LTC1603 High Speed, 16-Bit, 250ksps Sampling A/D Converter with Shutdown DESCRIPTIO U FEATURES ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ The LTC®1603 is a 250ksps, 16-bit sampling A/D converter that draws only 220mW from ±5V supplies. This high performance device includes a high dynamic range sample-and-hold, a precision reference and a high speed parallel output. Two digitally selectable power shutdown modes provide power savings for low power systems. A Complete, 250ksps 16-Bit ADC 90dB S/(N+D) and –100dB THD (Typ) Power Dissipation: 220mW (Typ) Nap (7mW) and Sleep (10µW) Shutdown Modes No Pipeline Delay No Missing Codes over Temperature Operates with Internal 15ppm/°C Reference or External Reference True Differential Inputs Reject Common Mode Noise 5MHz Full Power Bandwidth ±2.5V Bipolar Input Range Pin Compatible with LTC1604 and LTC1608 36-Pin SSOP Package The LTC1603’s full-scale input range is ± 2.5V. Outstanding AC performance includes 90dB S/(N+D) and – 100dB THD at a sample rate of 250ksps. The unique differential input sample-and-hold can acquire single-ended or differential input signals up to its 15MHz bandwidth. The 68dB common mode rejection allows users to eliminate ground loops and common mode noise by measuring signals differentially from the source. U APPLICATIO S ■ ■ ■ ■ ■ The ADC has µP compatible,16-bit parallel output port. There is no pipeline delay in conversion results. A separate convert start input and a data ready signal (BUSY) ease connections to FlFOs, DSPs and microprocessors. Telecommunications Digital Signal Processing Multiplexed Data Acquisition Systems High Speed Data Acquisition Spectrum Analysis Imaging Systems , LTC and LT are registered trademarks of Linear Technology Corporation. U ■ TYPICAL APPLICATIO 10µF 2.2µF 10Ω + 3 VREF 5V 10µF + 36 AVDD 5V 35 9 AVDD 10µF + 10 DVDD DGND SHDN 33 4 REFCOMP + 7.5k 4.375V 1.75X CONTROL LOGIC AND TIMING 2.5V REF CS 32 µP CONTROL LINES CONVST 31 RD 30 BUSY 27 47µF OVDD 29 + 1 AIN+ DIFFERENTIAL ANALOG INPUT ±2.5V 2 AIN– + – OGND 28 16-BIT SAMPLING ADC AGND 5 AGND 6 OUTPUT BUFFERS B15 TO B0 AGND 7 16-BIT PARALLEL BUS 11 TO 26 AGND VSS 8 D15 TO D0 5V OR 3V 10µF 1603 TA01 34 + 10µF –5V 1603f 1 LTC1603 W U U U W W W ABSOLUTE MAXIMUM RATINGS PACKAGE/ORDER INFORMATION AVDD = DVDD = OVDD = VDD (Notes 1, 2) ORDER PART NUMBER TOP VIEW Supply Voltage (VDD) ................................................ 6V Negative Supply Voltage (VSS) ............................... – 6V Total Supply Voltage (VDD to VSS) .......................... 12V Analog Input Voltage (Note 3) ......................... (VSS – 0.3V) to (VDD + 0.3V) VREF Voltage (Note 4) ................. – 0.3V to (VDD + 0.3V) REFCOMP Voltage (Note 4) ......... – 0.3V to (VDD + 0.3V) Digital Input Voltage (Note 4) ....................– 0.3V to 10V Digital Output Voltage .................. – 0.3V to (VDD + 0.3V) Power Dissipation ............................................. 500mW Operating Temperature Range LTC1603C .............................................. 0°C to 70°C LTC1603I ............................................ – 40°C to 85°C Storage Temperature Range ................ – 65°C to 150°C Lead Temperature (Soldering, 10 sec)................. 300°C AIN+ 1 36 AVDD AIN– 2 35 AVDD VREF 3 34 VSS REFCOMP 4 33 SHDN AGND 5 32 CS AGND 6 31 CONV AGND 7 30 RD AGND 8 29 OVDD DVDD 9 28 OGND DGND 10 27 BUSY D15 (MSB) 11 26 D0 D14 12 25 D1 D13 13 24 D2 D12 14 23 D3 D11 15 22 D4 D10 16 21 D5 D9 17 20 D6 D8 18 LTC1603CG LTC1603IG 19 D7 G PACKAGE 36-LEAD PLASTIC SSOP TJMAX = 125°C, θJA = 95°C/W Consult LTC Marketing for parts specified with wider operating temperature ranges. U CO VERTER CHARACTERISTICS The ● denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25°C. With Internal Reference (Notes 5, 6) PARAMETER CONDITIONS Resolution (No Missing Codes) Integral Linearity Error ● (Note 7) Transition Noise (Note 8) Offset Error (Note 9) Offset Tempco (Note 9) Full-Scale Error Internal Reference External Reference Full-Scale Tempco IOUT(Reference) = 0, Internal Reference MIN TYP 16 16 ● ±1 ● ±0.05 MAX UNITS Bits ±3 LSB 0.7 LSB ±0.125 0.5 ±0.125 % ppm/°C ±0.25 ±0.25 ±15 % % ppm/°C U U A ALOG I PUT The ● denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25°C. SYMBOL PARAMETER CONDITIONS VIN Analog Input Range (Note 2) 4.75 ≤ VDD ≤ 5.25V, – 5.25 ≤ VSS ≤ – 4.75V, VSS ≤ (AIN–, AIN+) ≤ AVDD IIN Analog Input Leakage Current CS = High CIN Analog Input Capacitance Between Conversions During Conversions tACQ MIN TYP MAX ±2.5 V ±1 ● UNITS µA 43 5 pF pF Sample-and-Hold Acquisition Time 380 ns tAP Sample-and-Hold Acquisition Delay Time – 1.5 ns tjitter Sample-and-Hold Acquisition Delay Time Jitter CMRR Analog Input Common Mode Rejection Ratio – 2.5V < (AIN– = AIN+) < 2.5V 5 psRMS 68 dB 1603f 2 LTC1603 W U DY A IC ACCURACY The ● denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25°C. (Note 5) SYMBOL PARAMETER CONDITIONS MIN TYP S/N 5kHz Input Signal 100kHz Input Signal ● 87 90 90 dB dB 5kHz Input Signal 100kHz Input Signal (Note 10) ● 84 90 89 dB dB Total Harmonic Distortion Up to 5th Harmonic 5kHz Input Signal 100kHz Input Signal ● SFDR Spurious Free Dynamic Range 100kHz Input Signal 96 IMD Intermodulation Distortion fIN1 = 29.37kHz, fIN2 = 32.446kHz – 88 Signal-to-Noise Ratio S/(N + D) Signal-to-(Noise + Distortion) Ratio THD – 100 – 94 Full Power Bandwidth Full Linear Bandwidth (S/(N + D) ≥ 84dB U U U I TER AL REFERE CE CHARACTERISTICS MAX UNITS dB dB – 88 dB dB 5 MHz 350 kHz (Note 5) PARAMETER CONDITIONS MIN TYP MAX UNITS VREF Output Voltage IOUT = 0 2.475 2.500 2.515 V VREF Output Tempco IOUT = 0 ±15 ppm/°C VREF Line Regulation 4.75 ≤ VDD ≤ 5.25V – 5.25V ≤ VSS ≤ – 4.75V 0.01 0.01 LSB/V LSB/V VREF Output Resistance 0 ≤ IOUT ≤ 1mA 7.5 kΩ REFCOMP Output Voltage IOUT = 0 4.375 V U U DIGITAL I PUTS A D DIGITAL OUTPUTS The ● denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25°C. (Note 5) SYMBOL PARAMETER CONDITIONS VIH High Level Input Voltage VDD = 5.25V ● VIL Low Level Input Voltage VDD = 4.75V ● IIN Digital Input Current VIN = 0V to VDD ● CIN Digital Input Capacitance VOH High Level Output Voltage VOL Low Level Output Voltage MIN VDD = 4.75V, IOUT = – 10µA VDD = 4.75V, IOUT = – 400µA ● VDD = 4.75V, IOUT = 160µA VDD = 4.75V, IOUT = 1.6mA ● VOUT = 0V to VDD, CS High ● ● TYP MAX 2.4 UNITS V 0.8 V ±1 0 µA 5 pF 4.5 V V 4.0 0.05 0.10 0.4 V V ±10 µA IOZ Hi-Z Output Leakage D15 to D0 COZ Hi-Z Output Capacitance D15 to D0 CS High (Note 11) ISOURCE Output Source Current VOUT = 0V –10 mA ISINK Output Sink Current VOUT = VDD 10 mA 15 pF 1603f 3 LTC1603 U W POWER REQUIRE E TS The ● denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25°C. (Note 5) SYMBOL PARAMETER CONDITIONS MIN MAX UNITS VDD Positive Supply Voltage (Notes 12, 13) 4.75 TYP 5.25 V VSS Negative Supply Voltage (Note 12) – 4.75 – 5.25 V IDD Positive Supply Current Nap Mode Sleep Mode CS = RD = 0V CS = 0V, SHDN = 0V CS = 5V, SHDN = 0V ● 18 1.5 1 30 2.4 100 mA mA µA ISS Negative Supply Current Nap Mode Sleep Mode CS = RD = 0V CS = 0V, SHDN = 0V CS = 5V, SHDN = 0V ● 26 1 1 40 100 100 mA µA µA PD Power Dissipation Nap Mode Sleep Mode CS = RD = 0V CS = 0V, SHDN = 0V CS = 5V, SHDN = 0V ● 220 7.5 0.01 350 12 1 mW mW mW WU TI I G CHARACTERISTICS The ● denotes the specifications which apply over the full operating temperature range, otherwise specifications are at TA = 25°C. (Note 5) SYMBOL PARAMETER CONDITIONS fSMPL(MAX) Maximum Sampling Frequency ● 250 tCONV Conversion Time ● 2.2 tACQ Acquisition Time tACQ+CONV Throughput Time (Acquisition + Conversion) t1 CS to RD Setup Time (Notes 11, 12) ● 0 ns t2 CS↓ to CONVST↓ Setup Time (Notes 11, 12) ● 10 ns t3 SHDN↓ to CS↑ Setup Time (Notes 11, 12) ● 10 t4 SHDN↑ to CONVST↓ Wake-Up Time CS = Low (Note 12) t5 CONVST Low Time (Note 12) t6 CONVST to BUSY Delay CL = 25pF (Note 11) MIN TYP MAX 3.3 3.8 µs 480 ns kHz ● 4 ● ns 40 ns 36 80 ● t7 Data Ready Before BUSY↑ µs ns 400 ● UNITS 60 ns ns ● 32 ns ns t8 Delay Between Conversions (Note 12) ● 200 ns t9 Wait Time RD↓ After BUSY↑ (Note 12) ● –5 ns t10 Data Access Time After RD↓ CL = 25pF 40 50 60 ns ns 45 60 75 ns ns 50 60 70 75 ns ns ns ● CL = 100pF ● t11 Bus Relinquish Time LTC1603C LTC1603I ● ● t12 RD Low Time (Note 12) ● t10 ns t13 CONVST High Time (Note 12) ● 40 ns t14 Aperture Delay of Sample-and-Hold 2 ns 1603f 4 LTC1603 WU TI I G CHARACTERISTICS (Note 5) Note 1: Absolute Maximum Ratings are those values beyond which the life of a device may be impaired. Note 2: All voltage values are with respect to ground with DGND, OGND and AGND wired together unless otherwise noted. Note 3: When these pin voltages are taken below VSS or above VDD, they will be clamped by internal diodes. This product can handle input currents greater than 100mA below VSS or above VDD without latchup. Note 4: When these pin voltages are taken below VSS, they will be clamped by internal diodes. This product can handle input currents greater than 100mA below VSS without latchup. These pins are not clamped to VDD. Note 5: VDD = 5V, VSS = – 5V, fSMPL = 250kHz, and t r = t f = 5ns unless otherwise specified. Note 6: Linearity, offset and full-scale specification apply for a singleended AIN+ input with AIN– grounded. Note 7: Integral nonlinearity is defined as the deviation of a code from a straight line passing through the actual endpoints of the transfer curve. The deviation is measured from the center of the quantization band. Note 8: Typical RMS noise at the code transitions. See Figure 17 for histogram. Note 9: Bipolar offset is the offset voltage measured from – 0.5LSB when the output code flickers between 0000 0000 0000 0000 and 1111 1111 1111 1111. Note 10: Signal-to-Noise Ratio (SNR) is measured at 5kHz and distortion is measured at 100kHz. These results are used to calculate Signal-to-Nosie Plus Distortion (SINAD). Note 11: Guaranteed by design, not subject to test. Note 12: Recommended operating conditions. Note 13: The falling CONVST edge starts a conversion. If CONVST returns high at a critical point during the conversion it can create small errors. For best performance ensure that CONVST returns high either within 250ns after conversion start or after BUSY rises. U W TYPICAL PERFORMANCE CHARACTERISTICS Integral Nonlinearity vs Output Code Differential Nonlinearity vs Output Code 2.0 1.0 1.5 0.8 0.6 1.0 DNL (LSB) INL (LSB) 0.4 0.5 0.0 –0.5 0.2 0.0 –0.2 –0.4 –1.0 –0.6 –1.5 –2.0 –32768 –0.8 –1.0 –16384 0 16384 32767 CODE 1603 G11 –32768 –16384 0 16384 32767 CODE 1603 G10 1603f 5 LTC1603 U U U PIN FUNCTIONS AIN+ (Pin 1): Positive Analog Input. The ADC converts the difference voltage between AIN+ and AIN– with a differential range of ±2.5V. AIN+ has a ±2.5V input range when AIN– is grounded. OVDD (Pin 29): Digital Power Supply for Output Drivers. Bypass to OGND with 10µF tantalum in parallel with 0.1µF ceramic. AIN– (Pin 2): Negative Analog Input. Can be grounded, tied to a DC voltage or driven differentially with AIN+ . RD (Pin 30): Read Input. A logic low enables the output drivers when CS is low. VREF (Pin 3): 2.5V Reference Output. Bypass to AGND with 2.2µF tantalum in parallel with 0.1µF ceramic. REFCOMP (Pin 4): 4.375V Reference Compensation Pin. Bypass to AGND with 47µF tantalum in parallel with 0.1µF ceramic. AGND (Pins 5 to 8): Analog Grounds. Tie to analog ground plane. DVDD (Pin 9): 5V Digital Power Supply. Bypass to DGND with 10µF tantalum in parallel with 0.1µF ceramic. DGND (Pin 10): Digital Ground for Internal Logic. Tie to analog ground plane. D15 to D0 (Pins 11 to 26): Three-State Data Outputs. D15 is the Most Significant Bit. BUSY (Pin 27): The BUSY output shows the converter status. It is low when a conversion is in progress. Data is valid on the rising edge of BUSY. OGND (Pin 28): Digital Ground for Output Drivers. CONVST (Pin 31): Conversion Start Signal. This active low signal starts a conversion on its falling edge when CS is low. CS (Pin 32): The Chip Select Input. Must be low for the ADC to recognize CONVST and RD inputs. SHDN (Pin 33): Power Shutdown. Drive this pin low with CS low for nap mode. Drive this pin low with CS high for sleep mode. VSS (Pin 34): – 5V Negative Supply. Bypass to AGND with 10µF tantalum in parallel with 0.1µF ceramic. AVDD (Pin 35): 5V Analog Power Supply. Bypass to AGND with 10µF tantalum in parallel with 0.1µF ceramic. AVDD (Pin 36): 5V Analog Power Supply. Bypass to AGND with 10µF tantalum in parallel with 0.1µF ceramic and connect this pin to Pin 35 with a 10Ω resistor. 1603f 6 LTC1603 W FU CTIO AL BLOCK DIAGRA U 10µF U 2.2µF 10Ω + 3 VREF 5V 10µF 5V + 36 35 9 AVDD AVDD 10µF + 10 DVDD DGND SHDN 33 4 REFCOMP + 7.5k 1.75X 4.375V CS 32 CONTROL LOGIC AND TIMING 2.5V REF µP CONTROL LINES CONVST 31 RD 30 BUSY 27 47µF OVDD 29 + 1 AIN+ DIFFERENTIAL ANALOG INPUT ± 2.5V OGND 28 + 2 AIN– 16-BIT SAMPLING ADC – AGND 5 AGND OUTPUT BUFFERS B15 TO B0 AGND 6 7 16-BIT PARALLEL BUS D15 TO D0 11 TO 26 AGND VSS 8 5V OR 3V 10µF 1603 TA01 34 + 10µF –5V TEST CIRCUITS Load Circuits for Access Timing Load Circuits for Output Float Delay 5V 5V 1k DN 1k DN 1k CL (A) Hi-Z TO VOH AND VOL TO VOH DN CL DN 1k (B) Hi-Z TO VOL AND VOH TO VOL 1603 TC01 (A) VOH TO Hi-Z CL CL (B) VOL TO Hi-Z 1603 TC02 1603f 7 LTC1603 U U W U APPLICATIONS INFORMATION CONVERSION DETAILS The LTC1603 uses a successive approximation algorithm and internal sample-and-hold circuit to convert an analog signal to a 16-bit parallel output. The ADC is complete with a sample-and-hold, a precision reference and an internal clock. The control logic provides easy interface to microprocessors and DSPs. (Please refer to the Digital Interface section for the data format.) Conversion start is controlled by the CS and CONVST inputs. At the start of the conversion the successive approximation register (SAR) resets. Once a conversion cycle has begun it cannot be restarted. During the conversion, the internal differential 16-bit capacitive DAC output is sequenced by the SAR from the Most Significant Bit (MSB) to the Least Significant Bit (LSB). Referring to Figure 1, the AIN+ and AIN– inputs are acquired during the acquire phase and the comparator offset is nulled by the zeroing switches. In this acquire phase, a duration of 480ns will provide enough time for the sample-and-hold capacitors to acquire the analog signal. During the convert phase the comparator zeroing switches open, putting the comparator into compare mode. The input switches connect the CSMPL capacitors to ground, transferring the differential analog input charge onto the The A/D converter is designed to interface with microprocessors as a memory mapped device. The CS and RD control inputs are common to all peripheral memory interfacing. A separate CONVST is used to initiate a conversion. Internal Clock The A/D converter has an internal clock that runs the A/D conversion. The internal clock is factory trimmed to achieve a typical conversion time of 3.3µs and a maximum conversion time of 3.8µs over the full temperature range. No external adjustments are required. The guaranteed maximum acquisition time is 480ns. In addition, a throughput time (acquisition + conversion) of 4µs and a minimum sampling rate of 250ksps are guaranteed. The LTC1603 operates on ±5V supplies, which makes the device easy to interface to 5V digital systems. This device can also talk to 3V digital systems: the digital input pins (SHDN, CS, CONVST and RD) of the LTC1603 recognize 3V or 5V inputs. The LTC1603 has a dedicated output supply pin (OVDD) that controls the output swings of the digital output pins (D0 to D15, BUSY) and allows the part to talk to either 3V or 5V digital systems. The output is two’s complement binary. SAMPLE HOLD ZEROING SWITCHES CSMPL AIN– DIGITAL INTERFACE 3V Input/Output Compatible CSMPL AIN+ summing junctions. This input charge is successively compared with the binary-weighted charges supplied by the differential capacitive DAC. Bit decisions are made by the high speed comparator. At the end of a conversion, the differential DAC output balances the AIN+ and AIN– input charges. The SAR contents (a 16-bit data word) which represent the difference of AIN+ and AIN– are loaded into the 16-bit output latches. HOLD SAMPLE HOLD HOLD +CDAC + –CDAC COMP – +VDAC Power Shutdown –VDAC 16 SAR OUTPUT LATCHES • • • D15 D0 1603 F01 Figure 1. Simplified Block Diagram The LTC1603 provides two power shutdown modes, Nap and Sleep, to save power during inactive periods. The Nap mode reduces the power by 95% and leaves only the digital logic and reference powered up. The wake-up time from Nap to active is 200ns. In Sleep mode all bias 1603f 8 LTC1603 U W U U APPLICATIONS INFORMATION currents are shut down and only leakage current remains (about 1µA). Wake-up time from Sleep mode is much slower since the reference circuit must power up and settle. Sleep mode wake-up time is dependent on the value of the capacitor connected to the REFCOMP (Pin 4). The wake-up time is 160ms with the recommended 47µF capacitor. SHDN t3 CS 1603 F02a Figure 2a. Nap Mode to Sleep Mode Timing Shutdown is controlled by Pin 33 (SHDN). The ADC is in shutdown when SHDN is low. The shutdown mode is selected with Pin 32 (CS). When SHDN is low, CS low selects nap and CS high selects sleep. SHDN t4 CONVST 1603 F02b Timing and Control Conversion start and data read operations are controlled by three digital inputs: CONVST, CS and RD. A falling edge applied to the CONVST pin will start a conversion after the ADC has been selected (i.e., CS is low). Once initiated, it cannot be restarted until the conversion is complete. Converter status is indicated by the BUSY output. BUSY is low during a conversion. Figure 2b. SHDN to CONVST Wake-Up Timing CS t2 CONVST t1 We recommend using a narrow logic low or narrow logic high CONVST pulse to start a conversion as shown in Figures 5 and 6. A narrow low or high CONVST pulse prevents the rising edge of the CONVST pulse from upsetting the critical bit decisions during the conversion time. Figure 4 shows the change of the differential nonlinearity error versus the low time of the CONVST pulse. As shown, if CONVST returns high early in the conversion (e.g., CONVST low time <500ns), accuracy is unaffected. Similarly, if CONVST returns high after the conversion is over (e.g., CONVST low time >tCONV), accuracy is unaffected. For best results, keep t 5 less than 500ns or greater than tCONV. RD 1603 F03 Figure 3. CS to CONVST Setup Timing CHANGE IN DNL (LSB) 4 3 2 tCONV tACQ 1 0 0 500 1000 1500 2000 2500 3000 3500 4000 CONVST LOW TIME, t5 (ns) 1603 F04 Figure 4. Change in DNL vs CONVST Low Time. Be Sure the CONVST Pulse Returns High Early in the Conversion or After the End of Conversion Figures 5 through 9 show several different modes of operation. In modes 1a and 1b (Figures 5 and 6), CS and RD are both tied low. The falling edge of CONVST starts the conversion. The data outputs are always enabled and data can be latched with the BUSY rising edge. Mode 1a shows operation with a narrow logic low CONVST pulse. Mode 1b shows a narrow logic high CONVST pulse. In mode 2 (Figure 7) CS is tied low. The falling edge of CONVST signal starts the conversion. Data outputs are in 1603f 9 LTC1603 U W U U APPLICATIONS INFORMATION t CONV CS = RD = 0 t5 CONVST t6 t8 BUSY t7 DATA DATA (N + 1) D15 TO D0 DATA N D15 TO D0 DATA (N – 1) D15 TO D0 1603 F05 Figure 5. Mode 1a. CONVST Starts a Conversion. Data Outputs Always Enabled (CONVST = ) tCONV CS = RD = 0 t8 t5 t13 CONVST t6 t6 BUSY t7 DATA (N – 1) D15 TO D0 DATA DATA N D15 TO D0 DATA (N + 1) D15 TO D0 1603 F06 Figure 6. Mode 1b. CONVST Starts a Conversion. Data Outputs Always Enabled (CONVST = ) t13 tCONV t5 CS = 0 t8 CONVST t6 BUSY t9 t 12 t 11 RD t 10 DATA DATA N D15 TO D0 1603 F07 Figure 7. Mode 2. CONVST Starts a Conversion. Data is Read by RD 1603f 10 LTC1603 U U W U APPLICATIONS INFORMATION t8 t CONV CS = 0 RD = CONVST t6 t 11 BUSY t 10 t7 DATA (N – 1) D5 TO D0 DATA DATA N D15 TO D0 DATA N D15 TO D0 DATA (N + 1) D15 TO D0 1603 F08 Figure 8. Mode 2. Slow Memory Mode Timing t CONV CS = 0 t8 RD = CONVST t6 t 11 BUSY t 10 DATA DATA N D15 TO D0 DATA (N – 1) D15 TO D0 1603 F09 Figure 9. ROM Mode Timing three-state until read by the MPU with the RD signal. Mode 2 can be used for operation with a shared data bus. In slow memory and ROM modes (Figures 8 and 9) CS is tied low and CONVST and RD are tied together. The MPU starts the conversion and reads the output with the combined CONVST-RD signal. Conversions are started by the MPU or DSP (no external sample clock is needed). In slow memory mode the processor applies a logic low to RD (= CONVST), starting the conversion. BUSY goes low, forcing the processor into a wait state. The previous conversion result appears on the data outputs. When the conversion is complete, the new conversion results appear on the data outputs; BUSY goes high, releasing the processor and the processor takes RD (= CONVST) back high and reads the new conversion data. In ROM mode, the processor takes RD (= CONVST) low, starting a conversion and reading the previous conversion result. After the conversion is complete, the processor can read the new result and initiate another conversion. DIFFERENTIAL ANALOG INPUTS Driving the Analog Inputs The differential analog inputs of the LTC1603 are easy to drive. The inputs may be driven differentially or as a singleended input (i.e., the AIN – input is grounded). The AIN+ and AIN – inputs are sampled at the same instant. Any unwanted signal that is common mode to both inputs will be reduced by the common mode rejection of the sampleand-hold circuit. The inputs draw only one small current spike while charging the sample-and-hold capacitors at the end of conversion. During conversion the analog inputs draw only a small leakage current. If the source impedance of the driving circuit is low, then the LTC1603 inputs can be driven directly. As source impedance increases so will acquisition time (see Figure 10). For minimum acquisition time with high source impedance, a buffer amplifier should be used. The only requirement is that the amplifier driving the analog input(s) must settle after the small current spike before the next conversion 1603f 11 LTC1603 U W U U APPLICATIONS INFORMATION ACQUISITION TIME (µs) 10 LT ® 1007: Low Noise Precision Amplifier. 2.7mA supply current, ±5V to ±15V supplies, gain bandwidth product 8MHz, DC applications. 1 0.1 LT1097: Low Cost, Low Power Precision Amplifier. 300µA supply current, ±5V to ±15V supplies, gain bandwidth product 0.7MHz, DC applications. 0.01 LT1227: 140MHz Video Current Feedback Amplifier. 10mA supply current, ±5V to ±15V supplies, low noise and low distortion. 1 10 100 1k SOURCE RESISTANCE (Ω) 10k 1603 F10 Figure 10. tACQ vs Source Resistance starts (settling time must be 200ns for full throughput rate). Choosing an Input Amplifier Choosing an input amplifier is easy if a few requirements are taken into consideration. First, to limit the magnitude of the voltage spike seen by the amplifier from charging the sampling capacitor, choose an amplifier that has a low output impedance (< 100Ω) at the closed-loop bandwidth frequency. For example, if an amplifier is used in a gain of +1 and has a unity-gain bandwidth of 50MHz, then the output impedance at 50MHz should be less than 100Ω. The second requirement is that the closed-loop bandwidth must be greater than 15MHz to ensure adequate small-signal settling for full throughput rate. If slower op amps are used, more settling time can be provided by increasing the time between conversions. The best choice for an op amp to drive the LTC1603 will depend on the application. Generally applications fall into two categories: AC applications where dynamic specifications are most critical and time domain applications where DC accuracy and settling time are most critical. The following list is a summary of the op amps that are suitable for driving the LTC1603. More detailed information is available in the Linear Technology databooks, the LinearViewTM CD-ROM and on our web site at: www.linear-tech. com. LT1360: 37MHz Voltage Feedback Amplifier. 3.8mA supply current, ±5V to ±15V supplies, good AC/DC specs. LT1363: 50MHz Voltage Feedback Amplifier. 6.3mA supply current, good AC/DC specs. LT1364/LT1365: Dual and Quad 50MHz Voltage Feedback Amplifiers. 6.3mA supply current per amplifier, good AC/ DC specs. Input Filtering The noise and the distortion of the input amplifier and other circuitry must be considered since they will add to the LTC1603 noise and distortion. The small-signal bandwidth of the sample-and-hold circuit is 15MHz. Any noise or distortion products that are present at the analog inputs will be summed over this entire bandwidth. Noisy input circuitry should be filtered prior to the analog inputs to minimize noise. A simple 1-pole RC filter is sufficient for many applications. For example, Figure 11 shows a 3000pF capacitor from AIN+ to ground and a 100Ω source resistor to limit the input bandwidth to 530kHz. The 3000pF capacitor also acts as a charge reservoir for the input sample-and-hold and isolates the ADC input from sampling glitch sensitive circuitry. High quality capacitors and resistors should be used since these components can add distortion. NPO and silver mica type dielectric capacitors have excellent linearity. Carbon surface mount resistors can also generate distortion from self heating and from damage that may occur during soldering. Metal film surface mount resistors are much less susceptible to both problems. LinearView is a trademark of Linear Technology Corporation. 1603f 12 LTC1603 U U W U APPLICATIONS INFORMATION 100Ω ANALOG INPUT 3000pF 1 AIN+ 2.500V 2 R1 7.5k 3 VREF BANDGAP REFERENCE AIN– LTC1603 3 4 VREF 4.375V 4 REFCOMP REFERENCE AMP REFCOMP R2 12k 47µF 47µF 5 AGND 1603 F11 R3 16k 5 AGND LTC1603 1603 F12a Figure 11. RC Input Filter Figure 12a. LTC1603 Reference Circuit Input Range The ±2.5V input range of the LTC1603 is optimized for low noise and low distortion. Most op amps also perform well over this same range, allowing direct coupling to the analog inputs and eliminating the need for special translation circuitry. 5V ANALOG INPUT VIN LT1019A-2.5 VOUT 1 AIN+ 2 AIN– 3 LTC1603 4 Some applications may require other input ranges. The LTC1603 differential inputs and reference circuitry can accommodate other input ranges often with little or no additional circuitry. The following sections describe the reference and input circuitry and how they affect the input range. Internal Reference The LTC1603 has an on-chip, temperature compensated, curvature corrected, bandgap reference that is factory trimmed to 2.500V. It is connected internally to a reference amplifier and is available at VREF (Pin 3) (see Figure 12a). A 7.5k resistor is in series with the output so that it can be easily overdriven by an external reference or other circuitry (see Figure 12b). The reference amplifier gains the voltage at the VREF pin by 1.75 to create the required internal reference voltage. This provides buffering between the VREF pin and the high speed capacitive DAC. The reference amplifier compensation pin (REFCOMP, Pin 4) must be bypassed with a capacitor to ground. The reference amplifier is stable with capacitors of 22µF or greater. For the best noise performance a 47µF ceramic or 47µF tantalum in parallel with a 0.1µF ceramic is recommended. VREF + 10µF 0.1µF 5 REFCOMP AGND 1603 F12b Figure 12b. Using the LT1019-2.5 as an External Reference The VREF pin can be driven with a DAC or other means shown in Figure 13. This is useful in applications where the peak input signal amplitude may vary. The input span of the ADC can then be adjusted to match the peak input signal, maximizing the signal-to-noise ratio. The filtering of the internal LTC1603 reference amplifier will limit the bandwidth and settling time of this circuit. A settling time of 20ms should be allowed for after a reference adjustment. Differential Inputs The LTC1603 has a unique differential sample-and-hold circuit that allows rail-to-rail inputs. The ADC will always convert the difference of AIN+ – AIN– independent of the common mode voltage (see Figure 15a). The common mode rejection holds up to extremely high frequencies (see Figure 14a). The only requirement is that both inputs 1603f 13 LTC1603 U U W U APPLICATIONS INFORMATION 1 ANALOG INPUT 2V TO 2.7V DIFFERENTIAL 2 AIN+ ANALOG INPUT – AIN 0V TO 5V ±2.5V LTC1603 LTC1450 2V TO 2.7V 3 4 + 1 AIN+ 2 AIN– 3 VREF – VREF LTC1603 4 REFCOMP REFCOMP 10µF 47µF 5 5 AGND AGND 1603 F13 Figure 14b. Selectable 0V to 5V or ±2.5V Input Range 80 Full-Scale and Offset Adjustment 70 Figure 15a shows the ideal input/output characteristics for the LTC1603. The code transitions occur midway between successive integer LSB values (i.e., – FS + 0.5LSB, – FS + 1.5LSB, – FS + 2.5LSB,... FS – 1.5LSB, FS – 0.5LSB). The output is two’s complement binary with 1LSB = FS – (– FS)/65536 = 5V/65536 = 76.3µV. 60 50 40 30 20 10 0 1k 10k 100k INPUT FREQUENCY (Hz) 1M 1603 G14a Figure 14a. CMRR vs Input Frequency can not exceed the AVDD or VSS power supply voltages. Integral nonlinearity errors (INL) and differential nonlinearity errors (DNL) are independent of the common mode voltage, however, the bipolar zero error (BZE) will vary. The change in BZE is typically less than 0.1% of the common mode voltage. Dynamic performance is also affected by the common mode voltage. THD will degrade as the inputs approach either power supply rail, from 96dB with a common mode of 0V to 86dB with a common mode of 2.5V or – 2.5V. Differential inputs allow greater flexibility for accepting different input ranges. Figure 14b shows a circuit that converts a 0V to 5V analog input signal with only an additional buffer that is not in the signal path. In applications where absolute accuracy is important, offset and full-scale errors can be adjusted to zero. Offset error must be adjusted before full-scale error. Figure 15b shows the extra components required for full-scale error adjustment. Zero offset is achieved by adjusting the offset applied to the AIN– input. For zero offset error apply 011...111 011...110 OUTPUT CODE COMMON MODE REJECTION (dB) Figure 13. Driving VREF with a DAC 1603 F14b 000...001 000...000 111...111 111...110 100...001 100...000 – (FS – 1LSB) FS – 1LSB INPUT VOLTAGE (AIN+ – AIN– ) 1603 F15a Figure 15a. LTC1603 Transfer Characteristics 1603f 14 LTC1603 U W U U APPLICATIONS INFORMATION –5V ANALOG INPUT R3 24k R8 50k 1 AIN+ 2 AIN– R4 100Ω 3 R5 R7 47k 50k R6 24k + 47µF 4 LTC1603 VREF REFCOMP 0.1µF 5 AGND 1603 F15b Figure 15b. Offset and Full-Scale Adjust Circuit – 38µV (i.e., – 0.5LSB) at AIN+ and adjust the offset at the AIN– input by varying R8 until the output code flickers between 0000 0000 0000 0000 and 1111 1111 1111 1111. For full-scale adjustment, an input voltage of 2.499886V (FS/2 – 1.5LSBs) is applied to AIN+ and R7 is adjusted until the output code flickers between 0111 1111 1111 1110 and 0111 1111 1111 1111. BOARD LAYOUT AND GROUNDING Wire wrap boards are not recommended for high resolution or high speed A/D converters. To obtain the best performance from the LTC1603, a printed circuit board with ground plane is required. Layout should ensure that digital and analog signal lines are separated as much as possible. Particular care should be taken not to run any digital track alongside an analog signal track or underneath the ADC.The analog input should be screened by AGND. An analog ground plane separate from the logic system ground should be established under and around the ADC. Pin 5 to Pin 8 (AGNDs), Pin 10 (ADC’s DGND) and all other analog grounds should be connected to this single analog ground point. The REFCOMP bypass capacitor and the DVDD bypass capacitor should also be connected to this analog ground plane. No other digital grounds should be connected to this analog ground plane. Low impedance analog and digital power supply common returns are essential to low noise operation of the ADC and the foil width for these tracks should be as wide as possible. In applications where the ADC data outputs and control signals are connected to a continuously active microprocessor bus, it is possible to get errors in the conversion results. These errors are due to feedthrough from the microprocessor to the successive approximation comparator. The problem can be eliminated by forcing the microprocessor into a WAIT state during conversion or by using three-state buffers to isolate the ADC data bus. The traces connecting the pins and bypass capacitors must be kept short and should be made as wide as possible. The LTC1603 has differential inputs to minimize noise coupling. Common mode noise on the AIN+ and AIN– leads will be rejected by the input CMRR. The AIN– input can be used as a ground sense for the AIN+ input; the LTC1603 will hold and convert the difference voltage between AIN+ and AIN– . The leads to AIN+ (Pin 1) and AIN– (Pin 2) should be kept as short as possible. In applications where this is not possible, the AIN+ and AIN– traces should be run side by side to equalize coupling. SUPPLY BYPASSING High quality, low series resistance ceramic, 10µF or 47µF bypass capacitors should be used at the VDD and REFCOMP pins as shown in Figure 16 and in the Typical Application on the first page of this data sheet. Surface mount ceramic capacitors such as Murata GRM235Y5V106Z016 provide excellent bypassing in a small board space. Alternatively, 10µF tantalum capacitors in parallel with 0.1µF ceramic capacitors can be used. Bypass capacitors must be located as close to the pins as possible. The traces connecting the pins and the bypass capacitors must be kept short and should be made as wide as possible. 1603f 15 LTC1603 U U W U APPLICATIONS INFORMATION 1 AIN+ DIGITAL SYSTEM LTC1603 – ANALOG INPUT CIRCUITRY 2 + – AIN V REF REFCOMP AGND 3 4 2.2µF 47µF 5 TO 8 VSS AVDD AVDD DGND OVDD OGND DVDD 34 36 35 9 10µF 10µF 10µF 10µF 10 29 28 10µF 1603 F16 Figure 16. Power Supply Grounding Practice DC PERFORMANCE COUNT 2000 1500 1000 500 0 –5 –4 –3 –2 –1 0 1 CODE 2 3 4 5 1603 F17 Figure 17. Histogram for 4096 Conversions 0 fSAMPLE = 250kHz fIN = 9.959kHz SINAD = 90.2dB THD = –103.2dB –20 AMPLITUDE (dB) The noise of an ADC can be evaluated in two ways: signalto-noise raio (SNR) in frequency domain and histogram in time domain. The LTC1603 excels in both. Figure 18a demonstrates that the LTC1603 has an SNR of over 90dB in frequency domain. The noise in the time domain histogram is the transition noise associated with a high resolution ADC which can be measured with a fixed DC signal applied to the input of the ADC. The resulting output codes are collected over a large number of conversions. The shape of the distribution of codes will give an indication of the magnitude of the transition noise. In Figure 17 the distribution of output codes is shown for a DC input that has been digitized 4096 times. The distribution is Gaussian and the RMS code transition noise is about 0.66LSB. This corresponds to a noise level of 90.9dB relative to full scale. Adding to that the theoretical 98dB of quantization error for 16-bit ADC, the resultant corresponds to an SNR level of 90.1dB which correlates very well to the frequency domain measurements in DYNAMIC PERFORMANCE section. 2500 –40 –60 –80 –100 DYNAMIC PERFORMANCE The LTC1603 has excellent high speed sampling capability. Fast fourier transform (FFT) test techniques are used to test the ADC’s frequency response, distortions and noise at the rated throughput. By applying a low distortion sine wave and analyzing the digital output using an FFT algorithm, the ADC’s spectral content can be examined for frequencies outside the fundamental. Figures 18a and 18b show typical LTC1603 FFT plots. –120 –140 0 20 40 80 100 60 FREQUENCY (kHz) 120 1603 F18a Figure 18a. This FFT of the LTC1603’s Conversion of a Full-Scale 10kHz Sine Wave Shows Outstanding Response with a Very Low Noise Floor When Sampling at 250ksps 1603f 16 LTC1603 U W U U APPLICATIONS INFORMATION 0 Signal-to-Noise Ratio –40 AMPLITUDE (dB) The signal-to-noise plus distortion ratio [S/(N + D)] is the ratio between the RMS amplitude of the fundamental input frequency to the RMS amplitude of all other frequency components at the A/D output. The output is band limited to frequencies from above DC and below half the sampling frequency. Figure 18a shows a typical spectral content with a 250kHz sampling rate and a 5kHz input. The dynamic performance is excellent for input frequencies up to and beyond the Nyquist limit of 125kHz. fSAMPLE = 250kHz fIN = 97.152kHz SINAD = 89dB THD = –96dB –20 –60 –80 –100 –120 –140 0 20 40 60 80 100 120 FREQUENCY (kHz) Effective Number of Bits The effective number of bits (ENOBs) is a measurement of the resolution of an ADC and is directly related to the S/(N + D) by the equation: 1603 F18b Figure 18b. Even with Inputs at 100kHz, the LTC1603’s Dymanic Linearity Remains Robust Total Harmonic Distortion V22 + V32 + V42 + ...Vn2 V1 where V1 is the RMS amplitude of the fundamental frequency and V2 through Vn are the amplitudes of the second through nth harmonics. THD vs Input Frequency is shown in Figure 20. The LTC1603 has good distortion performance up to the Nyquist frequency and beyond. THD = 20Log 15 92 14 86 13 80 12 74 11 68 10 62 9 56 8 1k 10k 100k FREQUENCY (Hz) 50 1M 1603 F19 Figure 19. Effective Bits and Signal/(Noise + Distortion) vs Input Frequency AMPLITUDE (dB BELOW THE FUNDAMENTAL) Total harmonic distortion (THD) is the ratio of the RMS sum of all harmonics of the input signal to the fundamental itself. The out-of-band harmonics alias into the frequency band between DC and half the sampling frequency. THD is expressed as: 98 SINAD (dB) where N is the effective number of bits of resolution and S/(N + D) is expressed in dB. At the maximum sampling rate of 250kHz the LTC1603 maintains above 14 bits up to the Nyquist input frequency of 125kHz (refer to Figure 19). EFFECTIVE BITS N = [S/(N + D) – 1.76]/6.02 16 0 –10 –20 –30 –40 –50 –60 –70 –80 THD 3RD –90 –100 –110 1k 2ND 10k 100k INPUT FREQUENCY (Hz) 1M 1603 F20 Figure 20. Distortion vs Input Frequency 1603f 17 LTC1603 U U W U APPLICATIONS INFORMATION Intermodulation Distortion If the ADC input signal consists of more than one spectral component, the ADC transfer function nonlinearity can produce intermodulation distortion (IMD) in addition to THD. IMD is the change in one sinusoidal input caused by the presence of another sinusoidal input at a different frequency. If two pure sine waves of frequencies fa and fb are applied to the ADC input, nonlinearities in the ADC transfer function can create distortion products at the sum and difference frequencies of mfa ±nfb, where m and n = 0, 1, 2, 3, 0 fSAMPLE = 250kHz fIN1 = 29.3kHz fIN2 = 32.4kHz –20 ( ) IMD fa ± fb = 20Log Amplitude at (fa ± fb) Amplitude at fa Peak Harmonic or Spurious Noise The peak harmonic or spurious noise is the largest spectral component excluding the input signal and DC. This value is expressed in decibels relative to the RMS value of a full-scale input signal. Full-Power and Full-Linear Bandwidth The full-power bandwidth is that input frequency at which the amplitude of the reconstructed fundamental is reduced by 3dB for a full-scale input signal. –40 AMPLITUDE (dB) etc. For example, the 2nd order IMD terms include (fa – fb). If the two input sine waves are equal in magnitude, the value (in decibels) of the 2nd order IMD products can be expressed by the following formula: –60 –80 –100 –120 –140 0 20 40 60 80 FREQUENCY (kHz) 100 120 1603 F21 Figure 21. Intermodulation Distortion Plot The full-linear bandwidth is the input frequency at which the S/(N + D) has dropped to 84dB (13.66 effective bits). The LTC1603 has been designed to optimize input bandwidth, allowing the ADC to undersample input signals with frequencies above the converter’s Nyquist Frequency. The noise floor stays very low at high frequencies; S/(N + D) becomes dominated by distortion at frequencies far beyond Nyquist. 1603f 18 LTC1603 U PACKAGE DESCRIPTION G Package 36-Lead Plastic SSOP (5.3mm) (Reference LTC DWG # 05-08-1640) 12.50 – 13.10* (.492 – .516) 1.25 ±0.12 7.8 – 8.2 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 5.3 – 5.7 7.40 – 8.20 (.291 – .323) 0.42 ±0.03 0.65 BSC 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 RECOMMENDED SOLDER PAD LAYOUT 5.00 – 5.60** (.197 – .221) 2.0 (.079) 0° – 8° 0.09 – 0.25 (.0035 – .010) 0.55 – 0.95 (.022 – .037) NOTE: 1. CONTROLLING DIMENSION: MILLIMETERS MILLIMETERS 2. DIMENSIONS ARE IN (INCHES) 0.65 (.0256) BSC 0.22 – 0.38 (.009 – .015) 0.05 (.002) G36 SSOP 0802 3. DRAWING NOT TO SCALE *DIMENSIONS DO NOT INCLUDE MOLD FLASH. MOLD FLASH SHALL NOT EXCEED .152mm (.006") PER SIDE **DIMENSIONS DO NOT INCLUDE INTERLEAD FLASH. INTERLEAD FLASH SHALL NOT EXCEED .254mm (.010") PER SIDE 1603f Information furnished by Linear Technology Corporation is believed to be accurate and reliable. However, no responsibility is assumed for its use. Linear Technology Corporation makes no representation that the interconnection of its circuits as described herein will not infringe on existing patent rights. 19 LTC1603 U TYPICAL APPLICATION Using the LTC1603 and Two LTC1391s as an 8-Channel Differential 16-Bit ADC System 5V 1 2 3 4 5 6 7 CH7 + 8 CH0 V+ CH1 CH2 D DOUT CH4 DIN CH5 CS CH6 CLK CH7 GND 16 VREF 2 3 4 5 6 7 CH7 – 8 V+ CH1 D CH2 V– CH3 DOUT CH4 DIN CH5 CS CH6 CLK CH7 GND AVDD 10 9 DVDD AVDD DGND SHDN 33 –5V 1µF + 13 4 REFCOMP 12 + 4.375V 47µF 11 7.5k 1.75X CONTROL LOGIC AND TIMING 2.5V REF CS 32 RD 30 BUSY 27 OVDD 29 + 1 AIN 9 2 AIN– + OGND 28 + – 16-BIT SAMPLING ADC OUTPUT BUFFERS B15 TO B0 15 AGND AGND AGND 5 6 7 8 34 –5V 14 + + 12 CS 10 11 TO 26 1603 TA03 10µF –5V DIN 11 LTC1603 AGND VSS 5V OR 3V 10µF 16-BIT PARALLEL BUS D15 TO D0 3000pF 1µF 16 13 µP CONTROL LINES CONVST 31 10 LTC1391 CH0 + 35 15 5V 1 5V 10µF + 36 3000pF CH0 – 5V 10µF 10Ω 3 1µF 14 V– CH3 10µF + + LTC1391 CH0+ 2.2µF CLK µP CONTROL LINES 9 RELATED PARTS SAMPLING ADCs PART NUMBER DESCRIPTION COMMENTS LTC1410 12-Bit, 1.25Msps, ±5V ADC 71.5dB SINAD at Nyquist, 150mW Dissipation LTC1415 12-Bit, 1.25Msps, Single 5V ADC 55mW Power Dissipation, 72dB SINAD LTC1418 14-Bit, 200ksps, Single 5V ADC 15mW, Serial/Para llel ±10V LTC1419 Low Power 14-Bit, 800ksps ADC True 14-Bit Linearity, 81.5dB SINAD, 150mW Dissipation LTC1604 16-Bit, 333ksps, ±5V ADC Pin Compatible with LTC1603 LTC1605 16-Bit, 100ksps, Single 5V ADC ±10V Inputs, 55mW, Byte or Parallel I/O LTC1608 16-Bit, 500ksps, ±5V ADC Pin Compatible with LTC1603 DESCRIPTION COMMENTS DACs PART NUMBER LTC1592 16-Bit Serial SoftSpan DAC ±1LSB Max INL/DNL, Software-Selectable Output Spans LTC1595 16-Bit Serial Multiplying IOUT DAC in SO-8 ±1LSB Max INL/DNL, Low Glitch, DAC8043 16-Bit Upgrade LTC1596 16-Bit Serial Multiplying IOUT DAC ±1LSB Max INL/DNL, Low Glitch, AD7543/DAC8143 16-Bit Upgrade LTC1597 16-Bit Parallel, Multiplying DAC ±1LSB Max INL/DNL, Low Glitch, 4 Quadrant Resistors LTC1650 16-Bit Serial VOUT DAC Low Power, Low Gritch, 4-Quadrant Multiplication TM SoftSpan is a trademark of Linear Technolology Corporation. 1603f 20 Linear Technology Corporation LT/TP 0503 1K • PRINTED IN USA 1630 McCarthy Blvd., Milpitas, CA 95035-7417 (408) 432-1900 ● FAX: (408) 434-0507 ● www.linear.com LINEAR TECHNOLOGY CORPORATION 2003