19-1647; Rev 1; 6/00 MAX1711 Voltage Positioning Evaluation Kit Ordering Information PART TEMP. RANGE MAX1711EVKIT 0°C to +70°C Features ♦ Output Voltage Positioned ♦ Reduces CPU Power Consumption ♦ Lowest Number of Output Capacitors (only 4) ♦ High Speed, Accuracy, and Efficiency ♦ Fast-Response Quick-PWM™ Architecture ♦ 7V to 24V Input Voltage Range ♦ 0.925V to 2V Output Voltage Range ♦ 12A Load-Current Capability (14.1A peak) ♦ 550kHz Switching Frequency ♦ Power-Good Output ♦ 24-Pin QSOP Package ♦ Low-Profile Components ♦ Fully Assembled and Tested IC PACKAGE 24 QSOP Quick-PWM is a trademark of Maxim Integrated Products. Component List DESIGNATION QTY C1–C4, C20 5 DESCRIPTION 10µF, 25V ceramic capacitors Taiyo Yuden TMK432BJ106KM, Tokin C34Y5U1E106Z, or United Chemi-Con/Marcon THCR50E1E106ZT 220µF, 2.5V, 25mΩ low-ESR polymer capacitors Panasonic EEFUEOE 221R DESIGNATION QTY DESCRIPTION D2 1 100mA Schottky diode Central Semiconductor CMPSH-3 D3 1 1A Schottky diode Motorola MBRS130LT3, International Rectifier 10BQ040, or Nihon EC10QS03 D4 1 200mA switching diode Central Semiconductor CMPD2838 J1 1 Scope-probe connector Berg Electronics 33JR135-1 C5, C6, C7, C16 4 C8 1 10µF, 6.3V ceramic capacitor Taiyo Yuden JMK325BJ106MN or TDK C3225X5R1A106M C9 1 0.1µF ceramic capacitor JU1 1 2-pin header C10 0 0.01µF ceramic capacitor (not installed) JU3–9 0 Not installed C11, C12 2 0.22µF ceramic capacitors L1 1 0.47µH power inductor Sumida CEP 125 series 4712-T006 C13 0 0.1µF ceramic capacitor (not installed) C14 1 470pF ceramic capacitor N1 1 C15 1 1µF ceramic capacitor N-channel MOSFET (SO-8) International Rectifier IRF7811 or IRF7811A C18 1 1000pF ceramic capacitor N2, N3 2 1 2A Schottky diode SGS-Thomson STPS2L25U or Nihon EC31QS03L N-channel MOSFET (SO-8) International Rectifier IRF7805 or IRF7811 or IRF 7811A D1 ________________________________________________________________ Maxim Integrated Products 1 For free samples and the latest literature, visit www.maxim-ic.com or phone 1-800-998-8800. For small orders, phone 1-800-835-8769. Evaluates: MAX1711 General Description The MAX1711 evaluation kit (EV kit) demonstrates the high-power, dynamically adjustable notebook CPU application circuit with voltage positioning. Voltage positioning decreases CPU power consumption and reduces output capacitance requirements. This DC-DC converter steps down high-voltage batteries and/or AC adapters, generating a precision, low-voltage CPU core VCC rail. The MAX1711 EV kit provides a digitally adjustable 0.925V to 2V output voltage from a 7V to 24V battery input range. It delivers sustained output current of 12A and 14.1A peaks, operating at a 550kHz switching frequency, and has superior line- and load-transient response. The MAX1711 EV kit is designed to accomplish output voltage transitions in a controlled amount of time with limited input surge current. This EV kit is a fully assembled and tested circuit board. Evaluates: MAX1711 MAX1711 Voltage Positioning Evaluation Kit Component Suppliers Component List (continued) DESIGNATION QTY DESCRIPTION SUPPLIER PHONE FAX Central Semiconductor 516-435-1110 516-435-1824 Dale-Vishay 402-564-3131 402-563-6418 Fairchild 408-721-2181 408-721-1635 N4, N5 (not installed) 0 N-channel MOSFETs Motorola 2N7002 or Central Semiconductor 2N7002 R1 1 20Ω ±5% resistor R2 0 Not installed R3 1 1MΩ ±5% resistor International Rectifier 310-322-3331 310-322-3332 R4 1 100kΩ ±5% resistor Kemet 408-986-0424 408-986-1442 R6 1 100kΩ ±1% resistor Motorola 602-303-5454 602-994-6430 R9 1 140kΩ ±1% resistor Nihon 847-843-7500 847-843-2798 R10 1 1kΩ ±5% resistor Panasonic 714-373-7939 714-373-7183 R11 1 100Ω ±5% resistor Sanyo 619-661-6835 619-661-1055 1 0.005Ω ±1%, 1W resistor Dale WSL-2512-R005F SGS-Thomson 617-259-0300 617-259-9442 Sumida 708-956-0666 708-956-0702 R12 R13 1 1MΩ ±1% resistor Taiyo Yuden 408-573-4150 408-573-4159 R14 1 10kΩ ±1% resistor TDK 847-390-4373 847-390-4428 SW1 1 DIP-10 dip switch Tokin 408-432-8020 408-434-0375 SW2 1 Momentary switch, normally open Digi-Key P8006/7S U1 1 MAX1711EEG (24-pin QSOP) U2 (not installed) 0 Exclusive-OR gate (5-Pin SSOP) Toshiba TC4S30F None 1 Shunt (JU1) None 1 MAX1711 PC board None 1 MAX1711 data sheet Recommended Equipment • 7V to 24V, >20W power supply, battery, or notebook AC adapter • DC bias power supply, 5V at 100mA • Dummy load capable of sinking 14.1A • Digital multimeter (DMM) • 100MHz dual-trace oscilloscope Quick Start 1) Ensure that the circuit is connected correctly to the supplies and dummy load prior to applying power. 2) Ensure that the shunt is connected at JU1 (SHDN = VCC). 3) Set switch SW1 per Table 1 to achieve the desired output voltage. 4) Connect +5V or ground to the AC Present pad to disable the transition detector circuit. See the Dynamic Output Voltage Transitions section for more information regarding the transition detector circuit. 2 Note: Please indicate that you are using the MAX1711 when contacting these component suppliers. 5) Turn on battery power prior to +5V bias power; otherwise, the output UVLO timer will time out and the FAULT latch will be set, disabling the regulator until +5V power is cycled or shutdown is toggled (press the RESET button). 6) Observe the output with the DMM and/or oscilloscope. Look at the LX switching-node and MOSFET gate-drive signals while varying the load current. Detailed Description This 14A buck-regulator design is optimized for a 550kHz frequency and output voltage settings around 1.6V. At VOUT = 1.6V, inductor ripple is approximately 35%, with a resulting pulse-skipping threshold at roughly ILOAD = 2.2A. Setting the Output Voltage Select the output voltage using the D0–D4 pins. The MAX1711 uses an internal 5-bit DAC as a feedback resistor voltage divider. The output voltage can be digitally set from 0.925V to 2V using the D0–D4 inputs. Switch SW1 sets the desired output voltage. See Table 1. _______________________________________________________________________________________ MAX1711 Voltage Positioning Evaluation Kit tial output voltage 20mV high, and R12 (5mΩ) causes the output voltage to drop with increasing load (60mV or about 4% of 1.6V at 12A). Setting the output voltage high allows a larger stepdown when the output current increases suddenly, and regulating at the lower output voltage under load allows a larger step-up when the output current suddenly decreases. Allowing a larger step size means that the output capacitance can be reduced and the capacitor’s ESR can be increased. If voltage positioning is not used, one additional output capacitor is required to meet the same transient specification. Reduced power consumption at high load currents is an additional benefit of voltage positioning. Because the output voltage is reduced under load, the CPU draws less current. This results in lower power dissipation in the CPU, though some extra power is dissipated in R12. For a 1.6V, 12A nominal output, reducing the output voltage 2.75% (1.25% - 4%) gives an output voltage of 1.556V and an output current of 11.67A. So the CPU power consumption is reduced from 19.2W to 18.16W. The additional power consumption of R12 is 5mΩ · 11.7A2 = 0.68W, and the overall power savings is 19.2 – (18.16 + 0.68) = 0.36W. In effect, 1W of CPU dissipation is saved and the power supply dissipates much of the savings, but both the net savings and the transfer of dissipation away from the hot CPU are beneficial. D4 D3 D2 D1 D0 OUTPUT VOLTAGE (V) 0 0 0 0 0 2.00 0 0 0 0 1 1.95 0 0 0 1 0 1.90 0 0 0 1 1 1.85 0 0 1 0 0 1.80 0 0 1 0 1 1.75 0 0 1 1 0 1.70 0 0 1 1 1 1.65 1 1 0 0 0 1.60 1 1 0 0 1 1.55 1 1 0 1 0 1.50 1 1 0 1 1 1.45 1 1 1 0 0 1.40 1 1 1 0 1 1.35 1 1 1 1 0 1.30 0 1 1 1 1 Shutdown 1 0 0 0 0 1.275 1 0 0 0 1 1.250 1 0 0 1 0 1.225 Dynamic Output Voltage Transitions 1 0 0 1 1 1.200 1 0 1 0 0 1.175 1 0 1 0 1 1.150 1 0 1 1 0 1.125 1 0 1 1 1 1.100 1 1 0 0 0 1.075 1 1 0 0 1 1.050 1 1 0 1 0 1.025 1 1 0 1 1 1.000 1 1 1 0 0 0.975 1 1 1 0 1 0.950 1 1 1 1 0 0.925 1 1 1 1 1 Shutdown If the DAC inputs (D0–D4) are changed, the output voltage will change accordingly. However, under some circumstances, the output voltage transition may be slower than desired. All transitions to a higher voltage will occur very quickly, with the circuit operating at the current limit set by the voltage at the ILIM pin. Transitions to a lower output voltage require the circuit or the load to sink current. If SKIP is held low (PFM mode), the circuit won’t sink current, so the output voltage will decrease only at the rate determined by the load current. This is often acceptable, but some applications require output voltage transitions to be completed within a set time limit. Powering CPUs with Intel’s Geyserville technology is such an application. The specification requires that output voltage transitions occur within 100µs after a DAC code change. This fast transition timing means that the regulator circuit must sink as well as source current. The simplest way of meeting this requirement is to use the MAX1711’s fixed-frequency PWM mode (set SKIP high), allowing the regulator to sink or source currents equally. This EV kit is shipped with SKIP set high. Although this results in a V DD quiescent current to 20mA or more, depending on the MOSFETs and Voltage Positioning The MAX1711 EV kit uses voltage positioning to minimize the output capacitor requirements of the Intel Coppermine CPU’s transient voltage specification (-7.5% to +7.5%). The output voltage is initially set slightly high (1.25%) and then allowed to regulate lower as the load current increases. R13 and R14 set the ini- _______________________________________________________________________________________ 3 Evaluates: MAX1711 Table 1. MAX1710/1711 Output Voltage Adjustment Settings Evaluates: MAX1711 MAX1711 Voltage Positioning Evaluation Kit switching frequency used, it is often an acceptable choice. A similar but more clever approach is to use PWM mode only during transitions. This approach allows the regulator to sink current when needed and to operate with low quiescent current the rest of the time, but it requires that the system know when the transitions will occur. Any system with a changing output voltage must know when its output voltage changes occur. Usually, it is the system that initiates the transition, either by driving the DAC inputs to new levels or by selecting new DAC inputs with a digital mux. While it is possible for the regulator to recognize transitions by watching for DAC code changes, the glue logic needed to add that feature to existing controllers is unnecessarily complicated (refer to the MAX1710/MAX1711 data sheet, Figure 10). It is easier to use the chipset signal that selects DAC codes at the mux, or some other system signal to inform the regulator that a code change is occurring. For easy modification, the MAX1711 EV kit is designed to use an external chipset signal to indicate DAC code transitions (install U2, R2, C10, C13; short JU9 and cut JU10). This signal connects to the EV kit’s AC Present pad and should have 5V logic levels. Logic edges on AC Present are detected by exclusive-OR gate U2, which generates a 60µs pulse on each edge (determined by R2 and C10). These pulses drive SKIP, allowing the regulator to sink current during transitions. Because U2 is powered by VCC (5V), the signal connected to AC Present must have 5V logic levels so that U2’s output pulses will be symmetric for positive- and negative-going transitions. If the signal that’s available to drive AC Present has a different logic level, either level-shift the signal or lift U2’s supply pin and power it from the appropriate supply rail. In addition to controlling SKIP, the pulses from U2 have two other functions, which are optional. U2’s output drives the gates of two small-signal MOSFETs, N4 and N5 (not installed). N4 is used to temporarily reduce the circuit’s current limit, in effect soft-starting the regulator. This reduces the battery surge current, which otherwise would discharge (upward transitions) or charge (downward transitions) the regulator input (battery) at a rate determined by the regulator’s maximum current limit. N5 pulls down on PGOOD during transitions, indicating that the output voltage is in transition. Load-Transient Measurement One interesting experiment is to subject the output to large, fast load transients and observe the output with an oscilloscope. This necessitates careful instrumentation of the output, using the supplied scope-probe jack. 4 Accurate measurement of output ripple and load-transient response invariably requires that ground clip leads be completely avoided and that the probe hat be removed to expose the GND shield, so the probe can be plugged directly into the jack. Otherwise, EMI and noise pickup will corrupt the waveforms. Most benchtop electronic loads intended for powersupply testing are unable to subject the DC-DC converter to ultra-fast load transients. Emulating the supply current di/dt at the CPU VCORE pins requires at least 10A/µs load transients. One easy method for generating such an abusive load transient is to solder a MOSFET, such as an MTP3055 or 12N05 directly across the scope-probe jack. Then drive its gate with a strong pulse generator at a low duty cycle (10%) to minimize heat stress in the MOSFET. Vary the high-level output voltage of the pulse generator to adjust the load current. To determine the load current, you might expect to insert a meter in the load path, but this method is prohibited here by the need for low resistance and inductance in the path of the dummy-load MOSFET. There are two easy alternative methods for determining how much load current a particular pulse-generator amplitude is causing. The first and best is to observe the inductor current with a calibrated AC current probe, such as a Tektronix AM503. In the buck topology, the load current is equal to the average value of the inductor current. The second method is to first put on a static dummy load and measure the battery current. Then, connect the MOSFET dummy load at 100% duty momentarily, and adjust the gate-drive signal until the battery current rises to the appropriate level (the MOSFET load must be well heatsinked for this to work without causing smoke and flames). Efficiency Measurements and Effective Efficiency Testing the power conversion efficiency POUT/PIN fairly and accurately requires more careful instrumentation than might be expected. One common error is to use inaccurate DMMs. Another is to use only one DMM, and move it from one spot to another to measure the various input/output voltages and currents. This second error usually results in changing the exact conditions applied to the circuit due to series resistance in the ammeters. It’s best to get four 3-1/2 digit, or better, DMMs that have been recently calibrated, and monitor VBATT, VOUT, IBATT, and ILOAD simultaneously, using separate test leads directly connected to the input and output PC board terminals. Note that it’s inaccurate to test efficiency at the remote VOUT and ground termi- _______________________________________________________________________________________ MAX1711 Voltage Positioning Evaluation Kit Remember to include the power consumed by the +5V bias supply when making efficiency calculations: Efficiency = VOUT × I LOAD (VBATT × I BATT ) + (5V × I BIAS ) The choice of MOSFET has a large impact on efficiency performance. The International Rectifier MOSFETs used were of leading-edge performance for the 12A application at the time this kit was designed. However, the pace of MOSFET improvement is rapid, so the latest offerings should be evaluated. Once the actual efficiency data has been obtained, some work remains before an accurate assessment of a voltage-positioned circuit can be made. As discussed in the Voltage Positioning section, a voltagepositioned power supply can dissipate additional power while reducing system power consumption. For this reason, we use the concept of effective efficiency, which allows the direct comparison of a positioned and nonpositioned circuit’s efficiency. Effective efficien- cy is the efficiency required of a nonvoltage-positioned circuit to equal the total dissipation of a voltage-positioned circuit for a given CPU operating condition. Calculate effective efficiency as follows: • Start with the efficiency data for the positioned circuit (VIN, IIN, VOUT, IOUT). • Model the load resistance for each data point (RLOAD = VOUT / IOUT). • Calculate the output current that would exist for each RLOAD data point in a nonpositioned application (INP = VNP / RLOAD, where VNP = 1.6V in this example). • Effective efficiency = (VNP ✕ INP) / (VIN ✕ IIN) = calculated nonpositioned power output divided by the measured voltage-positioned power input. • Plot the efficiency data point at the current INP. The effective efficiency of the voltage-positioned circuit will be less than that of the nonpositioned circuit at light loads where the voltage-positioned output voltage is higher than the nonpositioned output voltage. It will be greater than that of the nonpositioned circuit at heavy loads where the voltage-positioned output voltage is lower than the nonpositioned output voltage. _______________________________________________________________________________________ 5 Evaluates: MAX1711 nals, because doing this incorporates the parasitic resistance of the PC board output and ground buses in the measurement (a significant power loss). Evaluates: MAX1711 MAX1711 Voltage Positioning Evaluation Kit Jumper and Switch Settings Table 2. Jumper JU1 Functions (Shutdown Mode) Table 4. Jumper JU6 Functions (Fixed/Adjustable Current-Limit Selection) SHUNT LOCATION SHDN PIN MAX1711 OUTPUT Installed Connected to VCC MAX1711 enabled Not Installed Connected to GND Shutdown mode, VOUT = 0 SHUNT LOCATION Installed Connected to VCC Not Installed Table 3. Jumpers JU3/JU4/JU5 Functions (Switching-Frequency Selection) SHUNT LOCATION TON PIN FREQUENCY (kHz) Not Installed Connected to VCC 200 Installed Not Installed Connected to REF 400 Not Installed Not Installed Installed Connected to GND 550 Not Installed Not Installed Not Installed Floating 300 JU3 JU4 JU5 Installed Not Installed Not Installed CURRENT-LIMIT THRESHOLD ILIM PIN 100mV Connected to GND via an external resistor divider, Adjustable R6/R9. Refer to the Pin between 50mV Description ILIM section in and 200mV the MAX1711 data sheet for more information. Table 5. Jumpers JU9/JU10 Functions (FBS and FB Integrator Disable Selection) SHUNT LOCATION JU9 JU10 Installed Not Installed Not Installed Installed SKIP PIN Connected to VCC Connected to the output of U2 IMPORTANT: Don’t change the operating frequency without first recalculating component values because the frequency has a significant effect on the peak current-limit level, MOSFET heating, preferred inductor value, PFM/PWM switchover point, output noise, efficiency, and other critical parameters. Table 6. Troubleshooting Guide SYMPTOM Circuit won’t start when power is applied. POSSIBLE PROBLEM SOLUTION Power-supply sequencing: +5V bias supply was applied first. Press the RESET button. Output overvoltage due to shorted high-side MOSFET. Replace the MOSFET. Output overvoltage due to load Circuit won’t start when RESET is pressed, recovery overshoot. +5V bias supply cycled. Overload condition. Reduce the inductor value, raise the switching frequency, or add more output capacitance. Remove the excessive load. Troubleshoot the power stage. Are the DH and DL Broken connection, bad MOSFET, gate-drive signals present? Is the 2V VREF preor other catastrophic problem. sent? On-time pulses are erratic or have unexpected changes in period. 6 VBATT power source has poor impedance characteristic. Add a bulk electrolytic bypass capacitor across the benchtop power supply, or substitute a real battery. _______________________________________________________________________________________ REF 2V D4 D3 D2 D1 D0 SKIP SHDN GND VCC SW1E 5 7V TO 24V JU6 JU3 200kHz 6 R3 1M JU4 400kHz JU5 550kHz 6 8 9 5 16 17 18 19 20 21 2 VCC FLOAT = 300kHz R8 SHORT 7 8 9 10 JU1 C12 0.22µF C14 470pF SW1D 4 SW1C 3 SW1B 2 SW1A 1 SKIP SW2 RESET R10 1k ILIM TON REF CC D4 D3 D2 D1 D0 SKIP U1 10 AGND V+ PGOOD GNDS FBS FB PGND DL LX DH BST 15 VDD PGOOD MAX1711 ILIM SHDN VCC 7 VDD VCC C11 R1 0.22µF 20Ω 12 11 4 3 14 13 23 24 22 1 R4 100k R7 SHORT 4 8 VCC 7 C9 0.1µF C18 1000pF R14 10k 1% C2 10µF 25V N2 1 2 5 6 D2 CMPSH-3 C1 10µF 25V R13 1M 1% C15 1µF R11 100Ω VDD C3 10µF 25V 7 3 7 PGOOD REF 4 8 4 8 C4 10µF 25V N3 1 2 5 6 6 5 N1 1 2 C20 10µF 25V D1 L1 0.47µH C21 OPEN R12 0.005Ω 1% C5 220µF 2.5V C16 220µF 2.5V C6 220µF 2.5V C17 OPEN C7 220µF 2.5V C19 OPEN D3 D4 CMPD2838 C8 10µF 6.3V VOUT J1 SCOPE JACK +5V VBIAS Evaluates: MAX1711 VBATT MAX1711 Voltage Positioning Evaluation Kit Figure 1. MAX1711 Voltage Positioning EV Kit Schematic _______________________________________________________________________________________ 7 SKIP TC4S30F 3 AC PRESENT ILIM R2 10k C10 0.01µF U2 2 5 R9 140k 1% R6 100k 1% 1 VCC N4 4 3 2 1 C13 0.1µF NOT INSTALLED JU9 1 3 2 VCC N5 JU10 SHORT (PC TRACE) PGOOD Evaluates: MAX1711 MAX1711 Voltage Positioning Evaluation Kit Figure 1. MAX1711 Voltage Positioning EV Kit Schematic (continued) 8 _______________________________________________________________________________________ MAX1711 Voltage Positioning Evaluation Kit Evaluates: MAX1711 1.0" 1.0" Figure 2. MAX1711 Voltage Positioning EV Kit Component Placement Guide—Component Side 1.0" Figure 3. MAX1711 Voltage Positioning EV Kit Component Placement Guide—Solder Side 1.0" Figure 4. MAX1711 Voltage Positioning EV Kit PC Board Layout—Component Side Figure 5. MAX1711 Voltage Positioning EV Kit PC Board Layout—Internal GND Plane (Layer 2) _______________________________________________________________________________________ 9 Evaluates: MAX1711 MAX1711 Voltage Positioning Evaluation Kit 1.0" 1.0" Figure 6. MAX1711 Voltage Positioning EV Kit PC Board Layout—Internal GND Plane (Layer 3) Figure 7. MAX1711 Voltage Positioning EV Kit PC Board Layout—Solder Side Maxim cannot assume responsibility for use of any circuitry other than circuitry entirely embodied in a Maxim product. No circuit patent licenses are implied. Maxim reserves the right to change the circuitry and specifications without notice at any time. 10 ____________________Maxim Integrated Products, 120 San Gabriel Drive, Sunnyvale, CA 94086 408-737-7600 © 2000 Maxim Integrated Products Printed USA is a registered trademark of Maxim Integrated Products.