AD7541 S E M I C O N D U C T O R 12-Bit Multiplying D/A Converter December 1993 Features Description * 12-Bit Linearity 0.01% The AD7541 is a monolithic, low cost, high performance, 12bit accurate, multiplying digital-to-analog converter (DAC). * Pretrimmed Gain Harris' wafer level laser-trimmed thin-film resistors on CMOS circuitry provide true 12-bit linearity with TTL/CMOS compatible operation. * Low Gain and Linearity Tempcos * Full Temperature Range Operation * Full Input Static Protection Special tabbed-resistor geometries (improving time stability), full input protection from damage due to static discharge by diode clamps to V+ and ground, large IOUT1 and IOUT2 bus lines (improving superposition errors) are some of the features offered by Harris AD7541. * TTL/CMOS Compatible * +5V to +15V Supply Range * 20mW Low Power Dissipation * Current Settling Time 1s to 0.01% of FSR Pin compatible with AD7521, this DAC provides accurate four quadrant multiplication over the full military temperature range. * Four Quadrant Multiplication Ordering Information PART NUMBER NONLINEARITY AD7541AD TEMPERATURE RANGE 0.02% (11-Bit) AD7541BD AD7541JN 18 Lead Plastic DIP o o 18 Lead Plastic DIP o o 18 Lead Plastic DIP o o 18 Lead Plastic DIP o o 18 Lead Plastic DIP -55 C to +125 C 0.01% (12-Bit) Pinout o 0 C to +70 C 0.02% (11-Bit) AD7541TD 18 Lead Plastic DIP o 0 C to +70 C 0.01% (12-Bit) Guaranteed Monotonic AD7541SD o 0 C to +70 C 0.01% (12-Bit) AD7541LN 18 Lead Plastic DIP o -25 C to +85 C 0.02% (11-Bit) AD7541KN o -25 C to +85 C 0.01% (12-Bit) PACKAGE o -55 C to +125 C Functional Diagram AD7541 (PDIP) TOP VIEW VREF IN 18 RFEEDBACK IOUT2 2 17 VREF IN BIT 1 (MSB) 4 10k 20k 20k 10k 10k (17) IOUT1 1 GND 3 10k 20k 20k 20k 20k (3) 16 V+ 15 BIT 12 (LSB) BIT 2 5 14 BIT 11 BIT 3 6 13 BIT 10 BIT 4 7 12 BIT 9 BIT 5 8 11 BIT 8 BIT 6 9 10 BIT 7 SPDT NMOS SWITCHES IOUT2 (2) IOUT1 (1) 10k MSB (4) BIT 2 (5) BIT 3 (6) RFEEDBACK (18) Switches shown are for Digital Inputs "High" CAUTION: These devices are sensitive to electrostatic discharge. Users should follow proper I.C. Handling Procedures. Copyright (c) Harris Corporation 1993 8-21 File Number 3107 Specifications AD7541 Absolute Maximum Ratings Thermal Information Supply Voltage (V+ to GND). . . . . . . . . . . . . . . . . . . . . . . . . . . +17V VREF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25V Digital Input Voltage Range . . . . . . . . . . . . . . . . . . . . . . . V+ to GND Output Voltage Compliance . . . . . . . . . . . . . . . . . . . . -100mV to V+ Storage Temperature . . . . . . . . . . . . . . . . . . . . . . . -65oC to +150oC Lead Temperature (Soldering 10s) . . . . . . . . . . . . . . . . . . . . +300oC Thermal Resistance JA Plastic DIP Package . . . . . . . . . . . . . . . . . . . . . . . . . . . .90oC/W Operating Temperature JN, KN, LN Versions . . . . . . . . . . . . . . . . . . . . . . . . 0oC to +70oC AD, BD Versions . . . . . . . . . . . . . . . . . . . . . . . . . . -25oC to +85oC SD, TD Version . . . . . . . . . . . . . . . . . . . . . . . . . . -55oC to +125oC Maximum Power Dissipation Plastic DIP Package up to +70oC . . . . . . . . . . . . . . . . . . . 670mW Junction Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . +150oC CAUTION: Stresses above those listed in "Absolute Maximum Ratings" may cause permanent damage to the device. This is a stress only rating and operation of the device at these or any other conditions above those indicated in the operational sections of this specification is not implied. Electrical Specifications V+ = +15V, VREF = +10V, VOUT1 = VOUT2 = 0V, TA = +25oC, Unless Otherwise Specified TA +25oC PARAMETER TEST CONDITIONS TA MIN-MAX MIN TYP MAX MIN MAX UNITS 12 - - 12 - Bits - - 0.024 - 0.024 % of FSR - - 0.012 - 0.012 % of FSR - - 0.012 - 0.012 % of FSR SYSTEM PERFORMANCE Resolution Nonlinearity A, S, J B, T, K L -10V VREF +10V VOUT1 = VOUT2 = 0V See Figure 3 (Note 4) Monotonicity Guaranteed Gain Error -10V VREF +10V (Note 4) - - 0.3 - 0.4 % of FSR Output Leakage Current (Either Output) VOUT1 = VOUT2 = 0 - - 50 - 200 nA Power Supply Rejection V+ = 14.5V to 15.5V See Figure 5, (Note 4) - - 0.005 - 0.01 % of FSR/% of V+ Output Current Settling Time To 0.1% of FSR See Figure 9, (Note 5) - - 1 - 1 s Feedthrough Error VREF = 20VPP, 10kHz All Digital Inputs Low See Figure 8, (Note 5) - - 1 - 1 mVP-P All Digital Inputs High IOUT1 at Ground 5 10 20 5 20 k DYNAMIC CHARACTERISTICS REFERENCE INPUTS Input Resistance ANALOG OUTPUT Voltage Compliance Output Capacitance Both Outputs, See Maximum Ratings (Note 6) COUT1 COUT2 COUT1 COUT2 Output Noise (Both Outputs) All Digital Inputs High See Figure 7, (Note 5) All Digital Inputs Low) See Figure 7, (Note 5) -100mV to V+ - - 200 - 200 pF - - 60 - 60 pF - - 60 - 60 pF - - 200 - 200 pF See Figure 6 Equivalent to 10k Johnson Noise DIGITAL INPUTS Low State Threshold, VIL (Note 1, 5) High State Threshold, VIH 8-22 - - 0.8 - 0.8 V 2.4 - - 2.4 - V Specifications AD7541 Electrical Specifications V+ = +15V, VREF = +10V, VOUT1 = VOUT2 = 0V, TA = +25oC, Unless Otherwise Specified (Continued) TA +25oC PARAMETER TEST CONDITIONS Input Current VIN = 0V or V+ (Note 5) Input Coding See Tables 1 & 2 (Note 5) Input Capacitance (Note 5) TA MIN-MAX MIN TYP MAX MIN MAX UNITS - - 1 - 1 A 8 pF Binary/Offset Binary - - 8 - POWER SUPPLY CHARACTERISTICS Power Supply Voltage Range Accuracy is not guaranteed over this range +5 to +16 V I+ All Digital Inputs High or Low (Excluding Ladder Network) - - 2.0 - 2.5 mA Total Power Dissipation (Including Ladder Network) - 20 - - - mW NOTES: 1. The digital control inputs are zener protected; however, permanent damage may occur on unconnected units under high energy electrostatic fields. Keep unused units in conductive foam at all times. 2. Do not apply voltages higher than VDD or less than GND potential on any terminal except VREF and RFEEDBACK. 3. Full scale range (FSR) is 10V for unipolar and 10V for bipolar modes. 4. Using internal feedback resistor, RFEEDBACK. 5. Guaranteed by design or characterization and not production tested. 6. Accuracy not guaranteed unless outputs at ground potential. Definition of Terms Detailed Description Nonlinearity: Error contributed by deviation of the DAC transfer function from a "best fit straight line" function. Normally expressed as a percentage of full scale range. For a multiplying DAC, this should hold true over the entire VREF range. The AD7541 is a 12-bit, monolithic, multiplying D/A converter. A highly stable thin film R-2R resistor ladder network and NMOS SPDT switches form the basis of the converter circuit. CMOS level shifters provide low power TTL/CMOS compatible operation. An external voltage or current reference and an operational amplifier are all that is required for most voltage output applications. A simplified equivalent circuit of the DAC is shown on page 1, (Functional Diagram). The NMOS SPDT switches steer the ladder leg currents between IOUT1 and IOUT2 buses which must be held at ground potential. This configuration maintains a constant current in each ladder leg independent of the input code. Converter errors are further eliminated by using wider metal interconnections between the major bits and the outputs. Use of high threshold switches reduces the offset (leakage) errors to a negligible level. Resolution: Value of the LSB. For example, a unipolar converter with n bits has a resolution of LSB = (VREF)/2-N. A bipolar converter of n bits has a resolution of LSB = (VREF)/ 2-(N-1). Resolution in no way implies linearity. Settling Time: Time required for the output function of the DAC to settle to within 1/2 LSB for a given digital input stimulus, i.e., 0 to Full Scale. Gain Error: Ratio of the DAC's operational amplifier output voltage to the nominal input voltage value. Feedthrough Error: Error caused by capacitive coupling from VREF to output with all switches OFF. Output Capacitance: Capacitance from IOUT1, and IOUT2 terminals to ground. Output Leakage Current: Current which appears on IOUT1, terminal when all digital inputs are LOW or on IOUT2 terminal when all inputs are HIGH. Each circuit is laser-trimmed, at the wafer level, to better than 12-bits linearity. For the first four bits of the ladder, special trim-tabbed geometries are used to keep the body of the resistors, carrying the majority of the output current, undisturbed. The resultant time stability of the trimmed circuits is comparable to that of untrimmed units. The level shifter circuits are comprised of three inverters with a positive feedback from the output of the second to first (Figure 1). This configuration results in TTL/COMS compatible operation over the full military temperature range. With the ladder SPDT switches driven by the level shifter, each switch is binary weighted for an "ON" resistance proportional to the respective ladder leg current. This assures a constant voltage drop across each switch, creating equipotential terminations for the 2R ladder resistor, resulting in accurate leg currents. 8-23 AD7541 V+ +15V 1 3 4 6 VREF 10V TO LADDER BIT 1 (MSB) 8 TTL/CMOS INPUT 2 5 9 16 RFEEDBACK 18 IOUT1 5 AD7541 1 CR1 I 15 3 2 OUT2 17 4 DIGITAL INPUT 7 BIT 12 (LSB) IOUT2 IOUT1 6 VOUT A + GND FIGURE 1. CMOS SWITCH FIGURE 2. UNIPOLAR BINARY OPERATION (2-QUADRANT MULTIPLICATION) Typical Applications Zero Offset Adjustment General Recommendations 1. Connect all digital inputs to GND. Static performance of the AD7541 depends on IOUT1 and IOUT2 (pin 1 and pin 2) potentials being exactly equal to GND (pin3). 2. Adjust the offset zero adjust trimpot of the output operational amplifier for 0V 0.5mV (max) at VOUT. Gain Adjustment The output amplifier should be selected to have a low input bias current (typically less than 75nA), and a low drift (depending on the temperature range). The voltage offset of the amplifier should be nulled (typically less than 200V). 1. Connect all digital inputs to VDD. 2. Monitor VOUT for a -VREF (1 - 1/212) reading. 3. To increase VOUT, connect a series resistor, (0 to 250), in the IOUT1 amplifier feedback loop. The bias current compensation resistor in the amplifier's non-inverting input can cause a variable offset. Non-inverting input should be connected to GND with a low resistance wire. 4. To decrease VOUT, connect a series resistor, (0 to 250), between the reference voltage and the VREF terminal. Ground-loops must be avoided by taking all pins going to GND to a common point, using separate connections. TABLE 1. CODE TABLE - UNIPOLAR BINARY OPERATION DIGITAL INPUT The V+ (pin 18) power supply should have a low noise level and should not have any transients exceeding +17V. ANALOG OUTPUT 111111111111 -VREF (1 - 1/212) 100000000001 -VREF (1/2 + 1/212) A high value resistor (~1M) can be used to prevent static charge accumulation, when the inputs are open-circuited for any reason. 100000000000 -VREF/2 011111111111 -VREF (1/2 - 1/212) When gain adjustment is required, low tempco (approximately 50ppm/oC) resistors or trim-pots should be selected. 000000000001 -VREF (1/212) 000000000000 0 Unused digital inputs must be connected to GND or VDD for proper operation. Unipolar Binary Operation The circuit configuration for operating the AD7541 in unipolar mode is shown in Figure 2. With positive and negative VREF values the circuit is capable of 2-Quadrant multiplication. The "Digital Input Code/Analog Output Value" table for unipolar mode is given in Table 1. A Schottky diode (HP5082-2811 or equivalent) prevents IOUT1 from negative excursions which could damage the device. This precaution is only necessary with certain high speed amplifiers. Bipolar (Offset Binary) Operation The circuit configuration for operating the AD7541 in the bipolar mode is given in Figure 3. Using offset binary digital input codes and positive and negative reference voltage values Four-Quadrant multiplication can be realized. The "Digital Input Code/Analog Output Value" table for bipolar mode is given in Table 2. 8-24 AD7541 A "Logic 1" input at any digital input forces the corresponding ladder switch to steer the bit current to IOUT1 bus. A "Logic 0" input forces the bit current to IOUT2 bus. For any code the IOUT1 and IOUT2 bus currents are complements of one another. The current amplifier at IOUT2 changes the polarity of IOUT2 current and the transconductance amplifier at IOUT1 output sums the two currents. This configuration doubles the output range of the DAC. The difference current resulting at zero offset binary code, (MSB = "Logic 1", All other bits = "Logic 0"), is corrected by using an external resistive divider, from VREF to IOUT2. Gain Adjustment Offset Adjustment 1. Adjust VREF to approximately +10V. 1. Connect all digital inputs to VDD. 2. Monitor VOUT for a -VREF (1 - 1/211) volts reading. 3. To increase VOUT, connect a series resistor, (0 to 250), in the IOUT1 amplifier feedback loop. 4. To decrease VOUT, connect a series resistor, (0 to 250), between the reference voltage and the VREF terminal. TABLE 2. CODE TABLE - BIPOLAR (OFFSET BINARY) OPERATION DIGITAL INPUT ANALOG OUTPUT 2. Set R4 to zero. 111111111111 -VREF (1 - 1/211) 3. Connect all digital inputs to "Logic 1". 100000000001 -VREF (1/211) 4. Adjust IOUT1 amplifier offset zero adjust trimpot for 0V 0.1mV at IOUT2 amplifier output. 100000000000 0 5. Connect a short circuit across R2. 011111111111 VREF (1/211) 000000000001 VREF (1 - 1/211) 000000000000 VREF 6. Connect all digital inputs to "Logic 0". 7. Adjust IOUT2 amplifier offset zero adjust trimpot for 0V0.1mV at IOUT1 amplifier output. 8. Remove short circuit across R2. 9. Connect MSB (Bit 1) to "Logic 1" and all other bits to "Logic 0". 10. Adjust R4 for 0V 0.2mV at VOUT. 10V VREF +15V 17 BIT 1 (MSB) 16 4 18 1 IOUT1 6 + DIGITAL INPUT A1 VOUT AD7541 R1 10K R2 10K R5 10K 15 BIT 12 (LSB) 2 IOUT2 3 6 + GND R3 390k A2 NOTE: R1 AND R2 SHOULD BE 0.01%, LOW-TCR RESISTORS FIGURE 3. BIPOLAR OPERATION (4-QUADRANT MULTIPLICATION) 8-25 R4 500 AD7541 Test Circuits +15V VREF 4 17 16 18 5 1 BIT 1 (MSB) 12-BIT BINARY COUNTER RFEEDBACK IOUT1 - AD7541 AD7541 15 2 3 BIT 12 (LSB) HA2600 + 10K 0.01% IOUT2 1M GND CLOCK - VREF HA2600 + BIT 1 (MSB) 10K 0.01% LINEARITY ERROR X 100 14-BIT REFERENCE DAC BIT 12 BIT 13 BIT 14 FIGURE 4. NONLINEARITY TEST CIRCUIT +15V UNGROUNDED SINE WAVE GENERATION 40Hz 1.0VP-P VREF 500K +10V 5K 0.01% BIT 1 (MSB) BIT 12 (LSB) HA2600 + 17 4 16 RFEEDBACK 5K 0.01% 18 IOUT1 5 1 AD7541 HA2600 IOUT2 + 15 3 2 VERROR X 100 GND FIGURE 5. POWER SUPPLY REJECTION TEST CIRCUIT +11V (ADJUST FOR VOUT = 0V) +15V 1K 100 15F 17 4 5 16 2 VOUT AD7541 101ALN IOUT1 15 3 F = 1KHz BW = 1Hz 10K IOUT2 1 50K 1K -50V 0.1F FIGURE 6. NOISE TEST CIRCUIT 8-26 QUAN TECH MODEL 134D WAVE ANALYZER AD7541 Test Circuits (Continued) +15V NC BIT 1 (MSB) +15V 17 16 18 4 5 AD7541 1 17 3 2 BIT 12 (LSB) +15V VREF = 20VPP 10kHz SINE WAVE BIT 1 (MSB) NC 1K 100mVPP 1MHz SCOPE BIT 12 (LSB) 17 16 18 4 5 AD7541 IOUT1 3 1 IOUT2 HA2600 15 3 2 2 6 VOUT GND FIGURE 7. OUTPUT CAPACITANCE TEST CIRCUIT FIGURE 8. FEEDTHROUGH ERROR TEST CIRCUIT +15V VREF +10V EXTRAPOLATE BIT 1 (MSB) +5V 0V DIGITAL INPUT 17 16 4 5 AD7541 1 15 BIT 12 (LSB) OSCILLOSCOPE +100mV IOUT2 2 3 3t: 5% SETTLING 9t: 0.01% SETTLING 100 GND FIGURE 9. OUTPUT CURRENT SETTLING TIME TEST CIRCUIT Dynamic Performance Similarly the output capacitance varies between the minimum and the maximum values depending on the input code. These variations necessitate the use of compensation capacitors, when high speed amplifiers are used. The dynamic performance of the DAC, also depends on the output amplifier selection. For low speed or static applications, AC specifications of the amplifier are not very critical. For high-speed applications slew-rate, settling-time, openloop gain and gain/phase-margin specifications of the amplifier should be selected for the desired performance. A capacitor in parallel with the feedback resistor (as shown in Figure 10) provides the necessary phase compensation to critically damp the output. The output impedance of the AD7541 looking into IOUT1 varies between 10k (RFEEDBACK alone) and 5K (RFEEDBACK in parallel with the ladder resistance). A small capacitor connected to the compensation pin of the amplifier may be required for unstable situations causing oscillations. Careful PC board layout, minimizing parasitic capacitances, is also vital. +15V VREF +10V BIT 1 (MSB) BIT 2 BIT 12 (LSB) 17 16 4 18 5 AD7541 1 15 3 2 RFEEDBACK IOUT1 IOUT2 CC - A + VOUT GND FIGURE 10. GENERAL DAC CIRCUIT WITH COMPENSATION CAPACITOR, CC 8-27