LMF60 High Performance 6th-Order Switched Capacitor Butterworth Lowpass Filter General Description Features The LMF60 is a high performance, precision, 6th-order Butterworth lowpass active filter. It is fabricated using National’s LMCMOS process, an improved silicon-gate CMOS process specifically designed for analog products. Switchedcapacitor techniques eliminate external component requirements and allow a clock-tunable cutoff frequency. The ratio of the clock frequency to the low-pass cutoff frequency is internally set to 50:1 (LMF60-50) or 100:1 (LMF60-100). A Schmitt trigger clock input stage allows two clocking options, either self-clocking (via an external resistor and capacitor) for stand-alone applications, or for tighter cutoff frequency control, a TTL or CMOS logic compatible clock can be directly applied. The maximally flat passband frequency response together with a DC gain of 1V/V allows cascading LMF60 sections for higher-order filtering. In addition to the filter, two independent CMOS op amps are included on the die and are useful for any general signal conditioning applications. The LMF60 is pin- and functionally-compatible with the MF6, but provides improved performance. Y Y Y Y Y Y Y Y Y Y Y Cutoff frequency range of 0.1 Hz to 30 kHz Cutoff frequency accuracy of g 1.0%, maximum Low offset voltage g 100 mV, maximum, g 5V supply Low clock feedthrough of 10 mVp– p, typical Dynamic range of 88 dB, typical Two uncommitted op amps available No external components required 14-pin DIP or 14-pin wide-body S.O. package Single/Dual Supply Operation: a 4V to a 14V ( g 2V to g 7V) Cutoff frequency set by external or internal clock Pin-compatible with the MF6 Applications Y Y Y Y Y Y Communication systems Audio filtering Anti-alias filtering Data acquisition noise filtering Instrumentation High-order tracking filters Block and Connection Diagrams All Packages TL/H/9294 – 2 Top View Order Number LMF60CMJ-50, (5962-9096 701MCA or LMF60CMJ50/883), LMF60CMJ-100, or (5962-9096 702MCA or LMF60CMJ100/883) See NS Package Number J14A TL/H/9294 – 1 Order Number LMF60CIWM-50 or LMF60CIWM-100 See NS Package Number M14B Order Number LMF60CIN-50 or LMF60CIN-100 See NS Package Number N14A TRI-STATEÉ is a registered trademark of National Semiconductor Corporation. C1996 National Semiconductor Corporation TL/H/9294 RRD-B30M56/Printed in U. S. A. LMF60 High Performance 6th-Order Switched Capacitor Butterworth Lowpass Filter May 1996 Absolute Maximum Ratings (Note 1) Soldering Information: If Military/Aerospace specified devices are required, please contact the National Semiconductor Sales Office/Distributors for availability and specifications. Supply Voltage (V a b b V ) (Note 2) # N Package: 10 sec. # J Package: 10 sec. # SO Package: Vapor Phase (60 sec.) 15V Infrared (15 sec.) (Note 6) a Voltage at Any Pin V a 0.2V b V b 0.2V Input Current at Any Pin (Note 3) Package Input Current (Note 3) Power Dissipation (Note 4) Storage Temperature ESD Susceptibility (Note 5) CLK IN Pin 260§ C 300§ C 215§ C 220§ C Operating Ratings (Note 1) Temperature Range TMin s TA s TMax LMF60CIN-50, LMF60CIN-100 LMF60CIJ-50, LMF60CIJ-100, LMF60CIWM-50, b 40§ C s TA s a 85§ C LMF60CIWM-100 LMF60CMJ-50, LMF60CMJ-100, LMF60CMJ50/883, b 55§ C s TA s a 125§ C LMF60CMJ100/883 a 4V to 14V Supply Voltage (V b Vb) 5 mA 20 mA 500 mW b 65§ C to a 150§ C 2000V 1700V Filter Electrical Characteristics The following specifications apply for fCLK e 500 kHz (Note 7) unless otherwise specified. Boldface limits apply for TA e TJ e TMIN to TMAX; all other limits TA e TJ e 25§ C. Symbol V a Parameter Typical (Note 8) Conditions Limits (Note 9) Units (Limits) 1.5 Hz (Min) MHz (Max) e a 5V, V b e b 5V fCLK IS Ho Clock Frequency Range (Note 16) 5 Total Supply Current 7.0 Clock Feedthrough VIN e 0V Filter Opamp DC Gain RSource s 2 kX / Clock to Cutoff Frequency Ratio (Note 10) 0.10 / / 0.10 dB (Max) dB (Min) 49.00 g 0.8% /49.00 g 1.0% (Max) LMF60-100 98.10 g 0.8% /98.10 g 1.0% (Max) AMIN Stopband Attenuation VOS DC Offset Voltage 4 ppm/§ C At 2 c fC LMF60-50 LMF60-100 VOUT Output Voltage Swing (Note 2) ISC Output Short Circuit Current (Note 11) 36 dB (Min) g 100 mV (Max) mV (Max) g 150 a 3.9 b 4.2 Source Sink Dynamic Range (Note 12) http://www.national.com b 0.30 LMF60-50 Temperature Coefficient of fCLK/fC Additional Magnitude Response Test Points (Note 13) mA (Max) mVp-p mVp-p b 0.26 fCLK/fC 12.0 10 5 LMF60-50 LMF60-100 / / a 3.7 b 4.0 V (Min) V (Max) 90 2.2 mA mA 88 dB fIN e 12 kHz b 9.45 g 0.46 /b9.45 g 0.50 dB fIN e 9 kHz b 0.87 g 0.16 /b0.87 g 0.20 dB fIN e 6 kHz b 9.30 g 0.46 /b9.30 g 0.50 dB fIN e 4.5 kHz b 0.87 g 0.16 /b0.87 g 0.20 dB 2 Filter Electrical Characteristics (Continued) The following specifications apply for fCLK e 250 kHz (Note 7) unless otherwise specified. Boldface limits apply for TA e TJ e TMIN to TMAX; all other limits TA e TJ e 25§ C. Symbol Parameter Typical (Note 8) Conditions Limits (Note 9) Units (Limits) 750 Hz (Min) kHz (Max) V a e a 2.5V, Vb e b2.5V fCLK IS Ho Clock Frequency Range (Note 16) 5 Total Supply Current 5.0 Clock Feedthrough (Peak to Peak) VIN e 0V Filter Opamp DC Gain (with RSource s 2 kX) fCLK e 250 kHz Clock to Cutoff Frequency Ratio (Note 10) LMF60-50 0.10 fCLK e 250 kHz VOS DC Offset Voltage dB (Max) dB (Min) 49.00 g 0.8% /49.00 g 1.0% (Max) Output Voltage Swing (Note 2) ISC Output Short Circuit Current (Note 11) 98.10 g 0.8% /98.10 g 1.0% (Max) 4 ppm/§ C At 2 c fC a 1.4 b 2.0 Source Sink Dynamic Range (Note 12) LMF60-100 36 dB (Min) g 60 mV (Max) mV (Max) g 90 RL e 5 kX LMF60-50 0.10 dB LMF60-50 LMF60-100 VOUT Additional Magnitude Response Test Points (Note 13) b 0.30 98.10 g 0.6% Temperature Coefficient of fCLK/fC Stopband Attenuation / / 49.00 g 0.6% fCLK e 250 kHz fCLK e 500 kHz AMIN mA (Max) b 0.08 fCLK e 500 kHz LMF60-100 6.5 mV mV b 0.26 fCLK e 500 kHz fCLK/fC / 6 3 / / a 1.2 b 1.8 V (Min) V (Max) 42 0.9 mA mA 81 dB fIN e 6 kHz b 9.45 g 0.46 /b9.45 g 0.50 dB fIN e 4.5 kHz b 0.87 g 0.16 /b0.87 g 0.20 dB fIN e 3 kHz b 9.30 g 0.46 /b9.30 g 0.50 dB fIN e 2.25 kHz b 0.87 g 0.16 /b0.87 g 0.20 dB 3 http://www.national.com Op Amp Electrical Characteristics Boldface limits apply for TA e TJ e TMIN to TMAX; all other limits TA e TJ e 25§ C. Symbol V a Parameter Conditions Typical (Note 8) Limits (Note 9) Units (Limits) g 20 mV (Max) e a 5V, V b e b 5V VOS Input Offset Voltage IB Input Bias Current CMRR Common Mode Rejection Ratio (Op Amp Ý2 Only) Test Input Range e b 2.2V to a 1.8V VO Output Voltage Swing RL e 5 kX 10 pA 55 3.8 b 4.2 ISC Output Short Circuit Current (Note 13) Source Sink / / dB 3.6 b 4.0 90 2.1 V (Min) V (Max) mA mA SR Slew Rate 4 V/ms AVOL DC Open Loop Gain 80 dB (Min) GBW Gain Bandwidth Product 2.0 MHz V a e a 2.5V, V b e b 2.5V VOS Input Offset Voltage IB Input Bias Current g 20 CMRR Common Mode Rejection Ratio (Op Amp Ý2 Only) Test Input Range e b 0.9V to a 0.5V VO Output Voltage Swing RL e 5 kX pA 55 1.3 b 1.8 ISC mV (Max) 10 Output Short Circuit Current (Note 13) Source Sink / / dB 1.1 b 1.6 42 0.9 V (Min) V (Max) mA mA SR Slew Rate 3 V/ms AVOL DC Open Loop Gain 74 dB (Min) GBW Gain Bandwidth Product 2.0 MHz Logic Input-Output Characteristics The following specifications apply for Vb e 0V (Note 15), L.Sh e 0V unless otherwise specified. Boldface limits apply for TA e TJ e TMIN to TMAX; all other limits TA e TJ e 25§ C. Symbol Parameter Conditions Typical (Note 8) Limits (Note 9) Units (Limits) 2.0 0.8 V (Min) V (Max) 2.0 / 0.4 V (Min) V (Max) TTL CLOCK INPUT, CLK R PIN (NOTE 14) VIH VIL TTL Input Voltage Logical ‘‘1’’ Logical ‘‘0’’ V a e a 5V, V b e b 5V VIH VIL CLK R Input Voltage Logical ‘‘1’’ Logical ‘‘0’’ V a e a 2.5V, V b e b 2.5V 0.6 Maximum Leakage Current at CLK R http://www.national.com 2.0 4 mA Logic Input-Output Characteristics (Continued) The following specifications apply for V b e 0V (Note 15), L.Sh e 0V unless otherwise specified. Boldface limits apply for TA e TJ e TMIN to TMAX; all other limits TA e TJ e 25§ C. Symbol Parameter Conditions Typical (Note 8) Limits (Note 9) Units (Limits) SCHMITT TRIGGER VT a VTb VT a bVTb Positive Going Input Threshold Voltage Negative Going Input Threshold Voltage Hysteresis V a e 10V 6.1 8.8 / / 6.0 8.9 V (Min) V (Max) V a e 5V 3.0 4.3 / / 2.9 4.4 V (Min) V (Max) V a e 10V 1.4 3.8 / / 1.3 3.9 V (Min) V (Max) V a e 5V 0.7 1.9 / / 0.6 2.0 V (Min) V (Max) V a e 10V 2.3 7.4 / / 2.1 7.6 V (Min) V (Max) V a e 5V 1.1 3.6 / / 0.9 3.8 V (Min) V (Max) VOH Logical ‘‘1’’ Voltage IO e b10 mA, Pin 11 a V e a 10V a V e a 5V 9.1 4.6 / / 9.0 4.5 V (Min) V (Min) VOL Logical ‘‘0’’ Voltage IO e b10 mA, Pin 11 a V e a 10V a V e a 5V 0.9 0.4 / / 1.0 0.5 V (Max) V (Max) ISOURCE Output Source Current, Pin 11 CLK R to V a V e a 10V a V e a 5V 4.9 1.6 / / 3.7 1.2 mA (Min) mA (Min) Output Sink Current, Pin 11 CLK R to V a V e a 10V a V e a 5V 4.9 1.6 / / 3.7 1.2 mA (Min) mA (Min) ISINK b a Note 1: Absolute Maximum Ratings indicate limits beyond which damage to the device may occur. Operating Ratings indicate conditions for which the device is functional. Specified Electrical Characteristics do not apply when operating the device outside its specified conditions. Note 2: All voltages are measured with respect to AGND, unless otherwise specified. Note 3: When the input voltage (VIN) at any pin exceeds the power supply rails (VIN k Vb or VIN l V a ) the absolute value of current at that pin should be limited to 5 mA or less. The 20 mA package input current limits the number of pins that can exceed the power supply boundaries with 5 mA to four. Note 4: The Maximum power dissipation must be derated at elevated temperatures and is dictated by TJ Max, iJA, and the ambient temperature TA. The maximum allowable power dissipation is PD e (TJ Max b TA)/iJA or the number given in the absolute ratings, whichever is lower. For this device, TJ Max e 125§ C, and the typical junction-to-ambient thermal resistance of the LMF60CCN when board mounted is 67§ C/W. For the LMF60CIJ this number decreases to 62§ C/W. For the LMF60CIWM, iJA e 78§ C/W. Note 5: Human body model: 100 pF discharged through a 1.5 kX resistor. Note 6: See AN450 ‘‘Surface Mounting Methods and Their Effect on Product Reliability’’ or the section titled ‘‘Surface Mount’’ found in any current Linear Databook for other methods of soldering surface mount devices. Note 7: The specifications given are for a clock frequency (fCLK) of 500 kHz at a 5V and 250 kHz at g 2.5V. Above this frequency, the cutoff frequency begins to deviate from the specified error band over the temperature range but the filter still maintains its amplitude characteristics. See application hints. Note 8: Typicals are at 25§ C and represent the most likely parametric norm. Note 9: Guaranteed to National’s Average Outgoing Quality Level (AOQL). Note 10: The cutoff frequency of the filter is defined as the frequency where the magnitude response is 3.01 dB less than the DC gain of the filter. Note 11: The short circuit source current is measured by forcing the output to its maximum positive swing and then shorting that output to the negative supply. The short circuit sink current is measured by forcing the output being tested to its maximum negative voltage and then shorting that output to the positive supply. These are worst case conditions. Note 12: For g 5V supplies the dynamic range is referenced to 2.62 Vrms (3.7V peak), where the wideband noise over a 20 kHz bandwidth is typically 100 mV. For g 2.5V supplies the dynamic range is referenced to 0.849 Vrms (1.2V peak), where the wideband noise over a 20 kHz bandwidth is typically 75 mVrms. Note 13: The filter’s magnitude response is tested at the cutoff frequency, fC, at fIN e 2 fC, and at these two additional frequencies. Note 14: The LMF60 is operated with symmetrical supplies and L.Sh is tied to GND. Note 15: For simplicity all the logic levels (except for the TTL input logic levels) have been referenced to Vb e 0V. The logic levels will scale accordingly for g 5V and g 2.5V supplies. Note 16: The nominal ratio of the clock frequency to the low-pass cutoff frequency is internally set to 50-to-1 (LMF60-50) or 100-to-1 (LMF60-100). 5 http://www.national.com Typical Performance Characteristics fCLK/fC Deviation vs Power Supply Voltage fCLK/fC Deviation vs Temperature fCLK/fC Deviation vs Clock Frequency fCLK/fC Deviation vs Power Supply Voltage fCLK/fC Deviation vs Temperature fCLK/fC Deviation vs Clock Frequency DC Gain Deviation vs Power Supply Voltage DC Gain Deviation vs Temperature DC Gain Deviation vs Clock Frequency TL/H/9294 – 3 http://www.national.com 6 Typical Performance Characteristics (Continued) DC Gain Deviation vs Power Supply Voltage DC Gain Deviation vs Temperature DC Gain Deviation vs Clock Frequency DC Offset Voltage Deviation vs Power Supply Voltage Power Supply Current vs Power Supply Voltage Power Supply Current vs Temperature Positive Voltage Swing vs Power Supply Voltage Negative Voltage Swing vs Power Supply Voltage Positive Voltage Swing vs Temperature TL/H/9294 – 4 7 http://www.national.com Typical Performance Characteristics Negative Voltage Swing vs Temperature (Continued) CLK R Trigger Threshold vs Power Supply Voltage Crosstalk from Filter to Op Amps Schmitt Trigger Threshold vs Power Supply Voltage Crosstalk from Either Op Amp to Filter Equivalent Input Noise Voltage of Op Amps TL/H/9294 – 5 http://www.national.com 8 Crosstalk Test Circuits From Filter to Op-Amps TL/H/9294 – 6 From Either Op-Amp to Filter Output TL/H/9294 – 7 Pin Description (Pin Numbers) Pin FILTER OUT (3) FILTER IN (8) VOSADJ (7) AGND (5) Pin CLK IN (9) Description The output of the lowpass filter will typically swing to within 1V of each supply rail. The input to the lowpass filter. To minimize gain errors the source impedance that drives this input should be less than 2k (See Section 1.4). For single supply operation the input signal must be biased to mid-supply or AC coupled. CLK R (11) This pin is used to adjust the DC offset of the filter output; if not used it must be tied to the AGND potential. (See Section 1.3) The analog ground pin. This pin sets the DC bias level for the filter section and the noninverting input of Op-Amp Ý1 and must be tied to the system ground for split supply operation or to mid-supply for single supply operation (See Section 1.2). When tied to mid-supply this pin should be well bypassed. VO1(4), INV1 (13) VO1 is the output and INV1 is the inverting input of Op-Amp Ý1. The non-inverting input of this Op-Amp is internally connected to the AGND pin. VO2(2), INV2 (14), NINV2 (1) VO2 is the output, INV2 is the inverting input, and NINV2 is the non-inverting input of Op-Amp Ý2. V a (6), Vb (10) The positive and negative supply pins. The total power supply range is 4V to 14V. Decoupling these pins with 0.1 mF capacitors is highly recommended. L.Sh (12) 9 Description A CMOS Schmitt-trigger input to be used with an external CMOS logic level clock. Also used for self-clocking Schmitt-trigger oscillator (See Section 1.1). A TTL logic level clock input when in split supply operation ( g 2V to g 7V) and L. Sh tied to system ground. This pin becomes a low impedance output when L.Sh is tied to Vb. Also used in conjunction with the CLK IN pin for self clocking Schmitt-trigger oscillator (See Section 1.1). Level shift pin, selects the logic threshold levels for the desired clock. When tied to Vb it enables an internal TRISTATEÉ buffer stage between the Schmitt trigger and the internal clock level shift stage thus enabling the CLK IN Schmitt-trigger input and making the CLK R pin a low impedance output. When the voltage level at this input exb ceeds [25% (V a b Vb) a V ] the internal TRI-STATEÉ buffer is disabled allowing the CLK R pin to become the clock input for the internal clock level shift stage. The CLK R threshold level is now 2V above the voltage applied to the L.Sh pin. Driving the CLK R pin with TTL logic levels can be accomplished through the use of split supplies and by tying the L.Sh pin to system ground. http://www.national.com 1.0 LMF60 Application Hints frequency is dependent on the buffer’s threshold levels as well as on the resistor/capacitor tolerance (See Figure 1 ). The LMF60 is comprised of a non-inverting unity gain lowpass sixth-order Butterworth switched capacitor filter section and two undedicated CMOS Op-Amps. The switchedcapacitor topology makes the cutoff frequency (where the gain drops 3.01 dB below the DC gain) a direct ratio (100:1 or 50:1) of the clock frequency supplied to the lowpass filter. Internal integrator time constants set the filter’s cutoff frequency. The resistive element of these integrators is actually a capacitor which is ‘‘switched’’ at the clock frequency (for a detailed discussion see Input Impedance section). Varying the clock frequency changes the value of this resistive element and thus the time constant of the integrators. The clock to cutoff frequency ratio (fCLK/fC) is set by the ratio of the input and feedback capacitors in the integrators. The higher the clock to cutoff frequency ratio (or the sampling rate) the closer the approximation is to the theoretical Butterworth response. The LMF60 is available in fCLK/fC ratios of 50:1 (LMF60-50) or 100:1 (LMF60-100). Schmitt-trigger threshold voltage levels can vary significantly causing the R/C oscillator’s frequency to vary greatly from part to part. Where accuracy in fC is required an external clock can be used to drive the CLK R input of the LMF60. This input is TTL logic level compatible and also presents a very light load to the external clock source ( E 2 mA) with split supplies and L.Sh tied to system ground. The logic level is programmed by the voltage applied to level shift (L.Sh) pin (See the Pin Description for L.Sh pin). 1.2 POWER SUPPLY BIASING The LMF60 can be biased from a single supply or dual split supplies. The split supply mode shown in Figures 2 and 3 is the most flexible and easiest to implement. As discussed earlier split supplies, g 2V to g 7V, will enable the use of TTL or CMOS clock logic levels. Figure 4 shows two schemes for single supply biasing. In this mode only CMOS clock logic levels can be used. 1.1 CLOCK INPUTS The LMF60 has a Schmitt-trigger inverting buffer which can be used to construct a simple R/C oscillator. The oscillator fCLK e RC In Typically for fCLK e TL/H/9294 – 8 FIGURE 1. Schmitt Trigger R/C Oscillator http://www.national.com 10 Ð# 1 VCC b VTb b VCC VT a VCC e V a b 1 1.37 RC JV ( VT a Tb Vb e 10V: 1.0 LMF60 Application Hints (Continued) cies these leakage currents can cause millivolts of error, for example: If the LMF60-50 or the LMF60-100 were set up for a cutoff frequency of 10 kHz the input impedance would be: RIN e 1 c 1010 e 1 MX 10 kHz fCLK e 100 Hz, ILEAKAGE e 1 pA, C e 1 pF 1 pA e 10 mV Ve 1 pF (100 Hz) In this example with a source impedance of 10k the overall gain, if the LMF60 had an ideal gain of 1 (0 dB) would be: 1 MX e 0.99009 ( b 86.4 mdB) AV e 10 kX a 1 MX Since the maximum overall gain error for the LMF60 is a 0.1 dB, b0.3 dB with a RS s 2 kX the actual gain error for this case would be a 0.21 dB to b0.39 dB. The propagation delay in the logic and the settling time required to acquire a new voltage level on the capacitors increases as the LMF60 power supply voltage decreases. This causes a shift in the fCLK/fC ratio which will become noticeable when the clock frequency exceeds 500 kHz. The amplitude characteristic will stay within tolerance until fCLK exceeds 750 kHz and will peak at about 0.4 dB at the cutoff frequency with a 2 MHz clock. The response of the LMF60 is still a reasonable approximation of the ideal Butterworth lowpass characteristic as can be seen in Figure 7 . 1.5 CUTOFF FREQUENCY RANGE The filter’s cutoff frequency (fC) has a lower limit caused by leakage currents through the internal switches discharging the stored charge on the capacitors. At lower clock frequen- TL/H/9294 – 17 TL/H/9294 – 18 FIGURE 7a. LMF60-100 g 5V Supplies Amplitude Response FIGURE 7b. LMF60-50 g 5V Supplies Amplitude Response TL/H/9294 – 19 TL/H/9294 – 20 FIGURE 7c. LMF60-100 g 2.5V Supplies Amplitude Response FIGURE 7d. LMF60-50 g 2.5V Supplies Amplitude Response 11 http://www.national.com 1.0 LMF60 Application Hints (Continued) TL/H/9294–9 TL/H/9294 – 10 FIGURE 2. Dual Supply Operation LMF60 Driven with CMOS Logic Level Clock (VIH t V a b 0.3 VS and VIL s Vb a 0.3 VS where VS e V a b Vb) FIGURE 3. Dual Supply Operation LMF60 Driven with TTL Logic Level Clock TL/H/9294 – 11 a) Resistor Biasing of AGND TL/H/9294 – 12 b) Using Op-Amp 2 to Buffer AGND FIGURE 4. Single Supply Operation http://www.national.com 12 1.0 LMF60 Application Hints (Continued) TL/H/9294 – 13 TL/H/9294 – 14 FIGURE 5. VOS Adjust Schemes The equivalent input resistor (RIN) then can be defined as 1 RIN e VIN/IIN e CINfCLK The input capacitor is 2 pF for the LMF60-50 and 1 pF for the LMF60-100, so for the LMF60-100 1.3 OFFSET ADJUST The VOSADJ pin is used in adjusting the output offset level of the filter section. If this pin is not used it must be tied to the analog ground (AGND) level, either mid-supply for single ended supply operation or ground for split supply operation. This pin sets the zero reference for the output of the filter. The implementation of this pin can be seen in Figure 5 . In 5(a) DC offset is adjusted using a potentiometer; in 5(b) the Op-Amp integrator circuit keeps the average DC output level at AGND. The circuit in 5(b) is therefore appropriate only for AC-coupled signals and signals biased at AGND. RIN e 1 c 1012 1 c 1012 1 c 1010 e e fCLK fC c 100 fC and 5 c 1011 5 c 1011 1 c 1010 e e fCLK fC c 50 fC for the LMF60-50. As shown in the above equations, for a given cutoff frequency (fC) the input impedance remains the same for the LMF60-50 and the LMF60-100. The higher the clock to cutoff frequency ratio, the greater equivalent input resistance for a given clock frequency. As the cutoff frequency increases the equivalent input impedance decreases. This input resistance will form a voltage divider with the source impedance (RSOURCE). Since RIN is inversely proportional to the cutoff frequency, operation at higher cutoff frequencies will be more likely to load the input signal which would appear as an overall decrease in gain at the output of the filter. Since the filter’s ideal gain is unity, its overall gain is given by: RIN e 1.4 INPUT IMPEDANCE The LMF60 lowpass filter input (FILTER IN pin) is not a high impedance buffer input. This input is a switched capacitor resistor equivalent, and its effective impedance is inversely proportional to the clock frequency. The equivalent circuit of the input to the filter can be seen in Figure 6 . The input capacitor charges to the input voltage (VIN) during one half of the clock period, during the second half the charge is transferred to the feedback capacitor. The total transfer of charge in one clock cycle is therefore Q e CINVIN, and since current is defined as the flow of charge per unit time the average input current becomes IIN e Q/T (where T equals one clock period) or AV e CINVIN e CINVINfCLK IIN e T RIN RIN a RSOURCE TL/H/9294 – 15 a) Equivalent Circuit for LMF60 Filter Input TL/H/9294 – 16 b) Actual Circuit for LMF60 Filter Input FIGURE 6. LMF60 Filter Input 13 http://www.national.com 2.0 Designing with the LMF60 To implement this example for the LMF60-50 the clock frequency will have to be set to fCLK e 50(1.119 kHz) e 55.95 kHz or for the LMF60-100 fCLK e 100(1.119 kHz) e 111.9 kHz Given any lowpass filter specification, two equations will come in handy in trying to determine whether the LMF60 will do the job. The first equation determines the order of the lowpass filter required: log (100.1AMin b 1) b log(100.1AMax b 1) (1) 2 log (fs/fb) where n is the order of the filter, AMin is the minimum stopband attenuation (in dB) desired at frequency fs, and AMax is the passband ripple or attenuation (in dB) at frequency fb. If the result of this equation is greater than 6, then more than a single LMF60 is required. The attenuation at any frequency can be found by the following equation: Attn(f) e 10 log[1 a (100.1AMax b 1) (f/fb)2n]dB (2) 2.2 CASCADING LMF60s In the case where a steeper stopband attenuation rate is required two LMF60’s can be cascaded (Figure 9) yielding a 12th order slope of 72 dB per octave. Because the LMF60 is a Butterworth filter and therefore has no ripple in its passband, when LMF60’s are cascaded the resulting filter also has no ripple in its passband. Likewise the DC and passband gains will remain at 1V/V. The resulting response is shown in Figure 10 . In determining whether the cascaded LMF60’s will yield a filter that will meet a particular amplitude response specification, as above, equations 3 and 4 can be used, shown below. ne where n e 6 (the order of the filter). 2.1 A LOWPASS DESIGN EXAMPLE Suppose the amplitude response specification in Figure 8 is given. Can the LMF60 be used? The order of the Butterworth approximation will have to be determined using eq. 1: AMin e 30 dB, AMax e 1.0 dB, fs e 2 kHz, and fb e 1 kHz ne log(103 b 1) b log(100.1 b 1) e 5.96 2 log(2) Since n can only take on integer values, n e 6. Therefore the LMF60 can be used. In general, if n is 6 or less a single LMF60 stage can be utilized. Likewise, the attenuation at fs can be found using equation 2 with the above values and n e 6 giving: Atten (2 kHz) e 10 log [1 a (100.1 b 1) (2/1)12] 2.3 IMPLEMENTING A ‘‘NOTCH’’ FILTER WITH THE LMF60 A ‘‘notch’’ filter with 60 dB of attenuation can be obtained by using one of the Op-Amps available in the LMF60 and three external resistors. The circuit and amplitude response are shown in Figure 11 . The frequency where the ‘‘notch’’ will occur is equal to the frequency at which the output signal of the LMF60 will have the same magnitude but be 180 degrees out of phase with its input signal. For a sixth order Butterworth filter 180§ phase shift occurs where f e fn e 0.742 fC. The attenuation at this frequency is 0.12 dB which must be compensated for by making R1 e 1.014 c R2. Since R1 does not equal R2 there will be a gain inequality above and below the notch frequency. At frequencies below the notch frequency (f m fn), the signal through the filter has a gain of one and is non-inverting. Summing this with the input signal through the Op-Amp yields an overall gain of two or a 6 dB. For f n fn, the signal at the output of the filter is greatly attenuated thus only the input signal will appear at the output of the Op-Amp. With R3 e R1 e 1.014 R2 the overall gain is 0.986 or b0.12 dB at frequencies above the notch. e 30.26 dB This result also meets the design specification given in Figure 8 again verifying that a single LMF60 section will be adequate. TL/H/9294–21 FIGURE 8. Design Example Magnitude Response Specification Where the Response of the Filter Design Must Fall Within the Shaded Area of the Specification Since the LMF60’s cutoff freqency fC, which corresponds to a gain attenuation of b3.01 dB, was not specified in this example it needs to be calculated. Solving equation 2 where f e fC as follows: 100.1(3.01 dB) b 1) 1/(2n) fc e fb (100.1AMax b 1) 100.301 b 1 1/12 e1 100.1 b 1 e 1.119 kHz where fC e fCLK/50 or fCLK/100. http://www.national.com J (3) (4) Attn(f) e 10 log[1 a (100.05 AMax b 1) (f/fb)2n] dB where n e 6 (the order of each filter). Equation 3 will determine whether the order of the filter is adequate (n s 6) while equation 4 can determine if the required stopband attenuation is met and what actual cutoff frequency (fC) is required to obtain the particular frequency response desired. The design procedure would be identical to the one shown in Section 2.1. ne Ð # log (100.05 Amin b 1) b log(100.05 AMax b 1) 2 log (fs/fb) ( 14 2.0 Designing with the LMF60 (Continued) TL/H/9294 – 22 FIGURE 9. Cascading Two LMF60s TL/H/9294 – 24 TL/H/9294 – 23 FIGURE 10b. Phase Response of Two Cascaded LMF60-50s FIGURE 10a. One LMF60-50 vs. Two LMF60-50s Cascaded 15 http://www.national.com 2.0 Designing with the LMF60 (Continued) TL/H/9294 – 25 FIGURE 11a. ‘‘Notch’’ Filter TL/H/9294 – 26 FIGURE 11b. LMF60-50 ‘‘Notch’’ Filter Amplitude Response http://www.national.com 16 2.0 Designing with the LMF60 (Continued) component will be ‘‘reflected’’ about fCLK/2 into the frequency range below fCLK/2 as in Figure 14b . If this component is within the passband of the filter and of large enough amplitude it can cause problems. Therefore if frequency components in the input signal exceed fCLK/2 they must be attenuated before being applied to the LMF60 input. The necessary amount of attenuation will vary depending on system requirements. In critical applications the signal components above fCLK/2 will have to be attenuated at least to the filter’s residual noise level. An example circuit is shown in Figure 15 using one of the uncommitted Op-Amps available in the LMF60. 2.4 CHANGING CLOCK FREQUENCY INSTANTANEOUSLY The LMF60 will respond well to a sudden change in clock frequency. Distortion in the output signal occurs at the transition of the clock frequency and lasts approximately three cutoff frequency (fC) cycles. As shown in Figure 12 , if the control signal is low the LMF60-50 has a 100 kHz clock making fC e 2 kHz; when this signal goes high the clock frequency changes to 50 kHz yielding 1 kHz fC. The transient response of the LMF60 seen in Figure 13 is also dependent on the fc and thus the fCLK applied to the filter. The LMF60 responds as a classical sixth order Butterworth lowpass filter. TL/H/9294 – 28 TL/H/9294 – 27 FIGURE 13. LMF60-50 Step Input Response, Vertical e 2V/Div., Horizontal e 1 ms/Div., fCLK e 100 kHz fIN e 1.5 kHz (Scope Time Base e 2 ms/Div) FIGURE 12. LMF60-50 Abrupt Clock Frequency Change 2.5 ALIASING CONSIDERATIONS Aliasing effects have to be taken into consideration when input signal frequencies exceed half the sampling rate. For the LMF60 this equals half the clock frequency (fCLK). When the input signal contains a component at a frequency higher than half the clock frequency, as in Figure 14a , that TL/H/9294 – 29 TL/H/9294 – 30 (a) Input Signal Spectrum (b) Output Signal Spectrum. Note that the input signal at fs/2 a f causes an output signal to appear at fs/2 b f. FIGURE 14. The phenomenon of aliasing in sampled-data systems. An input signal whose frequency is greater than one-half the sampling frequency will cause an output to appear at a frequency lower than one-half the sampling frequency. In the LMF60, fs e fCLK. 17 http://www.national.com 2.0 Designing with the LMF60 (Continued) TL/H/9294 – 31 f0 e 1 2 q 0R 1 R 2C 1C 2 H0 e R4/R3 (H0 e 1 when R3 and R4 are omitted and VO2 is directly tied to INV2). Design Procedure: pick C1 R2 e 1 2QC100 for a 2nd Order Butterworth Q e 0.707 R2 e 0.113 C 1f 0 make R1 e R2 and C2 e 1 (2qf0R1)2C1 Note: The parallel combination of R4 (if used), R1 and R2 should be t 10 kX in order not to load Op-Amp Ý2. FIGURE 15. Second Order Butterworth Anti-Aliasing Filter Using Uncommitted Op-Amp Ý2 http://www.national.com 18 Physical Dimensions inches (millimeters) unless otherwise noted Cavity Dual-In-Line Package (J) Order Number LMF60CMJ-50, LMF60CMJ50/883, LMF60CMJ-100 or LMF60CMJ100/883 NS Package Number J14A Small Outline Wide Body (M) Order Number LMF60CIWM-50 or LMF60CIWM-100 NS Package Number M14B 19 http://www.national.com LMF60 High Performance 6th-Order Switched Capacitor Butterworth Lowpass Filter Physical Dimensions inches (millimeters) unless otherwise noted (Continued) Lit. Ý 108461 Molded Dual-In-Line Package (N) Order Number LMF60CIN-50 or LMF60CIN-100 NS Package Number N14A LIFE SUPPORT POLICY NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT OF NATIONAL SEMICONDUCTOR CORPORATION. As used herein: 1. Life support devices or systems are devices or systems which, (a) are intended for surgical implant into the body, or (b) support or sustain life, and whose failure to perform, when properly used in accordance with instructions for use provided in the labeling, can be reasonably expected to result in a significant injury to the user. National Semiconductor Corporation 1111 West Bardin Road Arlington, TX 76017 Tel: 1(800) 272-9959 Fax: 1(800) 737-7018 http://www.national.com 2. A critical component is any component of a life support device or system whose failure to perform can be reasonably expected to cause the failure of the life support device or system, or to affect its safety or effectiveness. National Semiconductor Europe Fax: a49 (0) 180-530 85 86 Email: europe.support @ nsc.com Deutsch Tel: a49 (0) 180-530 85 85 English Tel: a49 (0) 180-532 78 32 Fran3ais Tel: a49 (0) 180-532 93 58 Italiano Tel: a49 (0) 180-534 16 80 National Semiconductor Hong Kong Ltd. 13th Floor, Straight Block, Ocean Centre, 5 Canton Rd. Tsimshatsui, Kowloon Hong Kong Tel: (852) 2737-1600 Fax: (852) 2736-9960 National Semiconductor Japan Ltd. Tel: 81-043-299-2308 Fax: 81-043-299-2408 National does not assume any responsibility for use of any circuitry described, no circuit patent licenses are implied and National reserves the right at any time without notice to change said circuitry and specifications.