Super Sequencer™ with Margining Control and Auxiliary ADC Inputs ADM1066 FEATURES FUNCTIONAL BLOCK DIAGRAM AUX1 AUX2 REFIN REFOUT REFGND ADM1066 MUX VREF SDA SCL A1 A0 SMBus INTERFACE 12-BIT SAR ADC EEPROM CLOSED-LOOP MARGINING SYSTEM VX1 VX3 VX4 VX5 CONFIGURABLE OUTPUT DRIVERS DUALFUNCTION INPUTS VX2 PDO1 PDO2 PDO3 (LOGIC INPUTS OR SFDs) (HV CAPABLE OF DRIVING GATES OF N-CHANNEL FET) PDO4 CONFIGURABLE OUTPUT DRIVERS PDO7 (LV CAPABLE OF DRIVING LOGIC SIGNALS) PDO9 PDO5 PDO6 SEQUENCING ENGINE VP1 VP2 VP3 PROGRAMMABLE RESET GENERATORS VP4 (SFDs) VH AGND PDO8 PDO10 PDOGND VOUT DAC VOUT DAC VOUT DAC VOUT DAC VOUT DAC VOUT DAC DAC1 DAC2 DAC3 DAC4 DAC5 DAC6 VDD ARBITRATOR VDDCAP GND Figure 1. APPLICATIONS Central office systems Servers/routers Multivoltage system line cards DSP/FPGA supply sequencing In-circuit testing of margined supplies GENERAL DESCRIPTION The ADM1066 is a configurable supervisory/sequencing device that offers a single-chip solution for supply monitoring and sequencing in multiple supply systems. In addition to these functions, the ADM1066 integrates a 12-bit ADC and six 8-bit voltage output DACs. These circuits can be used to implement a closed-loop margining system, which enables supply adjustment by altering either the feedback node or reference of a dc/dc converter using the DAC outputs. (continued on Page 3) Rev. 0 Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners. 04609-001 Complete supervisory and sequencing solution for up to 10 supplies 10 supply fault detectors enable supervision of supplies to better than 1% accuracy 5 selectable input attenuators allow supervision: Supplies up to 14.4 V on VH Supplies up to 6 V on VP1–4 5 dual-function inputs, VX1–5: High impedance input to supply fault detector with thresholds between 0.573 V and 1.375 V General-purpose logic input 10 programmable output drivers (PDO1–10): Open collector with external pull-up Push/pull output, driven to VDDCAP or VPn Open collector with weak pull-up to VDDCAP or VPn Internally charge-pumped high drive for use with external N-FET (PDO1–6 only) Sequencing engine (SE) implements state machine control of PDO outputs: State changes conditional on input events Enables complex control of boards Power-up and power-down sequence control Fault event handling Interrupt generation on warnings Watchdog function can be integrated in SE Program software control of sequencing through SMBus Complete voltage margining solution for 6 voltage rails 6 voltage output 8-bit DACs (0.300 V to 1.551 V) allow voltage adjustment via dc/dc converter trim/feedback node 12-bit ADC for readback of all supervised voltages 2 auxiliary (single-ended) ADC inputs Reference input, REFIN, has 2 input options: Driven directly from 2.048V (±0.25%) REFOUT pin More accurate external reference for improved ADC performance Device powered by the highest of VP1–4, VH for improved redundancy User EEPROM: 256 bytes Industry-standard 2-wire bus interface (SMBus) Guaranteed PDO low with VH, VPn = 1.2 V 40-lead 6 mm × 6 mm LFCSP and 48-lead 7 mm × 7 mm TQFP packages One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.326.8703 © 2004 Analog Devices, Inc. All rights reserved. ADM1066 TABLE OF CONTENTS General Description ......................................................................... 3 Timeout Detector ....................................................................... 19 Specifications..................................................................................... 4 Fault Reporting........................................................................... 19 Pin Configurations and Function Descriptions ........................... 7 Voltage Readback............................................................................ 20 Absolute Maximum Ratings............................................................ 8 Supply Supervision with the ADC ........................................... 20 Thermal Characteristics .............................................................. 8 Supply Margining ........................................................................... 21 ESD Caution.................................................................................. 8 Overview ..................................................................................... 21 Typical Performance Characteristics ............................................. 9 Open-Loop Margining .............................................................. 21 Powering the ADM1066 ................................................................ 12 Closed-Loop Supply Margining ............................................... 21 Inputs................................................................................................ 13 Writing to the DACs .................................................................. 22 Supply Supervision..................................................................... 13 Choosing the Size of the Attenuation Resistor....................... 22 Programming the Supply Fault Detectors............................... 13 DAC Limiting/Other Safety Features ...................................... 22 Input Comparator Hysteresis.................................................... 14 Applications Diagram .................................................................... 23 Input Glitch Filtering ................................................................. 14 Communicating with the ADM1066 ........................................... 24 Supply Supervision with VXn Inputs....................................... 14 Configuration Download at Power-Up................................... 24 VXn Pins as Digital Inputs........................................................ 15 Updating the Configuration ..................................................... 24 Outputs ............................................................................................ 16 Updating the Sequencing Engine............................................. 25 Supply Sequencing through Configurable Output Drivers .. 16 Internal Registers........................................................................ 25 Sequencing Engine ......................................................................... 17 EEPROM ..................................................................................... 25 Overview...................................................................................... 17 Serial Bus Interface..................................................................... 25 Warnings...................................................................................... 17 Write Operations ........................................................................ 27 SMBus Jump/Unconditional Jump .......................................... 17 Read Operations......................................................................... 29 Sequencing Engine Application Example ............................... 18 Outline Dimensions ....................................................................... 31 Sequence Detector...................................................................... 19 Ordering Guide .......................................................................... 31 Monitoring Fault Detector ........................................................ 19 REVISION HISTORY 10/04—Revision 0: Initial Version Rev. 0 | Page 2 of 32 ADM1066 GENERAL DESCRIPTION (continued from Page 1) Supply margining can be performed with a minimum of external components. The margining loop can be used for incircuit testing of a board during production (for example, to verify the board’s functionality at −5% of nominal supplies), or can be used dynamically to accurately control the output voltage of a dc/dc converter. The logical core of the device is a sequencing engine. This statemachine-based construction provides up to 63 different states. This design enables very flexible sequencing of the outputs, based on the condition of the inputs. The device is controlled via configuration data that can be programmed into an EEPROM. The whole configuration can be programmed using an intuitive GUI-based software package provided by ADI. The device also provides up to ten programmable inputs for monitoring under, over, or out-of-window faults on up to ten supplies. In addition, ten programmable outputs can be used as logic enables. Six of them can also provide up to a 12 V output for driving the gate of an N-channel FET, which can be placed in the path of a supply. AUX2 AUX1 REFIN REFOUT REFGND SDA SCL A1 ADM1066 A0 SMBus INTERFACE VREF OSC 12-BIT SAR ADC DEVICE CONTROLLER EEPROM GPI SIGNAL CONDITIONING VX1 CONFIGURABLE O/P DRIVER (HV) SFD PDO1 PDO2 VX2 PDO3 VX3 PDO4 VX4 PDO5 GPI SIGNAL CONDITIONING SEQUENCING ENGINE VX5 VP1 CONFIGURABLE O/P DRIVER (HV) PDO6 CONFIGURABLE O/P DRIVER (LV) PDO7 SFD SELECTABLE ATTENUATOR SFD VP2 VP3 PDO8 VP4 PDO9 SELECTABLE ATTENUATOR CONFIGURABLE O/P DRIVER (LV) SFD PDOGND SFDGND VDDCAP PDO10 VDD ARBITRATOR GND REG 5.25V CHARGE PUMP VOUT DAC VCCP DAC1 VOUT DAC DAC2 DAC3 Figure 2. Detailed Block Diagram Rev. 0 | Page 3 of 32 DAC4 DAC5 DAC6 04609-002 VH ADM1066 SPECIFICATIONS VH = 3.0 V to 14.4 V1, VPn = 3.0 V to 6.0 V1, TA = −40°C to +85°C, unless otherwise noted. Table 1. Parameter POWER SUPPLY ARBITRATION VH, VPn VP VH VDDCAP CVDDCAP POWER SUPPLY Supply Current, IVH, IVPn Additional Currents All PDO FET Drivers On Min Typ Max Unit Test Conditions/Comments 4.75 6.0 14.4 5.4 V V V V µF Minimum supply required on one of VPn, VH Maximum VDDCAP = 5.1 V, typical VDDCAP = 4.75 V Regulated LDO output Minimum recommended decoupling capacitance 4.2 6 mA VDDCAP = 4.75 V, PDO1–10 off, DACs off, ADC off mA 3.0 2.7 10 1 2.2 1 10 mA mA mA VDDCAP = 4.75 V, PDO1-6 loaded with 1 µA each, PDO7–10 off Maximum additional load that can be drawn from all PDO pull-ups to VDDCAP 6 DACs on with 100 µA maximum load on each Running round-robin loop 1 ms duration only, VDDCAP = 3 V ±0.05 % Midrange and high range Current Available from VDDCAP DACs Supply Current ADC Supply Current EEPROM Erase Current SUPPLY FAULT DETECTORS VH Pin Input Attenuator Error Detection Ranges High Range Midrange VPn Pins Input Attenuator Error Detection Ranges Midrange Low Range Ultralow Range VX Pins Input Impedance Detection Ranges Ultralow Range Absolute Accuracy 2 6 2.5 ±0.05 2.5 1.25 0.573 Input Reference Voltage on REFIN Pin, VREFIN Resolution INL Gain Error 6 3 1.375 1 V V % Low range and midrange V V V No input attenuation error MΩ 0.573 Threshold Resolution Digital Glitch Filter ANALOG-TO-DIGITAL CONVERTER Signal Range 14.4 6 mA 1.375 ±1 8 0 100 0 V % Bits µs µs VREFIN 2.048 V No input attenuation error VREF error + DAC nonlinearity + comparator offset error + input attenuation error Minimum programmable filter length Maximum programmable filter length The ADC can convert signals presented to the VH, VPn, and VX_GPIn pins. VPn and VH input signals are attenuated depending on selected range. A signal at the pin corresponding to the selected range is from 0.573 V to 1.375 V at the ADC input. V 12 ±2.5 ±0.05 Rev. 0 | Page 4 of 32 Bits LSB % Endpoint corrected, VREFIN = 2.048 V VREFIN = 2.048 V ADM1066 Parameter Conversion Time Min Offset Error Input Noise BUFFERED VOLTAGE OUTPUT DACs Resolution Code 0x80 Output Voltage Range 1 Range 2 Range 3 Range 4 Output Voltage Range LSB Step Size INL DNL Gain Error Load Regulation Minimum Load Capacitance Load Regulation PSRR PROGRAMMABLE DRIVER OUTPUTS High Voltage (Charge Pump) Mode (PDO1–6) Output Impedance VOH IOUTAVG Standard (Digital Output) Mode (PDO1–10) VOH Max 0.25 Unit ms ms LSB LSBrms 8 Bits ±2 0.592 0.796 0.996 1.246 0.6 0.8 1 1.25 601.25 2.36 0.603 0.803 1.003 1.253 ±0.75 ±0.4 1 -4 2 50 2 2.5 60 40 2.043 2.048 −0.25 0.25 2.053 1 2 60 11 10.5 500 12.5 12 20 14 13.5 2.4 Three-State Output Leakage Current Oscillator Frequency VPU − 0.3 0 100 Endpoint corrected Sourcing Current, IREFOUTMA X= -200µA Sinking Current, IREFOUTMA X= 100µA Per mA DC 100 mV step in 20 ns with 50 pF load No load Sourcing current, IDACnMAX = −100 µA Sinking current, IDACnMAX = 100 µA Capacitor required for decoupling, stability Per 100 µA DC kΩ V V µA IOH = 0 IOH = 1µA 2 V < VOH < 7 V 10 110 µA kHz Rev. 0 | Page 5 of 32 Same range, independent of center point V mV mV µF mV dB 2 0.50 20 60 20 90 V V V V mV mV LSB LSB % mV mV pF µs mV dB dB V V V V mA mA kΩ mA 4.5 VOL IOL2 ISINK2 RPULL-UP ISOURCE (VPn)2 Test Conditions/Comments One conversion on one channel All 12 channels selected, 16x averaging enabled VREFIN = 2.048 V Direct input (no attenuator) 6 DACs are individually selectable for centering on one of four output voltage ranges Maximum Load Capacitance Settling Time into 50 pF Load Load Regulation PSRR REFERENCE OUTPUT Reference Output Voltage Load Regulation Typ 0.44 84 VPU (pull-up to VDDCAP or VPN) = 2.7 V, IOH = 0.5 mA VPU to Vpn = 6.0 V, IOH = 0 mA VPU ≤ 2.7 V, IOH = 0.5 mA IOL = 20 mA Maximum sink current per PDO pin Maximum total sink for all PDOs Internal pull-up Current load on any VPn pull-ups, that is, total source current available through any number of PDO pull-up switches configured onto any one VPDO = 14.4 V All on-chip time delays derived from this clock ADM1066 Parameter DIGITAL INPUTS (VXn, A0, A1) Input High Voltage, VIH Input Low Voltage, VIL Input High Current, IIH Input Low Current, IIL Input Capacitance Programmable Pull-Down Current, IPULL-DOWN SERIAL BUS DIGITAL INPUTS (SDA, SCL) Input High Voltage, VIH Input Low Voltage, VIL Output Low Voltage, VOL 2 SERIAL BUS TIMING Clock Frequency, fSCLK Bus Free Time, tBUF Start Setup Time, tSU;STA Start Hold Time, tHD;STA SCL Low Time, tLOW SCL High Time, tHIGH SCL, SDA Rise Time, tr SCL, SDA Fall Time, tf Data Setup Time, tSU;DAT Data Hold Time, tHD;DAT Input Low Current, IIL SEQUENCING ENGINE TIMING State Change Time 1 2 Min Typ Max 2.0 0.8 −1 1 5 20 2.0 0.8 0.4 400 4.7 4.7 4 4.7 4 1000 300 250 5 1 10 Unit Test Conditions/Comments V V µA µA pF µA Maximum VIN = 5.5 V Maximum VIN = 5.5 V VIN = 5.5 V VIN = 0 VDDCAP = 4.75, TA = 25°C, if known logic state is required V V V IOUT = −3.0 mA kHz µs µs µs µs µs µs µs ns ns µA VIN = 0 µs At least one of the VH, VP1-4 pins must be ≥3.0 V to maintain the device supply on VDDCAP. Specification is not production tested, but is supported by characterization data at initial product release. Rev. 0 | Page 6 of 32 ADM1066 PDOGND NC 31 VCCP PDOGND 32 A0 VCCP 33 A1 A0 34 SCL A1 35 SDA SCL 36 AUX2 SDA 37 AUX1 AUX2 38 VDDCAP AUX1 39 GND VDDCAP 40 NC GND PIN CONFIGURATIONS AND FUNCTION DESCRIPTIONS 48 47 46 45 44 43 42 41 40 39 38 37 NC 1 VX1 1 36 NC PIN 1 INDICATOR 30 PDO1 VX1 2 29 PDO2 VX2 3 34 PDO2 VX3 3 28 PDO3 VX3 4 33 PDO3 VX4 4 27 PDO4 VX4 5 26 PDO5 VX5 6 ADM1066 31 PDO5 25 PDO6 VP1 7 TOP VIEW (Not to Scale) 30 PDO6 VP2 7 24 PDO7 VP2 8 29 PDO7 VP3 8 23 PDO8 VP3 9 28 PDO8 VP4 9 22 PDO9 VP4 10 27 PDO9 VH 10 21 PDO10 VH 11 26 PDO10 NC 12 25 NC PIN 1 INDICATOR 14 15 16 17 18 19 20 21 22 23 24 DAC1 DAC2 DAC3 DAC4 DAC5 DAC6 NC 04609-004 13 REFOUT 20 REFIN 19 REFGND 18 32 PDO4 NC 17 DAC3 16 DAC2 15 DAC1 14 REFOUT REFGND 13 REFIN 12 AGND 11 35 PDO1 AGND TOP VIEW (Not to Scale) 04609-003 VP1 6 DAC6 ADM1066 DAC5 VX5 5 DAC4 VX2 2 NC = NO CONNECT Figure 3. LFCSP Pin Configuration Figure 4. TQFP Pin Configuration Table 2. Pin Function Descriptions Pin No. LFCSP TQFP 1, 12–13, 24–25, 36–37, 48 1–5 2–6 Mnemonic NC Description No connection. VX1–5 High Impedance Inputs to Supply Fault Detectors. Fault thresholds can be set from 0.573 V to 1.375 V. Alternatively, these pins can be used as general-purpose digital inputs. Low Voltage Inputs to Supply Fault Detectors. Three input ranges can be set by altering the input attenuation on a potential divider connected to these pins, the output of which connects to a supply fault detector. These pins allow thresholds from 2.5 V to 6.0 V, 1.25 V to 3.00 V, and 0.573 V to 1.375 V. High Voltage Input to Supply Fault Detectors. Three input ranges can be set by altering the input attenuation on a potential divider connected to this pin, the output of which connects to a supply fault detector. This pin allows thresholds from 6.0 V to 14.4 V and 2.5 V to 6.0 V. Ground Return for Input Attenuators. Ground Return for On-Chip Reference Circuits. Reference Input for ADC. Nominally, 2.048 V. 2.048 V Reference Output. Voltage Output DACs. These pins default to high impedance at power-up. Programmable Output Drivers. Ground Return for Output Drivers. Central Charge-Pump Voltage of 5.25 V. A reservoir capacitor must be connected between this pin and GND. Logic Input. This pin sets the seventh bit of the SMBus interface address. Logic Input. This pin sets the sixth bit of the SMBus interface address. SMBus Clock Pin. Open-drain output requires external resistive pull-up. SMBus Data I/O Pin. Open-drain output requires external resistive pull-up. Auxiliary, Single-Ended ADC Input. Auxiliary, Single-Ended ADC Input. Device Supply Voltage. Linearly regulated from the highest of the VP1–4, VH pins to a typical of 4.75 V. Supply Ground. 6–9 7–10 VP1–4 10 11 VH 11 12 13 14 15–20 21–30 31 32 14 15 16 17 18–23 26–35 38 39 AGND REFGND REFIN REFOUT DAC1–6 PDO10–1 PDOGND VCCP 33 34 35 36 37 38 39 40 40 41 42 43 44 45 46 47 A0 A1 SCL SDA AUX2 AUX1 VDDCAP GND Rev. 0 | Page 7 of 32 ADM1066 ABSOLUTE MAXIMUM RATINGS Table 3. Parameter Voltage on VH Pin Voltage on VP Pins Voltage on VX Pins Voltage on AUX1, AUX2, REFIN Pins Input Current at Any Pin Package Input Current Maximum Junction Temperature (TJ max) Storage Temperature Range Lead Temperature, Soldering Vapor Phase, 60 s Rating 16 V 7V −0.3 V to +6.5 V −0.3 V to +5 V ±5 mA ±20 mA 150°C −65°C to +150°C Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. THERMAL CHARACTERISTICS 40-lead LFCSP package: θJA = 25°C/W 215°C 48-lead TQFP package: θJA = 14.8°C/W ESD CAUTION ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on the human body and test equipment and can discharge without detection. Although this product features proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high energy electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance degradation or loss of functionality. Rev. 0 | Page 8 of 32 ADM1066 TYPICAL PERFORMANCE CHARACTERISTICS 180 6 160 5 140 120 IVP1 (µA) VVDDCAP (V) 4 3 100 80 60 2 40 0 0 1 2 3 4 5 04609-053 04609-050 1 20 0 0 6 1 2 3 4 5 6 VVP1 (V) VVP1 (V) Figure 5. VVDDCAP vs. VVP1 Figure 8. IVP1 vs. VVP1 (VP1 Not as Supply) 5.0 6 4.5 5 4.0 3.5 3.0 IVH (mA) 3 2 2.5 2.0 1.5 1.0 04609-051 1 0 0 2 4 6 8 10 12 14 04609-054 VVDDCAP (V) 4 0.5 0 0 16 2 4 6 8 10 12 14 16 VVH (V) VVH (V) Figure 9. IVH vs. VVH (VH as Supply) Figure 6. VVDDCAP vs. VVH 350 5.0 4.5 300 4.0 250 IVH (µA) 3.0 2.5 2.0 1.5 200 150 100 1.0 0.5 0 0 1 2 3 4 5 04609-055 50 04609-052 IVP1 (mA) 3.5 0 0 6 1 2 3 4 VVH (V) VVP1 (V) Figure 10. IVH vs. VVH (VH Not as Supply) Figure 7. IVP1 vs. VVP1 (VP1 as Supply) Rev. 0 | Page 9 of 32 5 6 ADM1066 14 1.0 0.8 12 0.4 0.2 DNL (LSB) VPDO1 CHARGE PUMPED 0.6 10 8 6 0 –0.2 –0.4 4 04609-056 0 0 2.5 5.0 7.5 10.0 12.5 04609-066 –0.6 2 –0.8 –1.0 15.0 0 1000 ILOAD CURRENT (µA) 0.8 4.0 0.6 3.5 0.4 VP1 = 5V 2.5 VP1 = 3V 2.0 0.2 0 –0.2 1.5 –0.4 1.0 –0.6 0.5 0 3 4 5 04609-063 INL (LSB) 3.0 04609-057 VPDO1 (V) 1.0 4.5 2 4000 Figure 14. DNL for ADC 5.0 1 3000 CODE Figure 11. VPDO1 (FET Drive Mode) vs. ILOAD 0 2000 –0.8 –1.0 6 0 1000 ILOAD (mA) 2000 3000 4000 CODE Figure 12. VPDO1 (Strong Pull-Up VP) vs. ILOAD Figure 15. INL for ADC 4.5 12000 4.0 9894 10000 3.5 VP1 = 5V HITS PER CODE 2.5 VP1 = 3V 2.0 1.5 8000 6000 4000 1.0 0.5 0 0 10 20 30 40 50 25 81 0 60 2047 ILOAD (µA) 2048 2049 CODE Figure 16. ADC Noise, Midcode Input, 10,000 Reads Figure 13. VPDO1 (Weak Pull-Up to VP) vs. ILOAD Rev. 0 | Page 10 of 32 04609-064 2000 04609-058 VPDO1 (V) 3.0 ADM1066 1.005 1.004 1.003 DAC 20kΩ BUFFER OUTPUT 47pF PROBE POINT DAC OUTPUT 1.002 1.001 VP1 = 3.0V 1.000 VP1 = 4.75V 0.999 0.998 04609-065 0.997 0.996 04609-059 1 CH1 200mV M1.00µs CH1 0.995 –40 –20 0 20 40 60 80 100 80 100 TEMPERATURE (°C) 756mV Figure 17. Transient Response of DAC Code Change into Typical Load Figure 19. DAC Output vs. Temperature 2.058 1V PROBE POINT CH1 200mV M1.00µs CH1 VP1 = 4.75V 2.043 04609-060 1 VP1 = 3.0V 2.048 944mV 2.038 –40 04609-061 DAC 100kΩ BUFFER OUTPUT REFOUT (V) 2.053 –20 0 20 40 60 TEMPERATURE (°C) Figure 18. Transient Response of DAC to Turn-On from HI-Z State Figure 20. REFOUT vs. Temperature Rev. 0 | Page 11 of 32 ADM1066 POWERING THE ADM1066 VDDCAP VP1 IN OUT 4.75V LDO EN VP2 IN OUT 4.75V LDO EN VP3 IN OUT 4.75V LDO EN VP4 IN OUT 4.75V LDO EN VH An external capacitor to GND is required to decouple the onchip supply from noise. This capacitor should be connected to the VDDCAP pin, as shown in Figure 21. The capacitor has another use during brownouts (momentary loss of power). Under these conditions, when the input supply (VPn or VH) dips transiently below VDD, the synchronous rectifier switch immediately turns off so that it does not pull VDD down. The VDD cap can then act as a reservoir to keep the device active until the next highest supply takes over the powering of the device. 10 µF is recommended for this reservoir/decoupling function. Note that when two or more supplies are within 100 mV of each other, the supply that takes control of VDD first keeps control. For example, if VP1 is connected to a 3.3 V supply, then VDD powers up to approximately 3.1 V through VP1. If VP2 is then connected to another 3.3 V supply, VP1 still powers the device, unless VP2 goes 100 mV higher than VP1. Rev. 0 | Page 12 of 32 IN OUT 4.75V LDO INTERNAL DEVICE SUPPLY EN SUPPLY COMPARATOR 04609-022 The ADM1066 is powered from the highest voltage input on either the positive-only supply inputs (VPn) or the high voltage supply input (VH). This technique offers improved redundancy as the device is not dependent on any particular voltage rail to keep it operational. The same pins are used for supply fault detection (discussed later in the next section). A VDD arbitrator on the device chooses which supply to use. The arbitrator can be considered an OR’ing of five LDOs together. A supply comparator chooses which of the inputs is highest and selects this one to provide the on-chip supply. There is minimal switching loss with this architecture (~0.2 V), resulting in the ability to power the ADM1066 from a supply as low as 3.0 V. Note that the supply on the VXn pins cannot be used to power the device. Figure 21. VDD Arbitrator Operation ADM1066 INPUTS SUPPLY SUPERVISION The resolution is given by The ADM1066 has ten programmable inputs. Five of these are dedicated supply fault detectors (SFDs). These dedicated inputs are called VH and VP1–4 by default. The other five inputs are labeled VX1–VX5 and have dual functionality. They can be used as either supply fault detectors, with similar functionality to VH and VP1–4, or CMOS/TTL-compatible logic inputs to the devices. Therefore, the ADM1066 can have up to ten analog inputs, a minimum of five analog inputs and five digital inputs, or a combination. If an input is used as an analog input, it cannot be used as a digital input. Therefore, a configuration requiring ten analog inputs has no digital inputs available. Table 5 shows the details of each of the inputs. RANGE SELECT ULTRA LOW + VPn VREF OV COMPARATOR – GLITCH FILTER FAULT OUTPUT Step Size = Threshold Range/255 Therefore, if the high range is selected on VH, the step size can be calculated as follows: (14.4 V − 4.8 V)/255 = 37.6 mV Table 4 lists the upper and lower limit of each available range, the bottom of each range (VB), and the range itself (VR). Table 4. Voltage Range Limits Voltage Range (V) 0.573 to 1.375 1.25 to 3.00 2.5 to 6.0 4.8 to 14.4 VB (V) 0.573 1.25 2.5 4.8 VR (V) 0.802 1.75 3.5 9.6 The threshold value required is given by VT = (VR × N)/255 + VB + LOW – where: FAULT TYPE UV SELECT COMPARATOR 04609-023 MID Figure 22. Supply Fault Detector Block PROGRAMMING THE SUPPLY FAULT DETECTORS The ADM1066 has up to ten supply fault detectors (SFDs) on its ten input channels. These highly programmable reset generators enable the supervision of up to ten supply voltages. The supplies can be as low as 0.573 V and as high as 14.4 V. The inputs can be configured to detect an undervoltage fault (the input voltage droops below a preprogrammed value), an overvoltage fault (the input voltage rises above a preprogrammed value) or an out-ofwindow fault (undervoltage or overvoltage). The thresholds can be programmed to an 8-bit resolution in registers provided in the ADM1066. This translates to a voltage resolution that is dependent on the range selected. VT is the desired threshold voltage (UV or OV). VR is the voltage range. N is the decimal value of the 8-bit code. VB is the bottom of the range. Reversing the equation, the code for a desired threshold is given by N = 255 × (VT − VB)/VR For example, if the user wants to set a 5 V OV threshold on VP1, the code to be programmed in the PS1OVTH register (discussed in the AN-698 application note) is given by N = 255 × (5 − 2.5)/3.5 Therefore, N = 182 (1011 0110 or 0xB6). Table 5. Input Functions, Thresholds, and Ranges Input VH Function High V Analog Input VPn Positive Analog Input VXn High Z Analog Input Digital Input Voltage Range (V) 2.5–6.0 4.8–14.4 0.573–1.375 1.25–3.00 2.5–6.0 0.573–1.375 0–5 Maximum Hysteresis 425 mV 1.16 V 97.5 mV 212 mV 425 mV 97.5 mV N/A Rev. 0 | Page 13 of 32 Voltage Resolution (mV) 13.7 37.6 3.14 6.8 13.7 3.14 N/A Glitch Filter (µs) 0–100 0–100 0–100 0–100 0–100 0–100 0–100 ADM1066 The UV and OV comparators shown in Figure 22 are always looking at VPn. To avoid chattering (multiple transitions when the input is very close to the set threshold level), these comparators have digitally programmable hysteresis. The hysteresis can be programmed up to the values shown in Table 5. The hysteresis is added after a supply voltage goes out of tolerance. Therefore, the user can program how much above the UV threshold the input must rise again before a UV fault is deasserted. Similarly, the user can program how much below the OV threshold an input must fall again before an OV fault is deasserted. The hysteresis figure is given by INPUT PULSE SHORTER THAN GLITCH FILTER TIMEOUT INPUT PULSE LONGER THAN GLITCH FILTER TIMEOUT PROGRAMMED TIMEOUT PROGRAMMED TIMEOUT INPUT T0 INPUT TGF T0 TGF OUTPUT T0 VHYST = VR × NTHRESH/255 TGF OUTPUT T0 TGF 04609-024 INPUT COMPARATOR HYSTERESIS Figure 23. Input Glitch Filter Function SUPPLY SUPERVISION WITH VXn INPUTS where: VHYST is the desired hysteresis voltage. NTHRESH is the decimal value of the 5-bit hysteresis code. Note that NTHRESH has a maximum value of 31. The maximum hysteresis for the ranges is listed in Table 5. INPUT GLITCH FILTERING The final stage of the SFDs is a glitch filter. This block provides time-domain filtering on the output of the SFD comparators. This allows the user to remove any spurious transitions such as supply bounce at turn-on. The glitch filter function is additional to the digitally programmable hysteresis of the SFD comparators. The glitch filter timeout is programmable up to 100 µs. For example, when the glitch filter timeout is 100 µs, any pulses appearing on the input of the glitch filter block that are less than 100 µs in duration are prevented from appearing on the output of the glitch filter block. Any input pulse that is longer than 100 µs does appear on the output of the glitch filter block. The output is delayed with respect to the input by 100 µs. The filtering process is shown in Figure 23. The VXn inputs have two functions. They can be used as either supply fault detectors or digital logic inputs. When selected as an analog (SFD) input, the VXn pins have very similar functionality to the VH and VPn pins. The major difference is that the VXn pins have only one input range: 0.573 V to 1.375 V. Therefore, these inputs can directly supervise only the very low supplies. However, the input impedance of the VXn pins is high, allowing an external resistor divide network to be connected to the pin. Thus, any supply can be potentially divided down into the input range of the VXn pin and supervised. This enables the ADM1066 to monitor other supplies such as +24 V, +48 V, and −5 V. An additional supply supervision function is available when the VXn pins are selected as digital inputs. In this case, the analog function is available as a second detector on each of the dedicated analog inputs, VP1–4 and VH. The analog function of VX1 is mapped to VP1, VX2 is mapped to VP2, and so on. VX5 is mapped to VH. In this case, these SFDs can be viewed as a secondary or warning SFD. The secondary SFDs are fixed to the same input range as the primary SFD. They are used to indicate warning levels rather than failure levels. This allows faults and warnings to be generated on a single supply using only one pin. For example, if VP1 is set to output a fault if a 3.3 V supply droops to 3.0 V, VX1 can be set to output a warning at 3.1 V. Warning outputs are available for readback from the status registers. They are also OR’ed together and fed into the sequencing engine (SE), allowing warnings to generate interrupts on the PDOs. Therefore, in the example above, if the supply droops to 3.1 V, a warning is generated, and remedial action can be taken before the supply drops out of tolerance. Rev. 0 | Page 14 of 32 ADM1066 As mentioned previously, the VXn input pins on the ADM1066 have dual functionality. The second function is as a digital input to the device. Therefore, the ADM1066 can be configured for up to five digital inputs. These inputs are TTL/CMOS-compatible. Standard logic signals can be applied to the pins: RESET from reset generators, PWRGOOD signals, fault flags, manual resets, and so on. These signals are available as inputs to the SE, and so can be used to control the status of the PDOs. The inputs can be configured to detect either a change in level or an edge. The digital blocks feature the same glitch filter function that is available on the SFDs. This enables the user to ignore spurious transitions on the inputs. For example, the filter can be used to debounce a manual reset switch. When configured as digital inputs, each of the VXn pins has a weak (10 µA) pull-down current source available for placing the input in a known condition, even if left floating. The current source, if selected, weakly pulls the input to GND. VXn (DIGITAL INPUT) When configured for level detection, the output of the digital block is a buffered version of the input. When configured for edge detection, once the logic transition is detected, a pulse of programmable width is output from the digital block. The width is programmable from 0 µs to 100 µs. + DETECTOR GLITCH FILTER – VREF = 1.4V Figure 24. VXn Digital Input Function Rev. 0 | Page 15 of 32 TO SEQUENCING ENGINE 04609-027 VXn PINS AS DIGITAL INPUTS ADM1066 OUTPUTS from a backplane supply (a PDO can sustain greater than 10.5 V into a 1 µA load). The pull-down switches can also be used to drive status LEDs directly. SUPPLY SEQUENCING THROUGH CONFIGURABLE OUTPUT DRIVERS Supply sequencing is achieved with the ADM1066 using the programmable driver outputs (PDOs) on the device as control signals for supplies. The output drivers can be used as logic enables or as FET drivers. The data driving each of the PDOs can come from one of three sources. The source can be enabled in the PnPDOCFG configuration register (see the AN-698 application note for details). The sequence in which the PDOs are asserted (and, therefore, the supplies are turned on) is controlled by the sequencing engine (SE). The SE determines what action is to be taken with the PDOs based on the condition of the inputs of the ADM1066. Therefore, the PDOs can be set up to assert when the SFDs are in tolerance, the correct input signals are received on the VXn digital pins, no warnings are received from any of the inputs of the device, and so on. The PDOs can be used for a variety of functions. The primary function is to provide enable signals for LDOs or dc/dc converters, which generate supplies locally on a board. The PDOs can also be used to provide a POWER_GOOD signal when all the SFDs are in tolerance, or a RESET output if one of the SFDs goes out of specification (this can be used as a status signal for a DSP, FPGA, or other microcontroller). The data sources are • Output from the SE. • Directly from the SMBus. A PDO can be configured so that the SMBus has direct control over it. This enables software control of the PDOs. Therefore, a microcontroller can be used to initiate a software power-up/power-down sequence. • On-Chip Clock. A 100 kHz clock is generated on the device. This clock can be made available on any of the PDOs. It can be used, for example, to clock an external device such as an LED. By default, the PDOs are pulled to GND by a weak (20 kΩ) onchip pull-down resistor. This is also the condition of the PDOs on power-up, until the configuration is downloaded from EEPROM and the programmed setup is latched. The outputs are actively pulled low once a supply of 1 V or greater is on VPn or VH. The outputs remain high impedance prior to 1 V appearing on VPn or VH. This provides a known condition for the PDOs during power-up. The internal pull-down can be overdriven with an external pull-up of suitable value tied from the PDO pin to the required pull-up voltage. The 20 kΩ resistor must be accounted for in calculating a suitable value. For example, if PDOn must be pulled up to 3.3 V, and 5 V is available as an external supply, the pull-up resistor value is given by The PDOs can be programmed to pull up to a number of different options. The outputs can be programmed as follows: • Open-drain (allowing the user to connect an external pull-up resistor) • Open-drain with weak pull-up to VDD • Push/pull to VDD • Open-drain with weak pull-up to VPn • Push/pull to VPn • Strong pull-down to GND 3.3 V = 5 V × 20 kΩ/(RUP + 20 kΩ) • Internally charge-pumped high drive (12 V, PDO1–6 only) Therefore, The last option (available only on PDO1–6) allows the user to directly drive a voltage high enough to fully enhance an external N-FET, which is used to isolate, for example, a card-side voltage RUP = (100 kΩ − 66 kΩ)/3.3 = 10 kΩ VFET (PDO1-6 ONLY) VDD VP4 10Ω 20kΩ 10Ω 10Ω 20kΩ VP1 SEL 20kΩ CFG4 CFG5 CFG6 SE DATA PDO SMBus DATA Figure 25. Programmable Driver Output Rev. 0 | Page 16 of 32 04609-028 20kΩ CLK DATA ADM1066 SEQUENCING ENGINE The ADM1066’s sequencing engine (SE) provides the user with powerful and flexible control of sequencing. The SE implements a state machine control of the PDO outputs, with state changes conditional on input events. SE programs can enable complex control of boards such as power-up and power-down sequence control, fault event handling, interrupt generation on warnings, and so on. A watchdog function that verifies the continued operation of a processor clock can be integrated into the SE program. The SE can also be controlled via the SMBus, giving software or firmware control of the board sequencing. • Transition from one state to the next is made in less than 20 µs, which is the time needed to download a state definition from EEPROM to the SE. MONITOR FAULT STATE TIMEOUT SEQUENCE 04609-029 OVERVIEW Figure 26. State Cell The SE state machine comprises 63 state cells. Each state has the following attributes: • Monitors signals indicating the status of the 10 input pins, VP1 to VP4, VH, and VX1 to VX5. • Can be entered from any other state. • Three exit routes move the state machine on to a next state: sequence detection, fault monitoring, and timeout. • Delay timers for the sequence and timeout blocks can be programmed independently, and change with each state change. The range of timeouts is from 0 ms to 400 ms. The ADM1066 offers up to 63 state definitions. The signals monitored to indicate the status of the input pins are the outputs of the SFDs. WARNINGS The SE also monitors warnings. These warnings can be generated when the ADC readings violate their limit register value or when the secondary voltage monitors on VP1–4 and VH. The warnings are all OR’ed together and are available as a single warning input to each of the three blocks that enable exiting from a state. SMBus JUMP/UNCONDITIONAL JUMP The SE can be forced to advance to the next state unconditionally. This enables the user to force the SE to advance. Examples of where this might be used include moving to a margining state or debugging a sequence. The SMBus jump or go-to command can be seen as another input to sequence and timeout blocks, which provide an exit from each state. • Output condition of the 10 PDO pins is defined and fixed within a state. Table 6. Sample Sequence State Entries State IDLE1 IDLE2 EN3V3 Sequence If VX1 is low , go to state IDLE2. If VP1 is okay, go to state EN3V3. If VP2 is okay, go to state EN2V5. DIS3V3 EN2V5 If VX1 is high, go to state IDLE1. If VP3 is okay, go to state PWRGD. DIS2V5 FSEL1 FSEL2 PWRGD If VX1 is high, go to state IDLE1. If VP3 is not okay, go to state DIS2V5. If VP2 is not okay, go to state DIS3V3. If VX1 is high, go to state DIS2V5. Timeout Monitor If VP2 is not okay after 10 ms, go to state DIS3V3. If VP1 is not okay, go to state IDLE1. If VP3 is not okay after 20 ms, go to state DIS2V5. If VP1 or VP2 is not okay, go to state FSEL2. If VP1 or VP2 is not okay, go to state FSEL2. If VP1 is not okay, go to state IDLE1. If VP1, VP2, or VP3 is not okay, go to state FSEL1. Rev. 0 | Page 17 of 32 ADM1066 SEQUENCE STATES SEQUENCING ENGINE APPLICATION EXAMPLE The application in this section demonstrates the operation of the SE. Figure 27 shows how the simple building block of a single SE state can be used to build a power-up sequence for a 3-supply system. IDLE1 VX1 = 0 Table 7 lists the PDO outputs for each state in the same SE implementation. In this system, the presence of a good 5 V supply on VP1 and the VX1 pin held low are the triggers required for a power-up sequence to start. The sequence intends to turn on the 3.3 V supply next, then the 2.5 V supply (assuming successful turn-on of the 3.3 V supply). Once all three supplies are good, the PWRGD state is entered, where the SE remains until a fault occurs on one of the three supplies, or it is instructed to go through a power-down sequence by VX1 going high. IDLE2 VP1 = 1 MONITOR FAULT STATES TIMEOUT STATES EN3V3 10ms VP1 = 0 VP2 = 1 EN2V5 Faults are dealt with throughout the power-up sequence on a case-by-case basis. The following sections, which describe the individual blocks, use this sample application to demonstrate the state machine’s actions. DIS3V3 20ms (VP1 + VP2) = 0 VX1 = 1 VP3 = 1 PWRGD DIS2V5 VP2 = 0 (VP1 + VP2 + VP3) = 0 VX1 = 1 FSEL1 VX1 = 1 (VP1 + VP2) = 0 VP3 = 0 FSEL2 VP1 = 0 04609-030 VP2 = 0 Figure 27. Sample Application Flow Diagram Table 7. PDO Outputs for Each State PDO Outputs PDO1 = 3V3ON PDO2 = 2V5ON PDO3 = FAULT IDLE1 0 0 0 IDLE2 0 0 0 EN3V3 1 0 0 EN2V5 1 1 0 DIS3V3 0 1 1 Rev. 0 | Page 18 of 32 DIS2V5 1 0 1 PWRGD 1 1 0 FSEL1 1 1 1 FSEL2 1 1 1 ADM1066 MONITORING FAULT DETECTOR SEQUENCE DETECTOR The sequence detector block is used to detect when a step in a sequence has been completed. It looks for one of the inputs to the SE to change state, and is most often used as the gate on successful progress through a power-up or power-down sequence. A timer block is included in this detector, which can insert delays into a power-up or power-down sequence, if required. Timer delays can be set from 10 µs to 400 ms. Figure 28 is a block diagram of the sequence detector. 1-BIT FAULT DETECTOR VP1 MASK SENSE 1-BIT FAULT DETECTOR VX5 VP1 VX5 SUPPLY FAULT DETECTION FAULT SUPPLY FAULT DETECTION FAULT LOGIC INPUT CHANGE OR FAULT DETECTION MASK SENSE SEQUENCE DETECTOR 1-BIT FAULT DETECTOR LOGIC INPUT CHANGE OR FAULT DETECTION TIMER FAULT WARNINGS INVERT MASK SELECT 04609-032 FORCE FLOW (UNCONDITIONAL JUMP) Figure 28. Sequence Detector Block Diagram The sequence detector can also help to identify monitoring faults. In the sample application shown in Figure 27, the FSEL1 and FSEL2 states first identify which of the VP1,VP2, or VP3 pins has faulted, and then they take the appropriate action. MONITORING FAULT DETECTOR The monitoring fault detector block is used to detect a failure on an input. The logical function implementing this is a wide OR gate, which can detect when an input deviates from its expected condition. The clearest demonstration of the use of this block is in the PWRGD state, where the monitor block indicates that a failure on one or more of the VP1,VP2, or VP3 inputs has occurred. No programmable delay is available in this block, because the triggering of a fault condition is likely to be caused when a supply falls out of tolerance. In this situation, the user would want to react as quickly as possible. Some latency occurs when moving out of this state, however, because it takes a finite amount of time (~20 µs) for the state configuration to download from EEPROM into the SE. Figure 29 is a block diagram of the monitoring fault detector. 04609-033 WARNINGS Figure 29. Monitoring Fault Detector Block Diagram TIMEOUT DETECTOR The timeout detector allows the user to trap a failure to make proper progress through a power-up or power-down sequence. In the sample application shown in Figure 27, the timeout nextstate transition is from the EN3V3 and EN2V5 states. For the EN3V3 state, the signal 3V3ON is asserted upon entry to this state (on the PDO1 output pin) to turn on a 3.3 V supply. This supply rail is connected to the VP2 pin, and the sequence detector looks for the VP2 pin to go above its UV threshold, which is set in the supply fault detector (SFD) attached to that pin. The power-up sequence progresses when this change is detected. If, however, the supply fails (perhaps due to a short circuit overloading this supply), then the timeout block traps the problem. In this example, if the 3.3 V supply fails within 10 ms, then the SE moves to the DIS3V3 state and turns off this supply by bringing PDO1 low. It also indicates that a fault has occurred by taking PDO3 high. Timeout delays of from 100 µs to 400 ms can be programmed. FAULT REPORTING The ADM1066 has a fault latch for recording faults. Two registers are set aside for this purpose. A single bit is assigned to each input of the device, and a fault on that input sets the relevant bit. The contents of the fault register can be read out over the SMBus to determine which input(s) faulted. The fault register can be enabled/disabled in each state. This ensures that only real faults are captured and not, for example, undervoltage trips when the SE is executing a power-down sequence. Rev. 0 | Page 19 of 32 ADM1066 VOLTAGE READBACK The ADM1066 has an on-board 12-bit accurate ADC for voltage readback over the SMBus. The ADC has a 12-channel analog mux on the front end. The twelve channels consist of the ten SFD inputs (VH, VP1-4, VX1-5) and two auxiliary (singleended) ADC inputs (AXU1, AUX2). Any or all of these inputs can be selected to be read, in turn, by the ADC. The circuit controlling this operation is called round-robin. The roundrobin circuit can be selected to run through its loop of conversions just once or continuously. Averaging is also provided for each channel. In this case, the round-robin circuit runs through its loop of conversions 16 times before returning a result for each channel. At the end of this cycle, the results are all written to the output registers. The ADC samples single-sided inputs with respect to the AGND pin. A 0 V input gives out Code 0, and an input equal to the voltage on REFIN gives out full code (4095 decimal). The inputs to the ADC come directly from the VXn pins and from the back of the input attenuators on the VPn and VH pins, as shown in Figure 30 and Figure 31. DIGITIZED VOLTAGE READING NO ATTENUATION 12-BIT ADC 04609-025 VXn 2.048V VREF ATTENUATION NETWORK (DEPENDS ON RANGE SELECTED) DIGITIZED VOLTAGE READING 04609-026 12-BIT ADC 2.048V VREF Figure 31. ADC Reading on VPn/VH Pins The voltage at the input pin can be derived from the following equation: V= ADC Code 4095 Attenuation Factor 1 2.181 4.363 10.472 ADC Input Voltage Range (V) 0–2.048 0–4.46 0–6.01 0–14.41 _______________________________________________ 1 The upper limit is the absolute maximum allowed voltage on these pins. The normal way to supply the reference to the ADC on the REFIN pin is to simply connect the REFOUT pin to the REFIN pin. REFOUT provides a 2.048 V reference. As such, the supervising range covers less than half of the normal ADC range. It is possible, however, to provide the ADC with a more accurate external reference for improved readback accuracy. Supplies can also be connected to the input pins purely for ADC readback, even though they might go above the expected supervisory range limits (but not above 6 V, because this violates the absolute maximum ratings on these pins). For instance, a 1.5 V supply connected to the VX1 pin can be correctly read out as an ADC code of approximately 3/4 full scale, but it always sits above any supervisory limits that can be set on that pin. The maximum setting for the REFIN pin is 2.048 V. SUPPLY SUPERVISION WITH THE ADC Figure 30. ADC Reading on VXn Pins VPn/VH Table 8. ADC Input Voltage Ranges SFD Input Range (V) 0.573–1.375 1.25–3 2.5–6 4.8–14.4 × Attenuation Factor × 2.048 V The ADC input voltage ranges for the SFD input ranges are listed in Table 8. In addition to the readback capability, a further level of supervision is provided by the on-chip 12-bit ADC. The ADM1066 has limit registers on which the user can program to a maximum or minimum allowable threshold. Exceeding the threshold generates a warning that can either be read back from the status registers or input into the SE to determine what sequencing action the ADM1066 should take. Only one register is provided for each input channel, so a UV or OV threshold (but not both) can be set for a given channel. The round-robin circuit can be enabled via an SMBus write, or it can be programmed to turn on in any state in the SE program. For example, it can be set to start once a power-up sequence is complete and all supplies are known to be within expected tolerance limits. Note that a latency is built into this supervision, dictated by the conversion time of the ADC. With all 12 channels selected, the total time for the round-robin operation (averaging off) is approximately 6 ms (500 µs per channel selected). Supervision using the ADC, therefore, does not provide the same real time response as the SFDs. Rev. 0 | Page 20 of 32 ADM1066 SUPPLY MARGINING OVERVIEW CLOSED-LOOP SUPPLY MARGINING It is often necessary for the system designer to adjust supplies, either to optimize their level or force them away from nominal values to characterize the system performance under these conditions. This is a function typically performed during an incircuit test (ICT), such as when the manufacturer, for example, wants to guarantee that the product under test functions correctly at nominal supplies minus 10%. A much more accurate and comprehensive method of margining is to implement a closed-loop system. The voltage on the rail to be margined can be read back so that the rail can be accurately margined to the target voltage. The ADM1066 incorporates all the circuits required to do this, with the 12-bit successive approximation ADC used to read back the level of the supervised voltages, and the six voltage output DACs, implemented as described in the Open-Loop Margining section, used to adjust supply levels. These circuits can be used along with some other intelligence such as a microcontroller to implement a closed-loop margining system that allows any dc/dc or LDO supply to be set to any voltage, accurate to within ±0.5% of the target. OPEN-LOOP MARGINING The simplest method of margining a supply is to implement an open-loop technique. A popular method for this is to switch extra resistors into the feedback node of a power module, such as a dc/dc converter or low dropout regulator (LDO). The extra resistor alters the voltage at the feedback or trim node and forces the output voltage to margin up or down by a certain amount. µCONTROLLER VIN ADM1066 VIN µCONTROLLER VOUT DEVICE CONTROLLER (SMBus) FEEDBACK GND ATTENUATION RESISTOR DACOUTn Figure 32. Open-Loop Margining System Using the ADM1066 The ADM1066 can be commanded to margin a supply up or down over the SMBus by updating the values on the relevant DAC output. ADC R1 DACOUTn FEEDBACK R2 GND DAC DEVICE CONTROLLER (SMBus) PCB TRACE NOISE DECOUPLING CAPACITOR Figure 33. Closed-Loop Margining System Using the ADM1066 To implement closed-loop margining: 1. Disable the six DACn outputs. 2. Set the DAC output voltage equal to the voltage on the feedback node. 3. Enable the DAC. 4. Read the voltage at the dc/dc output, which is connected to one of the VP1–4, VH, or VX1–5 pins. 5. If necessary, modify the DACn output code up or down to adjust the dc/dc output voltage; otherwise, stop, because the target voltage has been reached. 6. Set the DAC output voltage to a value that alters the supply output by the required amount (for example, ±5%). 7. Repeat from Step 4. DAC PCB TRACE NOISE DECOUPLING CAPACITOR 04609-067 DC/DC CONVERTER MUX ATTENUATION RESISTOR, R3 OUTPUT ADM1066 OUTPUT VH/VPn/VXn DC/DC CONVERTER 04609-034 The ADM1066 can perform open-loop margining for up to six supplies. The six on-board voltage DACs (DAC1–6) can drive into the feedback pins of the power modules to be margined. The simplest circuit to implement this function is an attenuation resistor, which connects the DACn pin to the feedback node of a dc/dc converter. When the DACn output voltage is set equal to the feedback voltage, no current flows in the attenuation resistor, and the dc/dc output voltage does not change. Taking DACn above the feedback voltage forces current into the feedback node, and the output of the dc/dc converter is forced to fall to compensate for this. The dc/dc output can be forced high by setting the DACn output voltage lower than the feedback node voltage. The series resistor can be split in two, and the node between them decoupled with a capacitor to ground. This can help to decouple any noise picked up from the board. Decoupling to a ground local to the dc/dc converter is recommended. Steps 1 to 3 ensure that when the DACn output buffer is turned on it has little effect on the dc/dc output. The DAC output buffer is designed to power up without glitching by first powering up the buffer to follow the pin voltage. It does not drive out onto the pin at this time. Once the output buffer is properly enabled, the buffer input is switched over to the DAC, and the output stage of the buffer is turned on. Output glitching is negligible. Rev. 0 | Page 21 of 32 ADM1066 WRITING TO THE DACs Four DAC ranges are offered. They can be placed with midcode (Code 0x7F) at 0.6 V, 0.8 V, 1.0 V, and 1.25 V. These voltages are placed to correspond to the most common feedback voltages. Centering the DAC outputs in this way provides the best use of the DAC resolution. For most supplies it is possible to place the DAC midcode at the point where the dc/dc output is not modified, thereby giving half of the DAC range to margin up and the other half to margin down. The DAC output voltage is set by the code written to the DACn register. The voltage is linear with the unsigned binary number in this register. Code 0x7F is placed at the midcode voltage, as described previously. The output voltage is given by the following equation: DAC Output = (DACn − 0x7F)/255 × 0.6015 + VOFF where VOFF is one of the four offset voltages. There are 256 DAC settings available. The midcode value is located at DAC code 0x7F as close as possible to the middle of the 256 code range. The full output swing of the DACs is +302 mV (+128 codes) and −300 mV (−127 codes) around the selected midcode voltage. The voltage range for each midcode voltage is shown in Table 9. Table 9. Ranges for Midcode Voltages Midcode Voltage (V) 0.6 0.8 1.0 1.25 Minimum Voltage Output (V) 0.300 0.500 0.700 0.950 Maximum Voltage Output (V) 0.902 1.102 1.302 1.552 CHOOSING THE SIZE OF THE ATTENUATION RESISTOR How much this DAC voltage swing affects the output voltage of the dc/dc converter that is being margined is determined by the size of the attenuation resistor, R3 (see Figure 33). Because the voltage at the feedback pin remains constant, the current flowing from the feedback node to GND via R2 is a constant. Also, the feedback node itself is high impedance. This means that the current flowing through R1 is the same as the current flowing through R3. Therefore, direct relationship exists between the extra voltage drop across R1 during margining and the voltage drop across R3. This relationship is given by the equation ∂VOUT = R1 (VFB − VDACOUT) R3 where: ∂VOUT is the change in VOUT. VFB is the voltage at the feedback node of the dc/dc converter. VDACOUT is the voltage output of the margining DAC. This equation demonstrates that, if the user wants the output voltage to change by ±300 mV, then R1 = R3. If the user wants the output voltage to change by ±600 mV, then R1 = 2 × R3, and so on. It is best to use the full DAC output range to margin a supply. Choosing the attenuation resistor in this way provides the most resolution from the DAC. In other words, with one DAC code change, the smallest effect on the dc/dc output voltage is induced. If the resistor is sized up to use a code such as 27(dec) to 227(dec) to move the dc/dc output by ±5%, then it takes 100 codes to move 5% (each code moves the output by 0.05%). This is beyond the readback accuracy of the ADC, but should not prevent the user from building a circuit to use the most resolution. DAC LIMITING/OTHER SAFETY FEATURES Limit registers (called DPLIMn and DNLIMn) on the device offer the user some protection from firmware bugs, which can cause catastrophic board problems by forcing supplies beyond their allowable output ranges. Essentially, the DAC code written into the DACn register is clipped such that the code used to set the DAC voltage is actually given by DAC Code = DACn, DACn ≥ DNLIMn and DACn ≤ DPLIMn = DNLIMn, DACn < DNLIMn = DPLIMn, DACn > DPLIMn In addition, the DAC output buffer is three-stated, if DNLIMn > DPLIMn. By programming the limit registers in this way, the user can make it very difficult for the DAC output buffers to be turned on at all during normal system operation (these are among the registers downloaded from EEPROM at startup). Rev. 0 | Page 22 of 32 ADM1066 APPLICATIONS DIAGRAM 12V IN 12V OUT 5V IN 5V OUT 3V IN 3V OUT IN DC-DC1 EN VH 5V OUT 3V OUT 3.3V OUT 2.5V OUT 1.8V OUT 1.2V OUT 0.9V OUT OUT 3.3V OUT ADM1066 VP1 VP2 VP3 VP4 VX1 VX2 VX3 PDO1 PDO2 VX4 PDO6 IN DC-DC2 PDO3 PDO4 PDO5 EN OUT 2.5V OUT POWER_GOOD POWER_ON SIGNAL_VALID RESET_L PDO7 IN SYSTEM RESET VX5 DC-DC3 PDO8 EN PDO9 PDO10 OUT 1.8V OUT 3.3V OUT DAC1* IN REFIN VCCP VDDCAP GND LDO 10µF 10µF EN 10µF OUT 0.9V OUT 3.3V OUT *ONLY ONE MARGINING CIRCUIT SHOWN FOR CLARITY. DAC1 TO DAC6 WILL ALLOW MARGINING FOR UP TO SIX VOLTAGE RAILS. IN OUT EN 1.2V OUT TRIM 04608-068 DC-DC4 Figure 34. Applications Diagram Rev. 0 | Page 23 of 32 ADM1066 COMMUNICATING WITH THE ADM1066 CONFIGURATION DOWNLOAD AT POWER-UP The configuration of the ADM1066 (UV/OV thresholds, glitch filter timeouts, PDO configurations, and so on) is dictated by the contents of RAM. The RAM is comprised of digital latches that are local to each of the functions on the device. The latches are double-buffered and have two identical latches, Latch A and Latch B. Therefore, when an update to a function occurs, the contents of Latch A are updated first, and then the contents of Latch B are updated with identical data. The advantages of this architecture are explained in detail in this section. The latches are volatile memory and lose their contents at power-down. Therefore, the configuration in the RAM must be restored at power-up by downloading the contents of the EEPROM (nonvolatile memory) to the local latches. This download occurs in steps, as follows: 1. With no power applied to the device, the PDOs are all high impedance. 2. When 1 V appears on any of the inputs connected to the VDD arbitrator (VH or VPn), the PDOs are all weakly pulled to GND with a 20 kΩ impedance. 3. When the supply rises above the undervoltage lockout of the device (UVLO is 2.5 V), the EEPROM starts to download to the RAM. 4. The EEPROM downloads its contents to all Latch As. 5. Once the contents of the EEPROM are completely downloaded to the Latch As, the device controller signals all Latch As to download to all Latch Bs simultaneously, completing the configuration download. 6. At 0.5 ms after the configuration download completes, the first state definition is downloaded from EEPROM into the SE. Note that any attempt to communicate with the device prior to the completion of the download causes the ADM1066 to issue a no acknowledge (NACK). UPDATING THE CONFIGURATION After power-up, with all the configuration settings loaded from EEPROM into the RAM registers, the user might need to alter the configuration of functions on the ADM1066, such as changing the UV or OV limit of an SFD, changing the fault output of an SFD, or adjusting the rise time delay of one of the PDOs. The ADM1066 provides several options that allow the user to update the configuration over the SMBus interface. The following options are controlled in the UPDCFG register: 1. Update the configuration in real time. The user writes to RAM across the SMBus and the configuration is updated immediately. 2. Update the Latch As without updating the Latch Bs. With this method, the configuration of the ADM1066 remains unchanged and continues to operate in the original setup until the instruction is given to update the Latch Bs. 3. Change EEPROM register contents without changing the RAM contents, and then download the revised EEPROM contents to the RAM registers. Again, with this method, the configuration of the ADM1066 remains unchanged and continues to operate in the original setup until the instruction is given to update the RAM. The instruction to download from the EEPROM in Option 3 is also a useful way to restore the original EEPROM contents, if revisions to the configuration are unsatisfactory. For example, if the user needs to alter an OV threshold, this can be done by updating the RAM register as described in Option 1. However, if the user is not satisfied with the change and wants to revert to the original programmed value, then the device controller can issue a command to download the EEPROM contents to the RAM again, as described in Option 3, restoring the ADM1066 to its original configuration. The topology of the ADM1066 makes this type of operation possible. The local, volatile registers (RAM) are all doublebuffered latches. Setting Bit 0 of the UPDCFG register to 1 leaves the double-buffered latches open at all times. If Bit 0 is set to 0, then, when a RAM write occurs across the SMBus, only the first side of the double-buffered latch is written to. The user must then write a 1 to Bit 1 of the UPDCFG register. This generates a pulse to update all the second latches at once. EEPROM writes occur in a similar way. The final bit in this register can enable or disable EEPROM page erasure. If this bit is set high, the contents of an EEPROM page can all be set to 1. If low, then the contents of a page cannot be erased, even if the command code for page erasure is programmed across the SMBus. The bitmap for the UPDCFG register is shown in the AN-698 application note. A flow chart for download at power-up and subsequent configuration updates is shown in Figure 35. Rev. 0 | Page 24 of 32 ADM1066 SMBus E E P R O M L D DEVICE CONTROLLER R A M L D D A T A U P D LATCH A LATCH B EEPROM FUNCTION (OV THRESHOLD ON VP1) 04609-035 POWER-UP (VCC > 2.5V) Figure 35. Configuration Update Flow Diagram UPDATING THE SEQUENCING ENGINE Sequencing engine (SE) functions are not updated in the same way as regular configuration latches. The SE has its own dedicated 512-byte EEPROM for storing state definitions, providing 63 individual states with a 64-bit word each (one state is reserved). At power-up, the first state is loaded from the SE EEPROM into the engine itself. When the conditions of this state are met, the next state is loaded from EEPROM into the engine, and so on. The loading of each new state takes approximately 10 µs. To alter a state, the required changes must be made directly to EEPROM. RAM for each state does not exist. The relevant alterations must be made to the 64-bit word, which is then uploaded directly to EEPROM. INTERNAL REGISTERS The ADM1066 contains a large number of data registers. The principal registers are the address pointer register and the configuration registers. Address Pointer Register This register contains the address that selects one of the other internal registers. When writing to the ADM1066, the first byte of data is always a register address, which is written to the address pointer register. The major differences between the EEPROM and other registers are • An EEPROM location must be blank before it can be written to. If it contains data, it must first be erased. • Writing to EEPROM is slower than writing to RAM. • Writing to the EEPROM should be restricted, because it has a limited write/cycle life of typically 10,000 write operations due to the usual EEPROM wear-out mechanisms. The first EEPROM is split into 16 (0 to 15) pages of 32 bytes each. Pages 0 to 6, starting at Address 0xF800, hold the configuration data for the applications on the ADM1066 (the SFDs, PDOs, and so on). These EEPROM addresses are the same as the RAM register addresses, prefixed by F8. Page 7 is reserved. Pages 8 to 15 are for customer use. Data can be downloaded from EEPROM to RAM in one of the following ways: • At power-up, when Pages 0 to 6 are downloaded. • By setting Bit 0 of the UDOWNLD register (0xD8), which performs a user download of Pages 0 to 6. Configuration Registers SERIAL BUS INTERFACE These registers provide control and configuration for various operating parameters of the ADM1066. The ADM1066 is controlled via the serial system management bus (SMBus). The ADM1066 is connected to this bus as a slave device, under the control of a master device. It takes approximately 1 ms after power-up for the ADM1066 to download from its EEPROM. Therefore, access to the ADM1066 is restricted until the download is completed. EEPROM The ADM1066 has two 512-byte cells of nonvolatile, electrically erasable, programmable read-only memory (EEPROM), from Register Addresses 0xF800 to 0xFBFF. The EEPROM is used for permanent storage of data that is not lost when the ADM1066 is powered down. One EEPROM cell contains the configuration data of the device; the other contains the state definitions for the SE. Although referred to as read-only memory, the EEPROM can be written to as well as read from via the serial bus in exactly the same way as the other registers. Identifying the ADM1066 on the SMBus The ADM1060 has a 7-bit serial bus slave address. The device is powered up with a default serial bus address. The five MSBs of the address are set to 01101; the two LSBs are determined by the logical states of Pins A1 and A0. This allows the connection of four ADM1066s to one SMBus. Rev. 0 | Page 25 of 32 ADM1066 The device also has several identification registers (read-only), which can be read across the SMBus. Table 10 lists these registers with their values and functions. All other devices on the bus remain idle while the selected device waits for data to be read from or written to it. If the R/W bit is a 0, the master writes to the slave device. If the R/W bit is a 1, the master reads from the slave device. Table 10. Identification Register Values and Functions Name MANID Address 0xF4 Value 0x41 REVID MARK1 MARK2 0xF5 0xF6 0xF7 0x00 0x00 0x00 Function Manufacturer ID for Analog Devices Silicon revision S/w brand S/w brand 2. Data is sent over the serial bus in sequences of nine clock pulses, eight bits of data followed by an acknowledge bit from the slave device. Data transitions on the data line must occur during the low period of the clock signal and remain stable during the high period, because a low-tohigh transition when the clock is high might be interpreted as a stop signal. If the operation is a write operation, the first data byte after the slave address is a command byte. This tells the slave device what to expect next. It might be an instruction telling the slave device to expect a block write, or it might simply be a register address that tells the slave where subsequent data is to be written. Because data can flow in only one direction, as defined by the R/W bit, sending a command to a slave device during a read operation is not possible. Before a read operation, it might be necessary to perform a write operation to tell the slave what sort of read operation to expect and/or the address from which data is to be read. 3. When all data bytes have been read or written, stop conditions are established. In write mode, the master pulls the data line high during the 10th clock pulse to assert a stop condition. In read mode, the master device releases the SDA line during the low period before the ninth clock pulse, but the slave device does not pull it low. This is known as no acknowledge. The master then takes the data line low during the low period before the tenth clock pulse, then high during the tenth clock pulse to assert a stop condition. General SMBus Timing Figure 36, Figure 37, and Figure 38 are timing diagrams for general read and write operations using the SMBus. The SMBus specification defines specific conditions for different types of read and write operations, which are discussed in the Write Operations and Read Operations sections. The general SMBus protocol operates as follows: The master initiates data transfer by establishing a start condition, defined as a high-to-low transition on the serial data-line SDA, while the serial clock-line SCL remains high. This indicates that a data stream follows. All slave peripherals connected to the serial bus respond to the start condition and shift in the next 8 bits, consisting of a 7-bit slave address (MSB first) plus a R/W bit. This bit determines the direction of the data transfer, that is, whether data is written to or read from the slave device (0 = write, 1 = read). The peripheral whose address corresponds to the transmitted address responds by pulling the data line low during the low period before the ninth clock pulse, known as the acknowledge bit, and holding it low during the high period of this clock pulse. 1 9 1 9 SCL 0 SDA 1 1 0 1 A1 A0 D7 R/W D6 D5 D4 D3 D2 D1 ACK. BY SLAVE START BY MASTER D0 ACK. BY SLAVE FRAME 1 SLAVE ADDRESS FRAME 2 COMMAND CODE 1 9 1 9 SCL (CONTINUED) SDA (CONTINUED) D7 D6 D5 D4 D3 D2 FRAME 3 DATA BYTE D1 D0 D7 ACK. BY SLAVE D6 D5 D4 D2 FRAME N DATA BYTE Figure 36. General SMBus Write Timing Diagram Rev. 0 | Page 26 of 32 D3 D1 D0 ACK. BY SLAVE STOP BY MASTER 04609-036 1. ADM1066 1 9 1 9 SCL 0 SDA 1 1 0 1 A1 A0 R/W D7 D6 D5 D4 D3 D2 D1 ACK. BY SLAVE START BY MASTER 1 D0 ACK. BY MASTER FRAME 1 SLAVE ADDRESS 9 1 FRAME 2 DATA BYTE 9 SCL (CONTINUED) D7 D6 D5 D4 D3 D2 FRAME 3 DATA BYTE D1 D0 D7 D6 D5 D4 ACK. BY MASTER D3 D2 D1 D0 NO ACK. FRAME N DATA BYTE STOP BY MASTER 04609-037 SDA (CONTINUED) Figure 37. General SMBus Read Timing Diagram tR tF t HD; STA t LO W SCL t HI G H t HD; STA t HD; DAT t SU; STA t SU; STO t SU; DAT t BUF P S S P 04609-038 SDA Figure 38. Serial Bus Timing Diagram SMBus Protocols for RAM and EEPROM Data can be written to and read from both RAM and EEPROM as single data bytes. Data can be written only to unprogrammed EEPROM locations. To write new data to a programmed location, it must first be erased. EEPROM erasure cannot be done at the byte level. The EEPROM is arranged as 32 pages of 32 bytes each, and an entire page must be erased. Page erasure is enabled by setting Bit 2 in the UPDCFG register (Address 0x90) to 1. If this bit is not set, page erasure cannot occur, even if the command byte (0xFE) is programmed across the SMBus. WRITE OPERATIONS The SMBus specification defines several protocols for different types of read and write operations. The following abbreviations are used in the diagrams: S P R W A Start Stop Read Write Acknowledge A No acknowledge The ADM1066 uses the following SMBus write protocols. Send Byte In a send byte operation, the master device sends a single command byte to a slave device, as follows: 1. The master device asserts a start condition on SDA. 2. The master sends the 7-bit slave address followed by the write bit (low). 3. The addressed slave device asserts ACK on SDA. 4. The master sends a command code. 5. The slave asserts ACK on SDA. 6. The master asserts a stop condition on SDA and the transaction ends. In the ADM1066, the send byte protocol is used for two purposes: • To write a register address to RAM for a subsequent single byte read from the same address, or a block read or write starting at that address, as shown in Figure 39. 1 S 2 SLAVE ADDRESS W 3 4 5 6 A REGISTER ADDRESS (0x00 TO 0xDF) A P 04609-039 The ADM1066 contains volatile registers (RAM) and nonvolatile registers (EEPROM). User RAM occupies address locations from 0x00 to 0xDF; EEPROM occupies addresses from 0xF800 to 0xFBFF. Figure 39. Setting a RAM Address for Subsequent Read Rev. 0 | Page 27 of 32 ADM1066 S 2 SLAVE ADDRESS W 3 4 5 6 A COMMAND BYTE (0xFE) A P To write a single byte of data to RAM. In this case, the command byte is the RAM address from 0x00 to 0xDF and the only data byte is the actual data, as shown in Figure 41. 1 • The master sends the 7-bit slave address followed by the write bit (low). 3. The addressed slave device asserts ACK on SDA. 4. The master sends a command code. 6 7 8 DATA A P 2 3 4 5 6 7 8 Figure 42. Setting an EEPROM Address Note, for page erasure, that because a page consists of 32 bytes, only the three MSBs of the address low byte are important. The lower five bits of the EEPROM address low byte specify the addresses within a page and are ignored during an erase operation. Write Byte/Word 2. 5 EEPROM EEPROM SLAVE ADDRESS ADDRESS S W A A A P ADDRESS HIGH BYTE LOW BYTE (0xF8 TO 0xFB) (0x00 TO 0xFF) As soon as the ADM1066 receives the command byte, page erasure begins. The master device can send a stop command as soon as it sends the command byte. Page erasure takes approximately 20 ms. If the ADM1066 is accessed before erasure is complete, it responds with a no acknowledge (NACK). The master device asserts a start condition on SDA. 4 To set up a 2-byte EEPROM address for a subsequent read, write, block read, block write, or page erase. In this case, the command byte is the high byte of the EEPROM address from 0xF8 to 0xFB. The only data byte is the low byte of the EEPROM address, as shown in Figure 42. 1 1. 3 Figure 41. Single Byte Write to RAM Figure 40. EEPROM Page Erasure In a write byte/word operation, the master device sends a command byte and one or two data bytes to the slave device, as follows: 2 RAM S SLAVE W A ADDRESS A ADDRESS (0x00 TO 0xDF) • To write a single byte of data to EEPROM. In this case, the command byte is the high byte of the EEPROM address from 0xF8 to 0xFB. The first data byte is the low byte of the EEPROM address, and the second data byte is the actual data, as shown in Figure 43. 1 2 3 4 5 6 7 8 9 10 EEPROM EEPROM SLAVE ADDRESS ADDRESS S W A A A DATA A P ADDRESS HIGH BYTE LOW BYTE (0xF8 TO 0xFB) (0x00 TO 0xFF) 04609-043 1 • 04609-040 The master sends a command code that tells the slave device to erase the page. The ADM1066 command code for a page erasure is 0xFE (1111 1110). Note that, for a page erasure to take place, the page address has to be given in the previous write word transaction (see the Write Byte/Word section). Also, Bit 2 in the UPDCFG register (Address 0x90) must be set to 1. In the ADM1066, the write byte/word protocol is used for three purposes: 04609-041 To erase a page of EEPROM memory. EEPROM memory can be written to only if it is unprogrammed. Before writing to one or more EEPROM memory locations that are already programmed, the page or pages containing those locations must first be erased. EEPROM memory is erased by writing a command byte. 04609-042 • Figure 43. Single Byte Write to EEPROM Block Write 5. The slave asserts ACK on SDA. 6. The master sends a data byte. 7. The slave asserts ACK on SDA. 8. The master sends a data byte (or asserts a stop condition at this point). 9. The slave asserts ACK on SDA. In a block write operation, the master device writes a block of data to a slave device. The start address for a block write must have been set previously. In the ADM1066, a send byte operation sets a RAM address, and a write byte/word operation sets an EEPROM address, as follows: 1. The master device asserts a start condition on SDA. 2. The master sends the 7-bit slave address followed by the write bit (low). 3. The addressed slave device asserts ACK on SDA. 4. The master sends a command code that tells the slave device to expect a block write. The ADM1066 command code for a block write is 0xFC (1111 1100). 10. The master asserts a stop condition on SDA to end the transaction. Rev. 0 | Page 28 of 32 ADM1066 The slave asserts ACK on SDA. 6. The master sends a data byte that tells the slave device how many data bytes are being sent. The SMBus specification allows a maximum of 32 data bytes in a block write. 7. The slave asserts ACK on SDA. 8. The master sends N data bytes. 9. The slave asserts ACK on SDA after each data byte. In the ADM1066, the receive byte protocol is used to read a single byte of data from a RAM or EEPROM location whose address has previously been set by a send byte or write byte/word operation, as shown in Figure 45. 1 2 S SLAVE ADDRESS R 3 4 5 6 A DATA A P 04609-045 5. Figure 45. Single Byte Read from EEPROM or RAM Block Read 10. The master asserts a stop condition on SDA to end the transaction. 2 3 4 5 6 7 8 9 10 S SLAVE W A COMMAND 0xFC A BYTE A DATA A DATA A DATA A P COUNT (BLOCK WRITE) ADDRESS 1 N 2 04609-044 1 Figure 44. Block Write to EEPROM or RAM Unlike some EEPROM devices that limit block writes to within a page boundary, there is no limitation on the start address when performing a block write to EEPROM, except • There must be at least N locations from the start address to the highest EEPROM address (0xFBFF), to avoid writing to invalid addresses. • If the addresses cross a page boundary, both pages must be erased before programming. Note that the ADM1066 features a clock extend function for writes to EEPROM. Programming an EEPROM byte takes approximately 250 µs, which would limit the SMBus clock for repeated or block write operations. The ADM1066 pulls SCL low and extends the clock pulse when it cannot accept any more data. In a block read operation, the master device reads a block of data from a slave device. The start address for a block read must have been set previously. In the ADM1066, this is done by a send byte operation to set a RAM address, or a write byte/word operation to set an EEPROM address. The block read operation itself consists of a send byte operation that sends a block read command to the slave, immediately followed by a repeated start and a read operation that reads out multiple data bytes, as follows: 1. The master device asserts a start condition on SDA. 2. The master sends the 7-bit slave address followed by the write bit (low). 3. The addressed slave device asserts ACK on SDA. 4. The master sends a command code that tells the slave device to expect a block read. The ADM1066 command code for a block read is 0xFD (1111 1101). 5. The slave asserts ACK on SDA. 6. The master asserts a repeat start condition on SDA. 7. The master sends the 7-bit slave address followed by the read bit (high). 8. The slave asserts ACK on SDA. 9. The ADM1066 sends a byte-count data byte that tells the master how many data bytes to expect. The ADM1066 always returns 32 data bytes (0x20), which is the maximum allowed by the SMBus 1.1 specification. READ OPERATIONS The ADM1066 uses the following SMBus read protocols. Receive Byte In a receive byte operation, the master device receives a single byte from a slave device, as follows: 1. The master device asserts a start condition on SDA. 2. The master sends the 7-bit slave address followed by the read bit (high). 3. The addressed slave device asserts ACK on SDA. 4. The master receives a data byte. 5. The master asserts no acknowledge on SDA. 6. The master asserts a stop condition on SDA, and the transaction ends. 10. The master asserts ACK on SDA. 11. The master receives 32 data bytes. 12. The master asserts ACK on SDA after each data byte. 13. The master asserts a stop condition on SDA to end the transaction. Rev. 0 | Page 29 of 32 ADM1066 2 3 4 5 6 7 8 9 10 11 12 2. S SLAVE W A COMMAND 0xFD A S SLAVE R A BYTE A DATA A ADDRESS (BLOCK READ) ADDRESS COUNT 1 Note that the PEC byte is calculated using CRC-8. The frame check sequence (FCS) conforms to CRC-8 by the polynomial P 04609-046 13 14 DATA A 32 Figure 46. Block Read from EEPROM or RAM C(x) = x8 + x2 + x1 + 1 See the SMBus 1.1 specification for details. Error Correction The ADM1066 provides the option of issuing a PEC (packet error correction) byte after a write to RAM, a write to EEPROM, a block write to RAM/EEPROM, or a block read from RAM/ EEPROM. This enables the user to verify that the data received by or sent from the ADM1066 is correct. The PEC byte is an optional byte sent after that last data byte has been written to or read from the ADM1066. The protocol is as follows: 1. A no acknowledge (NACK) is generated after the PEC byte to signal the end of the read. An example of a block read with the optional PEC byte is shown in Figure 47. 1 2 3 4 5 6 7 8 9 10 11 12 S SLAVE W A COMMAND 0xFD A S SLAVE R A BYTE A DATA A ADDRESS (BLOCK READ) ADDRESS COUNT 1 The ADM1066 issues a PEC byte to the master. The master checks the PEC byte and issues another block read, if the PEC byte is incorrect. Rev. 0 | Page 30 of 32 13 14 15 DATA 32 A PEC A P Figure 47. Block Read from EEPROM or RAM with PEC 04609-047 1 ADM1066 OUTLINE DIMENSIONS 6.00 BSC SQ 0.60 MAX 0.60 MAX PIN 1 INDICATOR 31 30 PIN 1 INDICATOR 0.50 BSC 5.75 BCS SQ TOP VIEW 1.00 0.85 0.80 1 4.25 4.10 SQ 3.95 EXPOSED PAD (BOTTO M VIEW) 0.50 0.40 0.30 12° MAX 40 21 20 10 11 0.25 MIN 4.50 REF 0.80 MAX 0.65 TYP 0.05 MAX 0.02 NOM 0.30 0.23 0.18 SEATING PLANE 0.20 REF COPLANARITY 0.08 COMPLIANT TO JEDEC STANDARDS MO-220-VJJD-2 Figure 48. 40-Lead Lead Frame Chip Scale Package [LFCSP] (CP-40) Dimensions shown in millimeters 0.75 0.60 0.45 1.20 MAX 9.00 BSC SQ 37 36 48 1 SEATING PLANE PIN 1 7.00 BSC SQ TOP VIEW (PINS DOWN) 0° MIN 1.05 1.00 0.95 0.15 0.05 SEATING PLANE VIEW A 0.20 0.09 7° 3.5° 0° 0.08 MAX COPLANARITY 12 25 24 13 0.50 BSC 0.27 0.22 0.17 VIEW A ROTATED 90° CCW COMPLIANT TO JEDEC STANDARDS MS-026ABC Figure 49. 48-Lead Thin Plastic Quad Flat Package [TQFP] (SU-48) Dimensions shown in millimeters ORDERING GUIDE Model ADM1066ACP ADM1066ACP-REEL ADM1066ACP-REEL7 ADM1066ASU ADM1066ASU-REEL ADM1066ASU-REEL7 EVAL-ADM1066LFEB EVAL-ADM1066TQEB Temperature Range −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C −40°C to +85°C Description 40-Lead LFCSP 40-Lead LFCSP 40-Lead LFCSP 48-Lead TQFP 48-Lead TQFP 48-Lead TQFP ADM1066 Evaluation Kit (LFSCP Version) ADM1066 Evaluation Kit (TQFP Version) Rev. 0 | Page 31 of 32 Package Option CP-40 CP-40 CP-40 SU-48 SU-48 SU-48 ADM1066 NOTES © 2004 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D04609–0–10/04(0) Rev. 0 | Page 32 of 32