LF to 750 MHz Digitally Controlled VGA AD8370 FUNCTIONAL BLOCK DIAGRAM FEATURES The AD8370 is a low cost, digitally controlled, variable gain amplifier that provides precision gain control, high IP3, and low noise figure. The excellent distortion performance and wide bandwidth make the AD8370 a suitable gain control device for modern receiver designs. For wide input, dynamic range applications, the AD8370 provides two input ranges: high gain mode and low gain mode. A vernier 7-bit transconductance (Gm) stage provides 28 dB of gain range at better than 2 dB resolution, and 22 dB of gain range at better than 1 dB resolution. A second gain range, 17 dB higher than the first, can be selected to provide improved noise performance. The AD8370 is powered on by applying the appropriate logic level to the PWUP pin. When powered down, the AD8370 consumes less than 4 mA and offers excellent input to output isolation. The gain setting is preserved when operating in a power-down mode. VCCO 6 BIAS CELL ICOM 2 INHI 1 PRE AMP OUTPUT AMP TRANSCONDUCTANCE INLO 16 ICOM 15 5 VOCM 7 OCOM 8 OPHI 9 OPLO 10 OCOM SHIFT REGISTER AND LATCHES 13 03692-0-001 AD8370 14 12 DATA CLCK LTCH Figure 1. 70 40 CODE = LAST 7 BITS OF GAIN CODE (NO MSB) 60 30 50 20 LOW GAIN MODE 40 ∆ GAIN ≅ 0.409 ∆ CODE HIGH GAIN MODE 30 10 0 20 –10 10 LOW GAIN MODE 0 0 10 20 30 40 50 60 70 80 ∆ GAIN ≅ 0.059 ∆ CODE –20 –30 90 100 110 120 130 GAIN CODE VOLTAGE GAIN (dB) HIGH GAIN MODE VOLTAGE GAIN (V/V) GENERAL DESCRIPTION 11 PWUP 4 APPLICATIONS Differential ADC drivers IF sampling receivers RF/IF gain stages Cable and video applications SAW filter interfacing Single-ended-to-differential conversion VCCO 3 03692-0-003 Programmable low and high gain (<2 dB resolution) Low range: −11 dB to +17 dB High range: +6 dB to +34 dB Differential input and output: 200 Ω differential input 100 Ω differential output 7 dB noise figure @ maximum gain Two-tone IP3 of +35 dBm @ 70 MHz −3 dB bandwidth of 750 MHz 40 dB precision gain range Serial 8-bit digital interface Wide input dynamic range Power-down feature Single 3 V to 5 V supply VCCI Figure 2. Gain vs. Gain Code at 70 MHz Gain control of the AD8370 is through a serial 8-bit gain control word. The MSB selects between the two gain ranges, and the remaining 7 bits adjust the overall gain in precise linear gain steps. Fabricated on the ADI high speed XFCB process, the high bandwidth of the AD8370 provides high frequency and low distortion. The quiescent current of the AD8370 is 78 mA typically. The AD8370 amplifier comes in a compact, thermally enhanced 16-lead TSSOP package and operates over the temperature range of −40°C to +85°C. Rev. 0 Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.326.8703 © 2004 Analog Devices, Inc. All rights reserved. AD8370 TABLE OF CONTENTS Specifications..................................................................................... 3 Choosing between Gain Ranges............................................... 15 Absolute Maximum Ratings............................................................ 5 Layout and Operating Considerations .................................... 16 ESD Caution.................................................................................. 5 Package Considerations............................................................. 17 Pin Configuration and Functional Descriptions.......................... 6 Single-Ended-to-Differential Conversion............................... 17 Typical Performance Characteristics ............................................. 7 DC-Coupled Operation............................................................. 18 Theory of Operation ...................................................................... 13 ADC Interfacing ......................................................................... 19 Block Architecture...................................................................... 13 3 V Operation ............................................................................. 20 Preamplifier................................................................................. 13 Evaluation Board and Software .................................................... 21 Transconductance Stage ............................................................ 13 Appendix ......................................................................................... 24 Output Amplifier ........................................................................ 14 Characterization Equipment..................................................... 24 Digital Interface and Timing .................................................... 14 Composite Waveform Assumption .......................................... 24 Applications..................................................................................... 15 Definitions of Selected Parameters.......................................... 24 Basic Connections ...................................................................... 15 Outline Dimensions ....................................................................... 28 Gain Codes .................................................................................. 15 Ordering Guide .......................................................................... 28 Power-Up Feature....................................................................... 15 REVISION HISTORY Revision 0: Initial Version Rev. 0 | Page 2 of 28 AD8370 SPECIFICATIONS VS = 5 V, T = 25°C, ZS = 200 Ω, ZL = 100 Ω at Gain Code HG127, 70 MHz, 1 V p-p differential output, unless otherwise noted. Table 1. Parameter DYNAMIC PERFORMANCE −3 dB Bandwidth Slew Rate INPUT STAGE Maximum Input Input Resistance Common-Mode Input Range CMRR Input Noise Spectral Density GAIN Maximum Voltage Gain High Gain Mode Conditions Min VOUT < 1 V p-p Gain Code HG127, RL = 1 kΩ, AD8370 in Compression Gain Code LG127, RL = 1 kΩ, VOUT = 2 V p-p Pins INHI and IHLO Gain Code LG2, 1 dB Compression Differential 3500 V/ns 3.2 200 3.2 77 1.9 V p-p Ω V p-p dB nV/√Hz 34 52 17 7.4 dB Volts/Volt dB Volts/Volt −8 0.4 −25 0.06 0.408 0.056 –2 20 dB Volts/Volt dB Volts/Volt (Volts/Volt)/Code (Volts/Volt)/Code RL ≥ 1 kΩ (1 dB compression) Differential VINHI = VINLO, over all gain codes 8.4 95 ±60 V p-p Ω mV Within ±10 MHz of 10 MHz ±0.01 7.2 −77 −77 35 17 dB dB dBc dBc dBm dBm Differential, f = 10 MHz, Gain Code LG127 Gain Code = HG127 Minimum Voltage Gain High Gain Mode Gain Code = HG1 Low Gain Mode Gain Code = LG1 Gain Step Size High Gain Mode Low Gain Mode Gain Code = HG127 For 6 dB gain step, settled to 10% of final value NOISE/HARMONIC PERFORMANCE 10 MHz Gain Flatness Noise Figure Second Harmonic1 Third Harmonic Output IP3 Output 1 dB Compression Point 1 Unit MHz V/ns Gain Code = LG127 OUTPUT INTERFACE Output Voltage Swing Output Resistance Output Differential Offset Max 750 5750 Low Gain Mode Gain Temperature Sensitivity Step Response Typ mdB/°C ns Pins OPHI and OPLO VOUT = 2 V p-p VOUT = 2 V p-p See footnotes on next page. Rev. 0 | Page 3 of 28 AD8370 Parameter NOISE/HARMONIC PERFORMANCE (cont.) 70 MHz Gain Flatness Noise Figure Second Harmonic Third Harmonic Output IP3 Output 1 dB Compression Point 140 MHz Gain Flatness Noise Figure Second Harmonic Third Harmonic Output IP3 Output 1 dB Compression Point 190 MHz Gain Flatness Noise Figure Second Harmonic Third Harmonic Output IP3 Output 1 dB Compression Point 240 MHz Gain Flatness Noise Figure Second Harmonic Third Harmonic Output IP3 Output 1 dB Compression Point 380 MHz Gain Flatness Noise Figure Output IP3 Output 1 dB Compression Point POWER-INTERFACE Supply Voltage 1 1 1 1 1 1 1 1 Quiescent Current3 vs. Temperature 4 Total Supply Current Power Down Current vs. Temperature4 POWER UP INTERFACE Power-Up Threshold Power-Down Threshold PWUP Input Bias Current GAIN CONTROL INTERFACE VIH VIL Input Bias Current 4 4 4 4 Conditions Min Within ±10 MHz of 70 MHz VOUT = 2 V p-p VOUT = 2 V p-p Within ±10 MHz of 140 MHz VOUT = 2 V p-p VOUT = 2 V p-p Within ±10 MHz of 240 MHz VOUT = 2 V p-p VOUT = 2 V p-p Within ±10 MHz of 240 MHz VOUT = 2 V p-p VOUT = 2 V p-p Within ±10 MHz of 240 MHz Typ 72.5 dB dB dBc dBc dBm dBm ±0.03 7.2 −54 −50 33 17 dB dB dBc dBc dBm dBm ±0.03 7.2 −43 −43 33 17 dB dB dBc dBc dBm dBm ±0.04 7.4 –28 –33 32 17 dB dB dBc dBc dBm dBm ±0.04 8.1 27 14 dB dB dBm dBm 79 Refer to Figure 20 for performance into a lighter load. See the 3 V Operation section for more information. Minimum and maximum specified limits for this parameter are guaranteed by production test. 4 Minimum or maximum specified limit for this parameter is a 6-sigma value and not guaranteed by production test. 2 3 Rev. 0 | Page 4 of 28 5.5 V 85.5 mA 105 mA mA 82 3.7 5 1.8 mA mA 0.8 V V nA 0.8 V V nA 400 1.8 900 1 Unit ±0.02 7.2 −65 −62 35 17 3.02 PWUP High, GC = LG127, RL = ∞, 4 seconds after power-on, thermal connection made to exposed paddle under device −40°C ≤ TA ≤ +85°C PWUP High, VOUT = 1 V p-p, ZL = 100 Ω reactive, GC = LG127 (includes load current) PWUP Low −40°C ≤TA ≤ +85°C Pin PWUP Voltage to enable the device Voltage to disable the device PWUP = 0 V Pins CLCK, DATA, and LTCH Voltage for a logic high Voltage for a logic low Max AD8370 ABSOLUTE MAXIMUM RATINGS Table 2. Parameter Supply Voltage, VS PWUP, DATA, CLCK, LTCH Differential Input Voltage, VINHI – VINLO Common-Mode Input Voltage, VINHI or VINLO, with respect to ICOM or OCOM Internal Power Dissipation θJA (Exposed paddle soldered down) θJA (Exposed paddle not soldered down) θJC (At exposed paddle) Maximum Junction Temperature Operating Temperature Range Storage Temperature Range Lead Temperature Range (Soldering 60 sec) Rating 5.5 V VS + 500 mV 2V VS + 500 mV (maximum), VICOM – 500 mV, VOCOM – 500 mV (minimum) 575 mW 30°C/W 95°C/W 9°C/W 150°C –40°C to +85°C –65°C to +150°C 235°C Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those listed in the operational sections of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. ESD CAUTION ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on the human body and test equipment and can discharge without detection. Although this product features proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high energy electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance degradation or loss of functionality. Rev. 0 | Page 5 of 28 AD8370 PIN CONFIGURATION AND FUNCTIONAL DESCRIPTIONS INHI 1 16 INLO ICOM 2 15 ICOM VCCI 3 14 DATA PWUP 4 13 CLCK VOCM 5 AD8370 LTCH TOP VIEW VCCO 6 (Not to Scale) 11 VCCO OCOM 7 10 OCOM OPHI 8 9 OPLO 03692-0-002 12 Figure 3.16-Lead TSSOP Table 3. Pin Function Descriptions Pin No. 1 2, 15, PADDLE 3 4 5 Mnemonic INHI ICOM VCCI PWUP VOCM 6, 11 7, 10 8 9 12 VCCO OCOM OPHI OPLO LTCH 13 14 16 CLCK DATA INLO Description Balanced Differential Input. Internally biased. Input Common. Connect to a low impedance ground. This node is also connected to the exposed pad on the bottom of the device. Input Positive Supply. 3.0 V to 5.5 V. Should be properly bypassed. Power Enable Pin. Device is operational when PWUP is pulled high. Common-Mode Output Voltage Pin. The midsupply ((VVCCO − VOCOM)/2) common-mode voltage is delivered to this pin for external bypassing for additional common-mode supply decoupling. This can be achieved with a bypass capacitor to ground. This pin is an output only and is not to be driven externally. Output Positive Supply. 3.0 V to 5.5 V. Should be properly bypassed. Output Common. Connect to a low impedance ground. Balanced Differential Output. Biased to midsupply. Balanced Differential Output. Biased to midsupply. Serial Data Latch Pin. Serial data is clocked into the shift register via the DATA pin when LTCH is low. Data in shift register is latched on the next high-going edge. Serial Clock Input Pin. Serial Data Input Pin. Balanced Differential Input. Internally biased. Rev. 0 | Page 6 of 28 AD8370 TYPICAL PERFORMANCE CHARACTERISTICS VS = 5 V, ZS = 200 Ω, ZL = 100 Ω, T = 25°C, unless otherwise noted. 40 40 CODE = LAST 7 BITS OF GAIN CODE (NO MSB) 60 0 20 –10 0 0 10 20 30 40 50 60 70 80 LG90 15 LG36 5 HG3 –20 LOW GAIN CODES SHOWN WITH SOLID LINES –10 10 100 Figure 7. Frequency Response vs. Gain Code 30 40 50 HIGH GAIN MODE +25°C 25 15 20 10 15 5 SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM TWO BATCH LOTS. 10 5 0 20 40 60 80 100 120 OUTPUT IP3 (dBm) +25°C 25 OUTPUT IP3 (dBVrms) 20 GAIN CODE 45 30 40 +85°C 25 35 20 30 –40°C 15 0 –5 140 UNIT CONVERSION NOTE FOR 100Ω LOAD: dBVrms = dBm–10dB 35 25 SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM TWO BATCH LOTS. 10 03692-0-024 OUTPUT IP3 (dBm) 30 1000 FREQUENCY (MHz) Figure 4. Gain vs. Gain Code at 70 MHz LOW GAIN MODE LG18 LG9 –5 GAIN CODE 35 HG18 HG9 10 0 –30 90 100 110 120 130 40 LG127 HG25 20 0 50 100 150 200 250 300 350 20 400 FREQUENCY (MHz) OUTPUT IP3 (dBm) –40°C, +85°C LOW GAIN MODE 25 03692-0-026 10 ∆ GAIN ≅ 0.059 ∆ CODE VOLTAGE GAIN (dB) ∆ GAIN ≅ 0.409 ∆ CODE HIGH GAIN MODE 10 HG102 HG51 VOLTAGE GAIN (dB) LOW GAIN MODE 30 HG77 30 20 03692-0-003 VOLTAGE GAIN (V/V) HIGH GAIN MODE 50 40 HIGH GAIN CODES SHOWN WITH DASHED LINES HG127 35 30 03692-0-072 70 Figure 8. Output Third-Order Intercept vs. Frequency at Maximum Gain Figure 5. Output Third-Order Intercept vs. Gain Code at 70 MHz 45 25 40 20 NOISE FIGURE (dB) 30 380 MHz 25 LOW GAIN MODE 20 70 MHz LG127 15 10 HG18 15 HG127 5 380 MHz 10 HIGH GAIN MODE 5 0 20 40 60 80 100 120 GAIN CODE 140 0 0 100 200 300 400 500 FREQUENCY (MHz) Figure 9. Noise Figure vs. Frequency at Various Gains Figure 6. Noise Figure vs. Gain Code at 70 MHz Rev. 0 | Page 7 of 28 600 03692-0-011 70 MHz 03692-0-012 NOISE FIGURE (dB) 35 AD8370 20 1kΩ LOAD HIGH GAIN MODE 8 4 UNIT CONVERSION NOTE: FOR 100Ω LOAD: dBVrms = dBm–10dB FOR 1kΩ LOAD: dBVrms = dBm –4 SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM TWO BATCH LOTS. –8 0 20 40 60 80 100 120 140 GAIN CODE 14 16 UNIT CONVERSION NOTE: RE 100Ω LOAD: dBVrms = dBm – 10dB RE 1kΩ LOAD: dBVrms = dBm –40°C, 100Ω LOAD +25°C, 1kΩ LOAD 14 12 8 +85°C, 1kΩ LOAD 8 SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM TWO BATCH LOTS. 6 0 50 100 150 200 OUTPUT IMD (dBc) +25°C OUTPUT IMD (dBc) –80 80 100 120 140 GAIN CODE –72 –82 –74 –84 –82 –92 –84 50 100 150 200 250 300 350 –86 400 Figure 14. Two-Tone Output IMD3 vs. Frequency at Maximum Gain, RL = 1 kΩ, VOUT = 1 V p-p Composite Differential 1.0 GAIN ERROR (dB) 0.5 –40°C 0 +85°C –1.0 ERROR AT –40°C AND +85°C WITH RESPECT TO 25°C. SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM ONE BATCH LOT. 0.5 –40°C 0 +85°C –0.5 –1.0 –1.5 FREQUENCY (MHz) –80 –90 1.0 1000 –78 +85°C –88 1.5 100 –76 –40°C –86 FREQUENCY (MHz) 03692-0-007 GAIN ERROR (dB) –70 –80 1.5 10 –68 –78 2.0 –2.0 –66 +25°C –76 2.0 –1.5 –64 –74 0 Figure 11. Two-Tone Output IMD3 vs. Gain Code at 70 MHz, RL = 1 kΩ, VOUT = 1 V p-p Composite Differential –0.5 –62 –94 03692-0-027 –85 60 4 400 –60 –72 HIGH GAIN MODE 40 350 SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM TWO BATCH LOTS. –70 –70 20 300 –68 LOW GAIN MODE 0 –40°C, 1kΩ LOAD 250 Figure 13. Output P1dB vs. Frequency –65 –75 6 FREQUENCY (MHz) SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM TWO BATCH LOTS. –60 10 10 Figure 10. Output P1dB vs. Gain Code at 70 MHz –55 12 OUTPUT IMD (dBc) –40°C, +85°C 0 16 03692-0-030 LOW GAIN MODE 12 +85°C, 100Ω LOAD 18 OUTPUT P1dB (dBm) +25°C OUTPUT P1dB (dBm) –40°C, +85°C HIGH GAIN MODE 03692-0-028 OUTPUT P1dB (dB) 16 18 +25°C, 100Ω LOAD 100Ω LOAD –2.0 ERROR AT –40°C AND +85°C WITH RESPECT TO 25°C. SHADING INDICATES ±3σ FROM THE MEAN. DATA BASED ON 30 PARTS FROM ONE BATCH LOT. 10 100 1000 FREQUENCY (MHz) Figure 15. Gain Error over Temperature vs. Frequency, RL = 1 kΩ Figure 12. Gain Error over Temperature vs. Frequency, RL = 100 Ω Rev. 0 | Page 8 of 28 03692-0-006 LOW GAIN MODE 03692-0-083 20 0 0 –10 –10 LOW GAIN RL = 1kΩ HARMONIC DISTORTION (dBc) HARMONIC DISTORTION (dBc) AD8370 –20 –30 LOW GAIN, RL = 1kΩ –40 HIGH GAIN, RL = 100Ω LOW GAIN, RL = 100Ω –50 –60 –70 –80 HIGH GAIN RL = 1kΩ –20 HIGH GAIN RL = 100Ω –30 LOW GAIN RL = 100Ω –40 –50 –60 –70 –80 HIGH GAIN, RL = 1kΩ 80 100 120 140 –90 GAIN CODE 0 60 HARMONIC DISTORTION (dBc) –10 150 30 5MHz 0 330 210 270 HD3 RL = 100Ω –30 –40 –50 HD3 RL = 1kΩ –60 –70 0 –50 100 RESISTANCE (Ω) 150 REACTANCE (j Ω) 50 –100 0 400 500 600 –150 700 FREQUENCY (MHz) 150 200 250 300 350 400 120 80 100 60 80 40 60 20 40 0 16 DIFFERENT GAIN CODES REPRESENTED R+jX FORMAT 20 03692-0-031 50 100 Figure 20. Harmonic Distortion vs. Frequency at Maximum Gain, VOUT = 2 V p-p Composite Differential 16 DIFFERENT GAIN CODES REPRESENTED R+jX FORMAT 200 50 FREQUENCY (MHz) 100 250 RESISTANCE (Ω) HD2 RL = 100Ω –20 0 Figure 17. Input and Output Reflection Coefficients, S11 and S22, ZO = 100 Ω Differential 300 140 –90 03692-0-059 300 200 120 –80 240 100 100 HD2 RL = 1kΩ S11 0 80 0 1GHz 180 60 Figure 19. Third-Order Harmonic Distortion vs. Gain Code at 70 MHz, VOUT = 2 V p-p Differential 90 S22 40 GAIN CODE Figure 16. Second-Order Harmonic Distortion vs. Gain Code at 70 MHz, VOUT = 2 V p-p Differential 120 20 03692-0-036 60 03692-0-029 40 Figure 18. Input Resistance and Reactance vs. Frequency –20 0 0 100 200 300 400 500 600 –40 700 FREQUENCY (MHz) Figure 21. Output Resistance and Reactance vs. Frequency Rev. 0 | Page 9 of 28 REACTANCE (j Ω) 20 03692-0-033 0 03692-0-057 –90 AD8370 1400 860 RL = 1kΩ 840 1300 HIGH GAIN MODE 1200 780 760 LOW GAIN MODE 1100 900 740 800 720 700 700 0 10 20 30 40 50 60 70 80 90 100 110 120 130 GAIN CODE RL = 100Ω 1000 600 0 100 200 300 400 500 600 700 800 900 FREQUENCY (MHz) 03692-0-034 GROUP DELAY (ps) 800 03692-0-032 GROUP DELAY (ps) 820 Figure 25. Group Delay vs. Frequency at Maximum Gain Figure 22. Group Delay vs. Gain Code at 70 MHz 120 80 110 70 LG32, LG127 100 60 HG32, HG127 80 CMRR (dB) PSRR (dB) 90 70 60 50 40 30 50 20 40 1 10 100 1000 FREQUENCY (MHz) 0 03692-0-013 20 10 Figure 23. Power Supply Rejection Ratio vs. Frequency at Maximum Gain 1000 Figure 26. Common-Mode Rejection Ratio vs. Frequency 0 12 NOISE SPECTRAL DENSITY (nV/ Hz) FORWARD TRANSMISSION, HG0 –20 FORWARD TRANSMISSION, LG0 –40 –60 –80 –100 FORWARD TRANSMISSION, PWUP LOW 10 8 LG127 6 4 HG18 2 HG127 –120 10 100 FREQUENCY (MHz) 1000 0 10 110 210 310 410 510 FREQUENCY (MHz) Figure 27. Input Referred Noise Spectral Density vs. Frequency at Various Gains Figure 24. Various Forms of Isolation vs. Frequency Rev. 0 | Page 10 of 28 610 03692-0-010 REVERSE TRANSMISSION, HG127 03692-0-009 ISOLATION (dB) 100 FREQUENCY (MHz) 03692-0-005 10 30 AD8370 VOUT DIFFERENTIAL VOPHI VOLTAGE (1V/DIV) DIFFERENTIAL VOUT DIFFERENTIAL VIN GND TIME (2ns/DIV) 03692-0-067 GND 03692-0-069 VOLTAGE (600mV/DIV) VOPLO TIME (2ns/DIV) Figure 28. DC-Coupled Large Signal Pulse Response Figure 31. Overdrive Recovery 85 DIFFERENTIAL OUTPUT (50mV/DIV) 80 SUPPLY CURRENT (mA) ZERO PWUP (2V/DIV) GAIN CODE HG127 75 70 LOW GAIN 65 HIGH GAIN 60 55 GND 03692-0-068 TIME (40ns/DIV) 50 0 16 32 48 64 80 96 112 128 GAIN CODE 03692-0-014 INPUT = –30dBm, 70MHz 100 AVERAGES Figure 32. Supply Current vs. Gain Code Figure 29. PWUP Time Domain Response 35 DIFFERENTIAL OUTPUT (10mV/DIV) MEAN: 51.9 σ: 0.518 30 ZERO 25 DATA FROM 136 PARTS FROM ONE BATCH LOT COUNT 6dB GAIN STEP (HG36 TO LG127) LTCH (2V/DIV) 20 15 10 TIME (20ns/DIV) Figure 30. Gain Step Time Domain Response 0 50 51 52 53 54 55 GAIN (V/V) Figure 33. Distribution of Voltage Gain, HG127, 70 MHz, RL = 100 Ω Rev. 0 | Page 11 of 28 03692-0-073 5 INPUT = –30dBm, 70MHz NO AVERAGING 03692-0-035 GND AD8370 2.75 2.70 +85°C 2.65 2.55 2.50 –40°C 2.45 LOW GAIN MODE 2.40 0 32 64 96 HIGH GAIN MODE 0 32 64 96 128 GAIN CODE 03692-0-071 VCM (V) +25°C 2.60 Figure 34. Common-Mode Output Voltage vs. Gain Code at Various Temperatures Rev. 0 | Page 12 of 28 AD8370 THEORY OF OPERATION INHI/INLO 2kΩ VCC/2 1mA Figure 36. INHI/INLO Simplified Schematic The digitally controlled gm section has 42 dB of controllable gain and makes gain the adjustments within each gain range. The step size resolution ranges from a fine ~ 0.07 dB up to a coarse 6 dB per bit, depending on the gain code. As shown in Figure 37, of the 42 dB total range, 28 dB has resolution of better than 2 dB, and 22 dB has resolution of better than 1 dB. The three basic building blocks of the AD8370 are a high/low gain selectable input preamplifier, a digitally controlled transconductance (gm) block, and a fixed gain output stage. VCCO 3 11 VCCO 6 5 VOCM 7 OCOM 8 OPHI INLO 16 9 OPLO ICOM 15 10 OCOM BIAS CELL ICOM 2 INHI 1 TRANSCONDUCTANCE OUTPUT AMP The curves in Figure 37 show typical input levels that can be applied to this amplifier at different gain settings. The maximum input was determined by finding the 1 dB compression or expansion point of the VOUT/VSOURCE gain. Note that this is not VOUT/VIN. In this way, the change in the input impedance of the device is also taken into account. 3.2 SHIFT REGISTER AND LATCHES DATA CLCK LTCH Figure 35. Functional Block Diagram PREAMPLIFIER There are two selectable input preamplifiers. Selection is made by the most significant bit (MSB) of the serial gain control dataword. In the high gain mode, the overall device gain is 7.1 Volts/ Volt (17 dB) above the low gain setting. The two preamplifiers give the AD8370 the ability to accommodate a wide range of input amplitudes. The overlap between the two gain ranges allows the user some flexibility based on noise and distortion demands. See the Choosing between Gain Ranges section for more information. <0.5dB RES 2.8 Rev. 0 | Page 13 of 28 34dB GAIN 2.0 LOW GAIN <1dB RES 2.4 VOUT [V peak] (V) 13 12 03692-0-001 AD8370 14 17dB GAIN HIGH GAIN <0.5dB RESOLUTION <2dB RES 1.6 12dB GAIN 1.2 0.1dB GAIN 6dB GAIN 0.8 –5dB GAIN 0.4 –8dB GAIN <1dB RES <2dB RES –11dB GAIN –25dB GAIN 0 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 VSOURCE [V peak] (V) Figure 37. Gain Resolution and Nominal Input and Output Range over the Gain Range 1.8 03692-0-023 VCCI PRE AMP 1mA TRANSCONDUCTANCE STAGE BLOCK ARCHITECTURE PWUP 4 The input impedance is approximately 200 Ω differential, regardless of which preamplifier is selected. Note that the input impedance is formed by using active circuit elements and is not set by passive components. See Figure 36 for a simplified schematic of the input interface. 03692-0-018 The AD8370 is a low cost, digitally controlled, fine adjustment variable gain amplifier that provides both high IP3 and low noise figure. The AD8370 is fabricated on an ADI proprietary high performance 25 GHz silicon bipolar process. The –3 dB bandwidth is approximately 750 MHz throughout the variable gain range. The typical quiescent current of the AD8370 is 78 mA. A power-down feature reduces the current to less than 4 mA. The input impedance is approximately 200 Ω differential, and the output impedance is approximately 100 Ω differential to be compatible with saw filters and matching networks used in intermediate frequency (IF) radio applications. Because there is no feedback between the input and output and stages within the amplifier, the input amplifier is isolated from variations in output loading and from subsequent impedance changes, and excellent input to output isolation is realized. Excellent distortion performance and wide bandwidth make the AD8370 a suitable gain control device for modern differential receiver designs. The AD8370 differential input and output configuration is ideally suited to fully differential signal chain circuit designs, although it can be adapted to single-ended system applications, if required. AD8370 OUTPUT AMPLIFIER Table 4. Serial Programming Timing Parameters The output impedance is approximately 100 Ω differential and, like the input preamplifier, this impedance is formed using active circuit elements. See Figure 38 for a simplified schematic of the output interface. Parameter Clock Pulse Width (TPW) Clock Period (TCK) Setup Time Data vs. Clock (TDS) Setup Time Latch vs. Clock (TES) Hold Time Latch vs. Clock (TEH) OPHI/OPLO Min 25 ns 50 10 20 10 ns ns ns ns 10µA 740Ω VCC/2 03692-0-017 03692-0-019 CLCK/DATA/LTCH/PWUP Figure 38. OPHI/OPLO Simplified Circuit Figure 40. Simplified Circuit for Digital Inputs The gain of the output amplifier, and thus the AD8370 as a whole, is load dependent. The following equation can be used to predict the gain deviation of the AD8370 from that at 100 Ω as the load is varied: 1.98 98 1+ RLOAD VOCM 75Ω VCC/2 For example, if RLOAD is 1 kΩ, the gain is a factor of 1.80 (5.12 dB) above that at 100 Ω, all other things being equal. If RLOAD is 50 Ω, the gain is a factor of 0.669 (3.49 dB) below that at 100 Ω. Figure 41. Simplified Circuit for VOCM Output DIGITAL INTERFACE AND TIMING The digital control port uses a standard TTL interface. The 8-bit control word is read in a serial fashion when the LTCH pin is held low. The levels presented to the DATA pin are read on each rising edge of the CLCK signal. Figure 39 illustrates the timing diagram for the control interface. Minimum values for timing parameters are presented in Table 4. Figure 40 is a simplified schematic of the digital input pins. TDS DATA (Pin 14) MSB MSB-1 MSB-2 MSB-3 LSB+3 LSB+2 LSB+1 TCK LSB TPW CLCK (Pin 13) TEH 03692-0-038 TES LTCH (Pin 12) 03692-0-020 GainDeviation = Figure 39. Digital Timing Diagram Rev. 0 | Page 14 of 28 Unit AD8370 APPLICATIONS BASIC CONNECTIONS Figure 42 shows the minimum connections required for basic operation of the AD8370. Supply voltages between 3.0 V and 5.5 V are allowed. The supply to the VCCO and VCCI pins should be decoupled with at least one low inductance, surfacemount ceramic capacitor of 0.1 µF placed as close as possible to the device. INLO ICOM DATA CLCK 12 11 10 9 OPLO 13 OCOM 14 LTCH 15 VCCO 16 RL INHI ICOM VCCI PWUP VOCM VCCO OCOM OPHI AD8370 RS AV = GainCode Vernier(1 + (PreGain − 1) MSB) 1nF RS BALANCED SOURCE The theoretical linear voltage gain can be expressed with respect to the gain code as where: SERIAL CONTROL INTERFACE 1nF 2 in that range is given by LG127. The same is true for the high gain range. Both LG0 and HG0 essentially turn off the variable transconductance stage, and thus no output is available with these codes. See Figure 24. 1 2 3 4 5 6 7 8 BALANCED LOAD AV is the linear voltage gain. GainCode is the digital gain control word minus the MSB (the final 7 bits). Vernier = 0.055744 V/V PreGain = 7.079458 V/V MSB is the most significant bit of the 8-bit gain control word. The MSB sets the device in either high gain mode (MSB = 1 ) or low gain mode (MSB = 0). 2 1nF 1nF 1nF 0.1µF 03692-0-037 0.1µF +VS (3.0V TO 5.0V) For example, a gain control word of HG45 (or 10101101 binary) results in a theoretical linear voltage gain of 17.76 Volts/Volt, calculated as 45 × 0.055744 × (1 + (7.079458 − 1) × 1) Figure 42. Basic Connections The AD8370 is designed to be used in differential signal chains. Differential signaling allows improved even-order harmonic cancellation and better common-mode immunity than can be achieved using a single-ended design. To fully exploit these benefits, it is necessary to drive and load the device in a balanced manner. This requires some care to ensure that the common-mode impedance values presented to each set of inputs and outputs are balanced. Driving the device with an unbalanced source can degrade the common-mode rejection ratio. Loading the device with an unbalanced load can cause degradation to even-order harmonic distortion and premature output compression. In general, optimum designs are fully balanced, although the AD8370 still provides impressive performance when used in an unbalanced environment. The AD8370 is a fine adjustment, variable gain amplifier. The gain control transfer function is linear in voltage gain. On a decibel scale, this results in the logarithmic transfer functions indicated in Figure 4. At the low end of the gain transfer function, the slope is steep, providing a rather coarse control function. At the high end of the gain control range, the decibel step size decreases, allowing precise gain adjustment. GAIN CODES The AD8370’s two gain ranges are referred to as high gain (HG) and low gain (LG). Within each range, there are 128 possible gain codes. Therefore, the minimum gain in the low gain range is given by the nomenclature LG0 whereas the maximum gain Increments or decrements in gain within either gain range are simply a matter of operating on the GainCode. Six –dB gain steps, which are equivalent to doubling or halving the linear voltage gain, are accomplished by doubling or halving the GainCode. When power is first applied to the AD8370, the device is programmed to code LG0 to avoid overdriving the circuitry following it. POWER-UP FEATURE The power-up feature does not affect the GainCode and the gain setting is preserved when in power-down mode. Powering down the AD8370 (bringing PWUP low while power is still applied to the device) does not erase or change the GainCode from the AD8370, and the same gain code is in place when the device is powered up, that is, when PWUP is brought high again. Removing power from the device all together and reapplying, however, reprograms to LG0. CHOOSING BETWEEN GAIN RANGES There is some overlap between the two gain ranges; users can choose which one is most appropriate for their needs. When deciding which preamp to use, consider resolution, noise, linearity, and spurious-free dynamic range (SFDR). The most important points to keep in mind are • The low gain range has better gain resolution. • The high gain range has a better noise figure. Rev. 0 | Page 15 of 28 AD8370 • The high gain range has better linearity and SFDR at higher gains. • Conversely, the low gain range has higher SFDR at lower gains. LAYOUT AND OPERATING CONSIDERATIONS Figure 43 provides a summary of noise, OIP3, IIP3, and SFDR as a function of device power gain. SFDR is defined as SFDR = 2 (IIP3 −NF − N S ) 3 When laying out an RF trace with a controlled impedance, consider the following: where: IIP3 is the input third-order intercept point, the output intercept point in dBm minus the gain in dB. NF is the noise figure in dB. NS is source resistor noise, –174 dBm for a 1 Hz bandwidth at 300°K (27°C). In general, NS = 10 log10(kTB), where k = 1.374 ×10−23 , T is the temperature in degrees Kelvin, and B is the noise bandwidth in Hertz. • Space the ground plane to either side of the signal trace at least 3 line-widths away to ensure that a microstrip (vertical dielectric) line is formed, rather than a coplanar (lateral dielectric) waveguide. • Ensure that the width of the microstrip line is constant and that there are as few discontinuities as possible , such as component pads, along the length of the line. Width variations cause impedance discontinuities in the line and may result in unwanted reflections. • Do not use silkscreen over the signal line because it alters the line impedance. • Keep the length of the input and output connection lines as short as possible. 180 NF LOW GAIN 40 OIP3 LOW GAIN OIP3 HIGH GAIN 160 IIP3 LOW GAIN IIP3 HIGH GAIN 150 140 10 NF HIGH GAIN 0 130 –10 120 –20 –30 –30 SFDR LOW GAIN 100 –10 0 10 20 30 40 POWER GAIN (dB) Figure 44 shows the cross section of a PC board and Table 5 show the dimensions that provide a 100 Ω line impedance for FR-4 board material with εr = 4.6. Table 5. 100 Ω 22 mils 53 mils 2 mils W H T 110 SFDR HIGH GAIN –20 SFDR (dB) 30 170 03692-0-004 NOISE FIGURE (dB), OIP3 AND IIP3 (dBm) 50 20 Each input and output pin of the AD8370 presents either a 100 Ω or 50 Ω impedance relative to their respective ac grounds. To ensure that signal integrity is not seriously impaired by the printed circuit board, the relevant connection traces should provide an appropriate characteristic impedance to the ground plane. This can be achieved through proper layout. 3W 50 Ω 13 mils 8 mils 2 mils W 3W Figure 43. OIP3, IIP3, NF, and SFDR Variation with Gain T H ER 03692-0-021 As the gain increases, the input amplitude required to deliver the same output amplitude is reduced. This results in less distortion at the input stage, and therefore the OIP3 increases. At some point, the distortion of the input stage becomes small enough such that the nonlinearity of the output stage becomes dominant. The OIP3 does not improve significantly as the gain is increased beyond this point, which explains the knee in the OIP3 curve. The IIP3 curve has a knee for the same reason; however, as the gain is increased beyond the knee, the IIP3 starts to decrease rather than increase. This is because in this region OIP3 is constant, therefore the higher the gain, the lower the IIP3. The two gain ranges have equal SFDR at approximately 13 dB power gain. Figure 44. Cross-Sectional View of a PC Board It possible to approximate a 100 Ω trace on a board designed with the 50 Ω dimensions above by removing the ground plane within 3 line-widths of the area directly below the trace. The AD8370 contains both digital and analog sections. Care should be taken to ensure that the digital and analog sections are adequately isolated on the PC board. The use of separate ground planes for each section connected at only one point via a ferrite bead inductor ensures that the digital pulses do not adversely affect the analog section of the AD8370. Rev. 0 | Page 16 of 28 AD8370 0.5 High transient and noise levels on the power supply, ground, and digital inputs can, under some circumstances, reprogram the AD8370 to an unintended gain code. This further reinforces the need for proper supply bypassing and decoupling. The user should also be aware that probing the AD8370 and associated circuitry during circuit debug may also induce the same effect. HIGH GAIN MODE (GAIN CODE HG255) 0 –0.5 LOW GAIN MODE (GAIN CODE LG127) –1.0 0 100 200 300 400 500 FREQUENCY (MHz) 03692-0-040 DIFFERENTIAL BALANCE (dB) Due to the nature of the AD8370’s circuit design, care must be taken to minimize parasitic capacitance on the input and output. The AD8370 could become unstable with more than a few pF of shunt capacitance on each input. Using resistors in series with input pins is recommended under conditions of high source capacitance. Figure 46. Differential Output Balance for a Single-Ended Input Drive at Maximum Gain (RL = 1 kΩ, CAC = 10 nF) PACKAGE CONSIDERATIONS The package of the AD8370 is a compact, thermally enhanced TSSOP 16-lead design. A large exposed paddle on the bottom of the device provides both a thermal benefit and a low inductance path to ground for the circuit. To make proper use of this packaging feature, the PCB needs to make contact directly under the device, connected to an ac/dc common ground reference with as many vias as possible to lower the inductance and thermal impedance. SINGLE-ENDED-TO-DIFFERENTIAL CONVERSION SERIAL CONTROL INTERFACE CAC Figure 46 illustrates the differential balance at the output for a single-ended input drive for multiple gain codes. The differential balance is better than 0.5 dB for signal frequencies less than 250 MHz. Figure 47 depicts the differential balance over the entire gain range at 10 MHz. The balance is degraded for lower gain settings because the finite common gain allows some of the input signal applied to INHI to pass directly through to the OPLO pin. At higher gain settings, the differential gain dominates and balance is restored. 0.6 LOW GAIN MODE CAC HIGH GAIN MODE 10 9 OPLO 11 OCOM 12 LTCH 13 VCCO 14 CLCK 15 DATA INLO 16 ICOM RS SINGLEENDED SOURCE DIFFERENTIAL BALANCE (dB) 0.5 RL INHI ICOM VCCI PWUP VOCM VCCO OCOM OPHI AD8370 1 2 3 4 5 6 7 8 0.4 0.3 0.2 0.1 0.1µF 1nF 0.1µF +VS 0 Figure 45. Single-Ended-to-Differential Conversion The AD8370 is primarily designed for differential signal interfacing. The device can be used for single-ended-to-differential conversion simply by terminating the unused input to ground using a capacitor as depicted in Figure 45. The ac coupling capacitors should be selected such that their reactance is negligible at the frequency of operation. For example, using 1 nF capacitors for CAC presents a capacitive reactance of –j1.6 Ω on each input node at 100 MHz. This attenuates the applied input voltage by 0.003 dB. If 10 pF capacitors had been selected, the voltage delivered to the input would be reduced by 2.1 dB when operating with a 200 Ω source impedance. 0 32 64 96 0 32 64 GAIN CODE 96 128 03692-0-041 CAC 03692-0-039 CAC Figure 47. Differential Output Balance at 10 MHz for a Single-Ended Drive vs. Gain Code (RL = 1 kΩ, CAC = 10 nF) Even though the amplifier is no longer being driven in a balanced manner, the distortion performance remains adequate for most applications. Figure 48 illustrates the harmonic distortion performance of the circuit in Figure 45 over the entire gain range. If the amplifier is driven in single-ended mode, the input impedance varies depending on the value of the resistor used to terminate the other input as follows: RinSE = RinDIFF + RTERM where RTERM is the termination resistor connected to the other input. Rev. 0 | Page 17 of 28 AD8370 The AD8370 is also a dc accurate variable gain amplifier. The common-mode dc voltage present at the output pins is internally set to midsupply using what is essentially a buffered resistive divider network connected between the positive supply rail and the common (ground) pins. The input pins are at a slightly higher dc potential, typically 250 mV to 550 mV above the output pins, depending on gain setting. In a typical single-supply application, it is necessary to raise the common-mode reference level of the source and load to roughly midsupply to maintain symmetric swing and to avoid sinking or sourcing strong bias currents from the input and output pins. It is possible to use balanced dual supplies to allow ground referenced source and load as indicated in Figure 49. By connecting the VOCM pin and unused input to ground, the input and output commonmode potentials are forced to virtual ground. This allows direct coupling of ground referenced source and loads. The initial differential input offset is typically only a few 100 µV. Over temperature, the input offset could be as high as a few tens of mVs. If precise dc accuracy is need over temperature and time, it may be necessary to periodically measure the input offset and to apply the necessary opposing offset to the unused differential input, canceling the resulting output offset. –40 HARMONIC DISTORTION (dBc) –50 –60 HD2 HD2 –70 –80 HD3 HD3 –90 LOW GAIN MODE 0 32 64 HIGH GAIN MODE 96 0 32 64 96 03692-0-042 –100 128 GAIN CODE Figure 48. Harmonic Distortion of the Circuit in Figure 45 DC-COUPLED OPERATION –2.5V SERIAL CONTROL INTERFACE 0V 1nF DATA CLCK 12 11 10 9 OPLO 13 OCOM 14 LTCH 15 VCCO 16 INLO RS SINGLEENDED GROUND REFERENCED SOURCE ICOM RT RL INHI ICOM VCCI PWUP VOCM VCCO OCOM OPHI AD8370 1 2 3 4 5 6 7 8 0V –2.5V 0.1µF 03692-0-043 1nF +2.5V 0.1µF To address situations where dual supplies are not convenient, a second option is presented in Figure 50. The AD8138 differential amplifier is used to translate the common-mode level of the driving source to midsupply, which allows dc accurate performance with a ground-referenced source without the need for dual supplies. The bandwidth of the solution in Figure 50 is limited by the gain-bandwidth product of the AD8138. The normalized frequency response of both implementations is shown in Figure 51. 10 SERIAL CONTROL INTERFACE 100Ω RT 2 499Ω 9 OPLO PWUP VOCM VCCO OCOM OPHI 2 3 4 5 6 7 8 AD8370 USING DUAL ±2.5V SUPPLY –4 –6 10 100 1k 10k 100k 1M 10M 100M 1G FREQUENCY (Hz) VOCM 1nF 03692-0-044 0.1µF 0 –2 –10 +5V SINGLE-ENDED GROUND REFERENCED SOURCE 2 1 100Ω 1nF AD8370 WITH AD8138 SINGLE +5V SUPPLY 4 –8 RL 1 6 Figure 50. DC Coupling the AD8370. The AD8138 is used as a unity gain level shifting amplifier to lift the common-mode level of the source to midsupply. Rev. 0 | Page 18 of 28 Figure 51. Normalized Frequency Response of the Two Solutions in Figure 49 and Figure 50 03692-0-045 10 OCOM LTCH VCCO 11 VCCI RS 12 ICOM 499Ω 13 AD8370 VOCM AD8138 14 INHI RT 15 CLCK INLO 16 DATA +5V 499Ω ICOM 499Ω VOCM NORMALIZED RESPONSE (dB) 8 Figure 49. DC Coupling the AD8370. Dual supplies are used to set the input and output common-mode levels to 0 V. AD8370 Although the AD8370 is designed to provide a 100 Ω output source impedance, the device is capable of driving a variety of loads while maintaining reasonable gain and distortion performance. A common application for the AD8370 is ADC driving in IF sampling receivers and broadband wide dynamic range digitizers. The wide gain adjustment range allows the use of lower resolution ADCs. Figure 52 illustrates a typical ADC interface network. ROP CAC ZS RIP AD8370 VIN ZP 100Ω RT ZIN ADC VOCM ROP CAC ZS RIP 03692-0-046 VIN Figure 52. Generic ADC Interface Many factors need to be considered before defining component values used in the interface network, such as the desired frequency range of operation, the input swing, and input impedance of the ADC. AC coupling capacitors, CAC, should be used to block any potential dc offsets present at the AD8370 outputs, which would otherwise consume the available low-end range of the ADC. The CAC capacitors should be large enough so that they present negligible reactance over the intended frequency range of operation. The VOCM pin may serve as an external reference for ADCs that do not include an on-board reference. In either case, it is suggested that the VOCM pin be decoupled to ground through a moderately large bypassing capacitor (1 nF to 10 nF) to help minimize wideband noise pick-up. After defining reasonable values for coupling capacitors, suppressing resistors, and the terminating resistor, it is time to design the intermediate filter network. The example in Figure 52 suggests a second-order low-pass filter network comprised of series inductors and a shunt capacitor. The order and type of filter network used depends on the desired high frequency rejection required for the ADC interface, as well as on pass-band ripple and group delay. In some situations, the signal spectra may already be sufficiently band-limited such that no additional filter network is necessary, in which case ZS would simply be a short and ZP would be an open. In other situations, it may be necessary to have a rather high-order antialiasing filter to help minimize unwanted high frequency spectra from being aliased down into the first Nyquist zone of the ADC. To properly design the filter network, it is necessary to consider the overall source and load impedance presented by the AD8370 and ADC input, including the additional resistive contribution of suppression and terminating resistors. The filter design can then be handled by using a single-ended equivalent circuit as shown in Figure 53. A variety of references that address filter synthesis are available. Most provide tables for various filter types and orders, indicating the normalized inductor and capacitor values for a 1 Hz cutoff frequency and 1 Ω load. After scaling the normalized prototype element values by the actual desired cut-off frequency and load impedance, it is simply a matter of splitting series element reactances in half to realize the final balanced filter network component values. SOURCE RS Depending on the input impedance presented by the input system of the ADC, it may be desirable to terminate the ADC input down to a lower impedance by using a terminating resistor, RT. The high frequency response of the AD8370 exhibits greater peaking when driving very light loads. In addition, the terminating resistor helps to better define the input impedance at the ADC input. Any part-to-part variability of ADC input impedance is reduced when shunting down the ADC inputs by using a moderate tolerance terminating resistor (typically a 1% value is acceptable). ZS SINGLE-ENDED EQUIVALENT VS Often it is wise to include input and output parasitic suppression resistors, RIP and ROP. Parasitic suppressing resistors help to prevent resonant effects that occur as a result of internal bondwire inductance, pad to substrate capacitance, and stray capacitance of the printed circuit board trace artwork. If omitted, undesirable settling characteristics may be observed. Typically, only 10 Ω to 25 Ω of series resistance is all that is needed to help dampen resonant effects. Considering that most ADCs present a relatively high input impedance, very little signal is lost across the RIP and ROP series resistors. LOAD RS 2 RS 2 RL ZS 2 BALANCED CONFIGURATION VS ZP ZP RL 2 RL 2 ZS 2 03692-0-047 ADC INTERFACING Figure 53. Single-Ended-to-Differential Network Conversion As an example, a second-order Butterworth low-pass filter design is presented where the differential load impedance is 1200 Ω, and the padded source impedance of the AD8370 is assumed to be 120 Ω. The normalized series inductor value for the 10-to-1 load-to-source impedance ratio is 0.074H, and the normalized shunt capacitor is 14.814 F. For a 70 MHz cutoff frequency, the single-ended equivalent circuit consists of a 200 nH series inductor followed by a 27 pF capacitor. To realize the balanced equivalent, simply split the 200 nH inductor in half to realize the network shown in Figure 54. Rev. 0 | Page 19 of 28 AD8370 RS = RS = 0.1 RL 0 LN = 0.074H –10 NORMALIZED SINGLE-ENDED EQUIVALENT VS CN –20 RL= 1Ω 14.814F –30 –40 fC = 1Hz dBFS RS = 120Ω –50 200nH –60 –70 –80 DE-NORMALIZED SINGLE-ENDED EQUIVALENT VS RL= 1200Ω 27pF –90 –100 –110 fC = 70MHz –130 100nH 0 BALANCED CONFIGURATION RS = 60Ω 2 20 27pF RL 2 = 600Ω 100nH 30 40 50 60 Figure 55. FFT Plot of Two-Tone Intermodulation Distortion at 42 MHz for the Circuit in Figure 56 In Figure 55, the intermodulation products are comparable to the noise floor of the ADC. The spurious-free dynamic range of the combination is better than 66 dB for a 70 MHz measurement bandwidth. Figure 54. Second-Order Butterworth Low-Pass Filter Design Example A complete design example is shown in Figure 56. The AD8370 is configured for single-ended-to-differential conversion with the input terminated down to present a single-ended 75 Ω input. A sixth-order Chebyshev differential filter is used to interface the output of the AD8370 to the input of the AD9430 170 MSPS 12-bit ADC. The filter minimizes aliasing effects and improves harmonic distortion performance. 3 V OPERATION It is possible to operate the AD8370 at voltages as low as 3 V with only minor performance degradation. Table 6 gives typical specifications for operation at 3 V. Table 6. Parameter Ouptut IP3 P1dB −3 dB Bandwidth IMD3 The input of the AD9430 is terminated with a 1.5 kΩ resistor so that the overall load presented to the filter network is ~1 kΩ. The variable gain of the AD8370 extends the useable dynamic range of the ADC. The measured intermodulation distortion of the combination is presented in Figure 55 at 42 MHz. Typical (70 MHz, RL = 100 Ω) +23.5 dBm +12.7 dBm 650 MHz (HG 127) −82 dBc (RL = 1 kΩ) SERIAL CONTROL INTERFACE FROM 75Ω Tx-LINE CAC CAC 100nF 68nH 180nH 220nH 25Ω VINA INLO DATA CLCK 12 11 10 9 OPLO 13 OCOM 14 LTCH 15 VCCO 16 ICOM 120Ω 100nF AD8370 ICOM VCCI PWUP VOCM VCCO OCOM OPHI 27pF INHI 1 2 3 4 5 6 7 8 CAC CAC 68nH 39pF 180nH 27pF 220nH 1.5kΩ AD9430 25Ω VINB 100nF 1nF 100nF 0.1µF 0.1µF 03692-0-049 RS 70 FREQUENCY (MHz) 03692-0-048 VS 10 RL 2 = 600Ω 03692-0-050 –120 RS = 60Ω 2 +VS Figure 56. ADC Interface Example Rev. 0 | Page 20 of 28 AD8370 EVALUATION BOARD AND SOFTWARE The evaluation board allows quick testing of the AD8370 by using standard 50 Ω test equipment. The schematic is shown in Figure 57. Transformers T1 and T2 are used to transform 50 Ω source and load impedances to the desired input and output reference levels. The top and bottom layers are shown in Figure 61 and Figure 62. The ground plane was removed under the traces between T1 and pins INHI and INLO to approximate a 100 Ω characteristic impedance. 1 2 3 14 15 4 16 5 The evaluation board comes with the AD8370 control software that allows serial gain control from most computers. The evaluation board is connected via a cable to the parallel port of the computer. Simply by adjusting the slider bar in the control software, the gain code is automatically updated to the AD8370. 6 17 18 7 19 8 20 9 21 10 22 11 23 12 24 13 25 D-SUB 25 PIN MALE L2* C9 R7 R6 R5 1kΩ 1kΩ 1kΩ C1 12 11 10 9 LTCH OCOM OPLO 13 VCCO R2 0Ω 14 CLCK 1:4 15 DATA T2 16 INLO T1 ICOM R4 0Ω PWUP VOCM VCCO OCOM OPHI 1 2 3 4 5 6 7 8 JTX-2-10T 2:1 AD8370 R1 0Ω VCCI 50Ω Tx LINE 1nF ICOM 50Ω Tx LINE OUT+ 50Ω Tx LINE OUT– R3 0Ω C2 C4 1nF 1nF C8 0.1µF C5 0.1µF C6 1µF SW1 PWUP L1* VOCM +VS R9 OPEN R8 49.9Ω C7 0.1µF GND C10 OPEN P2 1 2 3 4 5 VS GND *EMI SUPPRESSION FERRITE HZ1206E601R-00 Figure 57. AD8370 Evaluation Board Schematic Rev. 0 | Page 21 of 28 03692-0-051 IN– TC4-1W 50Ω Tx LINE C3 1nF INHI IN+ OPEN AD8370 Figure 58. Evaluation Software Table 7. AD8370 Evaluation Board Configuration Options Component VS, GND, VOCM SW1, R8, C10, PWUP P1, R5, R6, R7, C9 J1, J2, J6, J7 C1, C2, C3, C4 T1, T2 R1, R2, R3, R4 C5, C6, C7, C8 L1, L2 Function Power Interface Vector Pins. Apply supply voltage between VS and GND. The VOCM pin allows external monitoring of the common-mode input and output bias levels. Device Enable. Set to position B to power up the device. When in position A, the PWUP pin is connected to the PWUP vector pin. The PWUP pin allows external power cycling of the device. R8 and C10 are provided to allow for proper cable termination. Serial Control Interface. The evaluation board can be controlled using most PCs. Windows® based control software is shipped with the evaluation kit. A 25-pin D-sub connector cable is required to connect the PC to the evaluation board. It may be necessary to use a capacitor on the clock line, depending on the quality of the PC port signals. A 1 nF capacitor for C9 is usually sufficient for reducing clock overshoot. Input and Output Signal Connectors. These SMA connectors provide a convenient way to interface the evaluation board with 50 Ω test equipment. Typically the device is evaluated using a single-ended source and load. The source should connect to J1 (IN+), and the load should connect to J6 (OUT+). AC Coupling Capacitors. Provide ac coupling of the input and output signals. Impedance Transformers. T1 provides a 50 Ω to 200 Ω impedance transformation. T2 provides a 100 Ω to 50 Ω impedance transformation. Single-Ended or Differential. R2 and R4 are used to ground the center tap of the secondary windings on transformers T1 and T2. R1 and R3 should be used to ground J2 and J7 when used in single ended applications. Power Supply Decoupling. Nominal supply decoupling consists of a ferrite bead series inductor followed by a 1 µF capacitor to ground followed by a 0.1 µF capacitor to ground positioned as close to the device as possible. C7 provides additional decoupling of the input common-mode voltage. L1 provides high frequency isolation between the input and output power supply. L2 provides high frequency isolation between the analog and digital ground. Rev. 0 | Page 22 of 28 Default Condition Not applicable SW1 = installed R8 = 49.9 Ω (Size 0805) C10 = open (Size 0805) P1 = installed R5, R6, R7 = 1 kΩ (Size 0603) C9 = open (Size 0603) Not applicable C1, C2, C3, C4 = 1 nF (Size 0603) T1 = TC4 −1W (MiniCircuits) T2 = JTX−2−10T (MiniCircuits) R1, R2, R3, R4 = 0 Ω (Size 0603) C6 = 1 µF (Size 0805) C5, C7, C8 = 0.1 µF (Size 0603) L1, L2 = HZ1206E601R-00 (Steward, Size 1206) Figure 59. Evaluation Board Top Silkscreen 03692-0-077 03692-0-076 AD8370 03692-0-078 03692-0-079 Figure 61. Evaluation Board Top Figure 62. Evaluation Board Bottom Figure 60. Evaluation Board Bottom Silkscreen Rev. 0 | Page 23 of 28 AD8370 APPENDIX CHARACTERIZATION EQUIPMENT DEFINITIONS OF SELECTED PARAMETERS An Agilent N4441A Balanced Measurement System was used to obtain the gain, phase, group delay, reverse isolation, CMRR, and s-parameter information contained in this data sheet. With the exception for the s-parameter information, T-attenuator pads were used to match the 50 Ω impedance of this instrument’s ports to the AD8370. An Agilent 4795A Spectrum Analyzer was used to obtain nonlinear measurements IMD, IP3, and P1dB through matching baluns and/or attenuator networks. Various other measurements were taken with setups shown in this section. Common-mode rejection ratio (Figure 26) has been defined for this characterization effort as Differenti al Mode Gain Common Mode Gain where the numerator is the gain into a differential load at the output due to a differential source at the input, and the denominator is the gain into a differential-mode load at the output due to a common-mode source at the input. In terms of mixed-mode s-parameters, this equates to COMPOSITE WAVEFORM ASSUMPTION The nonlinear two-tone measurements made for this data sheet, i.e., IMD and IP3, are based on the assumption of a fixed value composite waveform at the output, generally 1 V p-p. The frequencies of interest dictate the use of RF test equipment, and because this equipment is generally not designed to work in units of volts, but rather watts and dBm, an assumption was made to facilitate equipment setup and operation. Two sinusoidal tones can be represented as SDD21 SDC21 More information on mixed-mode s-parameters can be obtained in a reference by Bockelman, D.E. and Eisenstadt, W.R., Combined Differential and Common-Mode Scattering Parameters: Theory and Simulation. IEEE Transactions on Microwave Theory and Techniques, v 43, n 7, 1530 (July 1995). Reverse isolation (Figure 24) is defined as SDD12. Power supply rejection ratio (PSRR) has been defined as V1 = V sin (2∏f1t) Adm As V2 = V sin (2∏f2t) The RMS average voltage of one tone is 1 T T ∫ (V1 ) 2 dt = 0 where Adm is the differential mode forward gain (SDD21), and As is the gain from the power supply pins (VCCI and VCCO, taken together) to the output (OPLO and OPHI, taken differentially), corrected for impedance mismatch. The following reference provides more information: Gray, P.R., Hurst, P.J., Lewis, S.H. and Meyer, R.G., Analysis and Design of Analog Integrated Circuits, 4th Edition, John Wiley & Sons, Inc., page 422. 1 2 where T is the period of the waveform. The RMS average voltage of the two-tone composite signal is 1 T T ∫ (V 1 + V2 ) 2 dt = 1 0 It can be shown that the average power of this composite waveform is twice (3 dB) that of the single tone. This also means that the composite peak-to-peak voltage is twice (6 dB) that of a single tone. This principle can be used to set correct input amplitudes from generators scaled in dBm and is correct if the two tones are of equal amplitude and are reasonably close in frequency. Rev. 0 | Page 24 of 28 AD8370 –22.5dB PORT 1 SERIAL DATA SOURCE VS 5.0V 1nF 1nF T1 T2 LTCH VCCO OCOM OPLO VCCO OCOM OPHI 1 2 3 4 5 6 7 8 ICOM CLCK 9 VOCM 10 PWUP 11 VCCI 12 ICOM 13 INHI 0Ω 14 MINICIRCUITS TC2-1T PORT 2 AGILENT 8753D NETWORK ANALYZER 15 INLO 16 DATA MINICIRCUITS TC4-1W AD8370 1nF 1nF VS 5.0V 1µF 1nF 1nF 1nF 03692-0-064 VS 5.0V 1µF Figure 63. PSRR Adm Test Setup PORT 1 BIAS TEE CONNECTION TO PORT 1 SERIAL DATA SOURCE 1nF VCCO OCOM OPLO OCOM OPHI 3 4 5 6 7 8 DATA 2 INLO 1 ICOM VCCO 9 LTCH 10 VOCM 11 CLCK 12 PWUP 13 VCCI 14 ICOM 15 INHI 16 MINICIRCUITS TC2-1T AD8370 1nF 1nF 1nF 03692-0-066 200Ω PORT 2 AGILENT 8753D NETWORK ANALYZER 1nF Figure 64. PSRR As Test Setup Rev. 0 | Page 25 of 28 AD8370 HP8133A 3GHz PULSE GENERATOR 50Ω INPUT AUX IN TEKTRONIX TDS5104 DPO OSCILLOSCOPE 50Ω INPUT TRIG 3dB ATTEN OUT 6dB SPLITTER 50Ω 50Ω INPUT INPUT SERIAL DATA SOURCE VS 5.0V 3dB ATTEN 475Ω 2dB ATTEN 52.3Ω LTCH VCCO OCOM OPLO OCOM OPHI 3 4 5 6 7 8 DATA 2 ICOM VCCO 9 CLCK 10 VOCM 11 PWUP 12 VCCI 13 1 INLO 200Ω 14 ICOM 3dB ATTEN AD8370 6dB SPLITTER OUT 15 INHI 16 475Ω 3dB ATTEN 2dB ATTEN 52.3Ω 1µF 1nF 1nF 1nF 03692-0-080 VS 5.0V 1µF VS 5.0V Figure 65. DC Pulse Response and Overdrive Recovery Test Setup AGILENT 8648D SIGNAL GENERATOR TEKTRONIX TDS5104 DPO OSCILLOSCOPE SERIAL DATA SOURCE TEKTRONIX P6205 ACTIVE FET PROBE RF OUT VS 5.0V 1nF 475Ω 1nF 14 13 12 11 10 9 DATA CLCK LTCH VCCO OCOM OPLO MINICIRCUITS JTX-2-10T 105Ω INHI ICOM VCCI PWUP VOCM VCCO OCOM OPHI AD8370 1 2 3 4 5 6 7 8 1nF 475Ω 1nF VS 5.0V VS 5.0V 1µF 50Ω INPUT 1nF 1nF 1nF 1µF Figure 66. Gain Step Time Domain Response Test Setup Rev. 0 | Page 26 of 28 03692-0-081 0Ω 15 INLO T2 16 ICOM T1 MINICIRCUITS TC4-1W 50Ω INPUT AD8370 AGILENT 8648D SIGNAL GENERATOR 10MHz REF OUT TEKTRONIX TDS5104 DPO OSCILLOSCOPE SERIAL DATA SOURCE RF OUT VS 5.0V 1nF 475Ω 1nF 11 10 9 OPLO CLCK 12 OCOM 13 LTCH 14 VCCO 15 DATA MINICIRCUITS JTX-2-10T 105Ω INHI ICOM VCCI PWUP VOCM VCCO OCOM OPHI AD8370 1 2 3 4 5 6 7 8 1nF 475Ω 1nF 10MHz IN OUTPUT AGILENT 33250A FUNCTION/ARBITRARY WAVEFORM GENERATOR VS 5.0V VS 5.0V 1µF 50Ω INPUT 1nF 1nF 52.3Ω TEKTRONIX P6205 ACTIVE FET PROBE 50Ω INPUT 1µF 1nF Figure 67. PWUP Response Time Domain Test Setup Rev. 0 | Page 27 of 28 03692-0-082 0Ω 16 INLO T2 MINICIRCUITS TC4-1W ICOM T1 AD8370 OUTLINE DIMENSIONS 5.10 5.00 4.90 16 BOTTOM VIEW 9 4.50 4.40 4.30 TOP VIEW 1 EXPOSED PAD (Pins Up) 6.40 BSC 3.00 SQ 8 1.05 1.00 0.80 1.20 MAX 0.15 0.00 SEATING 0.65 BSC PLANE 0.30 0.19 0.20 0.09 8° 0° 0.75 0.60 0.45 COMPLIANT TO JEDEC STANDARDS MO-153-ABT Figure 68. 16-Lead TSSOP (RE-16) ORDERING GUIDE Model AD8370ARE AD8370ARE-REEL7 AD8370-EVAL Temperature –40°C to +85°C –40°C to +85°C Package Description 16-lead TSSOP, Tube 16-lead TSSOP, 7” Reel Evaluation Board © 2004 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D03692-0-1/04(0) Rev. 0 | Page 28 of 28 Package Option RE-16 RE-16