19-4191; Rev 3; 10/96 in Atha ( EL} rc ats Peer Sofiware Ni PLld MAXIM RNC TTC and 8th-Order Continuous-Time General Description The MAX274 and MAX275 are continuous-time active filters consisting of independent cascadable 2nd-order sections. Each section can implement any all-pole bandpass or lowpass filter resoonse, such as Butter- worth, Bessel, and Chebyshev, and is programmed by four external resistors. The MAX274/MAX275 provide lower noise than switched-capacitor filters, as well as superior dynamic performance - both due to the con- tinuous-time design. Since continuous-time filters do not require a clock, aliased and clock noise are eliminated with the MAX274/MAX275. The MAX274 comprises four 2nd-order sections, permit- ting 8th-order filters to be realized. Center frequencies range up to 150kHz, and are accurate to within +1% over the full operating temperature range. Total harmonic distortion (THD) is typically better than -89dB The MAX275 comprises two 2nd-order sections, permit- ting 4th-order filters to be realized. Center frequencies range up to 300kHz, and are accurate to within +0.9% over the full operating temperature range. Total harmonic distortion (THD) is typically better than -86dB Both filters operate from a single +5V supply or from dual +5V supplies Applications Low-Distortion Anti-Aliasing Filters DAC Output Smocthing Filters Modems Audio/Sonar/Avionics Frequency Filtering Vibration Analysis Pin Configurations Active Filters Features @ Continuous-Time Filter - No Clock, No Clock Noise @ Implement Butterworth, Chebyshev, Bessel and Other Filter Responses @ Lowpass, Bandpass Outputs @ Operate from a Single +5V Supply or Dual +5V Supplies @ Design Software Available @ MAX274 Evaluation Kit Available @ 8th-Order Four 2nd-Order Sections (MAX274) 4th-Order Two 2nd-Order Sections (MAX275) # Center-Frequency Range: 150kHz for MAX274 300kHz for MAX275 @ Low Noise: -86dB THD Typical for MAX274 -89dB THD Typical for MAX275 @ Center-Frequency Accurate Over Temp: within +1% for MAX274 within +0.9% for MAX275 Ordering Information PART TEMP. RANGE PIN-PACKAGE MAX274ACNG OCto+70C =24 Narrow Plastic DIP MAX274BCNG OC to+70C = 24 Narrow Plastic DIP MAX274ACWI OC to+70C 28 WideSO MAX274BCWI 0C to+70C 28 Wide SO MAX274BC/D OC to +70C ~ Dice* Ordering Information continued on last page * Contact factory for dice specifications. Typical Operating Circuits ropwiew POA [7] 2a LPOD INA S faa} ND BPIA [3] AAAXLAA [75] BPID BPOA " MAX274 = BPOD V+ [5] 20] GND LPIA [6] fig} LPID LPB [7] rig] LPIC FC [a| 7] V- BPoB [ 9| Hie] BPOC apis [10| 15] BPIC INB [14] fra] INC Leo [12| 13] LPOC Pin Configurations continued on last page DIP IN +5V it BV wd 5 20-7 8 117 t= 24 OUT 20.1k 4 99 200k 195k 5 19 AY JY 3 21 195k 200k 4 23 102.9 40ke 41] MAAXLAA | 135 40k 239k 9 MAX274 15 200k 195k 95k 7 18 10 16 195k 200k 12 14 358k 40k 10kHz 8TH-ORDER BUTTERWORTH LOWPASS FILTER PIN NUMBERS ARE FOR DIP. Typical Operating Circuits continued on fast page MAXUM Maxim integrated Products 1 For pricing, delivery, and ordering information, please contact Maxim/Dallas Direct! at 1-888-629-4642, or visit Maxims website at www.maxim-ic.com. HM A3/24EMYOS/SZEXVN/VLEXVNMAX274/MAX275/Software/EV Kit 4th- and 8th-Order Continuous-Time Active Filters ABSOLUTE MAXIMUM RATINGS VHEtTOV~ ees -0.3V, 12V Operating Temperature Ranges: Input Voltage to GND (any input) ....... V--O.3V, V+ + 0.3V MAX27_ Ch ee OC to +70C Continuous Power Dissipation (TA = +70C) MAX27_ Eo we -40C to +85C MAX274 MAX27__MRG ... 2 ee -55C to +125C 24-Pin Narrow Plastic DIP Storage Temperature Range .............. -65C to +165C (derate 13.33mW/"C above +70C) ... 1067mW Lead Temperature (soldering, 10 sec) 28-Pin Wide SO (derate 12.50mW/"C above +70) . 1000mW 24-Pin CERDIP (derate 12.50mW/C above +70C) . 1000mW MAX275 20-Pin Plastic DIP(derate 11.11mW/C above +70C) 889mW 20-Pin Wide SO (derate 10.00mW/C above +70) .. 800mW 20-Pin CERDIP (derate 11.11mW/C above +70C) .. 889mMW Lette ee +300C Stresses beyond those listed under Absolute Maximum Ratings" may cause permanent damage to the device. These are stress ratings only, and functional operation of the device at these or any other conditions beyond those indicated in the operational sections of the specifications is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. ELECTRICAL CHARACTERISTICS MAX274 (V+ = 5V, V- = -5V, test circuit A of Figure ta, Ta = TMIN to Tmax, unless otherwise noted.) PARAMETER | SYMBOL | CONDITIONS MIN TYP MAX | UNITS FILTER CHARACTERISTICS Maximum Operating Frequency 10 MHz Center-Frequency Range Fo (Note 1) 120 Hz MAX274A -1.0 1.0 Center-Frequency Accuracy Fo % MAX274B -1.4 1.4 MAX274A -10 10 Q Accuracy - Unadjusted % MAX274B -15 15 Q Accuracy - Adjusted Scaled for bandwidth compensation +2.8 % Fo Temperature Coefficient AFo/AT (Note 2) -28 ppm/C Q Temperature Coefficient AQ/AT (Note 2) 160 ppm/"C . ; LPO_, Figure ta, |_1H2 to 10H2 23 Wideband Noise VNOISE ~ I S Vi test circuit B 10Hz to 10kHz 120 wees DC CHARACTERISTICS Assume ideal MAX274A 2 2 DC Lowpass Gain Accuracy HoOLP resistors MAXO74B 3 3 %o MAX274A -200 200 LPO_ MAX274B -300 300 Offset Voltage at Outputs Vos mV MAX274A -40 40 BPO_ MAX274B -80 80 Offset Voltage Drift AVos/AT 20 pvPC Leakage Current at FC Pin IFC -10 10 pA DYNAMIC FILTER CHARACTERISTICS Fest = 1kHz, Figure 1a, -86 Signal-to-Noise plus Distorti sinap est erroultB LPO dB ignal-to-Noise plus Distortion ! 9 P Frest = 10kHz, | VLPO = 8Vp-p Figure 1a, -82 test circuit C 2 MAXIM4th- and 8th-Order Continuous-Time Active Filters ELECTRICAL CHARACTERISTICS (continued) MAX274 (V+ = 5V, V- = -5V, test circuit A of Figure 1a, Ta = TMIN to TMAX, unless otherwise noted.) PARAMETER SYMBOL CONDITIONS MIN TYP MAX | UNITS Output Voltage Swing VOUT LPO_, BPO_.RLOAD = 5kQ +3.25 +450 V Slew Rate SR 10 V/us Gain-Bandwidth Product GBW 75 MHz POWER REQUIREMENTS Supply Voltage Range VSUPP (Note 3) 42.37 +5.50 V Supply Current Ic For V+, V- 20 30 mA Power-Supply Rejection Ratio PSRR V+ = 5V + 100mVp-p at 1kHz, V- = -5V -30 dB Note 1: Center frequencies (Fos) below 100Hz are possible at reduced dynamic range. Note 2: Assume no drift for external resistors. Note 3: See Figure 9 for single-supply operation. ELECTRICAL CHARACTERISTICS MAX275 (V+ = 5V, V- = -5V, test circuit A of Figure 1b, TA = TMIN to Tmax, unless otherwise noted.) PARAMETER | SYMBOL | CONDITIONS MIN TYP MAX | UNITS FILTER CHARACTERISTICS Maximum Operating Frequency 10 MHz 1 Center-Frequency Range Fo (Note 1) Sooke Hz MAX275A -0.9 0.9 Center-Frequency Accuracy Fo % MAX275B -1.4 1.4 ; MAX275A -8 8 Q Accuracy - Unadjusted % MAX275B -12 12 Q Accuracy Adjusted Scaled for bandwidth compensation +1 % Fo Temperature Coefficient AFof/AT (Note 2) -24 pom/C Q Temperature Coefficient AQ/AT (Note 2) 38 ppm/C ; ; LPO_, test circuit B | 142 to 10Hz 6 Wideband Noise VNOI ~ V NOISE | of Figure 1b, 10Hz to 10kHz 42 pyRMS DC CHARACTERISTICS Assume idea! MAX275A a 1 DC Lowpass Gain Accuracy HoLP resistors MAX275B 5 5 % MAX275A -125 125 LPO_ MAX275B -250 250 Offset Voltage at Outputs Vos mV MAX275A -50 50 BPO_ MAX275B -100 100 Offset Voltage Drift AVos/AT 20 uv/?C Leakage Current at FC Pin IFC -10 10 pA DYNAMIC FILTER CHARACTERISTICS FTEsT = 1kHz, test circuit B of -89 Signal-to-Noise plus Distorti sinap |-Figure : LPO dB ignal-to-Noise plus Distortion = 9 P Frest = 10kHz, | VLPO = 8Vp-p test circuit C of -83 Figure 1b, MA AMXIMA 3 HM A2/84lM}jJOS/SZEXVW/PLEXVINMAX274/MAX275/Software/EV Kit 4th- and 8th-Order Continuous-Time Active Filters ELECTRICAL CHARACTERISTICS (continued) MAX275 (V+ = 5V, V- = -5V, test circuit A of Figure 1b, Ta = TMIN to TMAX, unless otherwise noted.) PARAMETER SYMBOL CONDITIONS MIN TYP MAX | UNITS Output Voltage Swing VOUT LPO_, BPO_, RLOAD = 5kQ +3.25 +4.50 V Internal Amplifier Slew Rate SR 10 Vius Gain-Bandwidth Product GBW 15 MHz POWER REQUIREMENTS Supply Voltage Range VSUPP (Note 3) 2.37 +5.50 Vv Supply Current Ic For V+, V- 10 24 mA Power-Supply Rejection Ratio PSRR V+ = 5V + 100mVp-p at 1kHz, V- = -5V -35 dB Note 1: Center frequencies (Fos) below 100Hz are possible at reduced dynamic range. Note 2: Assume no drift for external resistors. Note 3: See Figure 9 for single-supply operation. FILTER WIDEBAND RESPONSE USING TYPICAL OPERATING CIRCUIT 7 O_ Ta=+25C -20 -40 GAIN (dB) -60 -80 -100 1k 40k 100k 1M FREQUENCY (Hz) Fo ERROR vs. FREQUENCY 0 05 -10 p> < = Q= 06) ce 15} =! / \ 2 Q=25 \ ~ -2.0 INN N to NX ~25 mH lA=Ht, SS -3.0 100 FREQUENCY (kHz) 150 GAIN (dB) 1 FILTER RESPONSE USING TYPICAL OPERATING CIRCUIT 00 1k 10k FREQUENCY (Hz) Q ERROR vs. FREQUENCY 100k 50 FC = V+ 40 | Ty =425C / 30 f _ Y z Pras = 20 =t Q=06 & 10 Z 0 = Li to pf a -10 10 100 150 FREQUENCY (kHz) Typical Operating Characteristics-MAX274 MAXIM4th- and 8th-Order Continuous-Time Active Filters Typical Operating Characteristics-MAX274 (continued) Fo ERROR vs. FREQUENCY Q ERROR vs. FREQUENCY Q=10 Q=5 Q=25 Q=1 Q=06 g Z 2 2 =) ol ud Q = 0. Q=2. FC =V- Q=5 Ta=+25C Q= 10 ~ 100 1 10 100 100 1 10 400 FREQUENCY (kHz) FREQUENCY (kHz) Fo ERROR vs. FREQUENCY Q ERROR vs. FREQUENCY 50 FC = GND 40 | Ta=+25C | 8230 -_ _ 2 Q- 20 ae - ae =10 ef 5 83 & Q=1 ea Q=25 5 Q=25 Ss Gt os3 0=06 0 -10 1 10 100 150 1 10 400 150 FREQUENCY (kHz) FREQUENCY (kHz) TYPICAL DISTRIBUTION OF CENTER-FREQUENCY ERRORS TYPICAL DISTRIBUTION OF Q ERRORS | 30 Ta= 425C Ta=425C = 10 4 Fo = 10kHz, Q= 10 = Fo = 10kHz, Q = 10 3 S 20 5 5 oa a & wo a S 54 S & S 10- a a > > ee Fe L | _ -.505 -104 0.500 2.100 5.175 Fo ERROR (%) Q ERROR (%) MAXLMA 5 UM A2/O4eMYOS/SZEXVW/PLEXVNMAX274/MAX275/Software/EV Kit 4th- and 8th-Order Continuous-Time Active Filters GAIN RELATIVE TO INPUT SIGNAL (dB) SINAD (dB) -20 -40 -60 40 -50 -60 -70 -80 -90 Typical Operating Characteristics-MAX274 (continued) FFT PLOT OF 1kHz TEST SIGNAL , we ee Cb Ves SV, Ve =-8V pp oot (PO_ : iN 8Vp.p. at 1kHz | _. RLoaD = 5) ake Ta = 425C FREQUENCY (kHz) SIGNAL TO NOISE + DISTORTION (SINAD) vs. OUTPUT SWING I oy] V+ = 5V, V- =-5V LPO_ FIN = 1kHz Ta=+25C \ N\ RLOAD = 5kQQ NY RLOAD = ee NN res_] 0 2 4 6 8 OUTPUT SWING, (Vp-p) -20 -40 -60 -80 -100 GAIN RELATIVE TO INPUT SIGNAL (dB) -120 40 35 30 25 20 SUPPLY CURRENT (mA) FFT PLOT OF 10kHz TEST SIGNAL [ i | | Vee 5V, V- = 5y hes 8Vp p. at 1kKH. he 0 Z RLOAD = Ha Tas 425C pote ; J} oj | Mm 20 40 60 80 FREQUENCY (kHz) SUPPLY CURRENT vs. SUPPLY VOLTAGE | Ta=+25C ny 6 8 10 SUPPLY VOLTAGE (V+ to V-} (V) SPECTRAL NOISE DENSITY vs. FREQUENCY NOISE DENSITY uVANH2 NOISE DENSITY uVANHZ LPO_ 1k 10k FREQUENCY (kHz) 100k MAXIM4th- and 8th-Order Continuous-Time Active Filters Typical Operating Characteristics-MAX275 FILTER WIDEBAND RESPONSE USING TYPICAL OPERATING CIRCUIT FILTER RESPONSE BPO_ Ta = 425C S z s 100kH7 0.1dB = CHEBYSHEV BANDPASS = Ta= 425C = -180 Si ~360 NN Pay eT 100k 1M sok 100k 1M FREQUENCY (Hz) FREQUENCY (Hz) CENTER-FREQUENCY ERROR vs. PROGRAMMED Q ERROR vs. PROGRAMMED CENTER FREQUENCY CENTER FREQUENCY 50 FC = V+ f v6 Ta= 425C } 25 40 Q=20 DO Ne ISTE Zot AN ERROR (%) ERROR (%} 2 iow \ \ \ booy ne NA im \ IM N 10 100 300 10 100 300 FREQUENCY (kHz) FREQUENCY (kHz) MAAXLAA 7 UM AF/e4eMYOS/GZEXVW/PLEXVNMAX274/MAX275/Software/EV Kit 4th- and 8th-Order Continuous-Time Active Filters ERROR (%) ERROR (%) TYPICAL DISTRIBUTION (%) OT 15 FC = GND Typical Operating Characteristics-MAX275 (continued) CENTER-FREQUENCY ERROR vs. PROGRAMMED CENTER FREQUENCY na NO wn Q= Q= O= Q= WN+O9 Doon wou a sh w ooo FC = GND Ta= 425C 1 10 100 300 FREQUENCY (kHz) CENTER-FREQUENCY ERROR vs. PROGRAMMED CENTER FREQUENCY Q=5 =25 FC =V- Ta=+25C ~ 100 1k 10k 30k FREQUENCY (Hz) TYPICAL DISTRIBUTION OF CENTER- FREQUENCY ERRORS Ta=+25C Fo = 10kHz, Q = 10 -.257 0 141 Fo ERROR (%) ERROR (%) ERROR (%) TYPICAL DISTRIBUTION (%) FC = GND TA = +25C Q ERROR vs. PROGRAMMED CENTER FREQUENCY Q=50 = 30 Q=25 Q=1 =0.75 Q=0.6 =5 =25 =0.75 =1 Q=30 Q=50 =06 1 10 100 300 FREQUENCY (kHz) Q ERROR vs. PROGRAMMED CENTER FREQUENCY 50 TT Fo=V, wt ap | Taz 425C . 4 = 20 A 30 4 4 M4 20 J y Aa 10 0-06 T7Ja=1070=5 cy iy bi | i 0 == ee Q=1 Q=25 -10 100 1k 10k 400k FREQUENCY (Hz} TYPICAL DISTRIBUTION OF Q ERRORS lf 4 Ta=+25C 15 Fo = 10kHz, Q= 10 FC =GND 10 4 5 4 0 -2.893 ~4.635 QERROR (%) MAXIM4th- and 8th-Order Continuous-Time Active Filters Typical Operating Characteristics-MAX275 (continued) FFT PLOT OF 1kHz TEST SIGNAL 0 = == T s a SV S -20 2 Ibo 5V,V-=-5V = Ite = 8Vp, atti & 40): RLOAD = 5kQ2 an | Ta=+25C 2 0 {ft | 2 ppt Qo 1 | ! r , | i i S -80 4 wide wee pe : | cw -100 i | z : a | -120 1: a 0 2 4 6 8 10 FREQUENCY (kHz) SUPPLY CURRENT vs. SUPPLY VOLTAGE 12 11 Ic (mA) oOo ep MN gm WN MAAXISAA Ta=+25C 3 4 5 6 SUPPLY VOLTAGE (V+ TO V-) (V) FFT PLOT OF 10kHz TEST SIGNAL 0 S = -20 <= 5 V+ = BV, V- = -5V 2 Vk = BVpo, at 1kH > Ra p.p. at 1kHz = +0 RLOAD = OK 2 Ta=+25C eS = -80 = m -100 = & -t29l 0 20 40 60 80 FREQUENCY (kHz) SIGNAL-TO-NOISE + DISTORTION (SINAD) vs. OUTPUT SWING 40 rT Vos 60 LPO_ J FIN = 1kHz FIGURE1, -60 TEST CIRCUIT Bd 3 \ Ta=+25C QoQ S an -70 NS INQ RLOAD = 5kQ -80 MN RLOAD = c2 r~_| -90 0 2 4 6 8 NOISE-SPECTRAL DENSITY vs. FREQUENCY NOISE DENSITY SNAP NOISE DENSITY pVNHz BPO_ toogr6icrs FREQUENCY (kHz) OUTPUT SWING (Vp-p) 10 UM AW/e4eM}jOS/SZEXVW/PLEXVANMAX274/MAX275/Software/EV Kit 4th- and 8th-Order Continuous-Time Active Filters GND 5V [= 45V SECTION A SECTION B Naf 7} of 8] SI a} * GND FC Io, RIA RIB we INA LPOD INA INB 4 BPA BPOA pip 122_ RP a nape R4A 8) py, AAAXIAA (Pip [2 SECTION D BPOA R3A ___4T peo ppopf@ sP8B_. gpg SECTIONA 3 4 R4A 5) AAaxim 16 R4B we BPIA BPOD BPD LPIA MAX275 LPiB___\\/ LPOA IND | Ty BPIA Pa} NB RW IND R2A 2 19 R2B 1 43 LPOA LPOA LPOB LPOB Ws INB LPOC | ; . ROC BPB 1 PoB apic HS cre a RAB O77 tpi LPiC 18 SECTION C = V- GND V+ = SECTIONB 10 a [9 7 BPIB BPOC | ay OND ASV LPOB INC PIN NUMBERS ARE FOR DIP. PIN NUMBERS ARE FOR DIP. TEST Fo Q R1X R2X RSX R4X FCPIN TEST Fo Q RIX R2X RSX R4X FC CIRCUIT (kHz) (kQ) (kQ) (KQ) (kK) CIRCUIT (kHz) (kQ) (kQ) (kQ))(kQ)~ PIN A 10 10 80 200 400 195 GND A 10 10 80 200 400 195 GND B 10 0.707 40 200 28 195 GND B 10 0.707 40 200 28 4195 GND Cc 100 0.707 40 20 48 15 V+ Cc 100 0.707 40 20 48 14 V+ Figure 1a. MAX274 Connection Diagram and Test Circuit Detailed Description The MAX274 contains four identical 2nd-order filter sec- tions while the MAX275 contains two sections. Figure 2 shows the state-variable topography employed in each filter section. This topography allows simultaneous low- pass and bandpass functions at separate outputs. The MAX274/MAX275 employ a four-amplifier design, chosen for its relative insensitivity to parasitic capacitan- ces and high bandwidth. The built-in capacitors and amplifiers, together with external resistors, form cas- caded integrators with feedback to provide simultaneous lowpass and bandpass filtered outputs. To maximize bandwidth, the highpass (HP) node is not accessible. A 5kQ resistor is connected in series with the input of the last stage amplifier to isolate the integration capacitor from external parasitic capacitances that could alter the filter's pole accuracy. Although a notch output pin is not available, a notch can be created at the pole frequency by summing the input 10 Figure 1b. MAX275 Connection Diagram and Test Circuit and bandpass output. See Creating a Notch Output Section Filter Design Procedure Figure 3 outlines the overall filter design procedure. Maxim's Filter Design Software is highly recommended. This software automatically calculates filter order, poles, and Qs based on the required filter shape, so no manual calculations are necessary. Menu-driven commands and on-screen filter response graphs take the user through the complete design process, including the selection of resistor values for implementing a filter with the MAX274/MAX275. See Maxim Filter Design Software section. If designing without the filter software, see the filter design references listed at the end of this data sheet. These references provide numerical tables and equations needed to translate a desired filter response into order, poles, and Q. Once these three parameters have been calculated, see the next section, Translating Fo/Q Pairs into MAX274/MAX275 Hardware (Resistor Selection). MAXIAA4th- and 8th-Order Continuous-Time Active Filters CONNECT | RY/RX* FC TO: (kQ) V+ 13/52 GND 65/13 V- 325/13 LOWPASS OUTPUT 79.575pF 50k >| WN RI 50k INPUT > Bg BPO | WA i R3 BANDPASS QUTPUT Fotto-( Ps ay exw Hoge = BP GAIN AT Fo = EQUATIONS DO NOT INCLUDE COMPENSATION FACTORS THAT CORRECT SMALL ERRORS IN Fo AND Q DUE TO AMPLIFIER BANDWIDTH LIMITATIONS. SEE TYPICAL ; Ry voc (VR OPERATING CHARACTERISTICS FOR CORRECTION FACTORS. -(V aeatray }*ofe] _ ser-tramnoe-le * WHILE THE RATIO RY/RX IS ACCURATELY CONTROLLED, PROCESS VARIATIONS AND TEMPERATURE DRIFT RESULT IN UP TO +30% VARIATION OF ACTUAL VALUES OF RX AND RY. HM AZ/e4eMYOS/SZLEXVW/PLEXVW LOWPASS OUTPUT BANDPASS OUTPUT L A Hop Hopp Howe | 0.707 Howe __ 0.707 Hosp = = = Zz 3 5 Fp. FC - FL FeK FH i? 2 F(LOG SCALE) F(LOG SCALE) Wo G(S) = Ho_p = Hotp = LOWPASS OUTPUT GAIN AT DC S? + S(t) + t09 G(S) = Hopp -,Se) _ Fo = wo/2n = POLE FREQUENCY S* + S(@00) + Wo Hopp = BP GAIN AT w= @o Fo = Wo/2n = THE CENTER FREQUENCY OF THE COMPLEX POLE PAIR. INPUT-OUTPUT PHASE SHIFT IS - 180 AT Fo. PPK THE FREQUENCY AT WHICH BPO_ GAIN IS THE GREATEST (MAY NOT BE EQUAL 0}. Q = THE QUALITY FACTOR OF THE COMPLEX POLE PAIR. ALSO THE RATIO OF Fo, TO -30B . mn BANDWIDTH OF THE SECOND-ORDER BANDPASS RESPONSE. Q= a oA Figure 2. Single 2nd-Order Filter Section MAXIMA 11MAX274/MAX275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters LOWPASS BANDPASS HIGHPASS NOTCH ALLPASS dB dB dB dB dB 0 a f f f f f NONSTANDARD GLASSIC" TYPE "CLASSIC" TYPE NONSTANDARD IDENTICAL DESIGN (standard, simplest WITH ZEROS DESIGN CASCADED SEE MAX260 Have one or to design) (when absolute @ Includes zeros SECTIONS FILTER more poles to Butterworth Steepest filter is in transter Sharp band for FAMILY be implemented, Chebyshev required) function passing asingle meres | | Sess +l eve) function ai N le , h f f f f f GO TO CASCADING scan MaKe IDENTICAL CREATING A ORMAX260. BANDPASS NOTCH OUTPUT FILTER FAMILY SECTIONS IN SECTION DATA SHEET Y q t BUTTERWORTH CHEBYSHEV BESSEL Flat passband Steeper rolloff Least steep rolloff Passhand sipple Constant delay from input to assband rat oufpul, ragpadiess of FEUER at) Fastest settling time response hah to step input- no overshoot TY) DN PSEA IN ouT} FA - TIME USE TABLE, EQUATIONS MAXIM FILTER DESIGN SOFTWARE (SEE REFERENCES) TO CALCULATE Fo/Q PAIRS t Selects gains for individual sections to provide desired overall fitter gain while Maximizing dynamic range LI FOLLOW INSTRUCTIONS IN DATA SHEET FOR RESISTOR SELECTION Selects resistors for each filter section Piots filter response, phase, and delay alculates order, Fo/Q pairs based on desired filter shape for Butterworth, Chebyshev, ssel types i BUILD FILTER Figure 3. General Filter Design Flowchart 12 MAAXLMA4th- and 8th-Order Continuous-Time Translating Calculated F,/Q Pairs into MAX274/MAX275 Hardware (Resistor Selection) If the filter design procedure has been completed as. outlined in Figure 3, with the exception of external resistor selection, follow these steps: 1. Check all F,/Q pairs for realizability. The MAX274/MAX275 have limits on which Fo/Q values can be implemented. These limits are bound by finite amplifier gain-bandwidth and amplifier load drive capability (which limit the highest frequency Fo/highest Qs) as well as amplifier noise pickup and susceptibility to errors caused by stray capacitance (which sets a low-frequency limit on the poles). Refer to Figure 4 to be sure each Fo/Q pair is within the "realizable" portion of the graph. If filter Qs are too high, reduce them by increasing the filter order (that is, increase the number of poles in the overall filter). High-frequency Fos (up to 400kHz) and high Qs outside of Figure 4s limits are also realizable, but Fp and Q will deviate significantly from the ideal. Adjust resistor values by prototyping. To implement Fos less than 100Hz, see High- Value Resis- tor Transformation section. 2. Calculate resistor values for each section (Fo/Q pair). Calculate resistor values using graphs and equa- tions in steps A through D of this section. Begin by estimating required values according to the graphs; then use the given equations to derive a precise value. USABLE CENTER-FREQUENCY, Q RANGE 50 MAX274 40 MAX275 Usable 30 Fo, Q | Range \ 20 FILTER (Q) 100 1 10 Fo (KHz) 100 300 Figure 4. Usable Fo, Q Range. See Translating Fo/Q Pairs into Hardware (Resistor Selection). MA MXILIA Active Filters Resistor values should not exceed 4MQ because parasitic capacitances shunting such high values cause excessive F,/Q errors. Values lower than 5kQ for R2 and R3 are not recommended due to limited amplifier output drive capability. For cases where larger values are un- avoidable (as in low-frequency sections) refer to the High-Value Resistor Transformation section. The Frequency Control (FC) pin is connected to V+, GND, or V- and scales R3 and R1 to accomodate a wide range of gains and Q values. Different FC settings may be chosen for each section. Refer to the FC Pin Connection section. The steps for calculating resistor values are given below. STEP A. CALCULATE R2. _ (2x 10%) R2 Fo RESISTOR R2 vs. DESIRED CENTER FREQUENCY 10M 1M R2 (Q) 100k 10k 100 tk 1k DESIRED Fo (Hz) o @ @ USE RESISTOR "T-NETWORK TO REDUCE VALUE (SEE HIGH-VALUE RESISTOR TRANSFORMATION SECTION) 100k 300k Resistors R2 and R4 set the center frequency. STEP B. CALCULATE R4. R4 = R2 - 5kQ R4 may be less than 5kQ because an internal series 5kQ resistor limits BPO_ loading 13 HM A3/e4eMY0S/SLEXVN/PLEXVNMAX274/MAX275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters STEP C. CALCULATE R3. R3 sets the Q for the section. R3 values are plotted assuming Q = 1; since R3 is proportional to Q, multiply the graphs value by the desired Q. Given Q, three choices exist for R3, depending on the FC setting. Choose a setting that provides a reasonable resistor value (5kQ < R3 < 4MQ). R3 > 4MQ may be used if unavoidable refer to the High-Value Resistor Transformation section for an explanation of resistor "Ts." RESISTOR R3 vs. CENTER FREQUENCY 10M =1,FC =V+ 1M S & 10k 100 tk Wk 100k 300k Fo (Hz) e@ @ @ USE RESISTOR T-NETWORK TO REDUCE VALUE (SEE HIGH-VALUE RESISTOR TRANSFORMATION SECTION) Scale R3 to desired Q pg - (2) (2X10) , (Fy) Fo RY CONNECT FoTO: | AX/RY V+ 4 GND 15 V- 4/25 14 STEP D. CALCULATE Rt. R1 sets the gain. If individual section gains have not yet been calculated, refer to Cascaded Filter Gain Optimiza- tion, Ordering of Sections. R1 is inversely proportional to LP gain. R1 values for gains of 1 and 10 are plotted; scale R1 according to desired gain. Lowpass Filters: The FC pin setting was chosen in Step C (or from previous section calculations). LP GAIN RESISTOR R1 vs. FREQUENCY 1. FC =V- Hop = 10, FC = Hotp= 10. FC = V- RI (Q) 100 1k 10k 100k 300k Fo (Hz) USE RESISTOR "T-NETWORK 10 REDUCE VALUE (SEE HIGH-VALUE RESISTOR TRANSFORMATION SECTION) 9 Ry = 2) 10") (Fr ~ (Fo) (HoLP) | RY CBRNE | mony V+ 4/1 GND 15 V- 1/25 where Hop is the gain at LPO_ at DC. MAXLAA4th- and 8th-Order Continuous-Time Bandpass Filters: BP GAIN RESISTOR R1 vs. RESISTOR R3 10M Ht Hosp = 0.5 Hogp = 1 Hopp = 1 1M R1 (Q) 100k 10k 100k 1M R3 (Q) e e USE RESISTOR T-NETWORK" TO REDUCE VALUE (SEE HIGH-VALUE RESISTOR TRANSFORMATION SECTION) __R3_ Hopp where Hopp is the gain at BPO_ at Fo. 3. Recalculate resistor values to compensate for filter amplifier bandwidth errors. Some of the Typical Operat- ing Characteristics graphs show deviations in Fp and Q compared with expected values, due to gain rolloff of the internal amplifiers. If desired, correct these deviations by recalculating values R1- R4. 4. Build a filter prototype. Build and test all filter designs! Refer to the Prototyping, PC-Board Layout sec- tion of this data sheet. For applications that require high accuracy (for example, those with filter sections containing Qs greater than 10) or those that use a ground plane, a final prototype tuning procedure is recommended. Build a prototype filter; then adjust resistor values of each section until desired ac- curacy is achieved. High-Value Resistor Transformation High-value resistors (greater than 4MQ) used in the MAX274/MAX275 filter circuit introduce excessive Fp and Q errors. To reduce the impedance of these feedback paths while maintaining equivalent feedback current, use the resistor "T" method shown in Figure 5. MA MXKIAA Active Filters Fos less than 100Hz can be realized using T-networks. T-networks provide the equivalent of large resistor values for R2, R3, and R4, necessary for low-frequency filters; however, T-networks reduce dynamic range by attenuat- ing the input signal level. Note that parasitic capacitan- ces across these high resistor values affect the filter response at high frequencies. For best results, build a prototype and check its performance thoroughly. Odd Number of Poles For lowpass designs containing an odd number of poles, add an RC lowpass filter after the final filter section. The value of RC should be: RC = 1/2nFo where Fo is the desired real pole frequency. If required, buffer the RC with an op amp. Inmany cases it may be advantageous to simply increase the filter order by 1, and implement it with an additional 2nd-order section. FC Pin Connection Connect FC to GND for all applications, except where resistor values fall below 5kQ (at high Fos, low Qs). In these cases connect FC to V+. For low Fes and high Qs, connect FC to V- to keep the value of R1 and R3 below 4MQ. Fg and Q errors are significantly higher when FC is con- nected to V+ or V- (see Typical Operating Charac- teristics). Adjusting resistor values compensates for these errors, since the errors are repeatable from part to part. Note that noise increases threefold when FC is connected to V+. Cascading Identical Sections for Simplest Bandpass If designing a bandpass filter where a single frequency (or a very narrow band of frequencies) must be passed, several 2nd-order sections with identical Fos and Qs may be cascaded. The resulting Q (selectivity) of the filter is a function of the individual sections Qs and the number of sections cascaded: Q Ot= ERT where Qt is the overall cascaded filter Q, Q is the Q of each individual section, and N is the number of sections. 15 UM AZ/e4eM1YOS/SZEXVN/PLEXVNMAX274/MAX 275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters NORMAL FEEDBACK RESISTOR R<4MQ R (R2, R3, R4) RESISTOR USE THE T-NETWORK FEEDBACK METHOD WHENEVER THE CALCULATED VALUE FOR R2, R3, OR R4 IS GREATER THAN 4MQ. A VOLTAGE DIVIDER CONSISTING OF RA AND RB DROPS THE VOLTAGE ACROSS _ PERMITTING A LOWER VALUE FOR RY WHILE MAINTAINING AN EQUIVALENT CURRENT INTO THE SUMMING NODE (i). Rea T-NETWORK FEEDBACK ALLOWS REDUCED VALUE FOR R V Reo= LET R <4MQ _pl, RA " Rea=R +55 (RB +R) (RA) (RB) RA +RB SOLVE FOR RA AND RB LET = 10kQ Figure 5. Resistor T-Networks Reduce Resistor Values 16 MAXIAA4th- and 8th-Order Continuous-Time Rin RG SIGNAL MAXIAN IN MAX274 MAX275 LPO_ z R2 BPI 1 $24.02 Fo =FNOTCH = \ aa @x104 8+ S(@0t) + 00 2 1 RY a= Vm Gam (r) WHERE 0 = 2x Fo How = GAIN = F& GCS) = Hon Figure Ga. Creating a Notch Output Creating a Notch Output Anotch (zero)can be created in the filter response by summing the input signal with BPO_ using an extemal op amp (Figure 6a). The notch will have the poles and Q characteristics of the 2nd-order section, as well as a zero at the pole frequency (transfer function given in Figure 6a). Hopp (BP gain at Fo) must be accurately set to unity so the input signal summed with BPO_ cancels precisely at the pole frequency. The notchs maximum attenuation is therefore a function of the accuracy of R1, R3, Rin, and Rep. Anotch can be used to create a null within the passband of a lowpass filter to reject specific frequencies (see Applications section). s = Rin, Rep, R3 = 20 1% TOL, () FNOTCH = 2kHz -30 Q=5 -40 100 1k 10k FREQUENCY (Hz) Figure 6b. Notch Response MA MXIAA Active Filters Cascaded Filter Gain Optimization, Ordering of Sections Gains across the individual sections in a filter may be set an infinite number of ways, as long as the total gain from filter input to output is correct. Often, gains cannot be equally divided among sections, since different Fos and Qs create gain peaks and valleys at different frequencies for each section. The goal in choosing gains is to prevent section outputs from swinging beyond the +3.25V limit (using +5V sup- plies) while the full input signal is applied. On the other hand, if section gains are set too low and only a small proportion of output range is used, the noise factor in- creases. An optimal gain distribution between sections allows each section to swing as close to +3.25V as possible in a wide range of frequencies. Check the unused output (BPO_ or LPO_), and the inter- nal HP node for overvoltage, since clipping at any node will cause distortion at the outputs. The HP node is not available for probing (Figure 2); however, its gain may approach RX / R1. Low R1 values and connecting FC to V+ (which sets RX as high as 64kQ) may cause this node to clip. Maxim's Filter Design Software allows optimum gain by plotting output gains of each successive cascaded filter section, including the internal node. Gains may be ad- justed manually and sections reordered for the best overall dynamic range. To optimize gain without the help of software, begin by ordering the sections from lowest Q to highest Q. Divide gains equally between sections, setting each section gain to: Ho = ACN) where A = overall filter gain Ho = Hopp for bandpass designs (gain at Fg) Ho = Hop for lowpass designs (gain at DC) N= total number of sections This approach offers a good first-pass solution to clip- ping problems in the high Q sections by keeping gains low in the first (low Q) sections. The gains may then be adjusted in hardware to maximize overall dynamic range. 17 UM AZ/e4eMY0S/SLEXVW/PLEXVNMAX274/MAX275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters Figure 7b. MAX275 Suggested PC-Board Layout for DIP 18 MAXIAA4th- and 8th-Order Continuous-Time Resistors Aside from accuracy, the most important criterion for resis- tor selection is parasitic capacitance across the resistor. Typical capacitance should be less than 1pF. Precision wire-wound resistors exhibit several picofarads, as well as unacceptable inductance - DO NOT USE THESE. Capacitance effectively reduces the resistance at high frequencies (especially when using high-value resistors), and causes phase shifts in feedback loops. Do not mount resistors in sockets. Socket capacitance appearing across resistors is often several picofarads, and will cause sig- nificant errors in Fo and Q. Metal-film resistors minimize noise better than carbon types. Prototyping, PC-Board Layout | For highest accuracy filters, build the filter prototype on a PC board with a layout as similar as possible to the final production circuit. If a ground plane will be used in production, build prototype filters on a copper board. Do not use push-in type breadboards for prototyping pin- to-pin capacitance is too high. For faster prototyping, the MAX274 evaluation kit includes a PC-board circuit to test designs. Layout-sensitive errors, though repeatable from part to part, vary according to resistor placement, trace routing, and ground-plane layout. For highest accuracy, use the recommended layout provided in Figures 7a and 7b. Keep ail traces, especially LPI_ and BPI_ , as short as possible. LPI_ and BPI_ are particularly sensitive to ground capacitance, and may cause errors in Q. If a ground plane is used, tune the prototype filter by adjust- ing resistor values to cancel errors caused by ground capacitance. Prevent capacitive coupling between pins. Coupling between BPI_ and BPO_ can cause Fo errors; capacitance across resistors connecting IN and BPO_ (R3), BPI_ and LPO_ (R2), and BPO_ and LPI_ (R4) cause Fy and Q errors. Minimize these errors with "tight" (shortest trace) layout practices. Measuring F. and Q For multiple-order filters, measure each section in- dividually, before cascading, to verify correct Fp and Q. For best results, measure BPO_ with a spectrum analyzer. Fa is the frequency at which the input and BPO_ are 180 out of phase. Q is the ratio of Fpx to BPO_'s - 3dB bandwidth (Figure 2), where Fpx is the frequency at which BPO_ gain is the greatest (which may not be equal to Fo). PAMXLAAN Active Filters Filter F. and Q Accuracy Fo sensitivity to external resistor tolerance is 1:1 for example, use of 1% tolerant resistors for R2 and R4 adds +1% error to Fo (which should be added to the +1% tolerance of the MAX274/MAX275, guaranteed over temperature). Q errors are of greater magnitude, since they are a function of the internal resistor divider (contro!- led by the FC pin) and also involve R3. Typical Q error distributions are given in the Typical Operating Charac- teristics, additional Q errors associated with resistor tolerances are a function of R2, R3, and R4, and must be calculated according to the values used. DC Offset Removal Figures 8a and 8b show methods for removing the DC offset voitage at LPO_. The first method shows adjustable DC nulling signals injected into either BPI_ or the filter inout. RtRinm Must be adjusted until DC offset is nulled at the LPO_ (Figure 8a). Figure 8b shows a trimless solution for lowpass filters that removes DC offset by AC coupling the LPO_ output, while allowing a DC path through R from the input. At DC and low frequencies, the output is equal to the prefiltered signal input (across R); at higher frequen- cies, C conducts and the output equals the signal at LPO_. The external RC pole should be set at least one frequency decade lower than the overall filter Fo. A iow offset amplifier can buffer the output signal, if desired. For bandpass filters, a simple buffered RC highpass filter at the output removes DC offset. Noise and Distortion Noise-spectral density is shown in the Typical Operating Characteristics. The noise frequency distribution is shaped by the filter gain and response (higher Q section will have a proportionally higher noise peak around the pole frequency), as well as by amplifier 1/f noise. With FC set to V+, noise is 3 times greater than if set to GND or V-; therefore, avoid this setting for noise-sensitive ap- plications. The noise density graphs from the Typical Operating Characteristics can be scaled to any gain or Q for an accurate noise estimation. The MAX274/MAX275 can drive 5kQ joads to typically within 500mvV of the supply rails with negligible distor- tion. The outputs can drive up to 100pF; however, filters with high Fos and Qs will undergo some phase shift (1 at 100kHz driving 130pF, Fp = 100kHz, Q = 10 section). 19 HM AZ/e42M}j0S/SZLEXVW/PLEXVNMAX274/MAX275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters +5V 45V OFFSET OFFSET TRIM TRIM AT INPUT TRIM TRIM AT BPI_ RRM <<t-AA,_ RrriM* R R2 R R2 SV -V BPI LPO. BP LPO_ SIGNAL 1 IN AAAY INL MAAXLIM MAXIM MAX274 MAX274 MAX275 MAX275 SIGNAL RA BPO_ IN IN. BPO_|- RrriM = Ra + Rp +5V __ { z - R3 GAIN OF APPLIED TRIM VOLTAGE Riri | -f- << . - yt - ATLPO_ = (| fr atipo_ = #2 { RY IRY . R -5V WHERE R = (Raft Rp) +R Figure 8a. Trimmed Offset Removal LOWPASS FILTER R 45V WW r |" BUFFERED MAX427->@_ OUTPUT c _ Vos = 15pV AANAXLM MAXIM BY ie tee Rt Rt MAX275 UNBUFFERED SIGNAL _| AAA IN. LPO_ fe AAA IN. LPO_ OUTPUT ZERO OFFSET BANDPASS FILTER +5V MAXI L_|> MAX274 BUFFERED at MAX275 : MAKA27>4 OUTPUT SIGNAL = 1p NOMA IN_ LPO_ + C 0 R : UNBUFFERED ot =RC OUTPUT on ZERO OFFSET Figure 8b. Trimless Offset Removal MAXUM4th- and 8th-Order Continuous-Time Active Filters S8TH-ORDER BUTTERWORTH LP 0 MAXI MAX274 HIGH- 20 FREQUENCY 27k ROLLOFF 40 Rt SIGNAL AAA IN.LPO_, BPO_ p VV GAIN (dB) Loe! + 560pF \ = -80 \. -100 3k 10k 100k 1M 10M FREQUENCY (Hz) Figure 10. External RC Lowpass for High-Frequency Rolloff SIGNAL iV = +5V IN 40k 17| 4, 5| V- GND FC V+ 24 INA LPOD FILTER RESPONSE 99 1M BPOA BPID 19 10k 10k LPIA LPID MAAXLAA BPIA MAX274 BPOD LPOA IND oureut | = & INB LPOC ? 15 200k BPOB BPIC 18 LPIB LPIC 195k 100 tk 10k 100k BPIB BPOC 14-78k FREQUENCY (Hz) LPOB INC 40k WN PIN NUMBERS ARE FOR DIP. Figure 11. 10kHz 6th-Order Butterworth Lowpass Filter with 2kHz Notch (MAX274) MAXIMA 21 HM A2/e4eM1j0S/SZEXVN/VLEXVNMAX274/MAX275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters Applications (continued) Q=258 I 30% 301k ' | 376 SIGNAL Wy AN iN 6 ay YY 16 MASA 2 MAXZ75 ; 5 i 1/2 MAXA12 RFBA VREFA WHERE Tk * TRUXTEDR x: Fo = BANDPASS CENTER FREQUENCY CODE = DIGITAL CODE PROGRAMMED INTO BOTH DACS, J6 25 256 256 Q TRACKS WITH Fo TO MAINTAIN CONSTANT 500Hz BANOWIDTH yi 5kHz TO 20kKHz W2MAXOTS ADJUSTABLE BANDPASSFILTER 19 6 q <4 @) <] FIXED 500Hz BANDWIDTH 16 18 4 Vv AWN W 376k | | 301k 301k PIN NUMBERS ARE FOR DIP. Q=258 Figure 12. Programmable Bandpass Filter (MAX275) Chip Topographies BPOA INA IND BPOD | BPIA | LOPA LPOD | Bro | |. } |__| | gaes aire Ve GND LPIA LPID 0.105 0.478" {2670 mm) (4.550mm) LPIB LIC Fe v- Pie ee) a 0.070" MAX275 Mute | elect eae | ap | LPOB | we Bpoc * BPOB INB LPOC BPIC MAX274 gq 0.140" CONNECTED TOV <" He (3.560mm) 22 MAXLAA4th- and 8th-Order Continuous-Time _ Ordering Information (continued) Active Filters ___ Pin Configurations (continued) PART TEMP. RANGE PIN-PACKAGE MAX274AENG -40C to +85C 24 Narrow Plastic DIP MAX274BENG -40C to +85C 24 Narrow Plastic DIP MAX274AEWI -40C to +85C 1 28WideSO MAX274BEwI -40C to +85C 28 Wide SO MAX274AMRG -55C to +125C 24 CERDIP** MAX274BMRG ss -65C to +125C 24 CERDIP* MAX274EV KIT- C 10 +70C Plastic DIP- Through Hole MAX274_SOFT _MAX274/MAX275 Design Software MAX275ACPP OCto+70C 20 Plastic DIP MAX275BCPP OC to +70C =.20 Plastic DIP MAX275ACWP OC to+70C + 20WideSO | MAX275BCWP OC to +70C 20WideSO MAX275BC/D OC to +70C ~_ Dice MAX275AEPP -40C to +85C 20 Plastic DIP MAX275BEPP -40C to +85C = 20 Plastic DIP MAX275AEWP ~~ -40C to +85C += 20 Wide SO MAX275BEWP = -40C to +85C 3- 20 Wide SO MAX275AMJP ~s -55C to +125C 20 CERDIP** MAX275BMJP ~s -55C. to +125C 20 CERDIP*" MAX274_SOFT MAX274/MAX275 Design Software * Contact factory for dice specifications. ** Contact factory for availability and processing to MIL-STD-683. Typical Operating Circuits (continued) 36.4k IN 31.5k | 16.2k MAAAXLAA MAX275 21.2k 100kHz 4TH-ORDER 0.1dB CHEBYSHEV BANDPASS FILTER lh 44 | 316k OUT 16 13.6k 15 19 18.6k 12 pL VA 8 4 6 n] 35] 20] 5V GND +5V PIN NUMBERS ARE FOR DIP MAAXILAN TOP VIEW LPOA [7] 2g] LPOD INA [2] f27)_ IND BPIA [3 | MAXIAA 26] BPID NC. [4] MAX274 [as] NC. BPOA [5 | 24] BPOD vs [6] 23] GND LPIA 22] LPID LPIB 21] LPIC Fc [9] 20] V- BPOB 19] BPOC ne. [14] a] NC. BPIB 17] BPiC INB 16] INC LPOB HS] tpoc WIDE SO . J Nc. [4 20] V+ LPOA [2 | fig] LPB GND [3 | AAAXIAA [18] 8P0B BPO [a] MAX275 [47] ono GND [5 | 6] LPIB LPIA [6 | 15] BPIB BPla [7 | 14] INB na [8 | 3] NC. FCA [9 | 2] FCB eno [10] Ha] v- DIP/SO 23 HM A2/e42MY0S/SLEXVW/VLEXVNMAX274/MAX275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters Power Supplies The MAX274/MAX275 can be operated from a single power supply or dual supplies (Figure 9). V+ and V- pins must be properly bypassed to GND with 4.7uF electrolytic (tantalum preferred) and 0.1uF ceramic capacitors in parallel. These should be as close as possible to the chip supply pins. For single-supply applications, GND must be centered between V+ and V- voltages so signals remain in the common-mode range of the internal amplifiers. SINGLE SUPPLY +5V V+ PAAXAAA MAX274 MAX275 GND DUAL SUPPLY V+ e e +5V MAX274 MAX275 GND tot 0.1yF 4.7uF v-] -5V | Figure 9. Power-Supply Configurations Design Software General Description Maxims filter software reduces the time required to design a continuous-time lowpass or bandpass filter using the MAX274 or MAX275. Starting from your basic filter requirements, using a "spreadsheet-style" format, the software calculates order, poles and Qs of classic filter types (Butterworth, Chebyshev, or Bessel), and resistor values required to implement the desired filter response. For hardware prototyping with the MAX274, the MAX274 evaluation kit is recommended, which includes a PC board and a MAX274 IC. 24 Features Calculates filter order, poles, and Qs from your filter requirements. @ Plots filter responses gain, phase, and grou delay for inspection BEFORE you build the filter. @ Calculates resistor values used to obtain desired filter response using the MAX274 or MAX275. Ordering Information PART DISK TYPE MAX274SOFT 54" Floppy MAX275SOFT 514" Floppy inthe USA and Canada, order directly from Maxim (1-800-998-8800). in other countries, call your local Maxim representative. Software Operation NOTE: CHECK FILE "README.DOC" FOR IMPOR- TANT CHANGES. installation You will need an IBM-compatible PC, DOS version 2.0 or later with a 514" floppy disk drive, and one of the following video displays: Hercules graphics, CGA, EGA, VGA, or compatible. Either a hard drive or an additional floppy drive is also required. To install the program, insert the floppy into your disk drive and type "A: INSTALL" (or B:INSTALL). Follow the instruc- tions on the screen. After installation, type "FILTER" to start the program. Be sure you are in the drive/directory where the software is installed. Help After installing the software, print a hard copy of the file FILTER.HLP by entering "TYPE FILTER.HLP > PRN" from DOS. This collection of helo screens serves as the in- struction manual for operating the software. Individual help screens may be printed while running the software by pressing F1, then following the instructions on the screen. References The following references contain information and tables to aid in filter designs: Carson, Chen. Active Filter Design, Hayden, 1982. Tedeschi, Franck. Active Filter Cookbook, Tao Books No 1133, 1979. Hilburn, Johnson. Manual of Active Filter Design, McGraw Hill, 1973. German Language: U. Tietze; Ch. Schenk. Halbleiter-Schaltungstecknik Springer-Verlag, Berlin Heidelberg, New York/Tokyo 1991. MAXIM4th- and 8th-Order Continuous-Time EV Kit General Description The MAX274 Evaluation Kit (EV Kit) shrinks the time required to design and implement a continuous-time lowpass or bandpass filter by providing a software design tool and a prototyping PC board complete with a MAX274 8th-order, continuous-time filter IC. Starting from your basic filter requirements, Maxim's Filter Design Software calculates filter order, poles, and Qs of classic filter types (Butterworth, Chebyshev, or Bessel), then calculates resistor values required to implement the complete filter. Installing these resis- tors on the PC board provided and cascading the required number of sections of the MAX274 filter yields a complete filter ready for testing eliminat- ing the need for expensive and time-consuming prototyping. The MAX274 PC board layout may be incorporated directly in production PC boards for absolutely consistent results from prototype to production. EV Kit Active Filters Feature @ Allows You to Design and Build Lowpass or Bandpass Filters @ Pole Frequencies (Fo) from 100Hz to 150kHz @ Kit Su ports Butterworth, Chebyshev, and Bessel Designs @ includes Design Software: Calculates filter order, poles, and Qs from your filter requirements ~ Plots filter responses gain, phase, and group delay for inspection BEFORE you uild the filter Calculates resistor values needed to build filter @ Includes PC Board for Evaluation: - PC board allows you to build filters immediately simply install proper resistor values on board Build up to 8th-order filters by cascading the four second-order sections or use sections individually for multiple filters Operates from single +5V or dual 5V supplies Ordering Information PART TEMP. RANGE BOARD TYPE MAX274EVKIT OCto+70C Plastic DIP ~ Through Hole Component List QUANTITY COMPONENT SYMBOL 1 MAX274ACNG Filter tC None 1 MAX274 Filter Circuit PC Board None 2 BNC Screw-in Connectors None 3 Banana Jacks None 4 Standoffs, 4-40 Screws None 2 10uF/16V Dipped Tantalum Ci, C4 or Capacitors CS1 2 0.1pF Ceramic Capacitors C2,C3 1 _ Say peansouse Nee Figure 1. The MAX274 EV kit includes filter design software, PC board with MAX274 IC, and full documentation. MA AXLAA 25 HM AZ/e4eM10S/SZEXVN/PLEXVNMAX274/MAX275/Software/EV kit 4th- and 8th-Order Continuous-Time Active Filters Filter Design Software NOTE: CHECK FILE "README.DOC" FOR IMPORTANT CHANGES Installation You will need an IBM-compatible PC, DOS version 2.0 or later, with a 5 1/4" floppy disk drive, and one of the following video displays: Hercules graphics, CGA, EGA, VGA or compatible. Either a hard drive or an additional floppy drive is also required. To install the program, insert the floppy into your disk drive and type "A:INSTALL" (or "B:INSTALL"). Follow the in- structions on the screen. After installation, tyoe "FILTER" to start the program. Be sure you are in the drive/direc- tory where the software is installed. Help After installing the software, print a hard copy of the file FILTER.HLP by entering TYPE FILTER.HLP > PRN from DOS. This collection of help screens serves as the instruction manual for operating the software. Individual help screens may be printed while running the software by pressing F1, then following instructions on the screen. Assembly Instructions 1. Install BNC connectors and banana jacks as shown in Figure 2. Connect wires from the V+, GND, and V- jacks to the nodes designated V+, GND, and V- on the circuit. Install MAX274 IC as shown (or install an IC socket if desired). 2. Install filter feedback resistors (R1A-R4A, R1B-R4B, R1C-R4C, R1D-R4D). The values of these resistors depend on the particular filter being built, and can be calculated from the data sheet or with the aid of Maxim's Filter Design Software. All resistors except RS1 and RS2 should be either carbon or metal-film type (not wire-wound). If using resistor T-networks (described in the High- Value Resistor Transformationsection of the MAX274 data sheet denoted as R5_ - R10_ in the filter software), perform the following: On the PC board, scratch off the green soldermask from the ground trace (Figure 2). Where required, substitute each normal resistor connection (R2, R3, or R4) with three T-network resistors on the underside of the board, using jumper wire to connect the T-net- work ground connection with the ground trace on the PC board (Figure 3). STANDOFF MAX274 8TH-ORDER CONTI NOUS FILTER REV Cc ? STANDOFF | GROUND STANDOFF TRACE Figure 2. MAX274 EV Kit Component Placement Diagram 26 MAXIMA4th- and 8th-Order Continuous-Time Active Filters ST I[-4 Fis VIEW FROM BOARD UNDERSIDE (TRACE SIDE) Figure 3. T-Network PC Board Connections Recommended Resistor Sources AVAILABLE TOLERANCES TYPE MANUFACTURER 1%, 0.1% . Dale , Metal Film Others PTE Series (402-371-0800 5-100ppm/C FAX: 402-644-4206 Surface Mount | KOA Speer 1%, 0.1% RN73H2B 814-362-5536 RN73E2B FAX: 814-362-8883 Precision Resistive Products oor ore to |PR,RLSeries [319-394-9131 Pp FAX: 319-394-9280 3. BE SURE IC IS INSTALLED ON BOARD BEFORE 4. FMAMXIAA POWER IS APPLIED. Single +5V or +10V operation: Install RS1 and RS2 (both 4.7kQ). Install C1, C2, and CS1 - omit C3 and C4. Apply positive supply to V+, supply ground to V-. Do not connect any supply to GND. For dual supply (5V) operation, omit RS1 and RS2, and apply positive supply, negative supply, and ground to V+, V-, and GND respectively. Perform a frequency-response test for each filter sec- tion used (A,B,C, and D) to verify Fp/Q accuracy, before connecting them in series (cascading). For highest-accuracy results, perform frequency- response tests using a spectrum analyzer. Use the Fo and Q measurement techniques described in the MAX274 data sheet in the Measuring Fp and Q sec- tion. These methods apply to both lowpass and bandpass designs. 5. To measure filter section As Fg and Q, apply the test sweep signal at INA, measure BPOA. Fo is the fre- quency at which INA and BPOA are exactly 180 out of phase. Q is determined by the formula: Q = FPK/-3dB bandwidth where FPK is the frequency at which the gain at BPOA is greatest (this may not be equal to Fg); and the -3dB bandwidih is the difference between the two frequen- cies at which BPOA is attenuated by 3dB from its peak gain. Repeat these measurements for filter sections B, C, and D. For more information, refer to Figure 2 of the MAX274 data sheet, Bandpass Out- put. If the filter operates from a single supply, signals applied to the filters must be "centered" between V+ and V- so that signals remain in the common-mode range of the internal amplifiers. Cascade the filter sections (connect them in series) to produce the desired filter response. For example, for an 8th-order bandpass filter, use jumper wires to connect BPOA to INB, BPOB to INC, and BPOC to IND. Apply the signal input at INA; the filter output is taken at OUTD. If desired, connect the INPUT and OUTPUT BNC connectors to the input and output of the cascaded filter using jumper wires. For lowpass filters, cascade the sections using LPO_ as the out- puts. For lower-order filters, omit unused sections; for higher than 8th-order filters, order additional MAX274 PC boards from Maxim to add the required number of sections. 27 HM A3/S4ARMYOS/SGZEXVW/PLOXVNMAX274/MAX275/Software/EV Kit 4th- and 8th-Order Continuous-Time Active Filters 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. 28 Maxim Integrated Products, 120 San Gabriel Drive, Sunnyvale, CA 94086 (408) 737-7600 1996 Maxim Integrated Products Printed USA MAAXIVAA is a registered trademark of Maxim integrated Products.