frame Gap Counter ‘‘IFG’’ register [PSR(6)]. 6.4 DESCRIPTION OF HARDWARE CONNECTION FOR MANAGEMENT INTERFACE The RIC has been designed so it may be connected to the Management bus directly or via external bus transceivers. The latter is advantageous in large repeaters. In this application the system backplane is often heavily loaded beyond the drive capabilities of the on-chip bus drivers. The uni-directional nature of information transfer on the MCRS, MRXD and MRXC signals, means a single open drain output pin is adequate for each of these signals. The Management Enable (MEN) RIC output pin performs the function of a drive enable for an external bus transceiver if one is required. In common with the Inter-RIC bus signals ACTN, ANYXN, COLN and IRE the MCRS active level asserted by the MCRS output is determined by the state of the BINV Mode Load configuration bit. Collision Bit Timer The Collision Timer counts in bit times the time between the start of repetition of the packet and the detection of the packet’s first collision. The Collision counter increments as the packet is repeated and freezes when a collision occurs. The value in the counter is only valid when the collision bit ‘‘COL’’ in [PSR(1)] is set. Repeat Byte Counter The Repeat Byte Counter is a 16 bit counter which can perform two functions. In cases where the transmitted packet possesses an SFD, the byte counter counts the number of received bytes after the SFD field. Alternatively if no SFD is repeated the counter reflects the length of the packet, counted in bytes, starting at the beginning of the preamble field. When performing the latter function the counter is shortened to 7 bits. Thus the maximum count value is 127 bytes. The mode of counting is indicated by the ‘‘NSFD’’ bit in [PSR(2)]. In order to check if the received packet was genuinely a Non-SFD packet, the status of the COL bit should be checked. During collisions SFD fields may be lost or created, Management software should be robust to this kind of behaviour. 7.0 Port Block Functions The RIC has 13 port logic blocks (one for each network connection). In addition to the packet repetition operations already described, the port block performs two other functions: 1. The physical connection to the network segment (transceiver function). 2. It provides a means to protect the network from malfunctioning segments (segment partition). Each port has its own status register. This register allows the user to determine the current status of the port and configure a number of port specific functions. Inter Frame Gap (IFG) Bit Timer The IFG counter counts in bit times the period in between repeater transmissions. The IFG counter increments whenever the RIC is not transmitting a packet. If the IFG is long, i.e., greater than 255 bits the counter sticks at this value. Thus an apparent count value of 255 should be interpreted as 255 or more bit times. 49 7.0 Port Block Functions (Continued) 7.1 TRANSCEIVER FUNCTIONS 7.2 SEGMENT PARTITION The RIC may connect to network segments in three ways: Each of the RIC’s ports has a dedicated state machine to perform the functions defined by the IEEE partition algorithm as shown in Figure 7.3 . To allow users to customize this algorithm for different applications a number of user selected options are available during device configuration at power up (the Mode Load Cycle). Five different options are provided: 1. Operation of the 13 partition state machines may be disabled via the disable partition DPART configuration bit (Pin D6). 2. The value of consecutive counts required to partition a segment (the CCLimit specification) may be set at either 31 or 63 consecutive collisions. 3. The use of the TW5 specification in the partition algorithm differentiates between collisions which occur early in a packet (before TW5 has elapsed) and those which occur late in the packet (after TW5 has elapsed). These late or ‘‘out of window’’ collisions can be regarded in the same manner as early collisions if the Out of Window Collision Enable OWCE option is selected. This configuration bit is applied to the D4 pin during the Mode Load operation. The use of OWCE delays until the end of the packet the operation of the state diagram branch marked (1) and enables the branch marked (2) in Figure 7.3 . 4. The operation of the ports’ state machines when reconnecting a segment may also be modified by the user. The Transmit Only TXONLY configuration bit allows the user to prevent segment reconnection unless the reconnecting packet is being sourced by the repeater. In this case the repeater is transmitting on to the segment, rather than the segment transmitting when the repeater is idle. The normal mode of reconnection does not differentiate between such packets. The TXONLY configuration bit is input on Pin D5 during the Mode Load cycle. If this option is selected the operation of the state machine branch marked (3) in Figure 7.3 is affected. 5. The RIC may be configured to use an additional criterion for segment partition. This is referred to as loop back partition. If this operation is selected the partition state machine monitors the receive and collision inputs from a network segment to discover if they are active when the port is transmitting. Thus determining if the network transceiver is looping back the data pattern from the cable. A port may be partitioned if no data or collision signals are seen by the partition logic in the following window: 61 to 96 network bit times after the start of transmission see data sheet Section 8 for details. A segment partitioned by this operation may be reconnected in the normal manner. In addition to the autonomous operation of the partition state machines, the user may reset these state machines. This may be done individually to each port by writing a logic one to the PART bit in its status register. The port’s partition state machine and associated counters are reset and the port is reconnected to the network. The reason why a port become partitioned may be discovered by the user by reading the port’s status register. 1. Over AUI cable to transceiver boxes, 2. Directly to board mounted transceivers, 3. To twisted pair cable via a simple interface. The first method is only supported by RIC port 1 (the AUI port). Options (2) and (3) are available on ports 2 to 13. The selection of the desired option is made at device initialization during the Mode Load operation. The Transceiver Bypass XBYPAS configuration bits are used to determine whether the ports will utilize the on-chip 10BASE-T transceiver or bypass these in favour of external transceivers. Four possible combinations of port utilization are supported: All ports (2 to 13) use the external Transceiver Interface. Ports 2 to 5 use the external interface, 6 to 13 use the internal 10BASE-T transceivers. Ports 2 to 7 use the external interface, 8 to 13 use the internal 10BASE-T transceivers. All ports (2 to 13) use the internal 10BASE-T transceivers. 10BASE-T Transceiver Operation The RIC contains virtually all the digital and analog circuits required for connection to 10BASE-T network segments. The only additional active component is an external driver packet. The connection for a RIC port to a 10BASE-T segment is shown in Figure 7.1 . The diagram shows the components required to connect one of the RIC’s ports to a 10BASE-T segment. The major components are the driver package, a member of the 74ACT family, and an integrated filter/choke network. The operation of the 10BASE-T transceiver’s logical functions may be modified by software control. The default mode of operation is for the transceivers to transmit and expect reception of link pulses. This may be modified if a logic one is written to the GDLNK bit of a port’s status register. The port’s transceiver will operate normally but will not transmit link pulses nor monitor their reception. Thus the entry to a link fail state and the associated modification of transceiver operation will not occur. The on-chip 10BASE-T transceivers automatically detect and correct the polarity of the received data stream. This polarity detection scheme relies upon the polarity of the received link pulses and the end of the packet waveform. Polarity detection and correction may be disabled under software control as follows: 1) Write the value 07H to the Page Select Register (address 10H). 2) Write the value 02H to the address 11H. (Note that address 11H will read back 00H after writing 02H to it). This is the only exception for accessing any of the reserved pages 4 to 7. External Transceiver Operation RIC ports 2 to 13 may be connected to media other than twisted-pair by opting to bypass the on-chip transceivers. When using external transceivers the user must have the external transceivers perform collision detection and the other functions associated with an IEEE 802.2 Media Access Unit. Figure 7.2 shows the connection between a repeater port and a coaxial transceiver using the AUI type interface. 50 7.0 Port Block Functions (Continued) Note that the link disable and port disable functions are mutually exclusive functions, i.e., disabling link does not affect receiving and transmitting from/to that port and disabling a port does not disable link. When a port is disabled packet transmission and reception between the port’s segment and the rest of the network is prevented. 7.3 PORT STATUS REGISTER FUNCTIONS Each RIC port has its own status register. In addition to providing status concerning the port and its network segment the register allows the following operations to be performed upon the port: 1. Port disable 2. Link Disable 3. Partition reconnection 4. Selection between normal and reduced squelch levels TL/F/11096 – 19 Note: For recommended modules, see ‘‘Ethernet Magnetics Vendors for 10BASE-T, 10BASE2, and 10BASE5’’ in Section 5 of this Databook. FIGURE 7.1. Port Connection to a 10BASE-T Segment 51 7.0 Port Block Functions (Continued) TL/F/11096 – 20 FIGURE 7.2. Port Connection to a 10BASE2 Segment (AUI Interface Selected) The preceding diagrams show a RIC port (Numbers 2 to 13) connected to a 10BASE-T and a 10BASE2 segment. The values of any components not indicated above are to be determined. 52 7.0 Port Block Functions (Continued) TL/F/11096 – 21 FIGURE 7.3. IEEE Segment Partition Algorithm 53 8.0 RIC Registers These registers may be directly accessed at any time via the RA(4:0) pins, (RA4 e 0). The upper half of the register map, (RA4 e 1), is organized as 4 pages of registers: RIC Register Address Map The RIC’s registers may be accessed by applying the required address to the five Register Address (RA(4:0)) input pins. Pin RA4 makes the selection between the upper and lower halves of the register array. The lower half of the register map consists of 16 registers: 1 RIC Real Time Status and Configuration register, 13 Port Real Time Status registers, 1 RIC Configuration Register 1 Real Time Interrupt Status Register. Event Count Configuration page (0), Event Record page (1), Lower Event Count page (2) Upper Event Count page (3) Register access within these pages is also performed using the RA(4:0) pins, (RA4 e 1). Page switching is performed by writing to the Page Selection bits (PSEL2, 1, 0). These bits are found in the Page Select Register, located at address 10 hex on each page of the upper half of the register array. AT power on these bits default to 0 Hex, i.e., page zero. 54 8.0 RIC Registers (Continued) Register Memory Map Name Address Page (0) Page (1) Page (2) 00H RIC Status and Configuration Register 01H Port 1 Status Register 02H Port 2 Status Register 03H Port 3 Status Register 04H Port 4 Status Register 05H Port 5 Status Register 06H Port 6 Status Register 07H Port 7 Status Register 08H Port 8 Status Register 09H Port 9 Status Register 0AH Port 10 Status Register 0BH Port 11 Status Register 0CH Port 12 Status Register 0DH Port 13 Status Register 0EH RIC Configuration Register 0FH Real Time Interrupt Register 10H Page Select Register 11H Device Type Register Port 1 Event Record Register (ERR) 12H Lower Event Count Mask Register (ECMR) Port 2 ERR Page (3) Port 1 Lower Event Count Register (ECR) Port 8 Lower ECR 13H Upper ECMR Port 3 ERR Port 1 Upper ECR Port 8 Upper ECR 14H Event Record Mask Register Port 4 ERR Port 2 Lower ECR Port 9 Lower ECR 15H resv Port 5 ERR Port 2 Upper ECR Port 9 Upper ECR 16H Management/Interrupt Configuration Register Port 6 ERR Port 3 Lower ECR Port 10 Lower ECR 17H RIC Address Register Port 7 ERR Port 3 Upper ECR Port 10 Upper ECR 18H Packet Compress Decode Register Port 8 ERR Port 4 Lower ECR Port 11 Lower ECR 19H resv Port 9 ERR Port 4 Upper ECR Port 11 Upper ECR 1AH resv Port 10 ERR Port 5 Lower ECR Port 12 Lower ECR 1BH resv Port 11 ERR Port 5 Upper ECR Port 12 Upper ECR 1CH resv Port 12 ERR Port 6 Lower ECR Port 13 Lower ECR 1DH resv Port 13 ERR Port 6 Upper ECR Port 13 Upper ECR 1EH resv Port 7 Lower ECR 1FH IFG Threshold Port 7 Upper ECR Note: All registers marked resv on pages 0 to 3 must not be accessed by the user. The other register pages, 4 to 7, are also reserved. 55 8.0 RIC Registers (Continued) Register Array Bit Map Addresses 00H to 10H Address (Hex) D7 D6 D5 D4 D3 D2 D1 D0 00 BINV BYPAS2 BYPAS1 APART JAB AREC ACOL resv 01 to 0D DISPT SQL PTYPE1 PTYPE0 PART REC COL GDLNK 0E MINMAX DPART TXONLY OWCE LPPART CCLIM Tw2 resv 0F IVCTR3 IVCTR2 IVCTR1 IVCTR0 ISRC3 ISRC2 ISRC1 ISRC0 Register Array Bit Map Addresses 10H to 1FH Page (0) Address (Hex) D7 D6 D5 D4 D3 D2 D1 D0 10 FC HC LC FF resv PSEL2 PSEL1 PSEL0 11 0 0 0 0 0 0 0 0 12 BDLNKC PARTC RECC SEC NSFDC PLERC ELBERC JABC 13 resv resv OWCC RXCOLC TXCOLC resv FWF ROR 14 BDLNKE PARTE OWCE SEE NSFDE PLERE ELBERE JABE 16 IFC IHC ILC IFF IREC ICOL IPART MIFCON 17 A5 A4 A3 A2 A1 A0 resv resv 18 PCD7 PCD6 PCD5 PCD4 PCD3 PCD2 PCD1 PCD0 1F IFGT7 IFGT6 IFGT5 IFGT4 IFGT3 IFGT2 IFGT1 IFGT0 Register Array Bit Map Addresses 10H to 1FH Page (1) Address (Hex) D7 D6 D5 D4 D3 D2 D1 D0 10 FC HC LC FF resv PSEL2 PSEL1 PSEL0 11 to 1D BDLNK PART OWC SE NSFD PLER ELBER JAB Register Array Bit Map Addresses 10H to 1FH Pages (2) and (3) Address (Hex) D7 D6 D5 D4 D3 D2 D1 D0 10 FC HC LC FF resv PSEL2 PSEL1 PSEL0 11 Ð Ð Ð Ð Ð Ð Ð Ð Even Locations EC7 EC6 EC5 EC4 EC3 EC2 EC1 EC0 Odd Locations EC15 EC14 EC13 EC12 EC11 EC10 EC9 EC8 56 8.0 RIC Registers (Continued) RIC Status and Configuration Register (Address 00H) The lower portion of this register contains real time information concerning the operation of the RIC. The upper three bits represent the chosen configuration of the transceiver interface employed. D7 D6 D5 D4 D3 D2 D1 D0 BINV BYPAS2 BYPAS1 APART JAB AREC ACOL resv Bit R/W Symbol Access D0 R resv D1 R ACOL ANY COLLISIONS: 0: A collision is occurring at one or more of the RIC’s ports. 1: No collisions. D2 R AREC ANY RECEIVE: 0: One of the RIC’s ports is the current packet or collision receiver. 1: No packet or collision reception within this RIC. D3 R JAB D4 R APART ANY PARTITION: 0: One or more ports are partitioned. 1: No ports are partitioned. D5 R BYPAS1 These bits define the configuration of ports 2 to 13 i.e., their use if the internal 10BASE-T transceivers or the external (AUI-like) transceiver interface. D6 R BYPAS2 D7 R BINV Description RESERVED FOR FUTURE USE: Reads as a logic 0. JABBER PROTECT: 0: The RIC has been forced into jabber protect state by one of its ports or by another port on the Inter-RIC bus, (Multi-RIC operations). 1: No jabber protect conditions exist. BUS INVERT: This register bit informs whether the Inter-RIC signals: IRE, ACTN, ANYXN, COLN and Management bus signal MCRS are active high or low. 0: Active high 1: Active low 57 8.0 RIC Registers (Continued) Port Real Time Status Registers (Address 01H to 0DH) D7 D6 D5 D4 D3 D2 DISPT EGP PTYPE1 Bit R/W Symbol D0 R/W GDLNK PTYPE0 PART REC D1 D0 COL GDLNK Description GOOD LINK: 0: Link pulses are being received by the port. 1: Link pulses are not being received by the port logic. Note: Writing a 1 to this bit will cause the 10BASE-T transceiver not the transmit or monitor the reception of link pulses. If the internal 10BASE-T transceivers are not selected or if port 1 (AUI port) is read, then this bit is undefined. D1 R COL COLLISION: 0: A collision is happening or has occurred during the current packet. 1: No collisions have occurred as yet during this packet. D2 R REC RECEIVE: 0: This port is now or has been the receive source of packet or collision information for the current packet. 1: This port has not been the receive source during the current packet. D3 R/W PART PARTITION: 0: This port is partitioned. 1: This port is not partitioned. Writing a logic one to this bit forces segment reconnection and partition state machine reset. Writing a zero to this bit has no effect. D(5, 4) R PTYPE0 PTYPE1 D6 R/W SQL D7 R/W DISPT PARTITION TYPE 0 PARTITION TYPE 1 The partition type bits provide information specifying why the port is partitioned. PTYPE1 PTYPE0 0 0 Consecutive Collision Limit Reached Information 0 1 Excessive Length of Collision Limit Reached 1 0 Failure to See Data Loopback from Transceiver in Monitored Window 1 1 Processor Forced Reconnection SQUELCH LEVEL: 0: Port operates with normal IEEE receive squelch level. 1: Port operates with reduced receive squelch level. Note: This bit has no effect when the external transceiver is selected. DISABLE PORT: 0: Port operates as defined by repeater operations. 1: All port activity is prevented. 58 8.0 RIC Registers (Continued) RIC Configuration Register (Address 0EH) This register displays the state of a number of RIC configuration bits loaded during the Mode Load operation. D7 D6 D5 D4 D3 D2 D1 D0 MINMAX DPART TX ONLY OWCE LPPART CCLIM Tw2 resv Bit R/W Symbol Description D0 R resv RESERVED FOR FUTURE USE: Value set at logic one. D1 R Tw2 CARRIER RECOVERY TIME: 0: Tw2 set at 5 bits. 1: Tw2 set at 3 bits. D2 R CCLIM D3 R LPPART D4 R OWCE OUT OF WINDOW COLLISION ENABLE: 0: Out of window collisions are treated as in window collisions by the segment partition state machines. 1: Out of window collisions are treated as out of window collisions by the segment partition state machines. D5 R TXONLY ONLY RECONNECT UPON SEGMENT TRANSMISSION: 0: A segment will only be reconnected to the network if a packet transmitted by the RIC onto that segment fulfills the requirements of the segment reconnection algorithm. 1: A segment will be reconnected to the network by any packet on the network which fullfills the requirements of the segment reconnection algorithm. D6 R DPART D7 R MINMAX CONSECUTIVE COLLISION LIMIT: 0: Consecutive collision limit set at 63 collisions. 1: Consecutive collision limit set at 31 collisions. LOOPBACK PARTITION: 0: Partitioning upon lack of loopback from transceivers is enabled. 1: Partitioning upon lack of loopback from transceivers is disabled. DISABLE PARTITION: 0: Partitioning of ports by on-chip algorithms is prevented. 1: Partitioning of ports by on-chip algorithms is enabled. MINIMUM/MAXIMUM DISPLAY MODE: 0: LED display set in minimum display mode. 1: LED display set in maximum display mode. 59 8.0 RIC Registers (Continued) Real Time Interrupt Register (Address 0FH) The Real Time Interrupt register (RTI) contains information which may change on a packet by packet basis. Any remaining interrupts which have not been serviced before the following packet is transmitted are cleared. Since multiple interrupt sources may be displayed by the RTI a priority scheme is implemented. A read cycle to the RTI gives the interrupt source and an address vector indicating the RIC port which generated the interrupt. The order of priority for the display of interrupt information is as follows: 1. The receive source of network activity (Port N), 2. The first RIC port showing collision 3. A port partitioned or reconnected. During the repetition of a single packet it is possible that multiple ports may be partitioned or alternatively reconnected. The ports have equal priority in displaying partition/reconnection information. This data is derived from the ports by the RTI register as it polls consecutively around the ports. Reading the RTI clears the particular interrupt. If no interrupt sources are active the RTI returns a no valid interrupt status. D7 D6 D5 D4 D3 D2 D1 D0 IVCTR3 IVCTR2 IVCTR1 IVCTR0 ISRC3 ISRC2 ISRC1 ISRC0 Symbol Access Bit R/W D(3:0) R ISCR(3:0) D(7:4) R IVCTR(3:0) Description INTERRUPT SOURCE: These four bits indicate the reason why the interrupt was generated. INTERRUPT VECTOR: This field defines the port address responsible for generating the interrupt. The following table shows the mapping of interrupt sources onto the D3 to D0 pins. Essentially each of the three interrupt sources has a dedicated bit in this field. If a read to the RTI produces a low logic level on one of these bits then the interrupt source may be directly decoded. Associated with the source of the interrupt is the port where the event is occurring. If no unmasked events (receive, collision, etc.), have occurred when the RTI is read then an all ones pattern is driven by the RIC onto the data pins. D7 D6 D5 D4 D3 D2 D1 D0 Comments PA3 PA2 PA1 PA0 1 1 0 1 First Collision PA(3:0) e Collision Port Address PA3 PA2 PA1 PA0 1 0 1 1 Receive PA(3:0) e Receive Port Address PA3 PA2 PA1 PA0 0 1 1 1 Partition Reconnection PA(3:0) e Partition Port Address 1 1 1 1 1 1 1 1 No Valid Interrupt 60 8.0 RIC Registers (Continued) Page Select Register ((All Pages) Address 10H) The Page Select register performs two functions: 1. It enables switches to be made between register pages, 2. It provides status information regarding the Event Logging Interrupts. D7 D6 D5 D4 D3 D2 D1 D0 FC HC LC FF resv PSEL2 PSEL1 PSEL0 Bit R/W Symbol Description D(2:0) R/W PSEL(2:0) PAGE SELECT BITS: When read these bits indicate the currently selected Upper Register Array Page. Write cycles to these locations facilitates page swapping. D3 R resv D4 R FF FLAG FOUND: This indicates one of the unmasked event recording latches has been set. D5 R LC LOW COUNT: This indicates one of the port event counters has a value less than 00FF Hex. D6 R HC HIGH COUNT: This indicates one of the port event counters has a value greater than C000 Hex. D7 R FC FULL COUNTER: This indicates one of the port event counters has a value equal to FFFF Hex. RESERVED FOR FUTURE USE Device Type Register (Page 0H Address 11H) This register may be used to distinguish different revisions of RIC. If this register is read it will return a different value each for DP83950 revisions. (Contact National Semiconductor for revision information.) Write operations to this register have no effect upon the contents. D7 D6 D5 D4 D3 D2 D1 D0 0 0 0 0 0 0 X X 61 8.0 RIC Registers (Continued) Lower Event Count Mask Register (Page 0H Address 12H) D7 D6 D5 D4 D3 D2 D1 D0 BDLNKC PARTC RECC SEC NSFDC PLERC ELBER C JABC Bit R/W Symbol D0 R/W JABC Description D1 R/W ELBERC D2 R/W PLERC PHASE LOCK ERROR COUNT ENABLE: Enables recording of Carrier Error events. D3 R/W NSFDC NON SFD COUNT ENABLE: Enables recording of Non SFD packet events. D4 R/W SEC D5 R/W RECC RECEIVE COUNT ENABLE: Enables recording of Packet Receive (port N status) events that do not suffer collisions. D6 R/W PARTC PARTITION COUNT ENABLE: Enables recording of Partition events. D7 R/W BDLNKC BAD LINK COUNT ENABLE: Enables recording of Bad Link events. JABBER COUNT ENABLE: Enables recording of Jabber Protect events. ELASTICITY BUFFER ERROR COUNT ENABLE: Enables recording of Elasticity Buffer Error events. SHORT EVENT COUNT ENABLE: Enables recording of Short events. Upper Event Count Mask Register (Page 0H Address 13H) D7 D6 D5 D4 D3 D2 D1 D0 resv resv OWCC RXCOLC TXCOLC resv FWF ROR Bit R/W Symbol D0 R/W ROR RESET ON READ: This bit selects the action a read operation has upon a port’s event counter: 0: No effect upon register contents. 1: The counter register is reset. D1 R/W FWF FREEZE WHEN FULL: This bit controls the freezing of the Event Count registers when the counter is full (FFFF Hex) RESERVED FOR FUTURE USE: This bit should be written with a low logic level. Description D2 R resv D3 R/W TXCOLC TRANSMIT COLLISION COUNT ENABLE: Enables recording of transmit collision events only. D4 R/W RXCOLC RECEIVE COLLISION COUNT ENABLE: Enables recording of receive collision events only. D5 R/W OWCC D(7: 6) R resv OUT OF WINDOW COLLISION COUNT ENABLE: Enables recording of out of window collision events only. RESERVED FOR FUTURE USE: These bits should be written with a low logic level. Note 1: To count all collisions then both the TXCOLC and RXCOLC bits must be set. The OWCC bit should not be set otherwise the port counter will be incremented twice when an out of collision window collision occurs. The OWCC bit alone should be set if only out of window collision are to be counted. Note 2: Writing a 1 enables the event to be counted. 62 8.0 RIC Registers (Continued) Event Record Mask Register (Page 0H Address 14H) D7 D6 D5 D4 D3 D2 D1 D0 BDLNKE PARTE OWCE SEE NSFDE PLERE ELBERE JABE Bit R/W Symbol D0 R/W JABE Description D1 R/W ELBERE D2 R/W PLERE PHASE LOCK ERROR ENABLE: Enables recording of Carrier Error events. D3 R/W NSFDE NON SFD ENABLE: Enables recording of Non SFD packet events. D4 R/W SEE D5 R/W OWCE OUT OF WINDOW COLLISION COUNT ENABLE: Enables recording of Out of Window Collision events only. D6 R/W PARTE PARTITION ENABLE: Enables recording of Partition events. D7 R/W BDLNKE BAD LINK ENABLE: Enables recording of Bad Link Events. JABBER ENABLE: Enables recording of Jabber Protect events. ELASTICITY BUFFER ERROR ENABLE: Enables recording of Elasticity Buffer Error events. SHORT EVENT ENABLE: Enables recording of Short Events. Note: Writing a 1 enables the event to be recorded. 63 8.0 RIC Registers (Continued) Interrupt and Management Configuration Register (Page 0H Address 16H) This register powers up with all bits set to one and must be initialized by a processor write cycle before any events will generate interrupts. D7 D6 D5 D4 D3 D2 D1 D0 IFC IHC ILC IFF IREC ICOL IPART MIFCON Bit R/W Symbol D0 R/W MIFCON Description D1 R/W IPART D2 R/W ICOL INTERRUPT ON COLLISION: 0: Interrupts will be generated(1) if this RIC has a port which experiences a collision, Single RIC applications, or contains a port which experiences a receive collision or is the first port to suffer a transmit collision in a packet in Multi-RIC applications. 1: No interrupts are generated by this condition. D3 R/W IREC INTERRUPT ON RECEIVE: 0: Interrupts will be generated(1) if this RIC contains the receive port for packet or collision activity. 1: No interrupts are generated by this condition. D4 R/W IFF INTERRUPT ON FLAG FOUND: 0: Interrupts will be generated(2) if one or more than one of the flags in the flag array is true. 1: No interrupts are generated by this condition. D5 R/W ILC INTERRUPT ON LOW COUNT: 0: Interrupt generated(2) when one or more of the Event Counters holds a value less than 256 counts. 1: No effect D6 R/W IHC INTERRUPT ON HIGH COUNT: 0: Interrupt generated(2) when one or more of the Event Counters holds a value in excess of 49152 counts. 1: No effect D7 R/W IFC INTERRUPT ON FULL COUNTER: 0: Interrupt generated(2) when one or more of the Event Counters is full. 1: No effect MANAGEMENT INTERFACE CONFIGURATION: 0: All Packets repeated are transmitted over the Management bus. 1: Packets repeated by the RIC which do not have a Start of Frame Delimiters are not transmitted over the Management bus. INTERRUPT ON PARTITION: 0: Interrupts will be generated(1) if a port becomes Partitioned. 1: No interrupts are generated by this condition. Note 1: (RTI pin goes active) Note 2: (ELI pin goes active) 64 8.0 RIC Registers (Continued) RIC Address Register (Page 0H Address 17H) This register may be used to differentiate between RICs in a multi-RIC repeater system. The contents of this register form part of the information available through the management bus. D7 D6 D5 D4 D3 D2 D1 D0 A5 A4 A3 A2 A1 A0 res res Packet Compress Decode Register (Page 0H Address 18H) This register is used to determine the number of bytes in the data field of a packet which are transferred over the management bus when the packet compress option is employed. The register bits perform the function of a direct binary decode. Thus up to 255 bytes of data may be transferred over the management bus if packet compression is selected. D7 D6 D5 D4 D3 D2 D1 D0 PCD7 PCD6 PCD5 PCD4 PCD3 PCD2 PCD1 PCD0 Inter Frame Gap Threshold Select Register (Page 0H Address 1FH) This register is used to configure the hub management interface to provide a certain minimum inter frame gap between packets transmitted over the management bus. The value written to this register, plus one, is the magnitude in bit times of the minimum IFG allowed on the management bus. D7 D6 D5 D4 D3 D2 D1 D0 IFGT7 IFGT6 IFGT5 IFGT4 IFGT3 IFGT2 IFGT1 IFGT0 Port Event Record Registers (Page 1H Address 11H to 1DH) These registers hold the recorded events for the specified RIC port. The flags are cleared when the register is read. D7 D6 D5 D4 D3 D2 D1 D0 BDLNK PART OWC SE NSFD PLER ELBER JAB Bit R/W Symbol D0 R JAB Description D1 R ELBER D2 R PLER PHASE LOCK ERROR: A Phase Lock Error event has occurred. D3 R NSFD NON SFD: A Non SFD packet event has occurred. D4 R SE D5 R OWC OUT OF WINDOW COLLISION: An out of window collision event has occurred. D6 R PART PARTITION: A partition event has occurred. D7 R BDLNK BAD LINK: A link failure event has occurred. JABBER: A Jabber Protect event has occurred. ELASTICITY BUFFER ERROR: A Elasticity Buffer Error has occurred. SHORT EVENT: A Short event has occurred. Port Event Count Register (Pages 2H and 3H) The Event Count (EC) register shows the instantaneous value of the specified port’s 16-bit counter. The counter increments when an enabled event occurs. The counter may be cleared when it is read and prevented from rolling over when the maximum count is reached by setting the appropriate control bits in the Upper Event Count mask register. Since the RIC’s processor port is octal and the counters are 16 bits long a temporary holding register is employed for register reads. When one of the counters is read, either high or low byte first, all 16 bits of the counter are transferred to a holding register. Provided the next read cycle to the counter array accesses the same counter’s, other byte, then the read cycle accesses the holding register. This avoids the problem of events occurring in between the two processor reads and indicating a false count value. In order to enter a new value to the holding register a different counter must be accessed or the same counter byte must be re-read. Lower Byte D7 D6 D5 D4 D3 D2 D1 D0 EC7 EC6 EC5 EC4 EC3 EC2 EC1 EC0 Upper Byte D7 D6 D5 D4 D3 D2 D1 D0 EC15 EC14 EC13 EC12 EC11 EC10 EC9 EC8 65 9.0 AC and DC Specifications Absolute Maximum Ratings Storage Temperature Range (TSTG) b 65§ C to a 150§ C If Military/Aerospace specified devices are required, please contact the National Semiconductor Sales Office/Distributors for availability and specifications. Supply Voltage (VCC) 0.5V to 7.0V Power Dissipation (PD) Lead Temperature (TL) (Soldering, 10 seconds) 260§ C DC Input Voltage (VIN) DC Output Voltage (VOUT) ESD Rating (Rzap e 1.5k, Czap e 120 pF) 1500V b 0.5V to VCC a 0.5V b 0.5V to VCC a 0.5V 2W DC Specifications TA e 0§ C to a 70§ C, VCC e 5V g 5% unless otherwise specified PROCESSOR, LED, TWISTED PAIR PORTS, INTER-RIC AND MANAGEMENT INTERFACES Symbol Description Conditions Min Max VOH Minimum High Level Output Voltage IOH e b8 mA VOL Minimum Low Level Output Voltage IOL e 8 mA VIH Minimum High Level Input Voltage VIL Maximum Low Level Input Voltage IIN Input Current VIN e VCC or GND IOZ Maximum TRI-STATE Output Leakage Current VOUT e VCC or GND ICC Average Supply Current VIN e VCC or GND VCC e 5.25V VOD Differential Output Voltage (TX g ) 78X Termination and 270X Pulldowns VOB Differential Output Voltage Imbalance (TX g ) 78X Termination and 270X Pulldowns Typical: 40 mV VU Undershoot Voltage (TX g ) 78X Termination and 270X Pulldowns Typical: 80 mV VDS Differential Squelch Threshold (RX g , CD g ) VCM Differential Input Common Mode Voltage (RX g , CD g ) (Note 1) 3.5 Units V 0.4 2.0 V V 0.8 V b 1.0 1.0 mA b 10 10 mA 380 mA g 1200 mV AUI (PORT 1) Note 1: This parameter is guaranteed by design and is not tested. 66 g 550 b 175 b 300 mV 0 5.5 V 9.0 AC and DC Specifications (Continued) DC Specifications TA e 0§ C to a 70§ C, VCC e 5V g 5% unless otherwise specified (Continued) Symbol Description Conditions Min Max Units g 450 g 1200 mV PSEUDO AUI (PORTS 2–13) VPOD Differential Output Voltage (TX g ) 270X Termination and 1 kX Pulldowns VPOB Differential Output Voltage Imbalance (TX g ) 270X Termination and 1 kX Pulldowns Typical: 40 mV VPU Undershoot Voltage (TX g ) 270X Termination and 1 kX Pulldowns Typical: 80 mV VPDS Differential Squelch Threshold (RX g , CD g ) VPCM Differential Input Common Mode Voltage (Rx g , CD g ) (Note 1) b 175 b 300 mV 0 5.5 V g 300 (Note 2) g 585 mV mV TWISTED PAIR (PORTS 2–13) VRON Minimum Receive Squelch Threshold Normal Mode Reduced Mode g 340 Note 1: This parameter is guaranteed by design and is not tested. Note 2: The operation in Reduced Mode is not guaranteed below 300 mV. AC Specifications PORT ARBITRATION TIMING TL/F/11096 – 22 Number Symbol Parameter Min Max Units T1 ackilackol ACKI Low to ACKO Low 24 ns T2 ackihackoh ACKI High to ACKO High 21 ns Note: Timing valid with no receive or collision activities. Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 67 9.0 AC and DC Specifications (Continued) RECEIVING TIMINGSÐAUI PORTS Receive activity propagation start up and end delays for ports in non 10BASE-T mode TL/F/11096 – 23 Number Symbol Max Units T3a T4a rxaackol rxiackoh RX Active to ACKO Low RX Inactive to ACKO High Parameter Min 66 325 ns ns T5a T6a rxaactna rxiactni RX Active to ACTNd Active RX Inactive to ACTNd Inactive 105 325 ns ns Note: ACKI assumed high. Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 68 9.0 AC and DC Specifications (Continued) RECEIVE TIMINGÐ10BASE-T PORTS Receive activity propagation start up and end delays for ports in 10BASE-T mode TL/F/11096 – 24 Number Symbol Max Units T3t T4t rxaackol rxiackoh RX Active to ACKO Low RX Inactive to ACKO High Parameter Min 240 255 ns ns T5t T6t rxaactna rxiactni RX Active to ACTNd Active RX Inactive to ACTNd Inactive 270 265 ns ns Note: ACKI assumed high. TRANSMIT TIMINGÐAUI PORTS Transmit activity propagation start up and end delays for ports in non 10BASE-T mode TL/F/11096 – 25 Number Symbol Parameter T15a actnatxa ACTNd Active to TX Active Min Max Units 585 ns Note: ACKI assumed high. Note: ACTNd and ACTNs are tied together. Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 69 9.0 AC and DC Specifications (Continued) TRANSMIT TIMINGÐ10BASE-T PORTS Receive activity propagation start up and end delays for ports in 10BASE-T mode TL/F/11096 – 26 Number Symbol Parameter T15t actnatxa ACTNd Active to TX Active Min Max Units 790 ns Note: ACKI assumed high. Note: ACTNd and ACTNs are tied together. COLLISION TIMINGÐAUI PORTS Collision activity propagation start up and end delays for ports in non 10BASE-T mode TRANSMIT COLLISION TIMING TL/F/11096 – 27 Number T30a T31a Symbol Parameter cdaanyxna cdianyxni CD Active to ANYXN Active CD Inactive to ANYXN Inactive (Notes 1, 2) Min Max Units 65 400 ns ns Note 1: TX collision extension has already been performed and no other port is driving ANYXN. Note 2: Includes TW2. Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 70 9.0 AC and DC Specifications (Continued) COLLISION TIMINGÐAUI PORTS Collision activity propagation start up and end delays for ports in non 10BASE-T mode. RECEIVE COLLISION TIMING TL/F/11096 – 28 Number Symbol Max Units T32a T33a cdacolna cdicolni CD Active to COLN Active (Note 1) CD Inactive to COLN Inactive Parameter Min 55 215 ns ns T39 T40 colnajs colnije COLN Active to Start of Jam COLN Inactive to End of Jam (Note 2) 400 800 ns ns Note 1: PKEN assumed high. Note 2: Assuming reception ended before COLN goes inactive. TW2 is included in this parameter. Assuming ACTNd to ACTNs delay is 0. Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. COLLISION TIMINGÐ10BASE-T PORTS Collision activity propagation start up and end delays for ports in 10BASE-T mode TL/F/11096 – 29 Number Symbol Parameter T30t T31t colaanya colianyi Collision Active to ANYXN Active Collision Inactive to ANYXN Inactive (Note 1) Min Max Units 800 400 ns ns Note 1: TX collision extension has alreay been performed and no other port is asserting ANYXN. Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 71 9.0 AC and DC Specifications (Continued) COLLISION TIMINGÐALL PORTS TL/F/11096 – 38 Number Symbol Parameter Min Max Units T34 T35 anyamin anyitxai ANYXN Active Time ANYXN Inactive to TX to all Inactive 96 120 170 Bits ns T38 anyasj ANYXN Active to Start of Jam 400 ns TL/F/11096 – 39 Number Symbol Parameter T36 T37 actnitxi anyitxoi ACTN Inactive to TX Inactive ANYXN Inactive to TX ‘‘One Port Left’’ Inactive Min Max Units 120 405 170 ns ns Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 72 9.0 AC and DC Specifications (Continued) INTER RIC BUS OUTPUT TIMING TL/F/11096 – 35 Parameter Min Max Units T101 Number ircoh Symbol IRC Output High Time 45 55 ns T102 ircol IRC Output Low Time 45 55 ns T103 ircoc IRC Output Cycle Time 90 110 ns T104 actndapkena ACTNd Active to PKEN Active 555 T105 actndairea ACTNd Active to IRE Active 560 T106 ireoairca IRE Output Active to IRC Active T107 irdov IRD Output Valid from IRC T108 irdos IRD Output Stable Valid Time 90 T109 ircohirei IRC Output High to IRE Inactive 30 T110 ircclks Number of IRCs after IRE Inactive 5 ns ns 1.8 ms 10 ns ns 70 ns clks Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. Note: In a Multi-RIC system, the PKEN signal is valid only for the first receiving RIC. 73 9.0 AC and DC Specifications (Continued) INTER RIC BUS INPUT TIMING TL/F/11096 – 40 Number Symbol Parameter Min T111 ircih IRC Input High Time 20 Max Units ns T112 ircil IRC Input Low Time 20 ns T114 irdisirc IRD Input Setup to IRC 5 ns T115 irdihirc IRD Input Hold from IRC 10 T116 irchiire IRC Input High to IRE Inactive 10 ns 90 ns Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 74 9.0 AC and DC Specifications (Continued) MANAGEMENT BUS TIMING TL/F/11096 – 30 Number Symbol Parameter Min Max Units ns T50 mrxch MRXC High Time 45 55 T51 mrxcl MRXC Low Time 45 55 ns T52 mrxcd MRXC Cycle Time 90 110 ns T53 actndamena ACTNd Active to MEN Active 715 ns T54 actndamcrsa ACTNd Active to MCRS Active T55 mrxds MRXD Setup T56 mrxdh MRXD Hold 45 T57 mrxclmcrsi MRXC Low to MCRS Inactive b5 T58 mcrsimenl MCRS Inactive to MEN Low T59 mrxcclks Min Number of MRXCs after MCRS Inactive 5 T60 pcompw PCOMP Pulse Width 20 720 40 ns ns ns 6 ns 510 ns 5 Clks ns Note: The preamble on this bus consists of the following string: 01011. Note: In these diagrams the Inter-RIC and Management Busses are shown using active high signals, active low signals may also be used. See Section 5.5 Mode Load Operation. 75 9.0 AC and DC Specifications (Continued) MLOAD TIMING TL/F/11096 – 31 Number Symbol Parameter Min Max Units T61 mldats Data Setup 10 T62 mldath Data Hold 10 T63 mlabufa MLOAD Active to BUFEN Active 35 ns T64 mlibufi MLOAD Inactive to BUFEN Inactive 35 ns T65 mlw MLOAD Width ns ns 800 ns STROBE TIMING TL/F/11096 – 32 Number Symbol Parameter Min Max Units T66 stradrs Strobe Address Setup 80 115 ns T67 strdats Strobe Data Setup 40 65 ns T68 strdath Strobe Data Hold 135 160 ns T69 strw Strobe Width 30 65 ns CDEC TIMING TL/F/11096 – 41 Min Max Units T70 Number cdecpw Symbol CDEC Pulse Width Parameter 20 100 ns T71 cdeccdec CDEC to CDEC Width 200 76 ns 9.0 AC and DC Specifications (Continued) REGISTER READ TIMING TL/F/11096 – 33 Number Symbol Parameter Min T80 T81 rdadrs rdadrh Read Address Setup Read Address Hold 0 0 T82 T83 rdabufa rdibufi Read Active to BUFEN Active Read Inactive to BUFEN Inactive 95 T84 T85 rdadatv rddath Read Active to Data Invalid Read Data Hold 245 75 T86 T87 rdardya rdirdyi Read Active to RDY Active Read Inactive to RDY Inactive 340 T88 rdw Read Width 600 Note: Minimum high time between read/write cycles is 100 ns. 77 Max Units ns ns 345 35 ns ns ns ns 585 30 ns ns ns 9.0 AC and DC Specifications (Continued) REGISTER WRITE TIMING TL/F/11096 – 34 Number Symbol T90 T91 wradrs wradrh Write Address Setup Write Address Hold Parameter Min 0 0 T92 T93 wrabufa wribufi Write Active to BUFEN Active Write Inactive to BUFEN Inactive 95 T94 T95 wradatv wrdath Write Active to Data Valid Write Data Hold T96 T97 wrardya wrirdyi Write Active to RDY Active Write Inactive to RDY Inactive 340 T98 wrw Write Width 600 T99 wradt Write Active to Data TRI-STATE Max Units ns ns 355 35 ns ns 275 ns ns 585 30 ns ns 0 ns 350 ns Note: Assuming zero propagation delay on external buffer. Note: Minimum high time between read/write cycles is 100 ns. Note: The data will always TRI-STATE before BUFEN goes active with a load of 100 pF on the data bus. Note: When RDY is used, the minimum 600 ns write width does not have to be maintained. 10.0 AC Timing Test Conditions All specifications are valid only if the mandatory isolation is employed and all differential signals are taken to be at the AUI side of the pulse transformer. Input Pulse Levels (TTL/CMOS) GND to 3.0V Input Rise and Fall Times (TTL/CMOS) 5 ns Input and Output Reference Levels (TTL/CMOS) 1.5V Input Pulse Levels (Diff.) 2.0 VP-P Input and Output Reference Levels (Diff.) TRI-STATE Reference Levels 50% Point of the Differential Float (DV) g 0.5V TL/F/11096 – 36 Note 1: 100 pF, includes scope and jig capacitance. Note 2: S1 e Open for timing tests for push pull outputs. S1 e VCC for VOL test. S1 e GND for VOH test. S1 e VCC for High Impedance to active low and active low to High Impedance measurements. Output Load (See Figure Below) S1 e GND for High Impedance to active high and active high to High Impedance measurements. 78 Capacitance TA e 25§ C, f e 1 MHz Typ Units CIN Symbol Input Capacitance Parameter 7 pF COUT Output Capacitance 7 pF TL/F/11096 – 37 Note: In the above diagram, the TX a and TX b signals are taken from the AUI side of the isolation (pulse transformer). The pulse transformer used for all testing is the Pulse Engineering PE64103. 11.0 Physical Dimensions inches (millimeters) JEDEC, Molded Plastic Quad Flat Package (VUL) Order Number DP83950BVQB NS Package Number VUL160A 79 DP83950B RIC Repeater Interface Controller 11.0 Physical Dimensions inches (millimeters) (Continued) Pin Grid Array (U) Order Number DP83950BNU NS Package Number UP159A LIFE SUPPORT POLICY NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT OF NATIONAL SEMICONDUCTOR CORPORATION. As used herein: 1. Life support devices or systems are devices or systems which, (a) are intended for surgical implant into the body, or (b) support or sustain life, and whose failure to perform, when properly used in accordance with instructions for use provided in the labeling, can be reasonably expected to result in a significant injury to the user. National Semiconductor Corporation 1111 West Bardin Road Arlington, TX 76017 Tel: 1(800) 272-9959 Fax: 1(800) 737-7018 2. A critical component is any component of a life support device or system whose failure to perform can be reasonably expected to cause the failure of the life support device or system, or to affect its safety or effectiveness. National Semiconductor Europe Fax: (a49) 0-180-530 85 86 Email: cnjwge @ tevm2.nsc.com Deutsch Tel: (a49) 0-180-530 85 85 English Tel: (a49) 0-180-532 78 32 Fran3ais Tel: (a49) 0-180-532 93 58 Italiano Tel: (a49) 0-180-534 16 80 National Semiconductor Hong Kong Ltd. 13th Floor, Straight Block, Ocean Centre, 5 Canton Rd. Tsimshatsui, Kowloon Hong Kong Tel: (852) 2737-1600 Fax: (852) 2736-9960 National Semiconductor Japan Ltd. Tel: 81-043-299-2309 Fax: 81-043-299-2408 National does not assume any responsibility for use of any circuitry described, no circuit patent licenses are implied and National reserves the right at any time without notice to change said circuitry and specifications.