19-1459; Rev 1; 5/99 KIT ATION EVALU LE B A IL A AV ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier The MAX104 PECL-compatible, 1Gsps, 8-bit analog-todigital converter (ADC) allows accurate digitizing of analog signals with bandwidths to 2.2GHz. Fabricated on Maxim’s proprietary advanced GST-2 bipolar process, the MAX104 integrates a high-performance track/hold (T/H) amplifier and a quantizer on a single monolithic die. The innovative design of the internal T/H, which has an exceptionally wide 2.2GHz full-power input bandwidth, results in high performance (greater than 7.5 effective bits) at the Nyquist frequency. A fully differential comparator design and decoding circuitry reduce out-ofsequence code errors (thermometer bubbles or sparkle codes) and provide excellent metastable performance of one error per 1016 clock cycles. Unlike other ADCs that can have errors resulting in false full- or zero-scale outputs, the MAX104 limits the error magnitude to 1LSB. The analog input is designed for either differential or single-ended use with a ±250mV input voltage range. Dual, differential, PECL-compatible output data paths ensure easy interfacing and include an 8:16 demultiplexer feature that reduces output data rates to one-half the sampling clock rate. The PECL outputs can be operated from any supply between +3V to +5V for compatibility with +3.3V or +5V referenced systems. Control inputs are provided for interleaving additional MAX104 devices to increase the effective system sampling rate. The MAX104 is packaged in a 25mm x 25mm, 192-contact Enhanced Super-Ball Grid Array (ESBGA™) and is specified over the commercial (0°C to +70°C) temperature range. Features ♦ 1Gsps Conversion Rate ♦ 2.2GHz Full-Power Analog Input Bandwidth ♦ >7.5 Effective Bits at fIN = 500MHz (Nyquist Frequency) ♦ ±0.25LSB INL and DNL ♦ 50Ω Differential Analog Inputs ♦ ±250mV Input Signal Range ♦ On-Chip, +2.5V Precision Bandgap Voltage Reference ♦ Latched, Differential PECL Digital Outputs ♦ Low Error Rate: 10-16 Metastable States at 1Gsps ♦ Selectable 8:16 Demultiplexer ♦ Internal Demux Reset Input with Reset Output ♦ 192-Contact ESBGA Package Ordering Information PART TEMP. RANGE MAX104CHC 0°C to +70°C PIN-PACKAGE 192 ESBGA 192-Contact ESBGA Ball Assignment Matrix TOP VIEW Applications Digital RF/IF Signal Processing Direct RF Downconversion High-Speed Data Acquisition Digital Oscilloscopes High-Energy Physics Radar/Sonar/ECM Systems ATE Systems MAX104 Typical Operating Circuit appears at end of data sheet. ESBGA is a trademark of Amkor/Anam. ESBGA ________________________________________________________________ Maxim Integrated Products For the latest literature: http://www.maxim-ic.com, or phone 1-800-998-8800. For small orders, phone 1-800-835-8769. 1 MAX104 General Description MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier ABSOLUTE MAXIMUM RATINGS VCCA to GNDA .........................................................-0.3V to +6V VCCD to GNDD.........................................................-0.3V to +6V VCCI to GNDI ............................................................-0.3V to +6V VCCO to GNDD ........................................-0.3V to (VCCD + 0.3V) AUXEN1, AUXEN2 to GND .....................-0.3V to (VCCD + 0.3V) VEE to GNDI..............................................................-6V to +0.3V Between GNDs......................................................-0.3V to +0.3V VCCA to VCCD .......................................................-0.3V to +0.3V VCCA to VCCI.........................................................-0.3V to +0.3V PECL Digital Output Current ...............................................50mA REFIN to GNDR ........................................-0.3V to (VCCI + 0.3V) REFOUT Current ................................................+100µA to -5mA ICONST, IPTAT to GNDI .......................................-0.3V to +1.0V TTL/CMOS Control Inputs (DEMUXEN, DIVSELECT)..........................................-0.3V to (VCCD + 0.3V) RSTIN+, RSTIN- ......................................-0.3V to (VCCO + 0.3V) VOSADJ Adjust Input ................................-0.3V to (VCCI + 0.3V) CLK+ to CLK- Voltage Difference..........................................±3V CLK+, CLK-.....................................(VEE - 0.3V) to (GNDD + 1V) CLKCOM.........................................(VEE - 0.3V) to (GNDD + 1V) VIN+ to VIN- Voltage Difference ............................................±2V VIN+, VIN- to GNDI................................................................±2V Continuous Power Dissipation (TA = +70°C) 192-Contact ESBGA (derate 61mW/°C above +70°C) ......4.88W (with heatsink and 200 LFM airflow, derate 106mW/°C above +70°C) ........................................8.48W Operating Temperature Range MAX104CHC........................................................0°C to +70°C Operating Junction Temperature.....................................+150°C Storage Temperature Range .............................-65°C to +150°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. DC ELECTRICAL CHARACTERISTICS (VCCA = VCCI = VCCD = +5.0V ±5%, VEE = -5.0V ±5%, VCCO = +3.0V to VCCD, REFIN connected to REFOUT, TA = TMIN to TMAX, unless otherwise noted. Typical values are at TA = +25°C.) PARAMETER SYMBOL CONDITIONS MIN TYP MAX UNITS 0.5 LSB ACCURACY Resolution RES Integral Nonlinearity (Note 1) INL TA = +25°C -0.5 ±0.25 Differential Nonlinearity (Note 1) DNL TA = +25°C -0.5 ±0.25 Missing Codes 8 Bits No missing codes guaranteed 0.5 LSB None Codes 525 mVp-p 51 Ω ANALOG INPUTS Full-Scale Input Range VFSR Note 1 Common-Mode Input Range VCM Signal + offset w.r.t. GNDI Input Resistance RIN VIN+ and VIN- to GNDI, TA = +25°C Input Resistance Temperature Coefficient TCR 475 500 ±0.8 49 50 V 150 ppm/°C VOS ADJUST CONTROL INPUT Input Resistance (Note 2) RVOS Input VOS Adjust Range VOSADJ = 0 to 2.5V 14 25 kΩ ±4 ±5.5 LSB 2.475 2.50 REFERENCE INPUT AND OUTPUT Reference Output Voltage Reference Output Load Regulation Reference Input Resistance 2 REFOUT Driving REFIN input only ∆REFOUT 0 < ISOURCE < 2.5mA RREF Referenced to GNDR 4 5 _______________________________________________________________________________________ 2.525 V 5 mV kΩ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier (VCCA = VCCI = VCCD = +5.0V ±5%, VEE = -5.0V ±5%, VCCO = +3.0V to VCCD, REFIN connected to REFOUT, TA = TMIN to TMAX, unless otherwise noted. Typical values are at TA = +25°C.) PARAMETER SYMBOL CONDITIONS MIN TYP MAX UNITS 48 50 52 Ω CLOCK INPUTS (Note 3) Clock Input Resistance RCLK Input Resistance Temperature Coefficient TCR CLK+ and CLK- to CLKCOM, TA = +25°C 150 ppm/°C TTL/CMOS CONTROL INPUTS (DEMUXEN, DIVSELECT) High-Level Input Voltage VIH Low-Level Input Voltage VIL 2.0 High-Level Input Current IIH VIH = 2.4V Low-Level Input Current IIL VIL = 0 V -1 0.8 V 50 µA 1 µA DEMUX RESET INPUT (Note 4) Digital Input High Voltage VIH Digital Input Low Voltage VIL -1.165 V -1.475 V PECL DIGITAL OUTPUTS (Note 5) Digital Output High Voltage VOH -1.025 -0.880 V Digital Output Low Voltage VOL -1.810 -1.620 V 480 780 mA 108 150 mA POWER REQUIREMENTS Positive Analog Supply Current ICCA Positive Input Supply Current ICCI Negative Input Supply Current IEE Digital Supply Current -290 ICCD -210 mA 205 340 mA 115 mA Output Supply Current (Note 6) ICCO 75 Power Dissipation (Note 6) PDISS 5.25 W Common-Mode Rejection Ratio (Note 7) CMRR VIN+ = VIN- = ±0.1V 40 68 dB Positive Power-Supply Rejection Ratio (Note 8) PSRR+ (Note 9) 40 73 dB Negative Power-Supply Rejection Ratio (Note 8) PSRR- (Note 10) 40 68 dB _______________________________________________________________________________________ 3 MAX104 DC ELECTRICAL CHARACTERISTICS (continued) MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier AC ELECTRICAL CHARACTERISTICS (VCCA = VCCI = VCCD = +5.0V, VEE = -5.0V, VCCO = +3.3V, REFIN connected to REFOUT, fS = 1Gsps, fIN at -1dBFS, TA = +25°C, unless otherwise noted.) PARAMETER SYMBOL CONDITIONS MIN TYP MAX UNITS ANALOG INPUT Analog Input Full-Power Bandwidth Analog Input VSWR Transfer Curve Offset BW-3dB VSWR VOS 2.2 fIN = 500MHz GHz 1.1:1 VOSADJ control input open -1.5 0 V/V +1.5 LSB DYNAMIC SPECIFICATIONS ENOB1000 Effective Number of Bits (Note 11) Signal-to-Noise Ratio (No Harmonics) Total Harmonic Distortion (Note 12) Spurious-Free Dynamic Range Signal-to-Noise Ratio and Distortion (Note 11) Two-Tone Intermodulation 4 fIN = 1000MHz ENOB500 fIN = 500MHz ENOB125 fIN = 125MHz SNR1000 fIN = 1000MHz SNR500 fIN = 500MHz SNR125 fIN = 125MHz THD1000 fIN = 1000MHz THD500 fIN = 500MHz THD125 fIN = 125MHz SFDR1000 fIN = 1000MHz SFDR500 fIN = 500MHz SFDR125 fIN = 125MHz SINAD1000 fIN = 1000MHz SINAD500 fIN = 500MHz SINAD125 fIN = 125MHz IMD Differential 7.52 Single-ended Differential 7.40 7.2 Single-ended Differential 7.55 7.49 7.4 7.74 Single-ended 7.73 Differential 46.4 Single-ended Differential 46.4 43.5 Single-ended Differential 47.0 47.1 44.2 47.4 Differential -52.6 Single-ended -49.6 -50 Single-ended Differential -52.5 -51.3 -61 -67.4 Differential 52.8 Single-ended 52.5 50 Single-ended Differential 52.3 52.3 62 69.5 Differential 47.0 Single-ended 46.3 45.1 Single-ended Differential Single-ended fIN1 = 124MHz, fIN2 = 126MHz, at -7dB below full-scale dB 68.9 Single-ended Differential dB -66.2 Single-ended Differential dB 47.4 Single-ended Differential Bits 47.2 46.9 46.3 dB 48.3 48.3 -57.7 _______________________________________________________________________________________ dB ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier (VCCA = VCCI = VCCD = +5.0V, VEE = -5.0V, VCCO = +3.3V, REFIN connected to REFOUT, fS = 1Gsps, fIN at -1dBFS, TA = +25°C, unless otherwise noted.) PARAMETER SYMBOL CONDITIONS MIN TYP MAX UNITS TIMING CHARACTERISTICS Maximum Sample Rate fMAX Clock Pulse Width Low tPWL Figure 17 0.45 Clock Pulse Width High tPWH Figure 17 0.45 Aperture Delay tAD Figure 4 100 ps Aperture Jitter tAJ Figure 4 <0.5 ps Reset Input Data Setup Time (Note 13) tSU Figure 15 0 ps Reset Input Data Hold Time (Note 13) tHD Figure 15 0 ps CLK to DREADY Propagation Delay tPD1 Figure 17 DREADY to DATA Propagation Delay (Note 14) tPD2 Figure 17 DATA Rise Time DATA Fall Time 1 Gsps ns 5 2.2 -50 150 ns ns 350 ps tRDATA 20% to 80%, CL = 3pF 420 ps tFDATA 20% to 80%, CL = 3pF 360 ps DREADY Rise Time tRDREADY 20% to 80%, CL = 3pF 220 ps DREADY Fall Time tFDREADY 20% to 80%, CL = 3pF Primary Port Pipeline Delay tPDP Figures 6, 7, 8 Auxiliary Port Pipeline Delay tPDA Figures 6, 7, 8 180 ps DIV1, DIV2 modes 7.5 DIV4 mode 7.5 Clock Cycles DIV1, DIV2 modes 8.5 DIV4 mode 9.5 Clock Cycles Static linearity parameters are computed from a “best-fit” straight line through the code transition points. The full-scale range (FSR) is defined as 256 · slope of the line. Note 2: The offset control input is a self-biased voltage divider from the internal +2.5V reference voltage. The nominal open-circuit voltage is +1.25V. It may be driven from an external potentiometer connected between REFOUT and GNDI. Note 3: The clock input’s termination voltage can be operated between -2.0V and GNDI. Observe the absolute maximum ratings on the CLK+ and CLK- inputs. Note 4: Input logic levels are measured with respect to the VCCO power-supply voltage. Note 5: All PECL digital outputs are loaded with 50Ω to VCCO - 2.0V. Measurements are made with respect to the VCCO powersupply voltage. Note 6: The current in the VCCO power supply does not include the current in the digital output’s emitter followers, which is a function of the load resistance and the VTT termination voltage. Note 7: Common-Mode Rejection Ratio is defined as the ratio of the change in the transfer-curve offset voltage to the change in the common-mode voltage, expressed in dB. Note 8: Power-Supply Rejection Ratio is defined as the ratio of the change in the transfer-curve offset voltage to the change in power-supply voltage, expressed in dB. Note 9: Measured with the positive supplies tied to the same potential; VCCA = VCCD = VCCI. VCC varies from +4.75V to +5.25V. Note 10: VEE varies from -5.25V to -4.75V. Note 1: _______________________________________________________________________________________ 5 MAX104 AC ELECTRICAL CHARACTERISTICS (continued) Note 11: Effective Number of Bits (ENOB) and Signal-to-Noise Plus Distortion (SINAD) are computed from a curve fit referenced to the theoretical full-scale range. Note 12: Total Harmonic Distortion (THD) is computed from the first five harmonics. Note 13: Guaranteed by design with a reset pulse one clock period long or greater. Note 14: The DREADY to DATA propagation delay is measured from the 50% point on the rising edge of the DREADY signal (when the output data changes) to the 50% point on a data output bit. This places the falling edge of the DREADY signal in the middle of the data output valid window, within the differences between the DREADY and DATA rise and fall times, which gives maximum setup and hold time for latching external data latches. Typical Operating Characteristics (VCCA = VCCI = VCCD = +5V, VEE = -5V, VCCO = +3.3V, REFIN connected to REFOUT, fS = 1Gsps, TA = +25°C, unless otherwise noted.) 7.75 8.00 -12dBFS 7.75 -6dBFS SINAD (dB) -1dBFS 7.25 -1dBFS 7.25 7.00 45 6.75 6.75 44 6.50 1000 1250 100 -1dBFS 46 7.00 10 -6dBFS 47 7.50 ENOB (Bits) 43 6.50 10 100 10 1000 1250 100 1000 1250 ANALOG INPUT FREQUENCY (MHz) ANALOG INPUT FREQUENCY (MHz) SIGNAL-TO-NOISE PLUS DISTORTION vs. ANALOG INPUT FREQUENCY (DIFFERENTIAL ANALOG INPUT DRIVE) SIGNAL-TO-NOISE RATIO vs. ANALOG INPUT FREQUENCY (SINGLE-ENDED ANALOG INPUT DRIVE) SIGNAL-TO-NOISE RATIO vs. ANALOG INPUT FREQUENCY (DIFFERENTIAL ANALOG INPUT DRIVE) -12dBFS 48 50 MAX104toc04 49 -1dBFS 46 -6dBFS 50 -1dBFS 46 -6dBFS -1dBFS -6dBFS 42 SNR (dB) 46 SNR (dB) 47 38 -12dBFS 45 34 44 43 100 1000 1250 ANALOG INPUT FREQUENCY (MHz) 42 38 -12dBFS 34 30 10 MAX104toc06 ANALOG INPUT FREQUENCY (MHz) MAX104toc05 ENOB (Bits) -12dBFS 48 -6dBFS 7.50 6 49 MAX104toc02 -12dBFS MAX104toc01 8.00 SIGNAL-TO-NOISE PLUS DISTORTION vs. ANALOG INPUT FREQUENCY (SINGLE-ENDED ANALOG INPUT DRIVE) EFFECTIVE NUMBER OF BITS vs. ANALOG INPUT FREQUENCY (DIFFERENTIAL ANALOG INPUT DRIVE) MAX104toc03 EFFECTIVE NUMBER OF BITS vs. ANALOG INPUT FREQUENCY (SINGLE-ENDED ANALOG INPUT DRIVE) SINAD (dB) MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier 30 10 100 ANALOG INPUT FREQUENCY (MHz) 1000 1250 10 100 ANALOG INPUT FREQUENCY (MHz) _______________________________________________________________________________________ 1000 1250 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier SPURIOUS-FREE DYNAMIC RANGE vs. ANALOG INPUT FREQUENCY (SINGLE-ENDED ANALOG INPUT DRIVE) -1dBFS 70 7.75 -6dBFS 50 40 60 7.50 ENOB (Bits) SFDR (dB) -12dBFS -12dBFS 50 30 6.75 6.50 30 1000 1250 100 10 ANALOG INPUT FREQUENCY (MHz) EFFECTIVE NUMBER OF BITS vs. VCCI = VCCA = VCCD (fIN = 125MHz, -1dBFS) DIFFERENTIAL CLOCK DRIVE 7.75 7.75 ENOB (Bits) SINGLE-ENDED CLOCK DRIVE 7.25 MAX104-13 8.00 MAX104toc12 8.00 ENOB (Bits) 1000 CLOCK FREQUENCY (MHz) ANALOG INPUT FREQUENCY (MHz) EFFECTIVE NUMBER OF BITS vs. CLOCK POWER (fIN = 125MHz, -1dBFS) 7.50 100 1000 1250 100 7.50 7.25 7.00 7.00 6.75 6.75 6.50 6.50 -12 -10 -8 -6 -4 -2 0 2 4 6 4.50 8 10 4.70 4.90 5.10 5.30 5.50 CLOCK POWER (dBm) PER SIDE VCC (V) EFFECTIVE NUMBER OF BITS vs. VEE (fIN = 125MHz, -1dBFS) SPURIOUS-FREE DYNAMIC RANGE vs. CLOCK POWER (fIN = 125MHz, -1dBFS) 75 MAX104-14 8.00 7.75 MAX104toc15 10 7.25 7.00 40 73 71 SINGLE-ENDED CLOCK DRIVE 69 7.50 SFDR (dB) ENOB (Bits) SFDR (dB) -6dBFS MAX104toc11 -1dBFS 70 8.00 MAX104toc10 80 MAX104toc09 80 60 EFFECTIVE NUMBER OF BITS vs. CLOCK FREQUENCY (fIN = 125MHz, -1dBFS) SPURIOUS-FREE DYNAMIC RANGE vs. ANALOG INPUT FREQUENCY (DIFFERENTIAL ANALOG INPUT DRIVE) 7.25 7.00 67 DIFFERENTIAL CLOCK DRIVE 65 63 61 59 6.75 57 55 6.50 -5.50 -5.30 -5.10 -4.90 VEE (V) -4.70 -4.50 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 CLOCK POWER (dBm) PER SIDE _______________________________________________________________________________________ 7 MAX104 Typical Operating Characteristics (continued) (VCCA = VCCI = VCCD = +5V, VEE = -5V, VCCO = +3.3V, REFIN connected to REFOUT, fS = 1Gsps, TA = +25°C, unless otherwise noted.) Typical Operating Characteristics (continued) (VCCA = VCCI = VCCD = +5V, VEE = -5V, VCCO = +3.3V, REFIN connected to REFOUT, fS = 1Gsps, TA = +25°C, unless otherwise noted.) 74 72 72 71 71 70 69 70 69 68 68 67 67 66 66 65 4.70 4.90 5.10 5.30 -76.8 -128.0 -5.50 -5.30 -5.10 -4.90 -4.70 0 -4.50 100 200 300 400 500 VCC (V) VEE (V) ANALOG INPUT FREQUENCY (MHz) FFT PLOT (fIN = 494.5068MHz, RECORD LENGTH 8192) FFT PLOT (fIN = 1005.0049MHz, RECORD LENGTH 8192) TWO-TONE INTERMODULATION FFT PLOT (fIN1 = 124MHz, fIN2 = 126MHz, 7dB BELOW FULL-SCALE, RECORD LENGTH 8192) -76.8 -102.4 -51.2 -76.8 -102.4 -128.0 100 200 300 400 500 fIN1 -51.2 -76.8 -102.4 -128.0 -128.0 0 100 200 300 400 500 0 100 200 300 400 500 ANALOG INPUT FREQUENCY (MHz) ANALOG INPUT FREQUENCY (MHz) ANALOG INPUT FREQUENCY (MHz) ANALOG INPUT BANDWIDTH -6dB BELOW FULL-SCALE ANALOG INPUT BANDWIDTH FULL-POWER INTEGRAL NONLINEARITY vs. OUTPUT CODE (LOW-FREQUENCY SERVO LOOP DATA) 0.3 0.2 -2 INL (LSB) AMPLITUDE (dB) -8 0.4 -1 -6 -7 0.5 MAX104toc24 0 MAX104toc22 -5 MAX104toc23 0 fIN2 -25.6 AMPLITUDE (dB) -51.2 0 MAX104toc21 ENOB = 7.51 BITS SNR = 46.3dB THD = -52.8dB SFDR = 53.7dB -25.6 AMPLITUDE (dB) -25.6 0 MAX104toc20 ENOB = 7.51 BITS SNR = 46.8dB THD = -51.9dB SFDR = 52.1dB MAX104toc19 0 AMPLITUDE (dB) 5.50 -51.2 -102.4 65 4.50 ENOB = 7.75 BITS SNR = 47.4dB THD = -66.2dB SFDR = 70.3dB -25.6 AMPLITUDE (dB) 73 SFDR (dB) 73 0 MAX104-17 74 SFDR (dB) 75 MAX104-16 75 FFT PLOT (fIN = 125.8545MHz, RECORD LENGTH 8192) SPURIOUS-FREE DYNAMIC RANGE vs. VEE (fIN = 125MHz, -1dBFS) MAX104toc18 SPURIOUS-FREE DYNAMIC RANGE vs. VCCI = VCCA = VCCD (fIN = 125MHz, -1dBFS) AMPLITUDE (dB) MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier -3 0.1 0 -0.1 -0.2 -4 -9 -0.4 -5 -10 500 1500 2500 ANALOG INPUT FREQUENCY (MHz) 8 -0.3 FULL-POWER BANDWIDTH = 2.2GHz SMALL-SIGNAL BANDWIDTH = 2.4GHz 500 1500 2500 ANALOG INPUT FREQUENCY (MHz) -0.5 0 32 64 96 128 160 192 224 OUTPUT CODE _______________________________________________________________________________________ 256 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier 0.4 0.3 1.4 DREADY (200mV/div) VSWR 0.2 0.1 0 1.3 1.2 -0.1 -0.2 DATA (200mV/div) -0.3 1.1 -0.4 1.0 -0.5 0 0 32 64 96 128 160 192 224 256 500 500ps/div 2000 2500 -61 -62 -63 -63 -64 -64 THD (dB) -62 -65 -66 -65 -66 -67 -67 -68 -68 -69 -69 -70 MAX104-29 -60 MAX104-28 -61 1500 TOTAL HARMONIC DISTORTION vs. VCCI = VCCA = VCCD (fIN = 125MHz, -1dBFS) TOTAL HARMONIC DISTORTION vs. VEE (fIN = 125MHz, -1dBFS) -60 1000 ANALOG INPUT FREQUENCY (MHz) OUTPUT CODE THD (dB) DNL (LSB) VSWR vs. ANALOG INPUT FREQUENCY 1.5 MAX104toc26 MAX104toc25 0.5 DREADY RISE/FALL TIME, DATA OUTPUT RISE/FALL TIME MAX104toc27 DIFFERENTIAL NONLINEARITY vs. OUTPUT CODE (LOW-FREQUENCY SERVO LOOP DATA) -70 -5.50 -5.30 -5.10 -4.90 -4.70 -4.50 4.50 4.70 4.90 VEE (V) 5.10 5.30 5.50 VCC (V) Pin Description CONTACT NAME FUNCTION A1–A4, A6, A7, B1, B2, C1, C2, D1, D2, D3, G1, H1, J2, J3, K1–K3, L2, L3, M1, N1, T2, T3, U1, V1, V2, W1–W4 GNDI Analog Ground. For T/H amplifier, clock distribution, bandgap reference, and reference amplifier. A5, B5, C5, H2, H3, M2, M3, U5, V5, W5 VCCI Analog Supply Voltage, +5V. Supplies T/H amplifier, clock distribution, bandgap reference, and reference amplifier. A8, B8, C8, U6, V6, W6 GNDA Analog Ground. For comparator array. A9, B9, C9, U7, V7, W7 VCCA Analog Supply Voltage, +5V. Supplies analog comparator array. Test Point. Do not connect. A10, E17, F2, P3, R17, R18 TESTPOINT (T.P.) A11, B11, B16, B17, C11, C16, U9, U17, V9, V17, V18, W9 GNDD Digital Ground A12–A19, B19, C19, D19, E19, F19, G19, H19, J19, K19, L19, M19, N19, P19, T19, U19, V19, W10-W19 VCCO PECL Supply Voltage, +3V to +5V _______________________________________________________________________________________ 9 MAX104 Typical Operating Characteristics (continued) (VCCA = VCCI = VCCD = +5V, VEE = -5V, VCCO = +3.3V, REFIN connected to REFOUT, fS = 1Gsps, TA = +25°C, unless otherwise noted.) MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier Pin Description (continued) CONTACT NAME FUNCTION B3, B4, C3, C4, E3, F3, G2, G3, N2, N3, U2, U3, U4, V3, V4 VEE Analog Supply Voltage, -5V. Supplies T/H amplifier, clock distribution, bandgap reference, and reference amplifier. B6, B7 GNDR Reference Ground. Must be connected to GNDI. B10, B18, C10, C17, C18, T17, T18, U8, U18, V8, W8 VCCD Digital Supply Voltage, +5V B12 P0+ Primary Output Data Bit 0 (LSB) B13 A0+ Auxiliary Output Data Bit 0 (LSB) B14 P1+ Primary Output Data Bit 1 B15 A1+ Auxiliary Output Data Bit 1 C6 REFIN 10 Reference Input C7 REFOUT C12 P0- Reference Output Complementary Primary Output Data Bit 0 (LSB) C13 A0- Complementary Auxiliary Output Data Bit 0 (LSB) C14 P1- Complementary Primary Output Data Bit 1 C15 A1- Complementary Auxiliary Output Data Bit 1 D17 DIVSELECT D18 AUXEN2 Tie to VCCO to power the auxiliary port. Tie to GNDD to power down. E1 ICONST Die Temperature Measurement Test Point. See Die Temperature Measurement section. E2 IPTAT Die Temperature Measurement Test Point. See Die Temperature Measurement section. E18 DEMUXEN F1 VOSADJ F17 P2- Complementary Primary Output Data Bit 2 F18 P2+ Primary Output Data Bit 2 G17 A2- Complementary Auxiliary Output Data Bit 2 G18 A2+ Auxiliary Output Data Bit 2 H17 P3- Complementary Primary Output Data Bit 3 H18 P3+ Primary Output Data Bit 3 TTL/CMOS Demux Divide Selection Input 1: Decimation DIV4 mode 0: Demultiplexed DIV2 mode TTL/CMOS Demux Enable Control 1: Enable Demux 0: Disable Demux Offset Adjust Input ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier CONTACT NAME FUNCTION J1 VIN- Differential Input Voltage (-) J17 A3- Complementary Auxiliary Output Data Bit 3 J18 A3+ Auxiliary Output Data Bit 3 K17 DREADY- Complementary Data-Ready Clock K18 DREADY+ Data-Ready Clock L1 VIN+ Differential Input Voltage (+) L17 P4- Complementary Primary Output Data Bit 4 L18 P4+ Primary Output Data Bit 4 M17 A4- Complementary Auxiliary Output Data Bit 4 M18 A4+ Auxiliary Output Data Bit 4 N17 P5- Complementary Primary Output Data Bit 5 N18 P5+ Primary Output Data Bit 5 P1 CLK- Complementary Sampling-Clock Input P2 TESTPOINT (T.P.) This contact must be connected to GNDI. P17 A5- Complementary Auxiliary Output Data Bit 5 Auxiliary Output Data Bit 5 P18 A5+ R1, R2, R3 CLKCOM 50Ω Clock Termination Return R19 AUXEN1 Tie to VCCO to power the auxiliary port. Tie to GNDD to power down. T1 CLK+ Sampling Clock Input U10 RSTIN- Complementary PECL Demux Reset Input U11 RSTOUT- Complementary PECL Reset Output U12 OR- Complementary PECL Overrange Bit U13 A7- Complementary Auxiliary Output Data Bit 7 (MSB) U14 P7- Complementary Primary Output Data Bit 7 (MSB) U15 A6- Complementary Auxiliary Output Data Bit 6 U16 P6- Complementary Primary Output Data Bit 6 V10 RSTIN+ V11 RSTOUT+ PECL Reset Output V12 OR+ PECL Overrange Bit V13 A7+ Auxiliary Output Data Bit 7 (MSB) V14 P7+ Primary Output Data Bit 7 (MSB) V15 A6+ Auxiliary Output Data Bit 6 V16 P6+ Primary Output Data Bit 6 PECL Demux Reset Input ______________________________________________________________________________________ 11 MAX104 Pin Description (continued) MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier REF REF OUT IN BANDGAP REFERENCE REFERENCE AMPLIFIER +2.5V GNDR VOSADJ DIFFERENTIAL PECL OUTPUTS MAX104 BIAS CURRENTS OVERRANGE BIT GNDI OR 2 50Ω T/H AMPLIFIER VIN+ VIN- AUXILIARY DATA PORT 2 8-BIT FLASH ADC 16 50Ω GNDI CLK+ CLKCOM 50Ω CLK- RSTIN+ RSTIN- PRIMARY DATA PORT 16 DATA READY CLOCK 2 DEMUX RESET OUTPUT 2 LOGIC CLOCK DRIVER 50Ω T/H CLOCK DRIVER RESET INPUT DUAL LATCH A0–A7 16 P0–P7 DREADY DEMUX CLOCK DRIVER ADC CLOCK DRIVER RESET PIPELINE DELAYED RESET DEMUX CLOCK GENERATOR RSTOUT DEMUXEN DIVSELECT Figure 1. Simplified Functional Diagram _______________Detailed Description The MAX104 is an 8-bit, 1Gsps flash analog-to-digital converter (ADC) with on-chip track/hold (T/H) amplifier and differential PECL-compatible outputs. The ADC (Figure 1) employs a fully differential 8-bit quantizer and a unique encoding scheme to limit metastable states to typically one error per 1016 clock cycles, with no error exceeding 1LSB max. An integrated 8:16 output demultiplexer simplifies interfacing to the part by reducing the output data rate to one-half the sampling clock rate. This demultiplexer has internal reset capability that allows multiple 12 MAX104s to be time-interleaved to achieve higher effective sampling rates. When clocked at 1Gsps, the MAX104 provides a typical effective number of bits (ENOB) of >7.5 bits at an analog input frequency of 500MHz. The analog input of the MAX104 is designed for differential or single-ended use with a ±250mV full-scale input range. In addition, this ADC features an on-chip +2.5V precision bandgap reference. If desired, an external reference can also be used. ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier MAX104 Principle of Operation OVERRANGE + 255 255 254 DIGITAL OUTPUT 129 128 127 126 3 2 1 0 The T/H amplifier buffers the input signal and allows a full-scale signal input range of ±250mV. The T/H amplifier’s differential 50Ω input termination simplifies interfacing to the MAX104 with controlled impedance lines. Figure 3 shows a simplified diagram of the T/H amplifier stage internal to the MAX104. Aperture width, delay, and jitter (or uncertainty) are parameters that affect the dynamic performance of high-speed converters. Aperture jitter, in particular, directly influences SNR and limits the maximum slew rate (dV/dt) that can be digitized without contributing significant errors. The MAX104’s innovative T/H amplifier design limits aperture jitter typically to less than 0.5ps. Aperture Width Aperture width (tAW) is the time the T/H circuit requires (Figure 4) to disconnect the hold capacitor from the input circuit (for instance, to turn off the sampling bridge and put the T/H unit in hold mode). Aperture Jitter Aperture jitter (tAJ) is the sample-to-sample variation (Figure 4) in the time between the samples. Aperture Delay Aperture delay (tAD) is the time defined between the rising edge of the sampling clock and the instant when an actual sample is taken (Figure 4). Internal Reference The MAX104 features an on-chip +2.5V precision bandgap reference, which can be used by connecting (+FS - 1LSB) +FS On-Chip Track/Hold Amplifier As with all ADCs, if the input waveform is changing rapidly during conversion, effective number of bits (ENOB) and signal-to-noise ratio (SNR) specifications will degrade. The MAX104’s on-chip, wide-bandwidth (2.2GHz) T/H amplifier reduces this effect and increases the ENOB performance significantly, allowing precise capture of fast analog data at high conversion rates. 0 (-FS + 1LSB) The MAX104’s flash or parallel architecture provides the fastest multibit conversion of all common integrated ADC designs. The key to this high-speed flash architecture is the use of an innovative, high-performance comparator design. The flash converter and downstream logic translate the comparator outputs into a parallel 8-bit output code and pass this binary code on to the optional 8:16 demultiplexer, where primary and auxiliary ports output PECL-compatible data at up to 500Msps per port (depending on how the demultiplexer section is set on the MAX104). The ideal transfer function appears in Figure 2. ANALOG INPUT Figure 2. Transfer Function ALL INPUTS ARE ESD PROTECTED (NOT SHOWN IN THIS INPUT SAMPLING SIMPLIFIED DRAWING). AMPLIFIER BRIDGE BUFFER AMPLIFIER VIN+ VIN- TO COMPARATORS 50Ω 50Ω CHOLD GNDI GNDI CLK+ CLK- CLOCK SPLITTER 50Ω TO COMPARATORS 50Ω CLKCOM Figure 3. Internal Structure of the 2.2GHz T/H Amplifier CLK CLK tAW ANALOG INPUT tAD tAJ SAMPLED DATA (T/H) T/H TRACK HOLD TRACK APERTURE DELAY (tAD) APERTURE WIDTH (tAW) APERTURE JITTER (tAJ) Figure 4. T/H Aperture Timing ______________________________________________________________________________________ 13 MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier REFOUT to REFIN. This connects the reference output to the positive input of the reference buffer. The buffer’s negative input is internally connected to GNDR. GNDR must be connected to GNDI on the user’s application board. REFOUT can source up to 2.5mA to supply external devices if required. An adjustable external reference can be used to adjust the ADC’s full-scale range. To use an external reference supply, simply connect a high-precision reference to the REFIN pin and leave the REFOUT pin floating. In this configuration, REFOUT must not be simultaneously connected to avoid conflicts between the two references. REFIN has a typical input resistance of 5kΩ and accepts input voltages of +2.5V ±200mV. Using the MAX104’s internal reference is recommended for best performance. Digital Outputs The MAX104 provides data in offset binary format to differential PECL outputs. A simplified circuit schematic of the PECL output cell is shown in Figure 5. All PECL outputs are powered from VCCO, which may be operated from any voltage between +3.0V to VCCD for flexible interfacing with either +3.3V or +5V systems. The nominal VCCO supply voltage is +3.3V. All PECL outputs on the MAX104 are open-emitter types and must be terminated at the far end of each transmission line with 50Ω to VCCO - 2V. Table 1 lists all MAX104 PECL outputs and their functions. VCCO 500Ω 500Ω A_+/P_+ DIFF. PAIR GNDD A_-/P_- 1.8mA GNDD GNDD Figure 5. Simplified PECL Output Structure Demultiplexer Operation The MAX104 features an internal data demultiplexer that provides for three different modes of operation (see the following sections on Demultiplexed DIV2 Mode, Non-Demultiplexed DIV1 Mode, and Decimation DIV4 Mode) controlled by two TTL/CMOS-compatible inputs: DEMUXEN and DIVSELECT. DEMUXEN enables or disables operation of the internal 1:2 demultiplexer. A logic high on DEMUXEN activates the internal demultiplexer, and a logic low deactivates it. With the internal demultiplexer enabled, DIVSELECT controls the selection of the operational mode. DIVSELECT low selects demultiplexed DIV2 mode, and DIVSELECT high selects decimation DIV4 mode (Table 2). Table 1. PECL Output Functions PECL OUTPUT SIGNALS P0+, P0-, P1+, P1-, P2+, P2-, P3+, P3-, P4+, P4-, P5+, P5-, P6+, P6-, P7+, P7- Primary-Port Differential Outputs from LSB to MSB. A “+” indicates the true outputs; a “-” denotes the complementary outputs. A0+, A0-, A1+, A1-, A2+, A2-, A3+, A3-, A4+, A4-, A5+, A5-, A6+, A6-, A7+, A7- Auxiliary-Port Differential Outputs from LSB to MSB. A “+” indicates the true outputs; a “-” denotes the complementary outputs. DREADY+, DREADYOR+, ORRSTOUT+, RSTOUT14 FUNCTIONAL DESCRIPTION Data-Ready Clock True and Complementary Outputs. These signal lines are used to latch the output data from the primary to the auxiliary output ports. Data changes on the rising edge of the DREADY clock. Overrange True and Complementary Outputs Reset Output True and Complementary Outputs ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier ADC SAMPLE NUMBER CLKCLK n n+1 n+2 PECL termination power supply (VCCO - 2V) may be removed from all auxiliary output ports. Demultiplexed DIV2 Mode The MAX104 features an internally selectable DIV2 mode (Table 2) that reduces the output data rate to one-half of the sample clock rate. The demultiplexed outputs are presented in dual 8-bit format with two consecutive samples appearing in the primary and auxiliary output ports on the rising edge of the data-ready clock (Figure 7). The auxiliary data port contains the previous sample, and the primary output contains the most recent data sample. AUXEN1 and AUXEN2 must be connected to VCCO to power-up the auxiliary port PECL output drivers. ADC SAMPLES ON THE RISING EDGE OF CLK+ n+3 n+4 n+5 n+6 n+7 n+8 n+9 n+10 n+11 n+12 n+13 AUXILIARY DATA PORT n n+1 n+2 n+3 n+4 PRIMARY DATA PORT n+1 n+2 n+3 n+4 n+5 n+10 n+11 n+12 n+13 CLK+ DREADY+ DREADY DREADY- NOTE: THE AUXILIARY PORT DATA IS DELAYED ONE ADDITIONAL CLOCK CYCLE FROM THE PRIMARY PORT DATA. GROUNDING AUXEN1 AND AUXEN2 WILL POWER DOWN THE AUXILIARY PORT TO SAVE POWER. Figure 6. Non-Demuxed, DIV1-Mode Timing Diagram ADC SAMPLE NUMBER CLKCLK n n+1 n+2 ADC SAMPLES ON THE RISING EDGE OF CLK+ n+3 n+4 n+5 n+6 n+7 n+8 n+9 CLK+ DREADY+ DREADY DREADYAUXILIARY DATA PORT n-1 n+1 n+3 PRIMARY DATA PORT n n+2 n+4 NOTE: THE LATENCY TO THE PRIMARY PORT IS 7.5 CLOCK CYCLES, AND THE LATENCY TO THE AUXILIARY PORT IS 8.5 CLOCK CYCLES. BOTH THE PRIMARY AND AUXILIARY DATA PORTS ARE UPDATED ON THE RISING EDGE OF THE DREADY+ CLOCK. Figure 7. Demuxed DIV2-Mode Timing Diagram ______________________________________________________________________________________ 15 MAX104 Non-Demultiplexed DIV1 Mode The MAX104 may be operated at up to 500Msps in nondemultiplexed DIV1 mode (Table 2). In this mode, the internal demultiplexer is disabled, and sampled data is presented to the primary port only, with the data repeated at the auxiliary port but delayed by one clock cycle (Figure 6). Since the auxiliary output port contains the same data stream as the primary output port, the auxiliary port can be shut down to save power by connecting AUXEN1 and AUXEN2 to digital ground (GNDD). This powers down the internal bias cells and causes both outputs (true and complementary) of the auxiliary port to pull up to a logic-high level. To save additional power, the external 50Ω termination resistors connected to the MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier Decimation DIV4 Mode The MAX104 also offers a special decimated, demultiplexed output (Figure 8) that discards every other input sample and outputs data at one-quarter the input sampling rate for system debugging at slower output data rates. With an input clock of 1GHz, the effective output data rate will be reduced to 250MHz per output port in the DIV4 mode (Table 2). Since every other sample is discarded, the effective sampling rate is 500Msps. the OR bit will flag an overrange condition if either the primary or auxiliary port contains an overranged sample (Table 2). In non-demultiplexed DIV1 mode, the OR port will flag an overrange condition only when the primary output port contains an overranged sample. Applications Information Single-Ended Analog Inputs The MAX104 T/H amplifier is designed to work at full speed for both single-ended and differential analog inputs (Figure 9). Inputs VIN+ and VIN- feature on-chip, laser-trimmed 50Ω termination resistors to provide excellent voltage standing wave ratio (VSWR) performance. Overrange Operation A single differential PECL overrange output bit (OR+, OR-) is provided for both primary and auxiliary demultiplexed outputs. The operation of the overrange bit depends on the status of the internal demultiplexer. In demultiplexed DIV2 mode and decimation DIV4 mode, ADC SAMPLE NUMBER CLK- n n+1 n+2 ADC SAMPLES ON THE RISING EDGE OF CLK+ n+3 n+4 n+5 n+6 n+7 n+8 n+9 n+10 n+11 n+12 n+13 CLK CLK+ DREADY+ DREADY DREADYAUXILIARY DATA PORT n-2 n+2 PRIMARY DATA PORT n n+4 NOTE: THE LATENCY TO THE PRIMARY PORT REMAINS 7.5 CLOCK CYCLES, WHILE THE LATENCY OF THE AUXILIARY PORT INCREASES TO 9.5 CLOCK CYCLES. THIS EFFECTIVELY DISCARDS EVERY OTHER SAMPLE AND REDUCES THE OUTPUT DATA RATE TO 1/4 THE SAMPLE CLOCK RATE. Figure 8. Decimation DIV4-Mode Timing Diagram Table 2. Demultiplexer Operation DEMUXEN DIVSELECT DEMUX MODE OVERRANGE BIT OPERATION Low X DIV1 500Msps (max) Flags overrange data appearing in primary port only. High Low DIV2 500Msps/port High DIV4 250Msps/port High Flags overrange data appearing in either the primary or auxiliary port. X = Don’t care 16 ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier MAX104 ANALOG INPUTS ARE ESD PROTECTED (NOT SHOWN IN THIS SIMPLIFIED DRAWING). +2.8V VIN+ +250mV VIN+ 500mVp-p FS ANALOG INPUT RANGE 50Ω 500mV OV VIN- GNDI -250mV t VIN = ±250mV 50Ω VIN- Figure 10a. Single-Ended Analog Input Signals VEE Figure 9. Simplified Analog Input Structure (Single-Ended/ Differential) In a typical single-ended configuration, the analog input signal (Figure 10a) enters the T/H amplifier stage at the in-phase input (VIN+), while the inverted phase input (VIN-) is reverse-terminated to GNDI with an external 50Ω resistor. Single-ended operation allows for an input amplitude of ±250mV. Table 3 shows a selection of input voltages and their corresponding output codes for single-ended operation. Differential Analog Inputs To obtain a full-scale digital output with differential input drive (Figure 10b), 250mVp-p must be applied between VIN+ and VIN- (VIN+ = +125mV and VIN- = -125mV). Midscale digital output codes (01111111 or 10000000) occur when there is no voltage difference between VIN+ and VIN-. For a zero-scale digital output code, the VIN+ VIN- +125mV ±250mV FS ANALOG INPUT RANGE 250mV -250mV -125mV OV t Figure 10b. Differential Analog Input Signals in-phase (VIN+) input must see -125mV and the inverted input (VIN-) must see +125mV. A differential input drive is recommended for best performance. Table 4 represents a selection of differential input voltages and their corresponding output codes. Table 3. Ideal Input Voltage and Output Code Results for Single-Ended Operation VIN+ VIN- OVERRANGE BIT OUTPUT CODE +250mV 0V 1 11111111 (full scale) +250mV - 1LSB 0V 0 11111111 0V 0V 0 01111111 toggles 10000000 -250mV + 1LSB 0V 0 0000001 -250mV 0V 0 00000000 (zero scale) ______________________________________________________________________________________ 17 MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier Table 4. Ideal Input Voltage and Output Code Results for Differential Operation VIN+ VIN- OVERRANGE BIT OUTPUT CODE +125mV -125mV 1 11111111 (full scale) +125mV - 0.5LSB -125mV + 0.5LSB 0 11111111 0V 0V 0 01111111 toggles 10000000 -125mV + 0.5LSB +125mV - 0.5LSB 0 00000001 -125mV +125mV 0 00000000 (zero scale) Offset Adjust The MAX104 provides a control input (VOSADJ) to compensate for system offsets. The offset adjust input is a self-biased voltage divider from the internal +2.5V precision reference. The nominal open-circuit voltage is onehalf the reference voltage. With an input resistance of typically 25kΩ, this pin may be driven by an external 10kΩ potentiometer (Figure 11) connected between REFOUT and GNDI to correct for offset errors. This control provides a typical ±5.5LSB offset adjustment range. Clock Operation The MAX104 clock inputs are designed for either singleended or differential operation (Figure 12) with flexible input drive requirements. Each clock input is terminated with an on-chip, laser-trimmed 50Ω resistor to CLKCOM (clock-termination return). The CLKCOM termination voltage can be connected anywhere between ground and -2V for compatibility with standard ECL drive levels. The clock inputs are internally buffered with a preamplifier to ensure proper operation of the data converter with even small-amplitude sine-wave sources. The MAX104 was designed for single-ended, low-phasenoise sine-wave clock signals with as little as 100mV amplitude (-10dBm). This eliminates the need for an external ECL clock buffer and its added jitter. Single-Ended Clock Inputs (Sine-Wave Drive) Excellent performance is obtained by AC- or DC-coupling a low-phase-noise sine-wave source into a single clock input (Figure 13a, Table 5). For proper DC balance, the undriven clock input should be externally 50Ω reverse-terminated to GNDI. The dynamic performance of the data converter is essentially unaffected by clock-drive power levels from REFOUT MAX104 POT 10k VOSADJ GNDI Figure 11. Offset Adjust with External 10kΩ Potentiometer CLK+ 50Ω +0.8V CLKCOM 50Ω GNDI CLK- CLOCK INPUTS ARE ESD PROTECTED (NOT SHOWN IN THIS SIMPLIFIED DRAWING). Figure 12. Simplified Clock Input Structure (Single-Ended/ Differential) 18 ______________________________________________________________________________________ VEE ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier Differential Clock Inputs (Sine-Wave Drive) The advantages of differential clock drive (Figure 13b, Table 5) can be obtained by using an appropriate balun or transformer to convert single-ended sine-wave sources into differential drives. The precision on-chip, laser-trimmed 50Ω clock-termination resistors ensure excellent amplitude matching. Refer to Single-Ended Clock Inputs (Sine-Wave Drive) for proper input amplitude requirements. Single-Ended Clock Inputs (ECL Drive) Configure the MAX104 for single-ended ECL clock drive by connecting the clock inputs as shown in Figure 13c (Table 5). A well bypassed VBB supply (-1.3V) is essential to avoid coupling noise into the undriven clock input, which would degrade dynamic performance. Differential Clock Inputs (ECL Drive) The MAX104 may be driven from a standard differential (Figure 13d, Table 5) ECL clock source by setting the clock termination voltage at CLKCOM to -2V. Bypass the clock-termination return (CLKCOM) as close as possible to the ADC with a 0.01µF capacitor connected to GNDI. AC-Coupling Clock Inputs The clock inputs CLK+ and CLK- can also be driven with positive referenced ECL (PECL) logic if the clock CLK+ CLK+ +0.5V CLK- +0.5V CLK- = 0V -0.5V t NOTE: CLKCOM = 0V -0.5V NOTE: CLKCOM = 0V Figure 13a. Single-Ended Clock Input Signals Figure 13b. Differential Clock Input Signals CLK+ CLK+ -0.8V t -0.8V CLK- CLK- = -1.3V -1.8V t NOTE: CLKCOM = -2V Figure 13c. Single-Ended ECL Clock Drive -1.8V t NOTE: CLKCOM = -2V Figure 13d. Differential ECL Clock Drive ______________________________________________________________________________________ 19 MAX104 -10dBm (100mV clock signal amplitude) to +10dBm (1V clock signal amplitude). The MAX104 dynamic performance specifications are determined by a singleended clock drive of +4dBm (500mV clock signal amplitude). To avoid saturation of the input amplifier stage, limit the clock power level to a maximum of +10dBm. MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier Table 5. DC-Coupled Clock Drive Options CLOCK DRIVE CLK+ CLK- CLKCOM REFERENCE Single-Ended Sine Wave -10dBm to +4dBm External 50Ω to GNDI GNDI Figure 13a Differential Sine Wave -10dBm to +4dBm -10dBm to +4dBm GNDI Figure 13b Single-Ended ECL ECL Drive -1.3V -2V Figure 13c Differential ECL ECL Drive ECL Drive -2V Figure 13d inputs are AC-coupled. Under this condition, connect CLKCOM to GNDI. Single-ended ECL/PECL/sine-wave drive is also possible if the undriven clock input is reverse-terminated to GNDI through a 50Ω resistor in series with a capacitor whose value is identical to that used to couple the driven input. VCCO Demux Reset Operation The MAX104 features an internal 1:2 demultiplexer that reduces the data rate of the output digital data to onehalf the sample clock rate. Demux reset is necessary when interleaving multiple MAX104s and/or synchronizing external demultiplexers. The simplified block diagram of Figure 1 shows that the demux reset signal path consists of four main circuit blocks. From input to output, they are the reset input dual latch, the reset pipeline, the demux clock generator, and the reset output. The signals associated with the demux reset operation and the control of this section are listed in Table 6. Reset Input Dual Latch The reset input dual-latch circuit block accepts differential PECL reset inputs referenced to the same VCCO power supply that powers the MAX104 PECL outputs. For applications that do not require a synchronizing reset, the reset inputs can be left open. In this case, they will self-bias to a proper level with internal 50kΩ resistors and 20µA current source. This combination creates a -1V difference between RSTIN+ and RSTINto disable the internal reset circuitry. When driven with PECL logic levels terminated with 50Ω to (VCCO - 2V), the internal biasing network can easily be overdriven. Figure 14 shows a simplified schematic of the reset input structure. To properly latch the reset input data, the setup time (tSU) and the data hold time (tHD) must be met with respect to the rising edge of the sample clock. The timing diagram of Figure 15 shows the timing relationship of the reset input and sampling clock. 50k 50k RSTIN+ RSTIN- 20µA RESET INPUTS ARE ESD PROTECTED (NOT SHOWN IN THIS SIMPLIFIED DRAWING). GNDD Figure 14. Simplified Reset Input Structure RSTIN+ 50% 50% RSTIN- tSU tHD CLK+ 50% CLK- Figure 15. Reset Input Timing Definitions 20 ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier SIGNAL NAME TYPE FUNCTION CLK+, CLK- Sampling clock inputs Master ADC timing signal. The ADC samples on the rising edge of CLK+. DREADY+, DREADY- Differential PECL outputs Data-Ready PECL Output. Output data changes on the rising edge of DREADY+. RSTIN+, RSTIN- Differential PECL inputs Demux reset input signals. Resets the internal demux when asserted. RSTOUT+, RSTOUT- Differential PECL outputs Reset outputs for resetting additional external demux devices Reset Pipeline The next section in the reset signal path is the reset pipeline. The purpose of this block is to add clock cycles of latency to the reset signal, to match the latency of the converted analog data through the ADC. In this way, when reset data arrives at the RSTOUT+/ RSTOUT- PECL output, it will be time-aligned with the analog data present in the primary and auxiliary ports at the time the reset input was deasserted at RSTIN+/ RSTIN-. Demux Clock Generator The demux clock generator creates the DIV1, DIV2, or DIV4 clocks required for the different modes of demux and non-demux operation. The TTL/CMOS control inputs DEMUXEN and DIVSELECT control the demuxed mode selection, as described in Table 2. The timing diagrams in Figure 16 and Figure 17 show the output timing and data alignment in DIV1, DIV2, and DIV4 modes, respectively. The phase relationship between the sampling clock at the CLK+/CLK- inputs and the data-ready clock at the DREADY+/DREADY- outputs will be random at device power-up. As with all divide-by-two circuits, two possible phase relationships exist between these clocks. The difference between the phases is simply the inversion of the DIV2-DREADY clock. The timing diagram in Figure 16 shows this relationship. Reset all MAX104 devices to a known DREADY phase after initial power-up for applications such as interleaving, where two or more MAX104 devices are used to achieve higher effective sampling rates. This synchronization is necessary to set the order of output samples between the devices. Resetting the converters accomplishes this synchronization. The reset signal is used to force the internal counter in the demux clock-generator block to a known phase state. tPWH CLK+ CLK+ CLK- CLK- 50% tPWL tPD1 tPD1 DREADY + DREADY"PHASE 1" DREADY+ DREADY - 50% tPD2 tFDREADY tRDREADY DREADY + AUXILIARY PORT DATA 80% "PHASE 2" 80% 20% 20% DREADY PRIMARY PORT DATA Figure 16. CLK and DREADY Timing in Demuxed DIV2 Mode Showing Two Possible DREADY Phases Figure 17. Output Timing for All Modes (DIV1, DIV2, DIV4) ______________________________________________________________________________________ 21 MAX104 Table 6. Demux Operating and Reset Control Signal MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier Reset Output Finally, the reset signal is presented in differential PECL format to the last block of the reset signal path. RSTOUT+/RSTOUT- output the time-aligned reset signal used for resetting additional external demuxes in applications that need further output data rate reduction. Many demux devices require that their reset signal be asserted for several clock cycles while they are clocked. To accomplish this, the MAX104 DREADY clock will continue to toggle while RSTOUT is asserted. When a single MAX104 device is used, no synchronizing reset is required, because the order of the samples in the output ports is unchanged regardless of the phase of the DREADY clock. In DIV2 mode, the data in the auxiliary port is delayed by 8.5 clock cycles, while the data in the primary port is delayed by 7.5 clock cycles. The older data is always in the auxiliary port, regardless of the phase of the DREADY clock. The reset output signal, RSTOUT, is delayed by one less clock cycle (6.5 clock cycles) than the primary ADC SAMPLE NUMBER CLKCLK CLK+ RESET INPUT n n+1 n+2 tSU RSTIN- port. The reduced latency of RSTOUT serves to mark the start of synchronized data in the primary and auxiliary ports. When the RSTOUT signal returns to a zero, the DREADY clock phase is reset. Since there are two possible phases of the DREADY clock with respect to the input clock, there are two possible timing diagrams to consider. The first timing diagram (Figure 18) shows the RSTOUT timing and data alignment of the auxiliary and primary output ports when the DREADY clock phase is already reset. For this example, the RSTIN pulse is two clock cycles long. Under this condition, the DREADY clock continues uninterrupted, as does the data stream in the auxiliary and primary ports. The second timing diagram (Figure 19) shows the results when the DREADY phase is opposite from the reset phase. In this case, the DREADY clock “swallows” a clock cycle of the sample clock, resynchronizing to the reset phase. Note that the data stream in the auxiliary and primary ports has reversed. Before reset was ADC SAMPLES ON THE RISING EDGE OF CLK+ n+3 n+4 n+5 n+6 n+7 n+8 n+9 n+10 n+11 n+12 n+13 tHD RSTIN+ DREADYDREADY DREADY+ AUXILIARY DATA PORT n-1 n+1 n+3 PRIMARY DATA PORT n n+2 n+4 RESET OUT DATA PORT RSTOUTRSTOUT+ NOTE: THE LATENCY TO THE RESET OUTPUT IS 6.5 CLOCK CYCLES. THE LATENCY TO THE PRIMARY PORT IS 7.5 CLOCK CYCLES, AND THE LATENCY TO THE AUXILIARY PORT IS 8.5 CLOCK CYCLES. ALL DATA PORTS ARE UPDATED ON THE RISING EDGE OF THE DREADY+ CLOCK. Figure 18. Reset Output Timing in Demuxed DIV2 Mode (DREADY Aligned) 22 ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier CLKCLK CLK+ RESET INPUT n n+1 n+2 tSU RSTIN- MAX104 ADC SAMPLE NUMBER ADC SAMPLES ON THE RISING EDGE OF CLK+ n+3 n+4 n+5 n+6 n+7 n+8 n+9 n+10 n+11 n+12 n+13 tHD RSTIN+ DREADY+ CLOCK PULSE “SWALLOWED” DREADY DREADY- OUT-OF-SEQUENCE SAMPLE AUXILIARY DATA PORT n-2 n n+2 PRIMARY DATA PORT n-1 n+1 n+4 RESET OUT DATA PORT RSTOUTRSTOUT+ NOTE: DREADY PHASE WAS ADJUSTED TO MATCH THE RESET PHASE BY “SWALLOWING” ONE INPUT CLOCK CYCLE. THE AUXILIARY PORT CONTAINS AN OUT-OF-SEQUENCE SAMPLE AS A RESULT OF THE DELAY. Figure 19. Reset Output Timing in Demuxed DIV2 Mode (DREADY Realigned) asserted, the auxiliary port contained “even” samples while the primary port contained “odd” samples. After the RSTOUT is deasserted (which marks the start of the DREADY clock’s reset phase), note that the order of the samples in the ports has been reversed. The auxiliary port also contains an out-of-sequence sample. This is a consequence of the “swallowed” clock cycle that was needed to resynchronize DREADY to the reset phase. Also note that the older sample data is always in the auxiliary port, regardless of the DREADY phase. These examples illustrate the combinations that result with a reset input signal of two clock cycles. It is also possible to reset the internal MAX104 demux successfully with a reset pulse of only one clock cycle, provided that the setup-time and hold-time requirements are met with respect to the sample clock. However, this is not recommended when additional external demuxes are used. Note that many external demuxes require that their reset signals be asserted while they are clocked, and may require more than one clock cycle of reset. More importantly, if the phase of the DREADY clock is such that a clock pulse will be “swallowed” to resynchronize, then no reset output will occur at all. In effect, the RSTOUT signal will be “swallowed” with the clock pulse. The best method to ensure complete system reset is to assert RSTIN for the appropriate number of DREADY clock cycles required to complete reset of the external demuxes. Die Temperature Measurement For applications that require monitoring of the die temperature, it is possible to determine the die temperature of the MAX104 under normal operating conditions by observing the currents ICONST and IPTAT, at contacts ICONST and IPTAT. ICONST and IPTAT are two 100µA (nominal) currents that are designed to be equal at +27°C. These currents are derived from the MAX104’s internal precision +2.5V bandgap reference. ICONST is designed to be temperature independent, while IPTAT is directly proportional to the absolute temperature. These currents are derived from pnp current sources referenced from VCCI and driven into two series diodes connected to GNDI. The contacts ICONST and IPTAT may be left open, because internal catch diodes prevent saturation of the current sources. The simplest method ______________________________________________________________________________________ 23 MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier of determining the die temperature is to measure each current with an ammeter (which shuts off the internal catch diodes) referenced to GNDI. The die temperature in °C is then calculated by the expression: I TDIE = 300 ⋅ PTAT − 273 ICONST Another method of determining the die temperature uses the operational amplifier circuit shown in Figure 20. The circuit produces a voltage that is proportional to the die temperature. A possible application for this signal is speed control for a cooling fan, to maintain constant MAX104 die temperature. The circuit operates by converting the ICONST and IPTAT currents to voltages VCONST and VPTAT, with appropriate scaling to account for their equal values at +27°C. This voltage difference is then amplified by two amplifiers in an instrumentation-amplifier configuration with adjustable gain. The nominal value of the circuit gain is 4.5092V/V. The gain of the instrumentation amplifier is given by the expression: VTEMP AV = VCONST − VPTAT AV = 1 + To calibrate the circuit, first connect pins 2-3 on JU1 to zero the input of the PTAT path. With the MAX104 powered up, adjust potentiometer R3 until the voltage at the VTEMP output is -2.728V. Connecting pins 1-2 on JU1 restores normal operation to the circuit after the calibration is complete. The voltage at the VTEMP node will then be proportional to the actual MAX104 die temperature according to the equation: TDIE (°C) = 100 · VTEMP The overall accuracy of the die temperature measurement using the operational-amplifier scaling circuitry is limited mainly by the accuracy and matching of the resistors in the circuit. Thermal Management Depending on the application environment for the ESBGA-packaged MAX104, the customer may have to apply an external heatsink to the package after board assembly. Existing open-tooled heatsinks are available from standard heatsink suppliers (see Heatsink Manufacturers). The heatsinks are available with preapplied adhesive for easy package mounting. R1 R1 + 2⋅ R2 R3 3.32k 6.65k 10-TURN R2 15k R1 7.5k IPTAT 5k R2 15k JU1 1 12.1k 6.65k 2 3 VPTAT 1/4 MAX479 1/4 MAX479 R1 7.5k 12.1k ICONST VCONST VTEMP 1/4 MAX479 1/4 MAX479 6.05k Figure 20. Die Temperature Acquisition Circuit with the MAX479 24 ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier THERMAL RESISTANCE vs. AIRFLOW 18 MAX104 θJA (°C/W) WITHOUT HEATSINK WITH HEATSINK 0 16.5 12.5 200 14.3 9.4 400 13 8.3 800 12.5 7.4 16 WITHOUT HEATSINK 14 Thermal Performance The MAX104 has been modeled to determine the thermal resistance from junction to ambient. Table 7 lists the ADC’s thermal performance parameters: Ambient Temperature: TA = +70°C Heatsink Dimensions: 25mm x 25mm x 10mm PC Board Size and Layout: 4" x 4" 2 Signal Layers 2 Power Layers Heatsink Manufacturers Aavid Engineering and IERC provide open-tool, lowprofile heatsinks, fitting the 25mm x 25mm ESBGA package. Aavid Engineering, Inc. Phone: 714-556-2665 Heatsink Catalog #: 335224B00032 Heatsink Dimensions: 25mm x 25mm x 10mm International Electronic Research Corporation (IERC) Phone: 818-842-7277 Heatsink Catalog #: BDN09-3CB/A01 Heatsink Dimensions: 23.1mm x 23.1mm x 9mm Bypassing/Layout/Power Supply Grounding and power-supply decoupling strongly influence the MAX104’s performance. At 1GHz clock frequency and 8-bit resolution, unwanted digital crosstalk may couple through the input, reference, power supply, and ground connections and adversely influence the dynamic performance of the ADC. Therefore, closely follow the grounding and power-supply decoupling guidelines (Figure 22). θJA (°C/W) AIRFLOW (linear ft./min.) MAX104 Table 7. Thermal Performance for MAX104 With or Without Heatsink 12 10 WITH HEATSINK 8 6 0 100 200 300 400 500 600 700 800 AIRFLOW (linear ft./min.) Figure 21. MAX104 Thermal Performance Maxim strongly recommends using a multilayer printed circuit board (PCB) with separate ground and powersupply planes. Since the MAX104 has separate analog and digital ground connections (GNDA, GNDI, GNDR, and GNDD, respectively), the PCB should feature separate analog and digital ground sections connected at only one point (star ground at the power supply). Digital signals should run above the digital ground plane, and analog signals should run above the analog ground plane. Keep digital signals far away from the sensitive analog inputs, reference inputs, and clock inputs. Highspeed signals, including clocks, analog inputs, and digital outputs, should be routed on 50Ω microstrip lines, such as those employed on the MAX104EVKIT. The MAX104 has separate analog and digital powersupply inputs: VEE (-5V analog and substrate supply) and VCCI (+5V) to power the T/H amplifier, clock distribution, bandgap reference, and reference amplifier; V CCA (+5V) to supply the ADC’s comparator array; VCCO (+3V to VCCD) to establish power for all PECLbased circuit sections; and VCCD (+5V) to supply all logic circuits of the data converter. The MAX104 V EE supply contacts must not be left open while the part is being powered up. To avoid this condition, add a high-speed Schottky diode (such as a Motorola 1N5817) between VEE and GNDI. This diode prevents the device substrate from forward biasing, which could cause latchup. ______________________________________________________________________________________ 25 MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier All supplies should be decoupled with large tantalum or electrolytic capacitors at the point where they enter the PCB. For best performance, bypass all power supplies to the appropriate ground with a 10µF tantalum capacitor, to filter power-supply noise, in parallel with a 0.01µF capacitor and a high-quality 47pF ceramic chip capacitor located very close to the MAX104 device, to filter very high-frequency noise. line can be either a best-straight-line fit or a line drawn between the endpoints of the transfer function, once offset and gain errors have been nullified. The static linearity parameters for the MAX104 are measured using the best-straight-line fit method. Differential Nonlinearity (DNL) Differential nonlinearity is the difference between an actual step width and the ideal value of 1LSB. A DNL error specification of less than 1LSB guarantees no missing codes and a monotonic transfer function. Static Parameter Definitions Integral Nonlinearity (INL) Integral nonlinearity is the deviation of the values on an actual transfer function from a straight line. This straight VCCO GNDD 10µF 10nF 10nF 47pF 10µF 10nF 10nF 47pF 47pF 47pF 47pF 47pF 47pF NOTE: LOCATE ALL 47pF CAPACITORS AS CLOSE AS POSSIBLE TO THE MAX104 DEVICE. VCCI GNDI 47pF VEE VCCA 10µF 1N5817 GNDI GNDA 10µF 10nF 10nF 47pF 10nF 47pF 47pF 47pF VCCA = +4.75V TO +5.25V VCCD = +4.75V TO +5.25V VCCI = +4.75V TO +5.25V VCCO = +3.0V TO VCCD VEE = -4.75V TO -5.25V VCCD GNDD 10nF 10µF 10nF 10nF 47pF 47pF 47pF 47pF Figure 22. MAX104 Bypassing and Grounding 26 ______________________________________________________________________________________ 47pF 47pF ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier Effective Number of Bits (ENOB) ENOB indicates the global accuracy of an ADC at a specific input frequency and sampling rate. An ideal ADC’s error consists of quantization noise only. ENOB is calculated from a curve fit referenced to the theoretical full-scale range. Dynamic Parameter Definitions Total Harmonic Distortion (THD) THD is the ratio of the RMS sum of the first five harmonics of the input signal to the fundamental itself. This is expressed as: Signal-to-Noise Ratio (SNR) For a waveform perfectly reconstructed from digital samples, the theoretical maximum SNR is the ratio of the full-scale analog input (RMS value) to the RMS quantization error (residual error). The ideal, theoretical minimum analog-to-digital noise is caused by quantization error only and results directly from the ADC’s resolution (N bits): SNR(MAX) = (6.02 · N + 1.76)dB In reality, there are other noise sources besides quantization noise: thermal noise, reference noise, clock jitter, etc. SNR is calculated by taking the ratio of the RMS signal to the RMS noise, which includes all spectral components minus the fundamental, the first five harmonics, and the DC offset. Signal-to-Noise Plus Distortion (SINAD) SINAD is calculated from the effective number of bits (ENOB) as follows: SINAD = (6.02 · ENOB) + 1.76 2 2 2 2 THD = 20 log V2 + V3 + V4 + V5 / V1 ⋅ where V1 is the fundamental amplitude, and V2 through V5 are the amplitudes of the 2nd- through 5th-order harmonics. Spurious-Free Dynamic Range (SFDR) SFDR is the ratio, expressed in decibels, of the RMS amplitude of the fundamental (maximum signal component) to the RMS value of the next largest spurious component, excluding DC offset. Intermodulation Distortion (IMD) The two-tone IMD is the ratio, expressed in decibels, of either input tone to the worst 3rd-order (or higher) intermodulation products. The input tone levels are at -7dB full scale. Chip Information TRANSISTOR COUNT: 20,486 SUBSTRATE CONNECTED TO VEE ______________________________________________________________________________________ 27 MAX104 Bit Error Rates (BER) Errors resulting from metastable states may occur when the analog input voltage (at the time the sample is taken) falls close to the decision point of any one of the input comparators. Here, the magnitude of the error depends on the location of the comparator in the comparator network. If it is the comparator for the MSB, the error will reach full scale. The MAX104’s unique encoding scheme solves this problem by limiting the magnitude of these errors to 1LSB; it reduces the probability of these errors occurring to typically one in every 1016 clock cycles. ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier MAX104 Typical Operating Circuit Z0 = 50Ω 50Ω +5V ANALOG -5V ANALOG +5V DIGITAL VEE VCCA VCCI +3.3V DIGITAL VCCD AUXEN1 VCCO AUXEN2 DIVSELECT +5V DIGITAL ALL PECL OUTPUTS MUST BE TERMINATED VCCO - 2V LIKE THIS. OR+/OR- DEMUXEN 2 2 P7+/P7- VOSADJ 2 P6+/P62 P5+/P5- Z0 = 50Ω Z0 = 50Ω VIN+ PRIMARY PECL OUTPUTS VIN- 2 2 P3+/P32 SAMPLE CLOCK 1GHz +4dBm MAX104 Z0 = 50Ω P4+/P4- P2+/P2- 2 P1+/P1- CLK+ 2 P0+/P02 A7+/A7- CLK- 2 50Ω A6+/A6- 2 A5+/A5- GNDI CLKCOM AUXILARY PECL OUTPUTS GNDI 2 A4+/A42 A3+/A32 A2+/A2- 2 A1+/A1RSTIN+ 2 A0+/A0- RSTINDREADY+/DREADYRSTOUT+/RSTOUTGNDA GNDR GNDI 28 2 2 GNDD REFOUT REFIN ______________________________________________________________________________________ TO MEMORY OR DIGITAL SIGNAL PROCESSOR DIFFERENTIAL ANALOG INPUT 500mVp-p FS ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier TOP VIEW MAX104 ______________________________________________________________________________________ 29 MAX104 192-Contact ESBGA PCB Land Pattern ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier SUPER BGA.EPS MAX104 Package Information 30 ______________________________________________________________________________________ ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier ______________________________________________________________________________________ 31 MAX104 Package Information (continued) MAX104 ±5V, 1Gsps, 8-Bit ADC with On-Chip 2.2GHz Track/Hold Amplifier NOTES Maxim cannot assume responsibility for use of any circuitry other than circuitry entirely embodied in a Maxim product. No circuit patent licenses are implied. Maxim reserves the right to change the circuitry and specifications without notice at any time. 32 ____________________Maxim Integrated Products, 120 San Gabriel Drive, Sunnyvale, CA 94086 408-737-7600 © 1999 Maxim Integrated Products Printed USA is a registered trademark of Maxim Integrated Products.