Low-Power, 12-Bit Sampling ADC with Internal Reference and Power-Down

Transcription

1 9-406; Rev 4; /97 EALUATION KIT MANUAL FOLLOWS DATA SHEET Low-Power, -Bit Sampling ADC General Description The is a monolithic, CMOS, -bit analog-todigital converter (ADC) featuring differential inputs, track/hold (T/H), internal voltage reference, internal or external clock, and parallel or serial µp interface. The has a 7.µs conversion time, a µs acquisition time, and a guaranteed 00ksps sample rate. The operates from a single + supply or from dual ± supplies, allowing ground-referenced bipolar input signals. The device features a logic power-down input, which reduces the 3mA DD supply current to 0µA max, including the internal-reference current. Decoupling capacitors are the only external components needed for the power supply and reference. This ADC operates with either an external reference, or an internal reference that features an adjustment input for trimming system gain errors. The provides three interface modes: two 8-bit parallel modes, and a serial interface mode that is compatible with SPI TM, QSPI TM, and MICROWIRE TM serialinterface standards. Applications Battery-Powered Data Logging PC Pen Digitizers High-Accuracy Process Control Electromechanical Systems Data-Acquisition Boards for PCs Automatic Testing Systems Telecommunications Digital Signal Processing (DSP) REF REFADJ AIN + AIN DD CLK/ REF 7 AGND DGND Functional Diagram OSC IN REF OUT -BIT SAR ADC SS 3-STATE OUTPUT 8-BIT BUS AND SERIAL I/O CONTROL LOGIC 8 PD PAR BIP D7/ D6/ OUT D/SSTRB D4 D3/D D/D0 D/D9 D0/D8 RD BUSY HBEN Features -Bit Resolution, /LSB Linearity + or ± Operation Built-In Track/Hold Internal Reference with Adjustment Capability Low Power: 3mA Operating Mode 0µA Power-Down Mode 00ksps Tested Sampling Rate Serial and 8-Bit Parallel µp Interface 4-Pin Narrow DIP and Wide SO Packages PART TEMP. RANGE PIN-PACKAGE ERROR (LSB) ACNG 0 C to +70 C 4 Narrow Plastic DIP ±/ BCNG ACWG 0 C to +70 C 0 C to +70 C 4 Narrow Plastic DIP 4 Wide SO ± ±/ BCWG 0 C to +70 C 4 Wide SO ± BC/D 0 C to +70 C Dice* ± AENG -40 C to +8 C 4 Narrow Plastic DIP ±/ BENG -40 C to +8 C 4 Narrow Plastic DIP ± AEWG -40 C to +8 C 4 Wide SO ±/ BEWG -40 C to +8 C 4 Wide SO ± AMRG - C to + C 4 Narrow CERDIP** ±/ BMRG - C to + C 4 Narrow CERDIP** ± * Dice are specified at T A = + C, DC parameters only. ** Contact factory for availability and processing to MIL-STD-883. TOP IEW PD SS AIN+ 3 AIN- 4 REF REFADJ 6 AGND 7 BIP 8 BUSY 9 D0/D8 0 D/D9 DGND Pin Configuration DIP/SO SPI and QSPI are trademarks of Motorola, Inc. MICROWIRE is a trademark of National Semiconductor Corp. Ordering Information 4 DD 3 CLK/ PAR HBEN 0 9 RD 8 D7/ 7 D6/ OUT 6 D/SSTRB D4 4 D3/D 3 D/D0 Maxim Integrated Products For free samples & the latest literature: or phone For small orders, phone ext

2 ABSOLUTE MAXIMUM RATINGS DD to DGND to +7 SS to AGND...-7 to +0.3 DD to SS... AGND, REF, REFADJ to DGND to ( DD + 0.3) AIN+, AIN-, PD to SS to ( DD + 0.3), RD, CLK, BIP, HBEN, PAR, to DGND to ( DD + 0.3) BUSY, D0 D7 to DGND to ( DD + 0.3) Continuous Power Dissipation (T A = +70 C) Narrow Plastic DIP (derate 3.33mW/ C above +70 C)...067mW Wide SO (derate.76mw/ C above +70 C)...94mW Narrow CERDIP (derate.0mw/ C above +70 C)...000mW Operating Temperature Ranges _C...0 C to +70 C _E C to +8 C _M...- C to + C Storage Temperature Range...-6 C to +60 C Lead Temperature (soldering, 0sec) C Stresses beyond those listed under Absolute Maximum Ratings" may cause permanent damage to the device. These are stress ratings only, and functional operation of the device at these or any other conditions beyond those indicated in the operational sections of the specifications is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. ELECTRICAL CHARACTERISTI ( DD = ±%, SS = 0 or - ±%, f CLK =.6MHz, 0% duty cycle, AIN- = AGND, BIP = 0, slow-memory mode, internal-reference mode, reference compensation mode external, synchronous operation, Figure 6, T A = T MIN to T MAX, unless otherwise noted.) (Note ) Offset Error PARAMETER SYMBOL CONDITIONS MIN TYP MAX UNITS DC ACCURACY (Note ) Resolution Bits Integral Nonlinearity B ± Differential Nonlinearity DNL No missing codes over temperature ± LSB Gain Error (Note 3) A ±/ A ± B ± A ± B ±3 Gain-Error Tempco (Note 4) Excludes internal-reference drift ±0. ppm/ C Signal-to-Noise plus Distortion Ratio Total Harmonic Distortion (up to the th Harmonic) khz input signal, T A = + C 70 db khz input signal, T A = + C -80 db Spurious-Free Dynamic Range SFDR khz input signal, T A = + C 80 db Conversion Time (Note ) INL DYNAMIC ACCURACY (sample rate = 00kHz, IN = 4p-p) CONERSION RATE SINAD THD t CON Synchronous CLK ( to 3 CLKs) Internal CLK, C L = 0pF 6 8 Track/Hold Acquisition Time µs Aperture Delay ns Aperture Jitter 0 ps External Clock Frequency f 0..6 MHz Range (Note 6) CLK LSB LSB LSB µs

3 ELECTRICAL CHARACTERISTI (continued) ( DD = ±%, SS = 0 or - ±%, f CLK =.6MHz, 0% duty cycle, AIN- = AGND, BIP = 0, slow-memory mode, internal-reference mode, reference compensation mode external, synchronous operation, Figure 6, T A = T MIN to T MAX, unless otherwise noted.) (Note ) Load Regulation Input Current Input Current PARAMETER ANALOG INPUT Input oltage Range (Note 7) Output Short-Circuit Current Capacitive Load Required Power-Supply Rejection REFADJ Input Adjustment Range (Note 0) REFADJ Disable Threshold REFADJ Output oltage REFADJ Input Current Input oltage Range Input Resistance Input Low oltage Input High oltage Input Current CLK Input Capacitance (Note 6) PD Input Low oltage PD Input High oltage PD Input Current SYMBOL IL IH ma ppm/ C I IN IN = 0 to DD ±0 µa I IN PD = high/float ±00 PD = low ±0. µa C IN 0 pf IL IH REFADJ = CONDITIONS T A = + C, I OUT = 0mA to ma Reference compensation mode external DD = ±%, SS = ±% External-reference mode External-reference = External-reference mode, RD, CLK, HBEN, PAR, BIP, RD, CLK, HBEN, PAR, BIP MIN TYP MAX Input Leakage Current IN = SS to DD ±0 µa Input Capacitance (Note 6) Small-Signal Bandwidth INTERNAL REFERENCE REF Output oltage REF Output Tempco (Note 8) Output Current Capability (Note 9) REFERENCE INPUT LOGIC INPUTS T A = + C _C _E _M T A = + C..0 0 I IN PD = 0 to DD (Note ) ±0 µa SS ±300.4 DD UNITS pf MHz ma m ma µf µ m µa ma kω PD External Leakage for Float State (Note ) Maximum current allowed for floating state ±00 na PD Floating-State oltage FLT Reference compensation mode external.8 3

4 ELECTRICAL CHARACTERISTI (continued) ( DD = ±%, SS = 0 or - ±%, f CLK =.6MHz, 0% duty cycle, AIN- = AGND, BIP = 0, slow-memory mode, internal-reference mode, reference compensation mode external, synchronous operation, Figure 6, T A = T MIN to T MAX, unless otherwise noted.) (Note ) PARAMETER LOGIC OUTPUTS Output Low oltage Output High oltage Three-State Leakage Current Three-State Output Capacitance (Note 6) POWER REQUIREMENTS Positive Supply oltage Negative Supply oltage Positive Supply Current Negative Supply Current Positive Supply Rejection (Note 3) Negative Supply Rejection (Note 3) SYMBOL OL OH I L C OUT DD SS I OUT =.6mA I OUT = -00µA D0/D8-D7/ CONDITIONS -. 0 = RD = DD, PD = high/float 3 ma I DD AIN =, D0/D8 D7/ PD = low 0 0 µa = 0 or DD, PD = high/float I SS HBEN = PAR = BIP 0 00 µa = 0 or DD PD = low 0 FS change, DD = ±% ±/ LSB FS change, SS = - ±% MIN TYP MAX ± ±/ UNITS µa pf LSB TIMING CHARACTERISTI (Figures 6 0) ( DD = ±%, SS = 0 or - ±%, T A = T MIN to T MAX, unless otherwise noted.) (Note 4) PARAMETER RD Pulse Width SYMBOL CONDITIONS 0 C/E MIN TYP MAX M MIN TYP MAX to RD Setup Time t ns RD to BUSY Delay Data Access Time (Note ) t t 3 t 4 t 6 UNITS C L = 0pF ns C L = 00pF ns 0 0 ns to RD Hold Time t ns Data Setup Time After BUSY (Note ) Bus-Relinquish Time (Note 6) HBEN to RD Setup Time ns t ns t ns HBEN to RD Hold Time t ns Delay Between Read Operations (Note 6) t ns Delay Between Conversions t µs Aperture Delay t Jitter < 0ps ns CLK to BUSY Delay (Note 6) t ns OUT to SSTRB Rise Delay OUT to SSTRB Fall Delay t 4 t T A = + C MIN TYP MAX ns ns 4

5 TIMING CHARACTERISTI (Figures 6 0) (continued) ( DD = ±%, SS = 0 or - ±%, T A = T MIN to T MAX, unless otherwise noted.) (Note 4) PARAMETER or RD Hold Time or RD Setup Time to OUT Delay SYMBOL t 6 t 7 CONDITIONS to Three-State t ns t 0 T A = + C MIN TYP MAX 0 60 C/E MIN TYP MAX 0 UNITS ns 80 M MIN TYP MAX to Delay t ns OUT to Delay t ns to SSTRB Delay t ns Note : Performance at power-supply tolerance limits guaranteed by power-supply rejection test. Note : DD =, SS = 0, FS = REF. Note 3: FS = REF, offset nulled, ideal last-code transition = FS - 3/ LSB. Note 4: Gain-Error Tempco = GE is the gain-error change from T A = + C to T MIN or T MAX. Note : Conversion time defined as the number of clock cycles times the clock period; clock has a 0% duty cycle. Note 6: Guaranteed by design, not production tested. Note 7: AIN+, AIN- must not exceed supplies for specified accuracy. Note 8: REF TC = T, where REF is reference-voltage change from T A = + C to T MIN or T MAX. Note 9: Output current should not change during conversion. This current is in addition to the current required by the internal DAC. Note 0: REFADJ adjustment range is defined as the allowed voltage excursion on REFADJ relative to its unadjusted value of.4. This will typically result in a.7 times larger change in the REF output (Figure 9a). Note : This current is included in the PD supply current specification. Note : Floating the PD pin guarantees external compensation mode. Note 3: REF = 4.096, external reference. Note 4: All input control signals are specified with t r = t f = ns (0% to 90% of ) and timed from a voltage level of.6. Note : t 3 and t 6 are measured with the load circuits of Figure and defined as the time required for an output to cross 0.8 or.4. Note 6: t 7 is defined as the time required for the data lines to change 0. when loaded with the circuits of Figure ns ns

6 Typical Operating Characteristics CLOCK FREQUENCY (MHz) 0 0. CLOCK FREQUENCY vs. TIMING CAPACITOR SEE FIGURE T A = + C GR9-A SUPPLY CURRENT (µa) 0 0 POWER-DOWN SUPPLY CURRENT vs. TEMPERATURE DD = + SS = - PD = 0 I DD GR9-B ISS (µa) 0 0 NEGATIE SUPPLY CURRENT vs. TEMPERATURE GR9-C I SS TIMING CAPACITOR (nf) TEMPERATURE ( C) TEMPERATURE ( C) IDD (ma) POSITIE SUPPLY CURRENT vs. TEMPERATURE GR9-D SIGNAL AMPLITUDE (db) khz FFT PLOT -0 f IN = khz f S = 00kHz -40 SNR = 7dB T A = + C -94.3dB -96.dB-98.0dB -93.8dB GR9-E SIGNAL AMPLITUDE (db) kHz FFT PLOT -86.0dB f IN = 0kHz f S = 00kHz SNR = 7.dB T A = + C -90.8dB GR9-F TEMPERATURE ( C) FREQUENCY (khz) FREQUENCY (khz) 6

7 PIN NAME PD SS Pin Description Power-Down Input. A logic low at PD deactivates the ADC only the bandgap reference is active. A logic high selects normal operation, internal-reference compensation mode. An open-circuit condition selects normal operation, external-reference compensation mode. Negative Supply, 0 to -. FUNCTION 3 AIN+ Sampled Analog Input 4 AIN- Analog Input Return. Pseudo-differential (see Gain and Offset Adjustment section). REF Reference-Buffer Output for Internal Reference. Input for external reference when REFADJ is connected to DD. 6 REFADJ Reference Adjust. Connect to DD to use an extended reference at REF. 7 AGND Analog Ground 8 BIP BIP = low selects unipolar mode BIP = high selects bipolar mode (see Gain and Offset Adjustment section) 9 BUSY BUSY Output is low during a conversion. 0 D0/D8 Three-State Data Outputs: LSB = D0 D/D9 Three-State Data Outputs DGND Digital Ground 3 D/D0 Three-State Data Outputs 4 D3/D Three-State Data Outputs: MSB = D D4 Three-State Data Output 6 D/SSTRB Three-State Data Output/Serial Strobe Output in serial mode 7 D6/ OUT Three-State Data Output/Serial Clock Output in serial mode 8 D7/ Three-State Data Output/Data Output in serial mode 9 RD 0 HBEN Read Input. In parallel mode, a low signal starts a conversion when and HBEN are low (memory mode). RD also enables the outputs when is low. In serial mode, RD = low enables OUT and SSTRB when is low. RD = high forces OUT and SSTRB into a high-impedance state. Chip-Select Input must be low for the ADC to recognize RD and HBEN inputs in parallel mode. The falling edge of starts a conversion in serial mode. = high in serial mode forces OUT, SSTRB, and into a high-impedance state. High-Byte Enable Input. In parallel mode, HBEN = high multiplexes the 4 MSBs of the conversion result into the lower bit outputs. HBEN = high also disables conversion starts. HBEN = low places the 8 LSBs onto the data bus. In serial mode, HBEN = low enables OUT to operate during the conversion only, HBEN = high enables OUT to operate continuously, provided is low. PAR Sets the output mode. PAR = high selects parallel output mode. PAR = low selects serial output mode. 3 CLK/ Clock Input/Serial Clock Input in serial mode. An external TTL-/CMOS-compatible clock may be applied to this pin, or a capacitor (0pF nominal) may be connected between CLK and DGND to operate the internal oscillator. 4 DD Positive Supply, + ±% 7

8 + 3k DN DN 3k C L C L DGND DGND a. High-Z to OH and OL to OH b. High-Z to OL and OH to OL 4.7µF 0.µF OPEN 0.µF PD DD CLK/ AIN+ AIN- REF REFADJ PAR HBEN RD AGND BIP D7/ D6/ OUT C SERIAL/PARALLEL INTERFACE MODE µp CONTROL INPUTS Figure. Load Circuits for Access Time + 3k OUTPUT STATUS 9 0 BUSY DO/DB D/D9 DGND SS D/SSTRB D4 D3/D D/D DN DN 0 TO - 3k 0pF 0pF µp DATA BUS DGND DGND NOTE: C 0pF GENERATES MHz NOMINAL CLOCK. a. OH to High-Z b. OL to High-Z Figure. Load Circuits for Bus-Relinquish Time Figure 3. Operational Diagram Detailed Description The uses successive approximation and input track/hold (T/H) circuitry to convert an analog input signal to a -bit digital output. Flexible control logic provides easy interface to microprocessors (µps), so most applications require only the addition of passive components. No external hold capacitor is required for the T/H. Figure 3 shows the in its simplest operational configuration. Pseudo-Differential Input The sampling architecture of the ADC s analog comparator is illustrated in the Equivalent Input Circuit (Figure 4). A capacitor switching between the AIN+ and AIN- inputs acquires the signal at the ADC s analog input. At the end of the conversion, the capacitor reconnects to AIN+ and charges to the input signal. An external input buffer is usually not needed for lowbandwidth input signals (<00Hz) because the ADC disconnects from the input during the conversion. In unbuffered applications, an input filter capacitor reduces conversion noise, but also may limit input bandwidth. When converting a single-ended input signal, AINshould be connected to AGND. If a differential signal is connected, consider that the configuration is pseudo differential only the signal side to the input channel is held by the T/H. The return side (AIN-) must remain stable within ±0.LSB (±0.LSB for best results) with respect to AGND during a conversion. Accomplish this by connecting a 0.µF capacitor from AIN- to AGND. Analog Input Track/Hold The T/H enters its tracking mode when the ADC is deselected ( pin is held high and BUSY pin is high). Hold mode starts approximately ns after a conversion is initiated. The variation in this delay from one conversion to the next (aperture jitter) is about 0ps. Figures 6 0 detail the T/H and interface timing for the 8

9 AIN + C PACKAGE pf TRACK HOLD C HOLD 3pF C SWITCH 0pF R IN HOLD COMPARATOR CLK CLOCK AIN - C EXT +.6 -BIT DAC DGND NOTE: C EXT = 0pF GENERATES MHz NOMINAL CLOCK Figure 4. Equivalent Input Circuit Figure. Internal Clock Circuit various interface modes. The time required for the T/H to acquire an input signal is a function of how quickly its input capacitance is charged. If the input signal s source impedance is high, the acquisition time lengthens and more time must be allowed between conversions. Acquisition time is calculated by: t ACQ = 0(R S + R IN )C HOLD (but never less than µs), where R IN = kω, R S = source impedance of the input signal, and C HOLD = 3pF (see Figure 4). Input Bandwidth The ADC s input tracking circuitry has a MHz typical large-signal bandwidth characteristic, and a 30/µs slew rate. It is possible to digitize high-speed transients and measure periodic signals with bandwidths exceeding the ADC s sample rate of 00ksps by using undersampling techniques. Note that if undersampling is used to measure high-frequency signals, special care must be taken to avoid aliasing errors. Without adequate input bandpass filtering, out-of-band signals and noise may be aliased into the measurement band. Input Protection Internal protection diodes, which clamp the analog input to DD and SS, allow AIN+ to swing from ( SS - 0.3) to ( DD + 0.3) with no risk of damage to the ADC. However, for accurate conversions near full scale, AIN+ should not exceed the power supplies by more than 0m because ADC accuracy is affected when the protection diodes are even slightly forward biased. Digital Interface Starting a Conversion In parallel mode, the ADC is controlled by the, RD, and HBEN inputs, as shown in Figure 6. The T/H enters hold mode and a conversion starts at the falling edge of and RD while HBEN (not shown) is low. BUSY goes low as soon as the conversion starts. On the falling edge of the 3th input clock pulse after the conversion starts, BUSY goes high and the conversion result is latched into three-state output buffers. In serial mode, the falling edge of initiates a conversion, and the T/H enters hold mode. Data is shifted out serially as the conversion proceeds (Figure 0). See the Parallel Digital-Interface Mode and Serial-Interface Mode sections for details. Internal/External Clock Figure shows the clock circuitry. The ADC includes internal circuitry to generate a clock with an external capacitor. As indicated in the Typical Operating Characteristics, a 0pF capacitor connected between the CLK and DGND pins generates a MHz nominal clock frequency (Figure ). Alternatively, an external clock (between 00kHz and.6mhz) can be applied to CLK. When using an external clock source, acceptable clock duty cycles are 9

10 CLK + RD t 6 t 7 t t 3 BUSY t CON t t CON Figure 6., RD, and CLK Synchronous Operation between 4% and %. Clock and Control Synchronization For best analog performance on the, the clock should be synchronized to the conversion start signals ( and RD) as shown in Figure 6. A conversion should not be started in the 0ns before a clock edge nor in the 00ns after it. This ensures that CLK transitions are not coupled to the analog input and sampled by the T/H. The magnitude of this feedthrough can be a few millivolts. When the clock and conversion start signals are synchronized, small end-point errors (offset and full-scale) are the most that can be generated by clock feedthrough. Even these errors (which can be trimmed out) can be avoided by ensuring that the start of a conversion (RD or falling edge) does not occur close to a clock transition (Figure 6), as described above. Parallel Digital-Interface Mode Output-Data Format The data output from the is straight binary in the unipolar mode. In the bipolar mode, the MSB is inverted (see Figure ). The data bits can be output either in two 8-bit bytes or as a serial output. Table shows the data-bus output format. A -byte read uses outputs D7 D0. Byte selection is controlled by HBEN. When HBEN is low, the lower 8 bits appear at the data outputs. When HBEN is high, the upper 4 bits appear at D0-D3 with the leading 4 bits low in locations D4 D7. Timing and Control Conversion-start and data-read operations are controlled by the HBEN,, and RD digital inputs. A logic low is required on all three inputs to start a conversion, and once the conversion is in progress it cannot be restarted. BUSY remains low during the entire conversion cycle. The timing diagrams of Figures 7 0 outline two parallel-interface modes and one serial mode. Slow-Memory Mode In slow-memory mode, the device appears to the µp as a slow peripheral or memory. Conversion is initiated with a read instruction (see Figure 7 and Table ). Set the PAR pin high for parallel interface mode. Beginning with HBEN low, taking and RD low starts the conversion. The analog input is sampled on the falling edge of RD. BUSY remains low while the conversion is in progress. The previous conversion result appears at the digital outputs until the end of conversion, when BUSY returns high. The output latches are then updated with the newest results of the 8 LSBs on D7 D0. A second read operation with HBEN high places the 4 MSBs, with 4 leading 0s, on data outputs D7 D0. The second read operation does not start a new conversion because HBEN is high. ROM Mode As in slow-memory mode, D7 D0 are used for -byte reads. A conversion starts with a read instruction with HBEN and low. The T/H samples the input on the falling edge of RD (see Figure 8 and Table 3). PAR is set high. At this point the data outputs contain the 8 LSBs from the previous conversion. Two more read operations are needed to access the conversion result. The first occurs with HBEN high, where the 4 MSBs with 4 leading 0s are accessed. The second read, with HBEN low, outputs the 8 LSBs and also starts a new conversion. Figure 9 and Table 4 show how to read output data within one conversion cycle without starting another conversion. Trigger the falling edge of a read on the ris- 0

11 HBEN RD t 8 t t t 9 t 8 t 9 t t 4 t t t CON t 0 t 0 BUSY t t 3 t 6 t 7 t 3 t 7 DATA OLD DATA D7 D0 NEW DATA D7 D0 NEW DATA D D8 HOLD* TRACK t t *INTERNAL SIGNAL. TRACKING INPUT SIGNAL WHEN HOLD = Low, HOLDING WHEN HOLD = High. Figure 7. Slow-Memory Mode Timing HBEN t 8 t 9 t 8 t 9 t 8 t 9 RD t t 4 t t t 4 t t t 4 t t t CON t 0 t BUSY t t 3 t 7 t 3 t 7 t 3 t 7 DATA OLD DATA D7 D0 NEW DATA D D8 NEW DATA D7 D0 HOLD* t t TRACK *INTERNAL SIGNAL. TRACKING INPUT SIGNAL WHEN HOLD = Low, HOLDING WHEN HOLD = High. Figure 8. ROM Mode Timing

12 HBEN CLK t 8 t 9 t 8 t t 4 t RD BUSY t t CON t 0 DATA t3 OLD DATA D7 D0 t 7 t 3 t 7 t 3 t 7 NEW DATA D7 D0 NEW DATA D D8 HOLD* TRACK t *INTERNAL SIGNAL. TRACKING INPUT SIGNAL WHEN HOLD = Low, HOLDING WHEN HOLD = High Figure 9. ROM Mode Timing, Reading Data without Starting a Conversion OUT THREE STATE t t 0 t 0 THREE STATE t 7 t 6 SSTRB t 3 t 3 THREE STATE t 4 t t THREE STATE t t t 9 HOLD TRACK CYCLES Figure 0. Serial-Interface Mode Timing Diagram (RD = low)

13 Table. Data-Bus Output, = RD = Low PIN NAME D7/ D6/ OUT D/SSTRB D4 D3/D D/D0 D/D9 D0/D8 HBEN = 0, PAR =, PARALLEL MODE D7 D6 D D4 D3 D D D0 HBEN =, PAR =, PARALLEL MODE Low Low Low Low D D0 D9 D8 HBEN = X, PAR = 0, SERIAL MODE, RD = 0 OUT SSTRB Low Low Low Low Low HBEN = X, PAR = 0, Three- Three- SERIAL MODE, RD = Stated Stated Low Low Low Low Low Note: D7/ D0/D8 are the ADC data output pins. D D0 are the -bit conversion results. D is the MSB. = Three-state data output. Data output in serial mode. OUT = Three-state data output. Clock output in serial mode. SSTRB = Three-state data output. Strobe output in serial mode. Table. Slow-Memory Mode, -Byte Read Data-Bus Status PIN NAME D7/ D6/ OUT D/SSTRB D4 D3/D D/D0 D/D9 D0/D8 FIRST READ (New Data) D7 D6 D D4 D3 D D D0 SECOND READ (New Data) Low Low Low Low D D0 D9 D8 Table 3. ROM Mode, -Byte Read Data-Bus Status PIN NAME D7/ D6/ OUT D/SSTRB D4 D3/D D/D0 D/D9 D0/D8 FIRST READ (Old Data) D7 D6 D D4 D3 D D D0 SECOND READ (New Data) Low Low Low Low D D0 D9 D8 THIRD READ (New Data) D7 D6 D D4 D3 D D D0 Table 4. ROM Mode, -Byte Read Data-Bus Status without Starting a Conversion Cycle PIN NAME D7/ D6/ OUT D/SSTRB D4 D3/D D/D0 D/D9 D0/D8 FIRST READ (Old Data) D7 D6 D D4 D3 D D D0 SECOND READ (New Data) D7 D6 D D4 D3 D D D0 THIRD READ (New Data) Low Low Low Low D D0 D9 D8 3

14 3 0 9 RD OUT HBEN A B CLOCK 74HC64 Q A Q B Q C Q D Q E Q F Q G SSTRB 6 9 CLEAR Q H 3 Q A A B CLOCK 74HC64 Q B Q C Q D Q E Q F Q G LOGIC INPUT 9 CLEAR Q H 3 OUT D DO t 9 SSTRB NOTE: USE SSTRB TO GATE PARALLEL DATA TRANSFER FROM SHIFT REGISTER, OR TO CLEAR SHIFT REGISTERS IF DESIRED. Figure. Simple Serial-to-Parallel Interface 4

15 ing edge of the first clock cycle after conversion end (when BUSY goes high). As mentioned previously, two more read operations (after BUSY goes high) are needed to access the conversion results. The only difference is that now the low byte can be read first. This happens by allowing the first read operation to occur with HBEN low, where the 8 LSBs are accessed. The second read, with HBEN high, accesses the 4 MSBs with 4 leading 0s. Serial-Interface Mode The serial mode is compatible with Microwire, SPI and QSPI serial interfaces. In addition, a framing signal (SSTRB) is provided that allows the devices to interface with the TMS30 family of DSPs. Set PAR low for serial mode. A falling edge on causes the T/H to sample the input (Figure 0). Conversion always begins on the next falling edge of, regardless of where occurs. The line remains high-impedance until a conversion begins. During the MSB decision, remains low (leading 0), while SSTRB goes high to indicate that a data frame is beginning. The data is available at on the rising edge of ( OUT when using an internal clock) and transitions on the falling edge. remains low after all data bits have been shifted out, inserting trailing 0s in the data stream until returns high. The OUT signal is synchronous with the internal or external clock. For interface flexibility,, OUT and SSTRB signals enter a high-impedance state when is high. When is low, RD controls the status of OUT and SSTRB outputs. A logic low RD enables OUT and SSTRB, while a logic high forces both outputs into a high-impedance state. Also, with low and HBEN high, OUT drives continuously, regardless of conversion status. This is useful with µps that require a continuous serial clock. If and HBEN are low, OUT is output only during the conversion cycle, while the converter internal clock runs continuously. This is useful for creating a simple serial-to-parallel interface without shift-register overflow (Figure ). Maximum Clock Rate in Serial Mode The maximum rate depends on the minimum setup time required at the serial data input to the µp and the ADC s to delay (t ) (see Figure ). The maximum f is as follows: a. SPI b. QSPI I/O SCK MISO SS SCK MISO SS I/O SK SI + + t c. MICROWIRE I/O CLKX f (MAX) = t SU (M) + t ( ) t SETUP (MIN) t SU (M) IS THE SETUP TIME REQUIRED AT THE SERIAL DATA INPUT TO THE µp. t IS THE MAXIMUM TO DELAY. Figure. f (MAX) is limited by the setup time required by the serial data input to the µp. CLKR DR FSR SSTRB d. TMS30 SERIAL INTERFACE Figure 3. Common Serial-Interface Connections to the

16 f (MAX) = (/) x / (tsu(m) + t) where t su (M) is the minimum data-setup time required at the serial data input to the µp. For example, Motorola s MC68HCA8 data book specifies a 00ns minimum data-setup time. Using the worst case for a military grade part of t = 80ns (see Timing Characteristics) and substituting in the above equation indicates a maximum frequency of.3mhz. Using the with SPI, QSPI and MICROWIRE Serial Interfaces Figure 3 shows interface connections to the for common serial-interface standards. SPI and MICROWIRE (CPOL=0, CPHA=0) The is compatible with SPI, QSPI and MICROWIRE serial-interface standards. When using SPI or QSPI, two modes are available to interface with the. You can set CPOL = 0 and CPHA = 0 (Figure 4a), or set CPOL = and CPHA = (Figure 4b). When using CPOL = 0 and CPHA = 0, the conversion begins on the first falling edge of following going low. Data is available from on the rising edge of, and transitions on the falling edge. Two consecutive -byte reads are required to get the full bits from the ADC. The first byte contains the following, in this order: a leading unknown bit ( will still be high-impedance on the first bit), a 0, and the six MSBs. The second byte contains the remaining six LSBs and two trailing 0s. SPI (CPOL=, CPHA=) Setting CPOL = and CPHA = starts the clock high during a read instruction. The will shift out a leading 0 followed by the data bits and three trailing 0s (Figure 4b). QSPI Unlike SPI, which requires two -byte reads to acquire the bits of data from the ADC, QSPI allows the minimum number of clock cycles required to clock in the data (Figure ). TMS30 Serial Interface Figure 3d shows the pin connections to interface the to the TMS30. Since the makes data available on the rising edge of and the TMS30 shifts data in on the falling edge of CLKR, use CLKX of the DSP to drive, and CLKX to drive the DSP s CLKR input. The inverter s propagation delay also provides more data-setup time at the DSP. For example, with no inverter delay, and using t = 80ns and f =.6MHz, the available setup time before the transition is: setup time = / ( x f ) - t = / ( x.6e6) - 80ns = 3ns This still exceeds the 3ns minimum DR setup time before the CLKR goes low (tsu(dr)), however, a generic 74HC04 provides an additional 0ns setup time (see Figure 3d). Figure 6 shows the DSP interface timing characteristics. The DSP begins clocking data in on the falling edge of CLKR after the falling edge of SSTRB. ST BYTE READ ND BYTE READ HIGH-Z a. CPOL = 0, CPHA = 0 LEADING ZERO MSB D0 D9 D8 D7 D6 D D4 D3 D D LSB HIGH-Z HIGH-Z b. CPOL =, CPHA = LEADING ZERO MSB D0 D9 D8 D7 D6 D D4 D3 D D LSB HIGH-Z Figure 4. SPI/MICROWIRE Serial-Interface Timing 6

17 HIGH-Z MSB D0 D9 D8 D7 D6 D D4 D3 D D LSB HIGH-Z a. CPOL = 0, CPHA = 0 HIGH-Z MSB D0 D9 D8 D7 D6 D D4 D3 D D LSB HIGH-Z b. CPOL =, CPHA = Figure. QSPI Serial-Interface Timing CLKR SSTRB HIGH-Z HIGH-Z HIGH-Z MSB D0 D9 D8 D7 D6 D D4 D3 D D LSB HIGH-Z Figure 6. TMS30 Interface Timing 7

18 Following the data transfer, the DSP receive shift register (RSR) contains a 6-bit word consisting of the data bits, MSB first, followed by four trailing 0s. Applications Information Power-On Initialization When the + power supply is first applied to the, perform a single conversion to initialize the ADC (the BUSY signal status is undefined at power-on). Disregard the data outputs. Power-Down Mode In some battery-powered systems, it is desirable to power down or remove power from the ADC during inactive periods. To power down the, drive PD low. In this mode, all internal ADC circuitry is off except the reference, and the ADC consumes less than 0µA max (assuming all signals, RD, CLK, and HBEN are static and within 00m of the supplies). Figure 7 shows a practical way to drive the PD pin. If using internal reference compensation, drive PD between DD and DGND with a µp I/O pin or other logic device (Figure 7a). For external-reference compensation mode, use the circuit in Figure 7b to drive PD between DGND and the floating voltage of PD. An alternative is to drive PD with three-state logic or a switch, provided the off leakage does not exceed 00nA. Internal Reference The internal reference is available at REF and must be bypassed to AGND with a 4.7µF low-esr capacitor (less than /Ω) in parallel with a 0.µF capacitor, unless internal-reference compensation mode is used (see the Internal Reference Compensation section). This minimizes noise and maintains a low reference impedance at high frequencies. The reference output can be disabled by connecting REFADJ to DD when using an external reference. Reference-Compensation Modes Power-down performance can be optimized for a given conversion rate by selecting either internal or external reference compensation. PD Internal Compensation The connection for internal compensation is shown in Figure 8a. In this mode, the reference stabilizes quickly enough so that a conversion typically starts within 3µs after the ADC is reactivated (PD pulled high). In this compensation mode, the reference buffer requires longer recovery time from SAR transients, therefore requiring a slower clock (and conversion time). With internal reference compensation, the typical conversion time rises to µs (Figure 8b). Figure 8c illustrates the typical average supply current vs. conversion rate, a. INTERNAL-REFERENCE COMPENSATION MODE + PD REF OPEN-DRAIN BUFFER PD 0.µF 6 REFADJ b. EXTERNAL-REFERENCE COMPENSATION MODE Figure 7. Drive Circuits for PD Pin Figure 8a. Internal-Compensation Mode Circuit 8

19 PD 0 REF µs 0µs µs SUPPLY CURRENT (µa) 0, fg8c RD k k 0k 00k CONERSIONS PER SECOND Figure 8b. Low Average-Power Mode Operation (Internal Compensation) which can be achieved using power-down between conversions. External Compensation Figure 9a shows the connection for external compensation with reference adjustment. In this mode, an external 4.7µF capacitor compensates the reference output amplifier, allowing for maximum conversion speed and lowest conversion noise. However, when reactivating the ADC after power-down, the reference takes typically ms to fully charge the 4.7µF capacitor, so more time is required before a conversion can start (Figure 9b). Thus, the average current consumed in power-up/powerdown operations is higher in external compensation mode than in internal compensation mode. Gain and Offset Adjustment Figure 0 depicts the nominal, unipolar input/output (I/O) transfer function, and Figure shows the bipolar I/O transfer function. Code transitions occur halfway between successive integer LSB values. Note that LSB =.00m (4.096/4096) for unipolar operation and LSB =.00m ((4.096/ /)/4096) for bipolar operation. Figures 9a and a show how to adjust the ADC gain in applications that require full-scale range adjustment. The connection shown in Figure a provides ±0.% for ±0LSBs of adjustment range and is recommended for applications that use an external reference. On the other hand, Figure 9a is recommended for applications that use the internal reference, because it uses fewer external components. If both offset and full scale need adjustment, the circuit in Figure b is recommended. For single-supply Figure 8c. Average Supply Current vs. Conversion Rate, Powering Down Between Conversions 0.µF 4.7µF k k k 00k 0.0µF REF REFADJ Figure 9a. External-Compensation Mode with Internal Reference Adjustment Circuit ADCs, it is virtually impossible to null system negative offset errors. However, the input configuration is pseudo-differential only the difference in voltage between AIN+ and AIN- will be converted into its digital representation. By applying a small positive voltage to AIN-, the 0 input voltage at AIN+ can be adjusted to above or below AIN- voltage, thus nulling positive or negative system offset errors. R9 and R0 can be removed for applications that require only positive system errors to be nulled. To trim the offset error of the, apply /LSB to the analog input and adjust R6 so the digital output code changes between 000 (hex) and 00 (hex). To adjust full scale, apply FS - /LSBs and adjust R until the output code changes 6 PD 9

20 PD 0 OPEN CIRCUIT (FLOAT) OUTPUT CODE FULL-SCALE TRANSITION REF... 0 ms 00ms RD.µs FS = REF LSB = FS Figure 9b. Low Average-Power Mode Operation (External Compensation) between FFE (hex) and FFF (hex). Because interaction occurs between adjustments, offset should be adjusted before gain. For an input gain of two, remove R7 and R FS AIN INPUT OLTAGE (LSB) FS LSB The accepts input voltages from AGND to DD while operating from a single supply, and SS to DD when operating from dual supplies. Figure shows the bipolar input transfer function with AIN- connected to midscale for single-supply operation and connected to GND operating from dual supplies. When operating from a single supply, the can be configured for bipolar operation on its pseudo-differential input. Instead of using AIN- as an analog input return, AINcan be set to a different positive potential voltage above ground (BIP pin is set high). The sampled analog input (AIN+) can swing to any positive voltage above and below AIN-, and the ADC performs bipolar conversions with respect to AIN-. When operating from dual supplies, the full-scale range is from -REF/ to +REF/. Figure 0. Unipolar Transfer Function IN MAX480 R3 0k R 00Ω R 49.9Ω R4 0k TO AIN+ Digital Bus Noise If the data bus connected to the ADC is active during a conversion, crosstalk from the data pins to the ADC comparator may generate errors. Slow-memory mode avoids this problem by placing the µp in a wait state during the conversion. In ROM mode, if the data bus is active during the conversion, it should be isolated from the ADC using three-state drivers. The ADC generates considerable digital noise in ROM mode when RD or go high and the output data drivers are disabled after a conversion has started. This noise can cause large errors if it occurs when the SAR latches a comparator decision. To avoid this problem, Figure a. Trim Circuit for Gain (±0.%) RD and should be active for less than one clock cycle. If this is not possible, RD or should go high at the rising edge of CLK, since the comparator output is always latched on falling edges of CLK. Layout, Grounding, Bypassing Use printed circuit boards for best system performance. 0

21 IN R7 0k R8 0k MAX480 R 0k AIN R 00Ω D0 D R6 0k REF R 0k 0.µF* R3 0k R4 49.9Ω REF R9* 0k R0* 49.9Ω AIN SINGLE SUPPLY AIN- = REF ( ) DUAL SUPPLY AIN- = 0 0 -REF REF 0 REF - LSB REF - LSB * CONNECT AIN- TO AGND WHEN USING DUAL SUPPLIES Figure b. Offset (±0m) and Gain (±%) Trim Circuit Wire-wrap boards are not recommended. Board layout should ensure that digital- and analog-signal lines are separated from each other. Do not run analog and digital (especially clock) lines parallel to one another, or digital lines underneath the ADC package. Figure 3 shows the recommended system ground connections. Establish a single-point ground ( star ground point) at AGND, separate from the logic ground. Connect all other analog grounds and DGND to it. No other digital-system ground should be connected to this single-point analog ground. The ground return to the power supply for this star ground should be low impedance and as short as possible for noisefree operation. High-frequency noise in the DD power supply may affect the high-speed comparator in the ADC. Bypass these supplies to the single-point analog ground with Figure. Bipolar Transfer Function SUPPLIES + - GND R* = 0Ω DD AGND SS DGND + DGND DIGITAL CIRCUITRY *OPTIONAL Figure 3. Power-Supply Grounding Connection

22 0.0µF and 0µF bypass capacitors. Minimize capacitor lead lengths for best supply-noise rejection. If the + power supply is very noisy, a 0Ω resistor can be connected as a lowpass filter to filter out supply noise (Figure 3). Dynamic Performance High-speed sampling capability and throughput make the ideal for wideband signal processing. To support these and other related applications, Fast Fourier Transform (FFT) test techniques guarantee the ADC's dynamic frequency response, distortion, and noise at the rated throughput. Specifically, this involves applying a low-distortion sine wave to the ADC input and recording the digital conversion results for a specified time. The data is then analyzed using an FFT algorithm, which determines its spectral content. Conversion errors are then seen as spectral elements outside the fundamental input frequency. FFT plots are shown in the Typical Operating Characteristics. ADCs have traditionally been evaluated by specifications such as zero and full-scale error, integral nonlinearity (INL), and differential nonlinearity (DNL). Such parameters are widely accepted for specifying performance with DC and slowly varying signals, but are less useful in signal-processing applications where the ADC s impact on the system transfer function is the main concern. The significance of various DC errors does not translate well to the dynamic case, so different tests are required. Signal-to-Noise Ratio (SNR) is the ratio between the RMS amplitude of the fundamental input frequency to the RMS amplitude of all other A/D output signals, except signal harmonics. Signal-to-Noise + Distortion ratio (SINAD) is the same as the SNR, but includes signal harmonics. The theoretical minimum A/D noise is caused by quantization error and is a direct result of the ADC s resolution: SNR = (6.0n +.76) db, where n is the number of bits of resolution. 74dB is the SNR of a perfect -bit ADC. By transposing the equation that converts resolution to SNR we can compute the effective resolution or the effective number of bits the ADC provides from the measured SNR: n = (SNR.76)/6.0 Total Harmonic Distortion Total Harmonic Distortion (THD) is the ratio of the RMS sum of all harmonics of the input signal (in the frequency band above DC and below one-half the sample rate) to the fundamental itself. This expressed as: THD = 0log [ ( n ) / ] where is the fundamental RMS amplitude and to n are the amplitudes of the nd through n th harmonics. Spurious-Free Dynamic Range Spurious-free dynamic range is the ratio of the fundamental RMS amplitude to the amplitude of the next largest spectral component (in the frequency band above DC and below one-half the sample rate). Usually this peak occurs at some harmonic of the input frequency. But if the ADC is exceptionally linear, it can occur at a random peak in the ADC s noise floor. Opto-Isolated A/D Interface Many industrial applications require isolation to prevent excessive current flow where ground disparities exist between the ADC and the rest of the system. In Figure 4, a MAX0 and four 6N36 opto-couplers create an

23 D CC D 4 3 T (SCHOTT) 00µF 6 IN IC 4 OUT 74L0 GND 00µF 6 IC -3 HCPL630 (QUALITY TECHNOLOGIES) 8 k 9 0 k 7 Q N DD 0.µF TTL/CMOS OUTPUTS IC MAX0 k 6 IC 4 3 k k Q N3906 AIN+ 3 IN 6 SSTRB AIN- 4 k SHDN IC k k 0 HBEN RD 9 TTL/CMOS INPUTS 6 k 4 3 IC 3 6 k 3 CLK PAR BIP 8 EN GND 4.7µF 0.µF REF AGND ISOLATION BARRIER 0.µF 6 REFADJ SS DGND Figure 4. Isolated Data-Acquisition Circuit 3

24 Chip Topography HBEN RD REF REFADJ AGND 0.98" (.09mm) BIP BUSY D0/D8 AIN- AIN+ AGND D/D9 DGND D/D0 D3/D D4 PD D/SSTRB DD CLK/ PAR D7/ D6/OUT 0.4" (3.606mm) SUBSTRATE CONNECTED TO DD Package Information PDIPN.EPS 4