LTC4266 Quad IEEE 802.3at Power Over Ethernet Controller Description Features n n n n n n n n n n n Four Independent PSE Channels Compliant with IEEE 802.3at Type 1 and 2 0.34 Total Channel Resistance 130mW/Port at 600mA Advanced Power Management 8-Bit Programmable Current Limit (ILIM) 7-Bit Programmable Overload Currents (ICUT) Fast Shutdown of Preselected Ports 14.5-Bit Port Current/Voltage Monitoring 2-Event Classification Very High Reliability 4-Point PD Detection: 2-Point Forced Voltage 2-Point Forced Current High Capacitance Legacy Device Detection LTC4259A-1 and LTC4258 Pin and SW Compatible 1MHz I2C Compatible Serial Control Interface Midspan Backoff Timer Supports Proprietary Power Levels Above 25W Available in 38-Pin 5mm x 7mm QFN and 36-Pin SSOP Packages Applications n n High Power PSE Switches/Routers High Power PSE Midspans L, LT, LTC, LTM, Linear Technology and the Linear logo are registered trademarks and ThinSOT is a trademark of Linear Technology Corporation. All other trademarks are the property of their respective owners. The LTC(R)4266 is a quad PSE controller designed for use in IEEE 802.3 Type 1 and Type 2 (high power) compliant Power over Ethernet systems. External power MOSFETs enhance system reliability and minimize channel resistance, cutting power dissipation and eliminating the need for heatsinks even at Type 2 power levels. External power components also allow use at very high power levels while remaining otherwise compatible with the IEEE standard. 80V-rated port pins provide robust protection against external faults. The LTC4266 includes advanced power management features, including current and voltage readback and programmable ICUT and ILIM thresholds. Available C libraries simplify power-management software development; an optional AUTO mode provides fully IEEE-compliant standalone operation with no software required. Proprietary 4-point PD detection circuitry minimizes false PD detection while supporting legacy phone operation. Midspan operation is supported with built-in 2-event classification and backoff timing. Host communication is via a 1MHz I2C serial interface. The LTC4266 is available in a 5mm x 7mm QFN package that significantly reduces board space compared with competing solutions. A legacy-compatible 36-pin SSOP package is also available. Typical Application Complete 4-Port Ethernet High Power Source 3.3V 0.1F INT SHDN1 SHDN2 SHDN3 SHDN4 VDD AUTO MSD RESET MID SCL SDAIN SDAOUT AD0 LTC4266 AD1 AD2 AD3 DGND AGND VEE SENSE1 GATE1 OUT1 SENSE2 GATE2 OUT2 SENSE3 GATE3 OUT3 SENSE4 GATE4 OUT4 0.22F 100V s4 S1B s4 S1B s4 -50V PORT1 -50V SMAJ58A PORT2 1F PORT3 4266 TA01 PORT4 4266fa LTC4266 Absolute Maximum Ratings Supply Voltages (Note 1) AGND - VEE. ............................................ -0.3V to 80V DGND - VEE. ................................................ -0.3V to 80V VDD - DGND............................................... -0.3V to 5.5V Digital Pins SCL, SDAIN, SDAOUT, INT, SHDNn, MSD, ADn, RESET, AUTO, MID........... DGND -0.3V to VDD + 0.3V Analog Pins GATEn, SENSEn, OUTn........... VEE -0.3V to VEE + 80V Operating Temperature Range LTC4266C................................................. 0C to 70C LTC4266I..............................................-40C to 85C Junction Temperature (Note 2).............................. 125C Storage Temperature Range................... -65C to 150C Lead Temperature (Soldering, 10 sec)................... 300C Pin Configuration TOP VIEW INT 3 34 OUT1 SCL 4 33 GATE1 SDAOUT 5 32 SENSE1 SDAIN 6 31 OUT2 OUT1 29 OUT2 30 GATE2 AD3 4 28 GATE2 27 SENSE2 AD2 8 29 SENSE2 AD1 9 28 VEE AD1 6 27 OUT3 NC 11 26 GATE3 NC 12 25 SENSE3 NC 13 24 OUT4 NC 14 23 GATE4 22 SENSE3 21 OUT4 NC 11 20 GATE4 NC 12 SENSE4 AGND SHDN4 13 14 15 16 17 18 19 SHDN3 19 SHDN3 23 GATE3 DGND 10 VDD 20 SHDN4 24 OUT3 NC 9 21 AGND SHDN2 18 25 VEE DNC 8 22 SENSE4 SHDN1 17 26 VEE 39 AD0 7 AD0 10 VDD 16 30 SENSE1 SDAIN 3 AD2 5 DGND 15 31 GATE1 NC 2 SHDN2 7 38 37 36 35 34 33 32 SDAOUT 1 SHDN1 AD3 AUTO 35 AUTO MSD 2 RESET MID MID 36 MSD INT 1 SCL TOP VIEW RESET UHF PACKAGE 38-LEAD (5mm s 7mm) PLASTIC QFN EXPOSED PAD IS VEE (PIN 39) MUST BE SOLDERED TO PCB TJMAX = 125C, QJA = 34C/W GW36 PACKAGE 36-LEAD PLASTIC WIDE SSOP TJMAX = 125C, QJA = 80C/W Order Information LEAD FREE FINISH LTC4266CGW#PBF LTC4266IGW#PBF LTC4266CUHF#PBF LTC4266IUHF#PBF TAPE AND REEL LTC4266CGW#TRPBF LTC4266IGW#TRPBF LTC4266CUHF#TRPBF LTC4266IUHF#TRPBF PART MARKING* LTC4266 LTC4266 4266 4266 PACKAGE DESCRIPTION TEMPERATURE RANGE 0C to 70C 36-Lead Plastic Wide SSOP -40C to 85C 36-Lead Plastic Wide SSOP 0C to 70C 38-Lead (5mm x 7mm) Plastic QFN -40C to 85C 38-Lead (5mm x 7mm) Plastic QFN Consult LTC Marketing for parts specified with wider operating temperature ranges. *The temperature grade is identified by a label on the shipping container. Consult LTC Marketing for information on non-standard lead based finish parts. For more information on lead free part marking, go to: http://www.linear.com/leadfree/ For more information on tape and reel specifications, go to: http://www.linear.com/tapeandreel/ 4266fa LTC4266 Electrical Characteristics The l denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25C. AGND - VEE = 54V, AGND = DGND, and VDD - DGND = 3.3V unless otherwise noted. (Notes 3, 4) SYMBOL PARAMETER CONDITIONS -48V Supply Voltage AGND - VEE For IEEE Type 1 Complaint Output For IEEE Type 2 Complaint Output Undervoltage Lock-out Level VDD VDD Supply Voltage VDD - DGND Undervoltage Lock-out MIN l l 45 51 l 20 l 3.0 DGND - VEE l IEE VEE Supply Current (AGND - VEE) = 55V l IDD VDD Supply Current (VDD - DGND) = 3.3V l Detection Current - Force Current First Point, AGND - VOUTn = 9V Second Point, AGND - VOUTn = 3.5V l l Detection Voltage - Force Voltage AGND - VOUTn, 5A IOUTn 500A First Point Second Point l l Detection Current Compliance AGND - VOUTn = 0V Detection Voltage Compliance Detection Voltage Slew Rate MAX UNITS 57 57 V V 25 30 V 3.3 4.3 2.2 l Allowable Digital Ground Offset TYP 25 V V 57 V -2.4 -5 mA 1.1 3 mA 220 140 240 160 260 180 A A 7 3 8 4 9 5 V V l 0.8 0.9 mA AGND - VOUTn, Open Port l 10.4 12 V AGND - VOUTn, CPORT = 0.15F l Detection VOC 0.01 V/s Min. Valid Signature Resistance l 15.5 17 18.5 k Max. Valid Signature Resistance l 27.5 29.7 32 k 16.0 Classification VCLASS VMARK Classification Voltage AGND - VOUTn, 0mA ICLASS 50mA l 20.5 V Classification Current Compliance VOUTn = AGND l 53 61 67 mA Classification Threshold Current Class 0 - 1 Class 1 - 2 Class 2 - 3 Class 3 - 4 Class 4 - Overcurrent l l l l l 5.5 13.5 21.5 31.5 45.2 6.5 14.5 23 33 48 7.5 15.5 24.5 34.9 50.8 mA mA mA mA mA Classification Mark State Voltage AGND - VOUTn, 0.1mA ICLASS 10mA l 7.5 9 10 V Mark State Current Compliance VOUTn = AGND l 53 61 67 mA GATE Pin Pull-Down Current Port Off, VGATEn = VEE + 5V Port Off, VGATEn = VEE + 1V l l 0.4 0.08 0.12 GATE Pin Fast Pull-Down Current VGATEn = VEE + 5V GATE Pin On Voltage VGATEn - VEE, IGATEn = 1A l 8 Power Good Threshold Voltage VOUTn - VEE l 2 OUT Pin Pull-Up Resistance to AGND 0V (AGND - VOUTn) 5V l 300 Gate Driver mA mA 30 mA 14 V 2.4 2.8 V 500 700 k Output Voltage Sense VPG 4266fa LTC4266 ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25C. AGND - VEE = 54V, AGND = DGND, and VDD - DGND = 3.3V unless otherwise noted. (Notes 3, 4) SYMBOL PARAMETER CONDITIONS MIN TYP MAX UNITS Overcurrent Sense Voltage VSENSEn - VEE, icut12 = icut34 = hpen = 00h hpen = 0Fh, cutn[5:0] 4 (Note 12) cutrng = 0 cutrng = 1 l 180 188 196 mV l l 9 4.5 9.38 4.69 9.75 4.88 mV/LSB mV/LSB Overcurrent Sense in Auto Mode Class 0, Class 3 Class 1 Class 2 Class 4 l l l l 90 26 49 152 94 28 52 159 98 30 55 166 mV mV mV mV Active Current Limit in 802.3af Compliant Mode VSENSEn - VEE, icut12 = icut34 = hpen = 00h VEE = 55V (Note 12) VEE < VOUT < AGND - 29V AGND - VOUT = 0V l l 204 40 212 220 100 mV mV hpen = 0Fh, limn = C0h, VEE = 55V VOUT - VEE = 0V to 10V VEE + 23V < VOUT < AGND - 29V AGND - VOUT = 0V l l l 204 100 20 212 106 221 113 50 mV mV mV VOUT - VEE = 0V to 10V, VEE = 55V Class 0 to Class 3 Class 4 l l 102 204 106 212 110 221 mV mV Current Sense VCUT VLIM VLIM VLIM Active Current Limit in High Power Mode Active Current Limit in Auto Mode VMIN DC Disconnect Sense Voltage VSENSEn - VEE, rdis = 0 VSENSEn - VEE, rdis = 1 l l 2.6 1.3 3.8 1.9 4.8 2.41 mV mV VSC Short-Circuit Sense VSENSEn - VEE - VLIM, rdis = 0 VSENSEn - VEE - VLIM, rdis = 1 l l 160 75 200 100 255 135 mV mV Port Current ReadBack Resolution No missing codes, fast_iv = 0 14 bits LSB Weight VSENSEn - VEE 50-60Hz Noise Rejection (Note 7) 30 dB Resolution No missing codes, fast_iv = 0 14 bits LSB Weight AGND - VOUTn 50-60Hz noise rejection (Note 7) 30.5 V/LSB Port Voltage ReadBack 5.835 mV/LSB 30 dB Digital Interface VILD Digital Input Low Voltage (Note 6) l 0.8 VIHD Digital Input High Voltage (Note 6) l Digital Output Low Voltage ISDAOUT = 3mA, IINT = 3mA ISDAOUT = 5mA, IINT = 5mA l l Internal Pull-Up to VDD ADn, SHDNn, RESET, MSD 50 k Internal Pull-Down to DGND AUTO, MID 50 k 2.2 V V 0.4 0.7 V V 4266fa LTC4266 ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25C. AGND - VEE = 54V, AGND = DGND, and VDD - DGND = 3.3V unless otherwise noted. (Notes 3, 4) SYMBOL PARAMETER CONDITIONS MIN TYP MAX UNITS 290 310 ms 470 ms Timing Characteristics tDET Detection Time Beginning to End of Detection (Note 7) l 270 tDETDLY Detection Delay From PD Connected to Port to Detection Complete (Note 7) l 300 tCLE1 First Class Event Duration (Note 7) l 11 12 13 ms tME1 First Mark Event Duration (Notes 7, 11) l 6.8 8.6 10.3 ms tCLE2 Second Class Event Duration (Note 7) l 11 12 13 ms tME2 Second Mark Event Duration (Note 7) l 19 22 tCLE3 Third Class Event Duration CPORT = 0.6F (Note 7) l 0.1 ms tPON Power On Delay in Auto Mode From End of Valid Detect to Application of Power to Port (Note 7) l 60 ms Turn On Rise Time (AGND - VOUT): 10% to 90% of (AGND - VEE), CPORT = 0.15F (Note 7) l Turn On Ramp Rate CPORT = 0.15F (Note 7) l Fault Delay From ICUT Fault to Next Detect l 1.0 Midspan Mode Detection Backoff Rport = 15.5k (Note 7) l 2.3 2.5 2.7 s Power Removal Detection Delay From Power Removal After tDIS to Next Detect (Note 7) l 1.0 1.3 2.5 s 15 ms 24 s 10 1.1 V/s s tSTART Maximum Current Limit Duration During Port tSTART1 = 0, tSTART0 = 0 (Notes 7, 12) Start-Up l 52 62.5 66 ms tLIM, tICUT Maximum Current Limit Duration After Port Start-Up tICUT1 = 0, tICUT0 = 0 (Notes 7, 12) l 52 62.5 66 ms Maximum Current Limit Duty Cycle (Note 7) l 5.8 6.3 6.7 % 3.6 ms 380 ms 6.5 s 6.5 s 3 s tMPS Maintain Power Signature (MPS) Pulse Width Current Pulse Width to Reset Disconnect Sensitivity Timer (Notes 7, 8) l 1.6 tDIS Maintain Power Signature (MPS) Dropout Time tconf [1:0] = 00b (Notes 5, 12) l 320 tMSD Masked Shut Down Delay (Note 7) l tSHDN Port Shut Down Delay (Note 7) l I2C Watchdog Timer Duration l 1.5 350 2 Minimum Pulse Width for Masked Shut Down (Note 7) l 3 s Minimum Pulse Width for SHDN (Note 7) l 3 s Minimum Pulse Width for RESET (Note 7) l 4.5 s 4266fa LTC4266 ELECTRICAL CHARACTERISTICS The l denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25C. AGND - VEE = 54V, AGND = DGND, and VDD - DGND = 3.3V unless otherwise noted. (Notes 3, 4) SYMBOL PARAMETER CONDITIONS MIN TYP MAX UNITS Clock Frequency (Note 7) l t1 Bus Free Time Figure 5 (Notes 7, 9) l 480 ns t2 Start Hold Time Figure 5 (Notes 7, 9) l 240 ns t3 SCL Low Time Figure 5 (Notes 7, 9) l 480 ns t4 SCL High Time Figure 5 (Notes 7, 9) l 240 ns t5 Data Hold Time Figure 5 (Notes 7, 9) Data into chip Data out of chip l l 60 t6 Data Set-Up Time Figure 5 (Notes 7, 9) l 80 ns t7 Start Set-Up Time Figure 5 (Notes 7, 9) l 240 ns t8 Stop Set-Up Time Figure 5 (Notes 7, 9) l 240 ns tr SCL, SDAIN Rise Time Figure 5 (Notes 7, 9) l 120 ns tf SCL, SDAIN Fall Time Figure 5 (Notes 7, 9) l 60 ns Fault Present to INT Pin Low (Notes 7, 9, 10) l 150 ns Stop Condition to INT Pin Low (Notes 7, 9, 10) l 1.5 s ARA to INT Pin High Time (Notes 7, 9) l 1.5 s SCL Fall to ACK Low (Notes 7, 9) l 120 ns I2C Timing Note 1: Stresses beyond those listed under Absolute Maximum Ratings may cause permanent damage to the device. Exposure to any Absolute Maximum Rating condition for extended periods may affect device reliability and lifetime. Note 2: This IC includes overtemperature protection that is intended to protect the device during momentary overload conditions. Junction temperature will exceed 140C when overtemperature protection is active. Continuous operation above the specified maximum operating junction temperature may impair device reliability. Note 3: All currents into device pins are positive; all currents out of device pins are negative. Note 4: The LTC4266 operates with a negative supply voltage (with respect to ground). To avoid confusion, voltages in this data sheet are referred to in terms of absolute magnitude. Note 5: tDIS is the same as tMPDO defined by IEEE 802.3at. 1 120 MHz ns ns Note 6: The LTC4266 digital interface operates with respect to DGND. All logic levels are measured with respect to DGND. Note 7: Guaranteed by design, not subject to test. Note 8: The IEEE 802.3af specification allows a PD to present its Maintain Power Signature (MPS) on an intermittent basis without being disconnected. In order to stay powered, the PD must present the MPS for tMPS within any tMPDO time window. Note 9: Values measured at VILD(MAX) and VIHD(MIN). Note 10: If fault condition occurs during an I2C transaction, the INT pin will not be pulled down until a stop condition is present on the I2C bus. Note 11: Load Characteristic of the LTC4266 during Mark: 7V < (AGND - VOUTn) < 10V or IOUT < 50A Note 12: See the LTC4266 Software Programming documentation for information on serial bus usage and device configuration and status registers. 4266fa LTC4266 Typical Performance Characteristics Power On Sequence in Auto Mode 10 0 GND PORT VOLTAGE (V) FORCED VOLTAGE DETECTION -40 PORT 1 VDD = 3.3V VEE = -54V -50 -60 VDD = 3.3V VEE = -54V PORT VOLTAGE 20V/DIV -10 -30 GND GND FORCED CURRENT DETECTION -20 802.3af Classification in Auto Mode Powering Up into a 180F Load LOAD FULLY CHARGED VEE PORT CURRENT 200 mA/DIV 802.3af CLASSIFICATION POWER ON FOLDBACK 0mA GATE VOLTAGE 10V/DIV VEE VEE -70 -18.4 FET ON VEE 5ms/DIV 100ms/DIV 5ms/DIV 4266 G02 4266 G01 4266 G03 Classification Transient Response to 40mA Load Step GND PORT CURRENT 20mA/DIV -17.6 40mA Classification Current Compliance 0 VDD = 3.3V VEE = -54V 0mA 1ST CLASS EVENT 2ND CLASS EVENT PORT 1 VDD = 3.3V VEE = -55V PD IS CLASS 4 PORT VOLTAGE 1V/DIV VEE -20V -8 -10 -12 -14 -16 2.4 1.5 1.3 1.2 1.1 1.0 0.9 2.7 2.9 3.1 3.3 3.5 3.7 3.9 VDD SUPPLY VOLTAGE (V) 4.1 4.3 4266 G07 20 30 40 50 60 CLASSIFICATION CURRENT 70 802.3at ILIM Threshold vs Temperature 215 214 2.3 2.2 2.1 -40C 25C 85C 2.0 -60 -55 -50 -45 -40 -35 -30 -25 -20 VEE SUPPLY VOLTAGE (V) 4266 G08 860 VDD = 3.3V VEE = -54V RSENSE = 0.25 REG 48h = C0h 856 213 852 212 848 211 844 210 -40 0 40 -80 TEMPERATURE (C) ILIM (mA) 1.4 10 4266 G06 VEE Supply Current vs Voltage VLIM (mV) 1.6 IEE SUPPLY CURRENT (mA) -40C 25C 85C 0 4266 G05 VDD Supply Current vs Voltage 1.7 IDD SUPPLY CURRENT (mA) -6 -20 50s/DIV 4266 G04 0.8 -4 -18 10ms/DIV 1.8 VDD = 3.3V VEE = -54V TA = 25C -2 CLASSIFICATION VOLTAGE (V) 2-Event Classification in Auto Mode PORT VOLTAGE 10V/DIV PORT 1 VDD = 3.3V VEE = -55V PD IS CLASS 1 PORT VOLTAGE 10V/DIV 425mA CURRENT LIMIT 840 120 4266 G09 4266fa LTC4266 TYPICAL PERFORMANCE CHARACTERISTICS 802.3af ILIM Threshold vs Temperature 163 162 429 426 423 105.75 105.00 -40 0 420 120 40 80 TEMPERATURE (C) 652 VDD = 3.3V VEE = -54V RSENSE = 0.25 REG 47h = E2h PORT 1 648 161 644 160 640 159 636 158 -40 0 40 80 TEMPERATURE (C) 4266 G10 DC Disconnect Threshold vs Temperature 384 2.0000 381 1.9375 378 93.75 375 93.00 -40 0 372 120 80 40 TEMPERATURE (C) 8.00 VDD = 3.3V VEE = -54V RSENSE = 0.25 REG 47h = E2h PORT 1 7.75 1.8750 7.50 1.8125 7.25 1.7500 -40 0 80 40 TEMPERATURE (C) 4266 G12 ADC Noise Histogram Current Readback in Fast Mode 400 200 350 175 300 150 500 125 400 100 300 75 200 50 100 25 50 0 0 0 -54 -45 -36 -18 -27 VOUTn (V) -9 0 4266 G14 BIN COUNT 600 VLIM (mV) ILIM (mA) 700 225 ADC Integral Nonlinearity Current Readback in Fast Mode 1.0 VSENSEn - VEE = 110.4mV ADC INTEGRAL NONLINEARITY (LSBs) VDD = 3.3V VEE = -54V RSENSE = 0.25 REG 48h = C0h 800 7.00 120 4266 G13 Current Limit Foldback 900 IMIN (mV) 94.50 VMIN (mV) VDD = 3.3V VEE = -54V RSENSE = 0.25 REG 47h = D4h PORT 1 ICUT (mA) VCUT (mV) 95.25 630 120 4266 G11 802.3af ICUT Threshold vs Temperature 96.00 ICUT (mA) 106.50 ILIM (mA) VLIM (mV) 107.25 432 VDD = 3.3V VEE = -54V RSENSE = 0.25 REG 48h = 80h PORT 1 VCUT (mV) 108.00 802.3at ICUT Threshold vs Temperature 250 200 150 100 191 192 193 194 ADC OUTPUT 195 196 4266 G15 0.5 0 -0.5 -1.0 0 50 100 150 200 250 300 350 400 450 500 CURRENT SENSE RESISTOR INPUT VOLTAGE (mV) 4266 G16 4266fa LTC4266 TYPICAL PERFORMANCE CHARACTERISTICS ADC Noise Histogram Current Readback in Slow Mode ADC INTEGRAL NONLINEARITY (LSBs) 250 BIN COUNT 200 150 100 50 6139 600 1.0 VSENSEn - VEE = 110.4mV 0 6141 6143 ADC OUTPUT 6145 6147 500 400 0 200 100 -1.0 AGND - VOUTn = 48.3V BIN COUNT 400 300 200 -0.5 100 20 40 30 PORT VOLTAGE (V) 50 60 0 8533 8532 8534 8535 ADC OUTPUT 8536 265 1.0 0.5 0 -0.5 -1.0 0 10 20 40 30 PORT VOLTAGE (V) 3 GND 2 60 4266 G22 VDD = 3.3V VEE = -54V PORT VOLTAGE 20V/DIV 2.5 50 MOSFET Gate Drive With Fast Pull Down INT and SDAOUT Pull Down Voltage vs Load Current PULL DOWN VOLTAGE (V) 264 4266 G21 4266 G20 VEE FAST PULL DOWN GATE VOLTAGE 10V/DIV VEE 1.5 1 PORT CURRENT 500mA/DIV 0mA 0.5 0 262 263 ADC OUTPUT ADC Integral Nonlinearity Voltage Readback in Slow Mode ADC INTEGRAL NONLINEARITY (LSBs) 600 500 10 261 260 4266 G19 ADC Noise Histogram Port Voltage Readback in Slow Mode 0.5 0 0 0 50 100 150 200 250 300 350 400 450 500 CURRENT SENSE RESISTOR INPUT VOLTAGE (mV) 4266 G18 1.0 -1.0 300 -0.5 ADC Integral Nonlinearity Voltage Readback in Fast Mode 0 AGND - VOUTn = 48.3V 0.5 4266 G17 ADC INTEGRAL NONLINEARITY (LSBs) ADC Noise Histogram Port Voltage Readback in Fast Mode BIN COUNT 300 ADC Integral Nonlinearity Current Readback in Slow Mode 0 5 10 15 20 25 30 LOAD CURRENT (mA) 35 40 4266 G23 50 FAULT APPLIED CURRENT LIMIT 50 FAULT REMOVED 100s/DIV 4266 G24 4266fa LTC4266 Test Timing Diagrams tDET CLASSIFICATION FORCED-CURRENT VPORTn FORCEDVOLTAGE 0V tME1 tME2 VOC VMARK 15.5V VCLASS 20.5V tCLE1 tCLE2 PD CONNECTED tCLE3 tPON tDETDLY VEE INT 4266 F01 Figure 1. Detect, Class and Turn-On Timing in Auto or Semiauto Modes VLIM VCUT VSENSEn TO VEE 0V VSENSEn TO VEE VMIN tSTART, tICUT INT INT tMPS tDIS 4266 F03 4266 F02 Figure 3. DC Disconnect Timing Figure 2. Current Limit Timing t3 VGATEn tr t4 tMSD tSHDN VEE SCL t2 MSD or SHDNn tf t5 t6 t7 t8 SDA 4266 F04 Figure 4. Shut Down Delay Timing t1 4266 F05 Figure 5. I2C Interface Timing 4266fa 10 LTC4266 I2C Timing Diagrams SCL SDA 0 1 0 AD3 AD2 AD1 AD0 R/W ACK A7 START BY MASTER A6 A5 A4 A3 A2 ACK BY SLAVE FRAME 1 SERIAL BUS ADDRESS BYTE A1 A0 ACK D7 D6 D5 D4 D3 D2 ACK BY SLAVE D1 D0 ACK STOP BY MASTER ACK BY SLAVE FRAME 2 REGISTER ADDRESS BYTE FRAME 3 DATA BYTE 4266 F06 Figure 6. Writing to a Register SCL SDA 0 1 0 AD3 AD2 AD1 AD0 R/W ACK A7 START BY MASTER FRAME 1 SERIAL BUS ADDRESS BYTE A6 ACK BY SLAVE A5 A4 A3 A2 A1 A0 ACK ACK BY SLAVE FRAME 2 REGISTER ADDRESS BYTE 0 1 0 AD3 AD2 AD1 AD0 R/W ACK D7 REPEATED START BY MASTER FRAME 1 SERIAL BUS ADDRESS BYTE D6 D5 D4 D3 ACK BY SLAVE D2 D1 D0 ACK NO ACK BY MASTER FRAME 2 DATA BYTE STOP BY MASTER 4266 F07 Figure 7. Reading from a Register 4266fa 11 LTC4266 I2C TiMING DIAGRAMS SCL SDA 0 1 0 AD3 AD2 AD1 AD0 R/W ACK D7 START BY MASTER D6 D5 D4 D3 ACK BY SLAVE D2 D1 D0 ACK STOP BY MASTER NO ACK BY MASTER FRAME 1 SERIAL BUS ADDRESS BYTE FRAME 2 DATA BYTE 4266 F08 Figure 8. Reading the Interrupt Register (Short Form) SCL SDA 0 0 0 1 1 0 0 R/W ACK START BY MASTER FRAME 1 ALERT RESPONSE ADDRESS BYTE 0 1 ACK BY SLAVE 0 AD3 AD2 AD1 AD0 1 NO ACK BY MASTER FRAME 2 SERIAL BUS ADDRESS BYTE ACK STOP BY MASTER 4266 F09 Figure 9. Reading from Alert Response Address 4266fa 12 LTC4266 Pin Functions RESET: Chip Reset, Active Low. When the RESET pin is low, the LTC4266 is held inactive with all ports off and all internal registers reset to their power-up states. When RESET is pulled high, the LTC4266 begins normal operation. RESET can be connected to an external capacitor or RC network to provide a power turn-on delay. Internal filtering of the RESET pin prevents glitches less than 1s wide from resetting the LTC4266. Internally pulled up to VDD. MID: Midspan Mode Input. When high, the LTC4266 acts as a midspan device. Internally pulled down to DGND. INT: Interrupt Output, Open Drain. INT will pull low when any one of several events occur in the LTC4266. It will return to a high impedance state when bits 6 or 7 are set in the Reset PB register (1Ah). The INT signal can be used to generate an interrupt to the host processor, eliminating the need for continuous software polling. Individual INT events can be disabled using the Int Mask register (01h). See Register Functions and Applications Information for more information. The INT pin is only updated between I2C transactions. SCL: Serial Clock Input. High impedance clock input for the I2C serial interface bus. SCL must be tied high if not used. SDAOUT: Serial Data Output, Open Drain Data Output for the I2C Serial Interface Bus. The LTC4266 uses two pins to implement the bidirectional SDA function to simplify optoisolation of the I2C bus. To implement a standard bidirectional SDA pin, tie SDAOUT and SDAIN together. SDAOUT should be grounded or left floating if not used. See Applications Information for more information. SDAIN: Serial Data Input. High impedance data input for the I2C serial interface bus. The LTC4266 uses two pins to implement the bidirectional SDA function to simplify optoisolation of the I2C bus. To implement a standard bidirectional SDA pin, tie SDAOUT and SDAIN together. SDAIN must be tied high if not used. See Applications Information for more information. AD3: Address Bit 3. Tie the address pins high or low to set the I2C serial address to which the LTC4266 responds. This address will be 010A3A2A1A0b. Internally pulled up to VDD. AD2: Address Bit 2. See AD3. AD1: Address Bit 1. See AD3. AD0: Address Bit 0. See AD3. NC, DNC: All pins identified with "NC" or "DNC" must be left unconnected. DGND: Digital Ground. DGND is the return for the VDD supply. VDD: Logic Power Supply. Connect to a 3.3V power supply relative to DGND. VDD must be bypassed to DGND near the LTC4266 with at least a 0.1F capacitor. SHDN1: Shutdown Port 1, Active Low. When pulled low, SHDN1 shuts down port 1, regardless of the state of the internal registers. Pulling SHDN1 low is equivalent to setting the Reset Port 1 bit in the Reset Pushbutton register (1Ah). Internal filtering of the SHDN1 pin prevents glitches less than 1s wide from reseting the port. Internally pulled up to VDD. SHDN2: Shutdown Port 2, Active Low. See SHDN1. SHDN3: Shutdown Port 3, Active Low. See SHDN1. SHDN4: Shutdown Port 4, Active Low. See SHDN1. AGND: Analog Ground. AGND is the return for the VEE supply. SENSE4: Port 4 Current Sense Input. SENSE4 monitors the external MOSFET current via a 0.5 or 0.25 sense resistor between SENSE4 and VEE. Whenever the voltage across the sense resistor exceeds the overcurrent detection threshold VCUT, the current limit fault timer counts up. If the voltage across the sense resistor reaches the current limit threshold VLIM, the GATE4 pin voltage is lowered to maintain constant current in the external MOSFET. See Applications Information for further details. If the port is unused, the SENSE4 pin must be tied to VEE. 4266fa 13 LTC4266 PIN FUNCTIONS GATE4: Port 4 Gate Drive. GATE4 should be connected to the gate of the external MOSFET for port 4. When the MOSFET is turned on, the gate voltage is driven to 13V (typ) above VEE. During a current limit condition, the voltage at GATE4 will be reduced to maintain constant current through the external MOSFET. If the fault timer expires, GATE4 is pulled down, turning the MOSFET off and recording a tCUT or tSTART event. If the port is unused, float the GATE4 pin. OUT4: Port 4 Output Voltage Monitor. OUT4 should be connected to the output port. A current limit foldback circuit limits the power dissipation in the external MOSFET by reducing the current limit threshold when the drain-tosource voltage exceeds 10V. The port 4 Power Good bit is set when the voltage from OUT4 to VEE drops below 2.4V (typ). A 500k resistor is connected internally from OUT4 to AGND when the port is idle. If the port is unused, OUT4 pin must be floated. SENSE3: Port 3 Current Sense Input. See SENSE4. GATE3: Port 3 Gate Drive. See GATE4. OUT3: Port 3 Output Voltage Monitor. See OUT4. VEE: Main Supply Input. Connect to a -45V to -57V supply, relative to AGND. SENSE2: Port 2 Current Sense Input. See SENSE4. GATE2: Port 2 Gate Drive. See GATE4. OUT2: Port 2 Output Voltage Monitor. See OUT4. SENSE1: Port 1 Current Sense Input. See SENSE4. GATE1: Port 1 Gate Drive. See GATE 4. OUT1: Port 1 Output Voltage Monitor. See OUT4. AUTO: Auto Mode Input. Auto mode allows the LTC4266 to detect and power up a PD even if there is no host controller present on the I2C bus. The voltage of the AUTO pin determines the state of the internal registers when the LTC4266 is reset or comes out of VDD UVLO (see the Register map). The states of these register bits can subsequently be changed via the I2C interface. The real-time state of the AUTO pin is read at bit 0 in the Pin Status register (11h). Internally pulled down to DGND. Must be tied locally to either VDD or DGND. MSD: Maskable Shutdown Input. Active low. When pulled low, all ports that have their corresponding mask bit set in the mconfig register (17h) will be reset, equivalent to pulling the SHDN pin low. Internal filtering of the MSD pin prevents glitches less than 1s wide from resetting ports. Internally pulled up to VDD. 4266fa 14 LTC4266 Operation Overview Power over Ethernet, or PoE, is a standard protocol for sending DC power over copper Ethernet data wiring. The IEEE group that administers the 802.3 Ethernet data standards added PoE powering capability in 2003. This original PoE spec, known as 802.3af, allowed for 48V DC power at up to 13W. This initial spec was widely popular, but 13W was not adequate for some requirements. In 2009, the IEEE released a new standard, known as 802.3at or PoE+, increasing the voltage and current requirements to provide 25W of power. channel; these minimize power loss compared to alternative designs with on-board MOSFETs and increase system reliability in the event a single channel is damaged. PoE Basics Common Ethernet data connections consist of two or four twisted pairs of copper wire (commonly known as CAT-5 cable), transformer-coupled at each end to avoid ground loops. PoE systems take advantage of this coupling arrangement by applying voltage between the center-taps of the data transformers to transmit power from the PSE to the PD without affecting data transmission. Figure 10 shows a high-level PoE system schematic. The IEEE standard also defines PoE terminology. A device that provides power to the network is known as a PSE, or power sourcing equipment, while a device that draws power from the network is known as a PD, or powered device. PSEs come in two types: Endpoints (typically network switches or routers), which provide data and power; and Midspans, which provide power but pass through data. Midspans are typically used to add PoE capability to existing non-PoE networks. PDs are typically IP phones, wireless access points, security cameras, and similar devices, but could be nearly anything that runs from 25W or less and includes an RJ45-style network connector. To avoid damaging legacy data equipment that does not expect to see DC voltage, the PoE spec defines a protocol that determines when the PSE may apply and remove power. Valid PDs are required to have a specific 25k common-mode resistance at their input. When such a PD is connected to the cable, the PSE detects this signature resistance and turns on the power. When the PD is later disconnected, the PSE senses the open circuit and turns power off. The PSE also turns off power in the event of a current fault or short circuit. The LTC4266 is a third-generation quad PSE controller that implements four PSE ports in either an endpoint or midspan design. Virtually all necessary circuitry is included to implement a IEEE 802.3at compliant PSE design, requiring only an external power MOSFET and sense resistor per When a PD is detected, the PSE optionally looks for a classification signature that tells the PSE the maximum power the PD will draw. The PSE can use this information to allocate power among several ports, police the current consumption of the PD, or to reject a PD that will draw PSE RJ45 4 5 GND DGND 3.3V INTERRUPT SMAJ58A I2C 1F 100V X7R -48V CAT 5 20 MAX ROUNDTRIP 0.05F MAX AGND VDD INT 1/4 SCL LTC4266 SDAIN SDAOUT VEE SENSE GATE OUT PD RJ45 4 5 1N4002 s4 SPARE PAIR 0.22F 100V X7R 1 1 Tx DATA PAIR 3 2 3 Rx 0.1F Tx 6 DATA PAIR 6 1N4002 s4 IRFM120A S1B GND RCLASS S1B 0.25 5F CIN 300F SMAJ58A 58V Rx 2 PWRGD LTC4265 7 7 8 8 -48VIN -48VOUT DC/DC CONVERTER + VOUT - SPARE PAIR 4266 F10 Figure 10. Power Over Ethernet System Diagram 4266fa 15 LTC4266 Operation more power that the PSE has available. The classification step is optional; if a PSE chooses not to classify a PD, it must assume that the PD is a 13W (full 802.3af power) device. compatible) devices can be substituted with the LTC4266 without software or PCB layout changes; only minor BOM changes are required to implement a fully compliant 802.3at design. New in 802.3at Because of the backwards compatibility features, some of the internal registers are redundant or unused when the LTC4266 is operated as recommended. For more details on usage in compatibility mode, refer to the LTC4258/ LTC4259A device datasheets. The newer 802.3at standard supersedes 802.3af and brings several new features: * A PD may draw as much as 25.5W. Such PDs (and the PSEs that support them) are known as Type 2. Older 13W 802.3af equipment is classified as Type 1. Type 1 PDs will work with all PSEs; Type 2 PDs may require Type 2 PSEs to work properly. The LTC4266 is designed to work in both Type 1 and Type 2 PSE designs, and also supports non-standard configurations at higher power levels. * The Classification protocol is expanded to allow Type 2 PSEs to detect Type 2 PDs, and to allow Type 2 PDs to determine if they are connected to a Type 2 PSE. Two versions of the new Classification protocol are available: an expanded version of the 802.3af Class Pulse protocol, and an alternate method integrated with the existing LLDP protocol (using the Ethernet data path). The LTC4266 fully supports the new Class Pulse protocol and is also compatible with the LLDP protocol (which is implemented in the data communications layer, not in the PoE circuitry). * Fault protection current levels and timing are adjusted to reduce peak power in the MOSFET during a fault; this allows the new 25.5W power levels to be reached using the same MOSFETs as older 13W designs. Backwards Compatibility The LTC4266 is designed to be backward compatible with earlier PSE chips in both software and pin functions. Existing systems using either the LTC4258 or LTC4259A (or Special Compatibility Mode Notes * The LTC4266 can use either 0.5 or 0.25 sense resistors, while the LTC425x chips always used 0.5. To maintain compatibility, if the AUTO pin is low when the LTC4266 powers up it assumes the sense resistor is 0.5; if it is high at power up, the LTC4266 assumes 0.25. The resistor value setting can be reconfigured at any time after power up. In particular, systems that use 0.25 sense resistors and have AUTO tied low must reconfigure the resistor settings after power up. * The LTC4259A included both AC and DC disconnect sensing circuitry, but the LTC4266 has only DC disconnect sensing. For the sake of compatibility, register bits used to enable AC disconnect in the LTC4259A are implemented in the LTC4266, but they simply mirror the bits used for DC disconnect. * The LTC4258 and LTC4259A required 10k resistors between the OUTn pins and the drains of the external MOSFETs. These resistors must be shorted or replaced with zero ohm jumpers when using the LTC4266. * The LTC4258 and LTC4259A included a BYP pin, decoupled to AGND with 0.1F. This pin changes to the MID pin on the LTC4266. The capacitor should be removed for Endspan applications, or replaced with a zero ohm jumper for Midspan applications. 4266fa 16 LTC4266 Applications Information Operating Modes The LTC4266 includes four independent ports, each of which can operate in one of four modes: Manual, Semiauto, Auto, or Shutdown. * In manual mode, the port waits for instructions from the host system before taking any action. It runs a single detection or classification cycle when commanded to by the host, and reports the result in its Port Status register. The host system can command the port to turn on or off the power at any time. * In semi-auto mode, the port repeatedly attempts to detect and classify any PD attached to it. It reports the status of these attempts back to the host, and waits for a command from the host before turning on power to the port. The host must enable detection (and optionally classification) for the port before detection will start. * Auto mode operates the same as Semi-auto mode except that it will automatically turn on the power to the port if detection is successful. * In shutdown mode, the port is disabled and will not detect or power a PD. Regardless of which mode it is in, the LTC4266 will remove power automatically from any port that generates a current limit fault. It will also automatically remove power from any port that generates a disconnect event if disconnect detection is enabled. The host controller may also command the port to remove power at any time. Power-On Reset and the AUTO/MID pins The initial LTC4266 configuration depends on the state of the AUTO and MID pins during reset. Reset occurs at power-up, or whenever the RESET pin is pulled low or the global Reset All bit is set. Note that the AUTO pin is only sampled when a reset occurs. Changing the state of AUTO or MID after power-up will not change the port behavior of the LTC4266 until a reset occurs. Although typically used with a host controller, the LTC4266 can also be used in a standalone mode with no connection to the serial interface. If there is no host present, the AUTO pin should be tied high so that, at reset, all ports will be configured to operate automatically. Each port will detect and classify repeatedly until a PD is discovered, set ICUT and ILIM according to the classification results, apply power after successful detection, and remove power when a PD is disconnected. Similarly, if the standalone application is a midspan, the MID pin should be tied high to enable correct midspan detection timing. Table 1 shows the ICUT and ILIM values that will be automatically set in standalone mode, based on the discovered class. Table 1. ICUT and ILIM Values in Standalone Mode CLASS ICUT ILIM Class 1 112mA 425mA Class 2 206mA 425mA Class 3 or Class 0 375mA 425mA Class 4 638mA 850mA The automatic setting of the ICUT and ILIM values only occurs if the LTC4266 is reset with the AUTO pin high. Detection Detection Overview To avoid damaging network devices that were not designed to tolerate DC voltage, a PSE must determine whether the connected device is a real PD before applying power. The IEEE specification requires that a valid PD have a commonmode resistance of 25k 5% at any port voltage below 10V. The PSE must accept resistances that fall between 19k and 26.5k, and it must reject resistances above 33k or below 15k (shaded regions in Figure 11). The PSE may choose to accept or reject resistances in the undefined areas between the must-accept and must-reject ranges. In particular, the PSE must reject standard computer network ports, many of which have 150 common-mode termination resistors that will be damaged if power is applied to them (the black region at the left of Figure 11). RESISTANCE 0 PD PSE 10k 20k 150 (NIC) 15k 30k 23.75k 26.25k 19k 26.5k 33k 4266 F11 Figure 11. IEEE 802.3af Signature Resistance Ranges 4266fa 17 LTC4266 APPLICATIONS INFORMATION 4-Point Detection Table 2. Detection Status MEASURED PD SIGNATURE DETECTION RESULT Incomplete or Not Yet Tested Detect Status Unknown <2.4k Short Circuit Capacitance > 2.7F CPD too High 2.4k < RPD < 17k RSIG too Low 17k < RPD < 29k Detect Good >29k RSIG too High >50k Open Circuit Voltage > 10V Port Voltage Outside Detect Range Operating Modes The port's operating mode determines when the LTC4266 runs a detection cycle. In manual mode, the port will idle until the host orders a detect cycle. It will then run detection, report the results, and return to idle to wait for another command. In semi-auto mode, the LTC4266 autonomously polls a port for PDs, but it will not apply power until commanded to do so by the host. The port status register is updated at the end of each detection cycle. If a valid signature 275 CURRENT (A) The LTC4266 uses a 4-point detection method to discover PDs. False-positive detections are minimized by checking for signature resistance with both forced-current and forced-voltage measurements. Initially, two test currents are forced onto the port (via the OUTn pin) and the resulting voltages are measured. The detection circuitry subtracts the two V-I points to determine the resistive slope while removing offset caused by series diodes or leakage at the port (see Figure 12). If the forced-current detection yields a valid signature resistance, two test voltages are then forced onto the port and the resulting currents are measured and subtracted. Both methods must report valid resistances for the port to report a valid detection. PD signature resistances between 17k and 29k (typically) are detected as valid and reported as Detect Good in the corresponding Port Status register. Values outside this range, including open and short circuits, are also reported. If the port measures less than 1V at the first forced-current test, the detection cycle will abort and Short Circuit will be reported. Table 2 shows the possible detection results. 25k SLOPE 165 VALID PD 0V-2V OFFSET FIRST DETECTION POINT SECOND DETECTION POINT VOLTAGE 4266 F12 Figure 12. PD Detection resistance is detected and classification is enabled, the port will classify the PD and report that result as well. The port will then wait for at least 100ms (or 2 seconds if midspan mode is enabled), and will repeat the detection cycle to ensure that the data in the port status register is up-to-date. If the port is in semi-auto mode and high power operation is enabled, the port will not turn on in response to a power-on command unless the current detect result is detect good. Any other detect result will generate a tSTART fault if a power-on command is received. If the port is not in high power mode, it will ignore the detection result and apply power when commanded, maintaining backwards compatibility with the LTC4259A. Behavior in auto mode is similar to semi-auto; however, after detect good is reported and the port is classified (if classification is enabled), it is automatically powered on without further intervention. In standalone mode, the ICUT and ILIM thresholds are automatically set in auto mode; see the power-on Reset and the AUTO Pin section for more information. The signature detection circuitry is disabled when the port is initially powered up with the AUTO pin low, in shutdown mode, or when the corresponding detect enable bit is cleared. Detection of Legacy PDs Proprietary PDs that predate the original IEEE 802.3af standard are commonly referred to today as legacy devices. One type of legacy PD uses a large common mode 4266fa 18 LTC4266 APPLICATIONS INFORMATION The LTC4266 can be configured to detect this type of legacy PD. Legacy detection is disabled by default, but can be manually enabled on a per-port basis. When enabled, the port will report detect good when it sees either a valid IEEE PD or a high-capacitance legacy PD. With legacy mode disabled, only valid IEEE PDs will be recognized. Classification 802.3af Classification A PD can optionally present a classification signature to the PSE to indicate the maximum power it will draw while operating. The IEEE specification defines this signature as a constant current draw when the PSE port voltage is in the VCLASS range (between 15.5V and 20.5V), with the current level indicating one of 5 possible PD classes. Figure 14 shows a typical PD load line, starting with the slope of the 25k signature resistor below 10V, then transitioning to the classification signature current (in this case, Class 3) in the VCLASS range. Table 3 shows the possible classification values. Table 3. Classification Values CLASS RESULT Class 0 No Class Signature Present; Treat Like Class 3 Class 1 3W Class 2 7W Class 3 13W Class 4 25.5W (Type 2) If classification is enabled, the port will classify the PD immediately after a successful detection cycle in semi-auto or auto modes, or when commanded to in manual mode. It measures the PD classification signature by applying 18V for 12ms (both values typical) to the port via the OUTn pin and measuring the resulting current; it then reports the discovered class in the port status register. If the LTC4266 was reset with the AUTO pin high and the port is in auto mode, it will additionally use the classification result to set 60 PSE LOAD LINE OVER CURRENT 50 48mA CURRENT (mA) capacitance (>10F) as the detection signature. Note that PDs in this range of capacitance are defined as invalid, so a PSE that detects legacy PDs is technically noncompliant with the IEEE spec. 40 CLASS 4 30 CLASS 3 23mA 20 TYPICAL CLASS 3 PD LOAD LINE 10 0 33mA 0 5 CLASS 2 CLASS 1 CLASS 0 10 15 VOLTAGE (VCLASS) 14.5mA 6.5mA 20 25 4266 F13 Figure 13. PD Classification the ICUT and ILIM thresholds. See the Power-On Reset and the AUTO/MID Pin section for more information. The classification circuitry is disabled when the port is initially powered up with the AUTO pin low, in shutdown mode, or when the corresponding class enable bit is cleared. 802.3at 2-Event Classification The 802.3at spec defines two methods of classifying a Type 2 PD. One method adds extra fields to the Ethernet LLDP data protocol; although the LTC4266 is compatible with this classification method, it cannot perform classification directly since it doesn't have access to the data path. LLDP classification requires the PSE to power the PD as a standard 802.3af (Type 1) device. It then waits for the host to perform LLDP communication with the PD and update the PSE port data. The LTC4266 supports changing the ILIM and ICUT levels on the fly, allowing the host to complete LLDP classification. The second 802.3at classification method, known as 2event classification or ping-pong, is fully supported by the LTC4266. A Type 2 PD that is requesting more than 13W will indicate Class 4 during normal 802.3af classification. If the LTC4266 sees Class 4, it forces the port to a specified lower voltage (called the mark voltage, typically 9V), pauses briefly, and then re-runs classification to verify the Class 4 reading (Figure 1). It also sets a bit in the high 4266fa 19 LTC4266 APPLICATIONS INFORMATION power status register to indicate that it ran the second classification cycle. The second cycle alerts the PD that it is connected to a Type 2 PSE which can supply Type 2 power levels. 2-event ping-pong classification is enabled by setting a bit in the port's high power mode register. Note that a pingpong enabled port only runs the second classification cycle when it detects a Class 4 device; if the first cycle returns Class 0 to 3, the port assumes it is connected to a Type 1 PD and does not run the second classification cycle. Invalid Type 2 Class Combinations The 802.3at spec defines a Type 2 PD class signature as two consecutive Class 4 results; a Class 4 followed by a Class 0-3 is not a valid signature. In auto mode, the LTC4266 will power a detected PD regardless of the classification results, with one exception: if the PD presents an invalid Type 2 signature (Class 4 followed by Class 0 to 3), the LTC4266 will not provide power and will restart the detection process. To aid in diagnosis, the port status register will always report the results of the last class pulse, so an invalid Class 4-Class 2 combination would report a second class pulse was run in the High Power Status register (which implies that the first cycle found Class 4), and Class 2 in the port status register. Power Control External MOSFET, Sense R Summary The primary function of the LTC4266 is to control the delivery of power to the PSE port. It does this by controlling the gate drive voltage of an external power MOSFET while monitoring the current via an external sense resistor and the output voltage at the OUT pin. This circuitry serves to couple the raw VEE input supply to the port in a controlled manner that satisfies the PD's power needs while minimizing power dissipation in the MOSFET and disturbances on the VEE backplane. The LTC4266 is designed to use 0.25 sense resistors to minimize power dissipation. It also supports 0.5 sense resistors, which are the default when LTC4258/LTC4259A compatibility is desired. Inrush Control Once the command has been given to turn on a port, the LTC4266 ramps up the GATE pin of that port's external MOSFET in a controlled manner. Under normal power-up circumstances, the MOSFET gate will rise until the port current reaches the inrush current limit level (typically 450mA), at which point the GATE pin will be servoed to maintain the specified IINRUSH current. During this inrush period, a timer (tSTART ) runs. When output charging is complete, the port current will fall and the GATE pin will be allowed to continue rising to fully enhance the MOSFET and minimize its on-resistance. The final VGS is nominally 13V. If the tSTART timer expires before the inrush period completes, the port will be turned back off and a tSTART fault reported. Current Limit Each LTC4266 port includes two current limiting thresholds (ICUT and ILIM), each with a corresponding timer (tCUT and tLIM). Setting the ICUT and ILIM thresholds depends on several factors: the class of the PD, the voltage of the main supply (VEE), the type of PSE (1 or 2), the sense resistor (0.5 or 0.25), the SOA of the MOSFET, and whether or not the system is required to implement class enforcement. Per the IEEE spec, the LTC4266 will allow the port current to exceed ICUT for a limited period of time before removing power from the port, whereas it will actively control the MOSFET gate drive to keep the port current below ILIM. The port does not take any action to limit the current when only the ICUT threshold is exceeded, but does start the tCUT timer. The tLIM timer starts when the ILIM threshold is exceeded and current limit is active. If the current drops below the ICUT current threshold before its timer expires, the tCUT timer counts back down, but at 1/16 the rate that it counts up. This allows the current limit circuitry to tolerate intermittent overload signals with duty cycles below about 6%; longer duty cycle overloads will turn the port off. ICUT is typically set to a lower value than ILIM to allow the port to tolerate minor faults without current limiting. 4266fa 20 LTC4266 APPLICATIONS INFORMATION Per the IEEE specification, the LTC4266 will automatically set ILIM to 425mA (shown in bold in Table 4) during inrush at port turn-on, and then switch to the programmed ILIM setting once inrush has completed. To maintain IEEE compliance, ILIM should kept at 425mA for all Type 1 PDs, and 850mA if a Type 2 PD is detected. ILIM is automatically reset to 425mA when a port turns off. even at extended 802.3at power levels. Current limit and foldback behavior are programmable on a per-port basis. Figure 14 shows MOSFET power dissipation with 802.3afstyle foldback compared with a typical MOSFET SOA curve; Figure 15 demonstrates how two-stage foldback keeps the FET within its SOA under the same conditions. Table 4 gives examples of recommended ILIM register settings. Table 4. Example Current Limit Settings The LTC4266 will support current levels well beyond the maximum values in the 802.3at specification. The shaded areas in Table 4 indicate settings that may require a larger external MOSFET, additional heat sinking, or a reduced tLIM setting. RSENSE = 0.5 53 88 106 08 159 89 213 80 266 8A 319 09 372 8B 425 00 478 8E 531 92 584 CB 638 10 90 744 D2 9A 850 40 C0 956 4A CA 1063 50 D0 1169 5A DA 1275 60 E0 1.0 1488 52 49 0.9 1700 40 0.8 1913 4A 0.7 2125 50 2338 5A 2550 60 2975 52 1.0 08 25 C 0.9 8A ILIM Foldback The LTC4266 features a two-stage foldback circuit that reduces the port current if the port voltage falls below the normal operating voltage. This keeps MOSFET power dissipation at safe levels for typical 802.3af MOSFETs, sA m 80 2.3 af SO A FO 75 0.7 0.6 0.5 2x PSE Current (A) 0.4 0.3 802.3af FOLDBACK 0.2 0.1 0.0 SOA DC AT 90C 0 10 30 40 50 20 PD Voltage (V) at VPSE = 58V 60 4266 F14 C OLDB ACK 25 T sA m 0.6 m A 0.5 SO 0.4 0.3 75 C 90 266 F T sA LTC4 75 A PSE Current (A) Figure 14. Turn On Currents vs FET Safe Operating Area at 90C Ambient SO 80 T 0.8 89 CK 88 BA RSENSE = 0.25 LD INTERNAL REGISTER SETTING (hex) ILIM (mA) 802.3af FOLDBACK 0.2 0.1 0.0 SOA DC AT 90C 0 10 30 40 50 20 PD Voltage (V) at VPSE = 58V 60 4266 F15 Figure 15. LTC4266 Foldback vs FET Safe Operating Area at 90C Ambient 4266fa 21 LTC4266 APPLICATIONS INFORMATION MOSFET Fault Detection LTC4266 PSE ports are designed to tolerate significant levels of abuse, but in extreme cases it is possible for the external MOSFET to be damaged. A failed MOSFET may short source to drain, which will make the port appear to be on when it should be off; this condition may also cause the sense resistor to fuse open, turning off the port but causing the LTC4266 SENSE pin to rise to an abnormally high voltage. A failed MOSFET may also short from gate to drain, causing the LTC4266 GATE pin to rise to an abnormally high voltage. The LTC4266 SENSE and GATE pins are designed to tolerate up to 80V faults without damage. If the LTC4266 sees any of these conditions for more than 180s, it disables all port functionality, reduces the gate drive pull-down current for the port and reports a FET Bad fault. This is typically a permanent fault, but the host can attempt to recover by resetting the port, or by resetting the entire chip if a port reset fails to clear the fault. If the MOSFET is in fact bad, the fault will quickly return, and the port will disable itself again. The remaining ports of the LTC4266 are unaffected. An open or missing MOSFET will not trigger a FET Bad fault, but will cause a tSTART fault if the LTC4266 attempts to turn on the port. Voltage and Current Readback The LTC4266 measures the output voltage and current at each port with an internal A/D converter. Port data is only valid when the port power is on. The converter has two modes: * Slow mode: 14 samples per second, 14.5 bits resolution * Fast mode: 440 samples per second, 9.5 bits resolution In fast mode, the least significant 5 bits of the lower byte are zeroes so that bit scaling is the same in both modes. Disconnect The LTC4266 monitors the port to make sure that the PD continues to draw the minimum specified current. A disconnect timer counts up whenever port current is below 7.5mA (typ), indicating that the PD has been disconnected. If the tDIS timer expires, the port will be turned off and the disconnect bit in the fault event register will be set. If the current returns before the tDIS timer runs out, the timer resets and will start counting from the beginning if the undercurrent condition returns. As long as the PD exceeds the minimum current level more often than tDIS, it will stay powered. Although not recommended, the DC disconnect feature can be disabled by clearing the corresponding enable bits. Note that this defeats the protection mechanisms built into the IEEE spec, since a powered port will stay powered after the PD is removed. If the still-powered port is subsequently connected to a non-PoE data device, the device may be damaged. The LTC4266 does not include AC disconnect circuitry, but includes AC disconnect enable bits to maintain compatibility with the LTC4259A. If the AC disconnect enable bits are set, DC disconnect will be used. Shutdown Pins The LTC4266 includes a hardware SHDN pin for each port. When a SHDN pin is pulled to DGND, the corresponding port will be shut off immediately. The port remains shut down until re-enabled via I2C or a device reset in auto mode. Masked Shutdown The LTC4266 provides a low latency port shedding feature to quickly reduce the system load when required. By allowing a pre-determined set of ports to be turned off, the current on an overloaded main power supply can be reduced rapidly while keeping high priority devices powered. Each port can be configured to high or low priority; all low-priority ports will shut down within 6.5s after the MSD pin is pulled low. If multiple ports in a LTC4266 device are shut down via MSD, they are staggered by at least 0.55s to help reduce voltage transients on the main supply. If a port is turned off via MSD, the corresponding detection and classification enable bits are cleared, so the port will remain off until the host explicitly re-enables detection. 4266fa 22 LTC4266 APPLICATIONS INFORMATION Serial Digital Interface Overview The LTC4266 communicates with the host using a standard SMBus/I2C 2-wire interface. The LTC4266 is a slave-only device, and communicates with the host master using the standard SMBus protocols. Interrupts are signaled to the host via the INT pin. The Timing Diagrams (Figures 6 through 10) show typical communication waveforms and their timing relationships. More information about the SMBus data protocols can be found at www.smbus.org. The LTC4266 requires both the VDD and VEE supply rails to be present for the serial interface to function. Bus Addressing The LTC4266's primary serial bus address is 010xxxxb, with the lower four bits set by the AD3-AD0 pins; this allows up to 16 LTC4266s on a single bus. All LTC4266s also respond to the address 0110000b, allowing the host to write the same command (typically configuration commands) to multiple LTC4266s in a single transaction. If the LTC4266 is asserting the INT pin, it will also respond to the alert response address (0001100b) per the SMBus spec. requires 3.3V (nominally) relative to DGND. VEE requires a negative voltage of between -44V and -57V for Type 1 PSEs, or -50V to -57V for Type 2 PSEs, relative to AGND. The relationship between the two grounds is not fixed; AGND can be referenced to any level from VDD to DGND, although it should typically be tied to either VDD or DGND. VDD provides power for most of the internal LTC4266 circuitry, and draws a maximum of 3mA. A ceramic decoupling cap of at least 0.1F should be placed from VDD to DGND, as close as practical to each LTC4266 chip. Figure 16 shows a three component low dropout regulator for a negative supply to DGND generated from the negative VEE supply. VDD is tied to AGND and DGND is negative referenced to AGND. This regulator drives a single LTC4266 device. In Figure 17, DGND is tied to AGND in this boost converter circuit for a positive VDD supply of 3.3V above AGND. This circuit can drive multiple LTC4266 devices and opto couplers. AGND D1 CMHZ4687-4.3V Register Description For information on serial bus usage and device configuration and status, refer to the LTC4266 Software Programming documentation. External Component Selection Power Supplies and Bypassing The LTC4266 requires two supply voltages to operate. VDD C1 0.1F LTC4266 DGND Q2 CMPTA92 Interrupts and SMBAlert Most LTC4266 port events can be configured to trigger an interrupt, asserting the INT pin and alerting the host to the event. This removes the need for the host to poll the LTC4266, minimizing serial bus traffic and conserving host CPU cycles. Multiple LTC4266s can share a common INT line, with the host using the SMBAlert protocol (ARA) to determine which LTC4266 caused an interrupt. AGND VDD R5 750k VEE 4266 F16 VEE Figure 16. Negative LDO to DGND VEE is the main supply that provides power to the PDs. Because it supplies a relatively large amount of power and is subject to significant current transients, it requires more design care than a simple logic supply. For minimum IR loss and best system efficiency, set VEE near maximum amplitude (57V), leaving enough margin to account for transient over- or undershoot, temperature drift, and the line regulation specs of the particular power supply used. Bypass capacitance between AGND and VEE is very important for reliable operation. If a short circuit occurs at one of the output ports it can take as long as 1s for the LTC4266 to begin regulating the current. During this time the current is limited only by the small impedances in the circuit and a high current spike typically occurs, causing a 4266fa 23 LTC4266 APPLICATIONS INFORMATION L3 100H SUMIDA CDRH5D28-101NC C77 0.22F 100V C76 10F 63V + C78 0.22F 100V R60 10 C74 100F 6.3V C75 10F 16V R53 4.7k 1% 5 VCC 1 R58 10 VEE R51 4.7k 1% R54 56k C79 2200pF ITH/RUN L4 10H SUMIDA CDRH4D28-100NC D28 B1100 C73 10F 6.3V R52 3.32k 1% Q13 FMMT723 NGATE 6 3.3V AT 400mA Q14 FMMT723 Q15 FDC2512 LTC3803 3 VFB SENSE 4 GND 2 R57 1k R55 806 1% R59 0.100 1%, 1W R56 47.5k 1% 4266 F17 Figure 17. Positive VDD Boost Converter voltage transient on the VEE supply and possibly causing the LTC4266 to reset due to a UVLO fault. A 1F, 100V X7R capacitor placed near the VEE pin is recommended to minimize spurious resets. of the LTC4266 is designed to drive the inputs of an optocoupler directly. Standard I2C/SMBus devices typically cannot drive opto-couplers, so U1 is used to buffer the signals from the host controller side. Isolating the Serial Bus External MOSFET The LTC4266 includes a split SDA pin (SDAIN and SDAOUT) to ease opto-isolation of the bidirectional SDA line. Careful selection of the power MOSFET is critical to system reliability. LTC recommends either Fairchild IRFM120A, FDT3612, FDMC3612 or Philips PHT6NQ10T for their proven reliability in Type 1 and Type 2 PSE applications. Non-standard applications that provide more current than the 850mA IEEE maximum may require heat sinking and other MOSFET design considerations. Contact LTC Applications before using a MOSFET other than one of these recommended parts. IEEE 802.3 Ethernet specifications require that network segments (including PoE circuitry) be electrically isolated from the chassis ground of each network interface device. However, network segments are not required to be isolated from each other, provided that the segments are connected to devices residing within a single building on a single power distribution system. For simple devices such as small PoE switches, the isolation requirement can be met by using an isolated main power supply for the entire device. This strategy can be used if the device has no electrically conducting ports other than twisted-pair Ethernet. In this case, the SDAIN and SDAOUT pins can be tied together and will act as a standard I2C/SMBus SDA pin. If the device is part of a larger system, contains additional external non-Ethernet ports, or must be referenced to protective ground for some other reason, the Power over Ethernet subsystem (including all LTC4266s) must be electrically isolated from the rest of the system. Figure 18 shows a typical isolated serial interface. The SDAOUT pin Sense R The LTC4266 is designed to use either 0.5 or 0.25 current sense resistors. For new designs 0.25 is recommended to reduce power dissipation; the 0.5 option is intended for existing systems where the LTC4266 is used as a drop-in replacement for the LTC4258 or LTC4259A. The lower sense resistor values reduce heat dissipation. Four commonly available 1 resistors (0402 or larger package size) can be used in parallel in place of a single 0.25 resistor. In order to meet the ICUT and ILIM accuracy required by the IEEE specification, the sense resistors should have 1% tolerance or better, and no more than 200ppm/C temperature coefficient. 4266fa 24 LTC4266 APPLICATIONS INFORMATION 0.1F 0.1F I2C ADDRESS VDD LTC4266 INT SCL SDAIN SDAOUT AD0 AD1 AD2 AD3 DGND AGND 0100000 VDD LTC4266 INT SCL SDAIN SDAOUT AD0 AD1 AD2 AD3 DGND AGND 0100001 VDD LTC4266 INT SCL SDAIN SDAOUT AD0 AD1 AD2 AD3 DGND AGND 0100010 0.1F VDD CPU 2k U2 200 U1 SCL 2k 200 SDA 0.1F HCPL-063L TO CONTROLLER U3 200 200 SMBALERT 0.1F 0.1F GND CPU U1: FAIRCHILD NC7WZ17 U2, U3: AGILENT HCPL-063L HCPL-063L 0.1F ISOLATED 3.3V 10F + ISOLATED GND * * * LTC4266 VDD INT SCL SDAIN SDAOUT AD0 AD1 AD2 AD3 DGND AGND 0101110 VDD LTC4266 INT SCL SDAIN SDAOUT AD0 AD1 AD2 AD3 DGND AGND 0101111 4266 F18 Figure 18. Opto-Isolating the I2C Bus 4266fa 25 LTC4266 APPLICATIONS INFORMATION Output Cap Each port requires a 0.22F cap across its outputs to keep the LTC4266 stable while in current limit during startup or overload. Common ceramic capacitors often have significant voltage coefficients; this means the capacitance is reduced as the applied voltage increases. To minimize this problem, X7R ceramic capacitors rated for at least 100V are recommended. ESD/Cable Discharge Protection Ethernet ports can be subject to significant ESD events when long data cables, each potentially charged to thousands of volts, are plugged into the low impedance of the RJ45 jack. To protect against damage, each port requires a pair of clamp diodes; one to AGND and one to VEE (Figure 10). An additional surge suppressor is required for each LTC4266 chip from VEE to AGND. The diodes at the ports steer harmful surges into the supply rails, where they are absorbed by the surge suppressor and the VEE bypass capacitance. The surge suppressor has the additional benefit of protecting the LTC4266 from transients on the VEE supply. S1B diodes work well as port clamp diodes, and an SMAJ58A or equivalent is recommended for the VEE surge suppressor. Layout Guidelines Standard power layout guidelines apply to the LTC4266: place the decoupling caps for the VDD and VEE supplies near their respective supply pins, use ground planes, and use wide traces wherever there are significant currents. The main layout challenge involves the arrangement of the current sense resistors, and their connections to the LTC4266. Because the sense resistor values are very low, layout parasitics can cause significant errors. Care is required to achieve specified accuracy, particularly with disconnect currents. Figure 19 illustrates the problem. In the example on the left, two ports have load currents I1 and I2 that return to the VEE power supply through a mutual resistance RM. RM represents the combined resistances of any traces, planes, and vias in the PCB that I1 and I2 share as they return to the VEE supply. The LTC4266 measures the voltage difference between its SENSE and VEE pins to sense the voltage drop across RS1, but as the example shows, RM introduces errors. The example on the right shows how errors can be minimized with a good layout. The circuit is rearranged so that RM no longer affects VS, and the VEE connection to the LTC4266 is used as a Kelvin sense trace. VEE is not a perfect Kelvin connection because all four ports controlled by the LTC4266 share the same sense trace, and because the current through the trace (IEE) is not zero. However, as the equation shows, the remaining error is a small offset term. Figure 20 shows two LTC4266 chips controlling eight ports (A though H). The ports are separated into two groups of four; each has its own trace on the top PCB layer that I1 I2 I1 I2 LTC4266 LTC4266 GATE GATE SENSE + VS VEE - IEE RM I1 + I2 + IEE RS1 RS2 MUTUAL RESISTANCE SENSE + VS VEE - IEE RK KELVIN SENSE LINE VS = I1RS1 + I1RM + I2RM SIGNAL SCALE ERROR CROSSTALK ERROR RS1 RS2 RM I1 + I2 + IEE VS = I1RS1 - IEERK SIGNAL SMALL OFFSET ERROR 4266 F19 Figure 19. Layout Affects Current Readback Accuracy 4266fa 26 LTC4266 APPLICATIONS INFORMATION connects to the VEE plane with a via. Currents from the U1 sub-circuit are effectively isolated from the U2 sub-circuit, reducing the layout problem down to 4-port chunks; this arrangement can be expanded for any number of ports. Figure 21 shows an example of good 4-port layout. Each 0.25 sense resistor consists of four 1 resistors in parallel. The four groups of resistors are arranged to minimize the overlap in their current flows, which minimizes mutual resistance. The horizontal slits cut in the copper help to keep the currents separate. Wide copper paths connect each group of resistors to the vias at the center, so the resistance is very low. Proper connection of the sense line is also important. In Figure 21, U1 is not connected directly to the VEE plane but is connected instead to a Kelvin sense trace that leads to the sense resistor array. Similarly, the via at the center of the sense resistor array has a matching hole in the VEE plane. This arrangement prevents the mutual resistance of the four large vias from influencing the current measurements. PORTS A THROUGH D U1 LTC4266 PORTS E THROUGH H U2 LTC4266 SENSE1 SENSE2 SENSE3 SENSE4 SENSE1 SENSE2 SENSE3 SENSE4 VEE VEE THIS TRACE PROVIDES VEE TO U1 BUT ALSO ACTS AS A KELVIN SENSE LINE FOR PORTS A-D RSENSE VIA VIA RETURN TO VEE POWER SUPPLY VEE COPPER FILL ON SURFACE LAYER VEE PLANE ON INNER LAYER BY KEEPING THESE COPPER FILLS SEPARATE ON THE SURFACE, MUTUAL RESISTANCE BETWEEN PORTS A-D AND E-H IS ELIMINATED 4266 F20 Figure 20. Layout Strategy to Reduce Mutual Resistance EDGE OF VEE PLANE (ON SOME INNER LAYER) KELVIN SENSE TRACE CONNECTS U1 TO VEE THROUGH THE VIAS ON THE RIGHT PORT A RSENSE PORT B RSENSE PIN 1 FOUR LARGE VIAS TO VEE PLANE HOLE IN VEE PLANE U1 THE PADDLE IS CONNECTED TO VEE PINS PORT C RSENSE PORT D RSENSE VIAS TO SOURCE PIN OF THE PORT D MOSFET LOCATED ON THE OPPOSITE SIDE OF THE BOARD 4266 F21 Figure 21. Good PCB Layout Example 4266fa 27 LTC4266 Package Description GW Package 36-Lead Plastic SSOP (Wide .300 Inch) (Reference LTC DWG # 05-08-1642) 36 19 1.40 0.127 10.804 MIN 15.291 - 15.545* (.602 - .612) 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 7.75 - 8.258 10.11 - 10.55 (.398 - .415) 1 0.520 0.0635 18 0.800 BSC RECOMMENDED SOLDER PAD LAYOUT 7.417 - 7.595** (.292 - .299) 0.254 - 0.406 s 45 (.010 - .016) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 0.355 REF 2.286 - 2.388 (.090 - .094) 2.44 - 2.64 (.096 - .104) 0 - 8 TYP 0.231 - 0.3175 (.0091 - .0125) 0.40 - 1.27 (.015 - .050) NOTE: 1. CONTROLLING DIMENSION: MILLIMETERS MILLIMETERS 2. DIMENSIONS ARE IN (INCHES) 0.800 (.0315) BSC 0.28 - 0.51 (.011 - .02) TYP 0.1 - 0.3 (.004 - .0118) GW36 SSOP 0204 *DIMENSION DOES NOT INCLUDE MOLD FLASH. MOLD FLASH SHALL NOT EXCEED 0.152mm (0.006") PER SIDE **DIMENSION DOES NOT INCLUDE INTERLEAD FLASH. INTERLEAD FLASH SHALL NOT EXCEED 0.254mm (0.010") PER SIDE 4266fa 28 LTC4266 PACKAGE DESCRIPTION UHF Package 38-Lead Plastic QFN (5mm x 7mm) (Reference LTC DWG # 05-08-1701 Rev C) 0.70 p 0.05 5.50 p 0.05 5.15 0.05 4.10 p 0.05 3.00 REF 3.15 0.05 PACKAGE OUTLINE 0.25 p 0.05 0.50 BSC 5.5 REF 6.10 p 0.05 7.50 p 0.05 RECOMMENDED SOLDER PAD LAYOUT APPLY SOLDER MASK TO AREAS THAT ARE NOT SOLDERED 5.00 p 0.10 0.75 p 0.05 PIN 1 NOTCH R = 0.30 TYP OR 0.35 s 45o CHAMFER 3.00 REF 37 0.00 - 0.05 38 0.40 p0.10 PIN 1 TOP MARK (SEE NOTE 6) 1 2 5.15 0.10 5.50 REF 7.00 p 0.10 3.15 0.10 (UH) QFN REF C 1107 0.200 REF 0.25 p 0.05 0.50 BSC R = 0.125 TYP R = 0.10 TYP BOTTOM VIEW--EXPOSED PAD NOTE: 1. DRAWING CONFORMS TO JEDEC PACKAGE OUTLINE M0-220 VARIATION WHKD 2. DRAWING NOT TO SCALE 3. ALL DIMENSIONS ARE IN MILLIMETERS 4. DIMENSIONS OF EXPOSED PAD ON BOTTOM OF PACKAGE DO NOT INCLUDE MOLD FLASH. MOLD FLASH, IF PRESENT, SHALL NOT EXCEED 0.20mm ON ANY SIDE 5. EXPOSED PAD SHALL BE SOLDER PLATED 6. SHADED AREA IS ONLY A REFERENCE FOR PIN 1 LOCATION ON THE TOP AND BOTTOM OF PACKAGE 4266fa Information furnished by Linear Technology Corporation is believed to be accurate and reliable. However, no responsibility is assumed for its use. Linear Technology Corporation makes no representation that the interconnection of its circuits as described herein will not infringe on existing patent rights. 29 LTC4266 Typical Application ISOLATED 3.3V ISOLATED GND 0.1F DGND AGND 2k U2 200 VDD CPU FB1 SCL 1/4 SDAIN LTC4266 SDAOUT INT U1 SCL VEE 2k 200 TO CONTROLLER 0.22F 100V X7R VDD 1F 100V X7R SENSE GATE OUT FB2 S1B RS 0.25 SDA Q1 HCPL-063L U3 SMAJ58A 200 -48V ISOLATED S1B RJ45 CONNECTOR T1 * 200 INTERRUPT * * * * 0.01F 200V 75 0.01F 200V 75 75 75 1 2 3 4 5 6 7 8 * PHY HCPL-063L DTSS: DIODES INC SMAJ58A Q1: FAIRCHILD IRFM120A OR PHILIPS PHT6NQ10T U1: FAIRCHILD NC7WZ17 U2, U3: AGILENT HCPL-063L FB1, FB2:TDK MPZ2012S601A T1: PULSE H6096NL OR COILCRAFT ETH1-230LD (NETWORK PHYSICAL LAYER CHIP) * * * GND CPU * * 0.1F 0.01F 200V 0.01F 200V * 4266 F22 1000pF 2000V Figure 22. One Complete Isolated Powered Etherent Port Related Parts PART NUMBER DESCRIPTION LT3803 Constant Frequency Current Mode Flyback DC/DC Controller in ThinSOTTM COMMENTS 200kHz Operation, Adjustable Slope Compensation LTC4258 Quad IEEE 802.3af PoE PSE Controller DC Disconnect Sensing Only LTC4263 Single IEEE 802.3af PSE Controller Internal FET Switch LTC4263-1 High Power Single PoE PSE Controller With Internal FET Switch LTC4265 IEEE 802.3at PD Interface Controller 100V, 1A Internal Switch, 2-Event Classification Recognition LTC4267 IEEE 802.3af PD Interface With Integrated Switching Regulator Internal 100V, 400mA Switch, Dual Inrush Current, Programmable Class LTC4268-1 High Power PD With Synchronous Flyback Controller No Opto-coupler Required LTC4269-1 IEEE 802.3at PD Interface Integrated Switching Regulator 2-Event Classification, Programmable Classification, Synchronous No-Opto Flyback Controller, 50kHz to 250kHz LTC4269-2 IEEE 802.3at PD Interface Integrated Switching Regulator 2-Event Classification, Programmable Classification, Synchronous Forward Controller, 100kHz to 500kHz 4266fa 30 Linear Technology Corporation 1630 McCarthy Blvd., Milpitas, CA 95035-7417 (408) 432-1900 FAX: (408) 434-0507 www.linear.com LT 0810 REV A * PRINTED IN USA LINEAR TECHNOLOGY CORPORATION 2009