MF10 Universal Monolithic Dual Switched Capacitor Filter

December 1994

MF10
Universal Monolithic Dual
Switched Capacitor Filter
General Description
The MF10 consists of 2 independent and extremely easy to Any of the classical filter configurations (such as Butter-
use, general purpose CMOS active filter building blocks. worth, Bessel, Cauer and Chebyshev) can be formed.
Each block, together with an external clock and 3 to 4 resis- For pin-compatible device with improved performance refer
tors, can produce various 2nd order functions. Each building to LMF100 datasheet.
block has 3 output pins. One of the outputs can be config-
ured to perform either an allpass, highpass or a notch func-
tion; the remaining 2 output pins perform lowpass and band-
Features
Y Easy to use
pass functions. The center frequency of the lowpass and
bandpass 2nd order functions can be either directly depen-
Y Clock to center frequency ratio accuracy g 0.6%
dent on the clock frequency, or they can depend on both Y Filter cutoff frequency stability directly dependent on
clock frequency and external resistor ratios. The center fre- external clock quality
quency of the notch and allpass functions is directly depen- Y Low sensitivity to external component variation
dent on the clock frequency, while the highpass center fre- Y Separate highpass (or notch or allpass), bandpass, low-
quency depends on both resistor ratio and clock. Up to 4th pass outputs
order functions can be performed by cascading the two 2nd Y fO c Q range up to 200 kHz
order building blocks of the MF10; higher than 4th order Y Operation up to 30 kHz
functions can be obtained by cascading MF10 packages. Y 20-pin 0.3 wide Dual-In-Line package
Y 20-pin Surface Mount (SO) wide-body package

System Block Diagram Connection Diagram

Surface Mount and Dual-In-Line
Package

TL/H/10399 4
Top View

Order Number MF10AJ or MF10CCJ

See NS Package Number J20A
Order Number MF10ACWM or
MF10CCWM
See NS Package Number M20B
Order Number MF10ACN or
MF10CCN
TL/H/10399 1 See NS Package Number N20A

C1995 National Semiconductor Corporation TL/H/10399 RRD-B30M115/Printed in U. S. A.

 Absolute Maximum Ratings (Note 1)
If Military/Aerospace specified devices are required, Soldering Information
please contact the National Semiconductor Sales N Package: 10 sec. 260 C
Office/Distributors for availability and specifications. J Package: 10 sec. 300 C
Supply Voltage (V a b Vb) 14V SO Package: Vapor Phase (60 Sec.) 215 C
Infrared (15 Sec.) 220 C
Voltage at Any Pin V a a 0.3V
Vb b 0.3V See AN-450 Surface Mounting Methods and Their Effect
on Product Reliability (Appendix D) for other methods of
Input Current at Any Pin (Note 2) 5 mA
soldering surface mount devices.
Package Input Current (Note 2) 20 mA
Power Dissipation (Note 3) 500 mW Operating Ratings (Note 1)
Storage Temperature 150 C Temperature Range TMIN s TA s TMAX
ESD Susceptability (Note 11) 2000V MF10ACN, MF10CCN 0 C s TA s 70 C
MF10CCWM, MF10ACWM 0 C s TA s 70 C
MF10CCJ b 40 C s TA s 85 C
MF10AJ b 55 C s TA s 125 C

Electrical Characteristics V a e a 5.00V and Vb e b5.00V unless otherwise specified. Boldface limits
apply for TMIN to TMAX; all other limits TA e TJ e 25 C.
MF10ACN, MF10CCN,
MF10CCJ, MF10AJ
MF10ACWM, MF10CCWM
Symbol Parameter Conditions Tested Design Tested Design Units
Typical Typical
Limit Limit Limit Limit
(Note 8) (Note 8)
(Note 9) (Note 10) (Note 9) (Note 10)
V a b Vb Supply Voltage Min 9 9 V
Max 14 14 V
IS Maximum Supply Clock Applied to Pins 10 & 11
8 12 12 8 12 mA
Current No Input Signal
fO Center Frequency Min fO c Q k 200 kHz 0.1 0.2 0.1 0.2 Hz
Range Max 30 20 30 20 kHz
fCLK Clock Frequency Min 5.0 10 5.0 10 Hz
Range Max 1.5 1.0 1.5 1.0 MHz
fCLK/fO 50:1 Clock to MF10A Q e 10 Vpin12 e 5V g 0.2 g 0.6 g 0.6 g 0.2 g 1.0 %
Center Frequency MF10C Mode 1 fCLK e 250 kHz
Ratio Deviation g 0.2 g 1.5 g 1.5 g 0.2 g 1.5 %

fCLK/fO 100:1 Clock to MF10A Q e 10 Vpin12 e 0V g 0.2 g 0.6 g 0.6 g 0.2 g 1.0 %
Center Frequency MF10C Mode 1 fCLK e 500 kHz
Ratio Deviation g 0.2 g 1.5 g 1.5 g 0.2 g 1.5 %

Clock Feedthrough Q e 10
10 10 mV
Mode 1
Q Error (MAX) Q e 10 Vpin12 e 5V
g2 g6 g6 g2 g 10 %
(Note 4) Mode 1 fCLK e 250 kHz
Vpin12 e 0V
g2 g6 g6 g2 g 10 %
fCLK e 500 kHz
HOLP DC Lowpass Gain Mode 1 R1 e R2 e 10k 0 g 0.2 g 0.2 0 g 0.2 dB
VOS1 DC Offset Voltage (Note 5) g 5.0 g 20 g 20 g 5.0 g 20 mV
VOS2 DC Offset Voltage Min Vpin12 e a 5V SA/B e V a b 150 b 185 b 185 b 150 b 185
mV
(Note 5) Max (fCLK/fO e 50) b 85 b 85 b 85
Min Vpin12 e a 5V SA/B e Vb b 70 b 70 mV
Max (fCLK/fO e 50)

VOS3 DC Offset Voltage Min Vpin12 e a 5V All Modes b 70 b 100 b 100 b 70 b 100
mV
(Note 5) Max (fCLK/fO e 50) b 20 b 20 b 20
VOS2 DC Offset Voltage Vpin12 e 0V SA/B e V a b 300 b 300 mV
(Note 5) (fCLK/fO e 100)
Vpin12 e 0V SA/B e Vb b 140 b 140 mV
(Note 5) (fCLK/fO e 100)
VOS3 DC Offset Voltage Vpin12 e 0V All Modes b 140 b 140 mV
(Note 5) (fCLK/fO e 100)

2
 Electrical Characteristics (Continued) V a e a 5.00V and Vb e b5.00V unless otherwise specified.
Boldface limits apply for TMIN to TMAX; all other limits TA e TJ e 25 C.
MF10ACN, MF10CCN,
MF10CCJ, MF10AJ
MF10ACWM, MF10CCWM
Symbol Parameter Conditions Tested Design Tested Design Units
Typical Typical
Limit Limit Limit Limit
(Note 8) (Note 8)
(Note 9) (Note 10) (Note 9) (Note 10)
VOUT Minimum Output BP, LP Pins RL e 5k g 4.25 g 3.8 g 3.8 g 4.25 g 3.8 V
Voltage Swing N/AP/HP Pin RL e 3.5k g 4.25 g 3.8 g 3.8 g 4.25 g 3.6 V
GBW Op Amp Gain BW Product 2.5 2.5 MHz
SR Op Amp Slew Rate 7 7 V/ms
Dynamic Range Vpin12 e a 5V
83 83 dB
(Note 6) (fCLK/fO e 50)
Vpin12 e 0V
80 80 dB
(fCLK/fO e 100)
ISC Maximum Output Short Source 20 20 mA
Circuit Current (Note 7) Sink 3.0 3.0 mA

Logic Input Characteristics Boldface limits apply for TMIN to TMAX; all other limits TA e TJ e 25 C
MF10ACN, MF10CCN,
MF10CCJ, MF10AJ
MF10ACWM, MF10CCWM
Parameter Conditions Tested Design Tested Design Units
Typical Typical
Limit Limit Limit Limit
(Note 8) (Note 8)
(Note 9) (Note 10) (Note 9) (Note 10)
CMOS Clock Min Logical 1 V a e a 5V, Vb e b5V, a 3.0 a 3.0 a 3.0 V
Input Voltage Max Logical 0 VLSh e 0V b 3.0 b 3.0 b 3.0 V
Min Logical 1 V a e a 10V, Vb e 0V, a 8.0 a 8.0 a 8.0 V
V e a 5V
Max Logical 0 LSh a 2.0 a 2.0 a 2.0 V
TTL Clock Min Logical 1 V a e a 5V, Vb e b5V, a 2.0 a 2.0 a 2.0 V
Input Voltage Max Logical 0 VLSh e 0V a 0.8 a 0.8 a 0.8 V
Min Logical 1 V a e a 10V, Vb e 0V, a 2.0 a 2.0 a 2.0 V
V
Max Logical 0 LSh a 0.8 a 0.8 a 0.8 V
Note 1: Absolute Maximum Ratings indicate limits beyond which damage to the device may occur. DC and AC electrical specifications do not apply when operating
the device beyond its specified operating conditions.
Note 2: 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 a 5 mA current limit to four.
Note 3: The maximum power dissipation must be derated at elevated temperatures and is dictated by TJMAX, iJA, and the ambient temperature, TA. The maximum
allowable power dissipation at any temperature is PD e (TJMAX b TA)/iJA or the number given in the Absolute Maximum Ratings, whichever is lower. For this
device, TJMAX e 125 C, and the typical junction-to-ambient thermal resistance of the MF10ACN/CCN when board mounted is 55 C/W. For the MF10AJ/CCJ, this
number increases to 95 C/W and for the MF10ACWM/CCWM this number is 66 C/W.
Note 4: The accuracy of the Q value is a function of the center frequency (fO). This is illustrated in the curves under the heading Typical Performance
Characteristics.
Note 5: VOS1, VOS2, and VOS3 refer to the internal offsets as discussed in the Applications Information Section 3.4.
Note 6: For g 5V supplies the dynamic range is referenced to 2.82V rms (4V peak) where the wideband noise over a 20 kHz bandwidth is typically 200 mV rms for
the MF10 with a 50:1 CLK ratio and 280 mV rms for the MF10 with a 100:1 CLK ratio.
Note 7: The short circuit source current is measured by forcing the output that is being tested to its maximum positive voltage swing and then shorting that output to
the negative supply. The short circuit sink current is measured by forcing the output that is being tested to its maximum negative voltage swing and then shorting
that output to the positive supply. These are the worst case conditions.
Note 8: Typicals are at 25 C and represent most likely parametric norm.
Note 9: Tested limits are guaranteed to Nationals AOQL (Average Outgoing Quality Level).
Note 10: Design limits are guaranteed but not 100% tested. These limits are not used to calculate outgoing quality levels.
Note 11: Human body model, 100 pF discharged through a 1.5 kX resistor.

3
Typical Performance Characteristics
Positive Output Voltage Swing Negative Output Voltage
Power Supply Current vs Load Resistance Swing vs Load
vs Power Supply Voltage (N/AP/HP Output) Resistance (N/AP/HP Output)

Negative Output Positive Output Swing Crosstalk vs Clock

Swing vs Temperature vs Temperature Frequency

Q Deviation vs Q Deviation vs Q Deviation vs

Temperature Temperature Clock Frequency

Q Deviation vs fCLK/fO Deviation fCLK/fO Deviation

Clock Frequency vs Temperature vs Temperature

TL/H/10399 2

4
Typical Performance Characteristics (Continued)

fCLK
fCLK/fO Deviation fCLK/fO Deviation Deviation of
vs Clock Frequency vs Clock Frequency vs Nominal Q fO

fCLK
Deviation of
vs Nominal Q fO

TL/H/10399 3

Pin Descriptions
LP(1,20), BP(2,19), The second order lowpass, bandpass SA/B(6) This pin activates a switch that connects
N/AP/HP(3,18) and notch/allpass/highpass outputs. one of the inputs of each filters second
These outputs can typically sink 1.5 mA summer to either AGND (SA/B tied to
and source 3 mA. Each output typically Vb) or to the lowpass (LP) output (SA/B
swings to within 1V of each supply. tied to V a ). This offers the flexibility
INV(4,17) The inverting input of the summing op- needed for configuring the filter in its
amp of each filter. These are high im- various modes of operation.
pedance inputs, but the non-inverting in- VA a (7),VD a (8) Analog positive supply and digital posi-
put is internally tied to AGND, making tive supply. These pins are internally
INVA and INVB behave like summing connected through the IC substrate and
junctions (low impedance, current in- therefore VA a and VD a should be de-
puts). rived from the same power supply
S1(5,16) S1 is a signal input pin used in the all- source. They have been brought out
pass filter configurations (see modes 4 separately so they can be bypassed by
and 5). The pin should be driven with a separate capacitors, if desired. They
source impedance of less than 1 kX. If can be externally tied together and by-
S1 is not driven with a signal it should be passed by a single capacitor.
tied to AGND (mid-supply). VAb(14), VDb(13) Analog and digital negative supplies.
a
The same comments as for VA and
a
VD apply here.

5
Pin Descriptions (Continued) 1.0 Definition of Terms
LSh(9) Level shift pin; it accommodates various fCLK: the frequency of the external clock signal applied to
clock levels with dual or single supply pin 10 or 11.
operation. With dual g 5V supplies, the fO: center frequency of the second order function complex
MF10 can be driven with CMOS clock pole pair. fO is measured at the bandpass outputs of the
levels ( g 5V) and the LSh pin should be MF10, and is the frequency of maximum bandpass gain.
tied to the system ground. If the same (Figure 1)
supplies as above are used but only TTL
fnotch: the frequency of minimum (ideally zero) gain at the
clock levels, derived from 0V to a 5V
notch outputs.
supply, are available, the LSh pin should
be tied to the system ground. For single fz: the center frequency of the second order complex zero
supply operation (0V and a 10V) the pair, if any. If fz is different from fO and if QZ is high, it can be
VAb, VDb pins should be connected to observed as the frequency of a notch at the allpass output.
the system ground, the AGND pin (Figure 10)
should be biased at a 5V and the LSh Q: quality factor of the 2nd order filter. Q is measured at
pin should also be tied to the system the bandpass outputs of the MF10 and is equal to fO divided
ground for TTL clock levels. LSh should by the b3 dB bandwidth of the 2nd order bandpass filter
be biased at a 5V for CMOS clock lev- (Figure 1) . The value of Q determines the shape of the 2nd
els in 10V single-supply applications. order filter responses as shown in Figure 6 .
CLKA(10), Clock inputs for each switched capaci- QZ: the quality factor of the second order complex zero pair,
CLKB(11) tor filter building block. They should both if any. QZ is related to the allpass characteristic, which is
be of the same level (TTL or CMOS). written:
The level shift (LSh) pin description dis-
s0O
# J
cusses how to accommodate their lev-
HOAP s2 b a 0O2
els. The duty cycle of the clock should QZ
be close to 50% especially when clock HAP(s) e
s0O
frequencies above 200 kHz are used. s2 a a 0O2
Q
This allows the maximum time for the
internal op-amps to settle, which yields where QZ e Q for an all-pass response.
optimum filter operation. HOBP: the gain (in V/V) of the bandpass output at f e fO.
50/100/CL(12) By tying this pin high a 50:1 clock-to-fil- HOLP: the gain (in V/V) of the lowpass output as f x 0 Hz
ter-center-frequency ratio is obtained. (Figure 2) .
Tying this pin at mid-supplies (i.e, analog
ground with dual supplies) allows the fil- HOHP: the gain (in V/V) of the highpass output as f x
ter to operate at a 100:1 clock-to-cen- fCLK/2 (Figure 3) .
ter-frequency ratio. When the pin is tied HON: the gain (in V/V) of the notch output as f x 0 Hz
low (i.e., negative supply with dual sup- and as f x fCLK/2, when the notch filter has equal gain
plies), a simple current limiting circuit is above and below the center frequency (Figure 4) . When the
triggered to limit the overall supply cur- low-frequency gain differs from the high-frequency gain, as
rent down to about 2.5 mA. The filtering in modes 2 and 3a (Figures 11 and 8) , the two quantities
action is then aborted. below are used in place of HON.
AGND(15) This is the analog ground pin. This pin HON1: the gain (in V/V) of the notch output as f x 0 Hz.
should be connected to the system HON2: the gain (in V/V) of the notch output as f x fCLK/2.
ground for dual supply operation or bi-
ased to mid-supply for single supply op-
eration. For a further discussion of mid-
supply biasing techniques see the Appli-
cations Information (Section 3.2). For
optimum filter performance a clean
ground must be provided.

6
1.0 Definition of Terms (Continued)
0O
HOBP S
Q
HBP(s) e
s0O
s2 a a 0O2
Q

fO
Qe ; f O e 0f Lf H
fH b fL
2

# 2Q 0# 2Q J J
b1 1
fL e fO a a1

TL/H/103995 TL/H/10399 6 2

# 0# J J
1 1
(b) fH e fO a a1
(a) 2Q 2Q

0 O e 2 q fO

FIGURE 1. 2nd-Order Bandpass Response

HOLP0O2
HLP(s) e
s0O
s2 a a 0O2
Q

fC e fO c
0#1 b
1
2Q2 J 0# 1
a b
1 2
J
2Q2
a1

0
1
fp e fO 1b
2Q2

1
HOP e HOLP c

0
TL/H/103997 1 1
TL/H/10399 8 1b
(a) Q 4Q2
(b)
FIGURE 2. 2nd-Order Low-Pass Response

TL/H/10399 9 TL/H/10399 10
(a) (b)

HOHPs2
HHP(s) e
s0O
s2 a a 0O2
Q

0#1
b1

(
1 2
J 0# 1 J
1
fC e fO c b a b a1
2Q2 2Q2

b1

01 (
1
fp e fO c b
2Q2

1
HOP e HOHP c

0
1 1
1b
Q 4Q2
FIGURE 3. 2nd-Order High-Pass Response

7
1.0 Definitions of Terms (Continued)
HON(s2 a 0O2)
HN(s) e
s0O
s2 a a 0O2
Q

fO
Qe ; fO e 0fL fH
fH b fL

# 2Q 0# 2Q J J
b1 1
fL e fO a a1

1 2
# 0# J J
TL/H/1039911 1
TL/H/10399 12 fH e fO a a1
(a) 2Q 2Q
(b)
FIGURE 4. 2nd-Order Notch Response

s0O

HAP(s) e
HOAP s2 b
# Q
a 0O2
J
s0O
s2 a a 0O2
Q

TL/H/1039913 TL/H/10399 14
(a) (b)
FIGURE 5. 2nd-Order All-Pass Response

(a) Bandpass (b) Low Pass (c) High-Pass

(d) Notch (e) All-Pass

TL/H/10399 15
FIGURE 6. Response of various 2nd-order filters as a function of Q.
Gains and center frequencies are normalized to unity.

8
2.0 Modes of Operation
The MF10 is a switched capacitor (sampled data) filter. To fO R3
fully describe its transfer functions, a time domain approach Q e e
BW R2
is appropriate. Since this is cumbersome, and since the e quality factor of the complex pole pair
MF10 closely approximates continuous filters, the following
discussion is based on the well know frequency domain. BW e the b3 dB bandwidth of the bandpass output.
Each MF10 can produce a full 2nd order function. See Ta- Circuit dynamics:
ble I for a summary of the characteristics of the various HOBP
modes. HOLP e or HOBP e HOLP c Q
Q
MODE 1: Notch 1, Bandpass, Lowpass Outputs: e HON c Q.
fnotch e fO (See Figure 7 )
HOLP(peak) j Q c HOLP (for high Qs)
fO e center frequency of the complex pole pair
MODE 1a: Non-Inverting BP, LP (See Figure 8 )
f f
e CLK or CLK f f
100 50 fO e CLK or CLK
100 50
fnotch e center frequency of the imaginary zero pair e fO.
R3
R2 Q e
HOLP e Lowpass gain (as f x 0) e b R2
R1
HOLP e b1; HOLP(peak) j Q c HOLP (for high Qs)
R3
HOBP Bandpass gain (at f
e e fO) e b R3
R1 HOBP1e b
R2

(
f x 0 b R2
HON e Notch output gain as f x f e HOBP2e 1 (Non-Inverting)
CLK/2 R1
Circuit Dynamics: HOBP1 e Q
Note: VIN should be driven from a low impedance ( k 1 kX) source.

TL/H/10399 16
FIGURE 7. MODE 1

TL/H/10399 17
FIGURE 8. MODE 1a

9
2.0 Modes of Operation (Continued)
MODE 2: Notch 2, Bandpass, Lowpass: fnotch k fO MODE 3: Highpass, Bandpass, Lowpass Outputs
(See Figure 9 ) (See Figure 10 )

0 0
fO e center frequency f R2 fCLK R2
fO e CLK c or c
100 R4 50 R4
0 0
f R2 f R2
e CLK a 1 or CLK a1
100 R4 50 R4 Q e quality factor of the complex pole pair

0
fCLK fCLK R2 R3
fnotch e or e c
100 50 R4 R2

# J
Q e quality factor of the complex pole pair fCLK R2
HOHP e Highpass Gain as f x eb
0R2/R4 a 1 2 R1
e

# J
R2/R3 R3
HOBP Lowpass Gain at f
e e fO e b
HOLP e Lowpass output gain (as f x 0) R1

# J
R2/R1 R4
e b HOLP e Lowpass Gain as f x 0 e b
R2/R4 a 1 R1
HOBP e Bandpass output gain (at f e fO) e bR3/R1 R2 HOHP
Circuit dynamics: e ;
HON1 e Notch output gain (as f x 0) R4 HOLP
R2/R1 HOBP e 0HOHP c HOLP c Q
e b
R2/R4 a 1 HOLP(peak) j Q c HOLP (for high Qs)
HOHP(peak) j Q c HOHP (for high Qs)
# J
fCLK
HON2 e Notch output gain as f x e b R2/R1
2
Filter dynamics: HOBP e Q 0HOLP HON2 e 0HON1 HON2

TL/H/10399 18
FIGURE 9. MODE 2

TL/H/10399 19
*In Mode 3, the feedback loop is closed around the input summing amplifier; the finite GBW product of this op amp causes a slight Q enhancement. If this is a
problem, connect a small capacitor (10 pF b 100 pF) across R4 to provide some phase lead.
FIGURE 10. MODE 3

10
2.0 Modes of Operation (Continued)
MODE 3a: HP, BP, LP and Notch with External Op Amp MODE 4: Allpass, Bandpass, Lowpass Outputs
(See Figure 11 ) (See Figure 12 )

0 0
f R2 fCLK R2 fO e center frequency
fO e CLK c or c
100 R4 50 R4 f f
e CLK or CLK;

0
R2 R3 100 50
Q e c
R4 R2 fz* e center frequency of the complex zero & fO
R2 f R3
HOHP e b Q e O e ;
R1 BW R2
R3 R3
HOBP e b QZ e quality factor of complex zero pair e
R1 R1
R4 For AP output make R1 e R2
HOLP e b
R1
# J
fCLK R2
HOAP* e Allpass gain at 0 k f k eb e b1

0 0
f Rh fCLK Rh 2 R1
fn e notch frequency e CLK or
100 RI 50 RI HOLP e Lowpass gain (as f x 0)
HON e gain of notch at
# J
R2
eb a 1 e b2

# J
Rg Rg R1
f e fO e Q HOLP b HOHP
RI Rh HOBP e Bandpass gain (at f e fO)
Rg
# J # J
R3 R2 R3
Hn1 e gain of notch (as f x 0) e c HOLP eb 1a eb2
RI R2 R1 R2

# as f x J
fCLK Circuit Dynamics: HOBP e (HOLP) c Q e (HOAP a 1)Q
Hn2 e gain of notch
2 *Due to the sampled data nature of the filter, a slight mismatch of fz and fO
occurs causing a 0.4 dB peaking around fO of the allpass filter amplitude
R
e b g c HOHP response (which theoretically should be a straight line). If this is unaccept-
Rh able, Mode 5 is recommended.

TL/H/10399 20
FIGURE 11. MODE 3a

TL/H/10399 21
FIGURE 12. MODE 4

11
2.0 Modes of Operation (Continued)
MODE 5: Numerator Complex Zeros, BP, LP MODE 6a: Single Pole, HP, LP Filter (See Figure 14 )
(See Figure 13 ) fc e cutoff frequency of LP or HP output

0 0
R2 fCLK R2 fCLK e R2 fCLK R2 fCLK
fO e 1a c or 1a c or
R4 100 R4 50 R3 100 R3 50

01 01
R2 f R1 fCLK R3
fz e b c CLK or b c HOLP e b
R4 100 R4 50 R1
R3 R2
Q e 01 a R2/R4 c HOHP e b
R2 R1
R3
QZ e 01 b R1/R4 c
R1 MODE 6b: Single Pole LP Filter (Inverting and Non-In-
verting) (See Figure 15 )
H0z1 e gain at C.Z. output (as f x 0 Hz) fc e cutoff frequency of LP outputs
b R2(R4 b R1)
R2 fCLK R2 fCLK
j or
R1(R2 a R4) R3 100 R3 50

# as f x 2 J
fCLK b R2
H0z2 e gain at C.Z. output e
R1 HOLP1 e 1 (non-inverting)

# J
R2 R3 R3
HOBP e b a1 c HOLP2 e b
R1 R2 R2

# R2 R4 J
R2 R1
a R4
HOLP e b c
a R1

TL/H/10399 22
FIGURE 13. MODE 5

TL/H/10399 23
FIGURE 14. MODE 6a

TL/H/10399 24
FIGURE 15. MODE 6b

12
 2.0 Modes of Operation (Continued)
TABLE I. Summary of Modes. Realizable filter types (e.g. low-pass) denoted by asterisks.
Unless otherwise noted, gains of various filter outputs are inverting and adjustable by resistor ratios.
Number of Adjustable
Mode BP LP HP N AP Notes
Resistors fCLK/fO
1 * * * 3 No
(2) May need input buffer.
1a HOBP1 e bQ HOLP a 1 2 No Poor dynamics for
HOBP2 e a 1 high Q.
Yes (above fCLK/50
2 * * * 3
or fCLK/100)
Universal State-Variable
3 * * * 4 Yes
Filter. Best general-purpose mode.
As above, but also includes
3a * * * * 7 Yes
resistor-tuneable notch.
Gives Allpass response with
4 * * * 3 No
HOAP e b1 and HOLP e b2.
Gives flatter allpass response
5 * * * 4
than above if R1 e R2 e 0.02R4.
6a * * 3 Single pole.
(2)
6b HOLP1 e a 1 2 Single Pole.
b R3
HOLP2 e
R2

3.0 Applications Information

The MF10 is a general-purpose dual second-order state filter. For the Chebyshev filter defined above, such a table
variable filter whose center frequency is proportional to the yields the following characteristics:
frequency of the square wave applied to the clock input f0A e 529 Hz QA e 0.785
(fCLK). By connecting pin 12 to the appropriate DC voltage,
f0B e 993 Hz QB e 3.559
the filter center frequency fO can be made equal to either
fCLK/100 or fCLK/50. fO can be very accurately set (within For unity gain at DC, we also specify:
g 6%) by using a crystal clock oscillator, or can be easily H0A e 1
varied over a wide frequency range by adjusting the clock H0B e 1
frequency. If desired, the fCLK/fO ratio can be altered by
The desired clock-to-cutoff-frequency ratio for the overall
external resistors as in Figures 9, 10, 11, 13, 14 and 15 . The
filter of this example is 100 and a 100 kHz clock signal is
filter Q and gain are determined by external resistors. available. Note that the required center frequencies for the
All of the five second-order filter types can be built using two second-order sections will not be obtainable with clock-
either section of the MF10. These are illustrated in Figures 1 to-center-frequency ratios of 50 or 100. It will be necessary
through 5 along with their transfer functions and some relat- fCLK
ed equations. Figure 6 shows the effect of Q on the shapes to adjust externally. From Table I, we see that Mode 3
of these curves. When filter orders greater than two are f0
desired, two or more MF10 sections can be cascaded. can be used to produce a low-pass filter with resistor-adjust-
able center frequency.
3.1 DESIGN EXAMPLE In most filter designs involving multiple second-order
In order to design a second-order filter section using the stages, it is best to place the stages with lower Q values
MF10, we must define the necessary values of three param- ahead of stages with higher Q, especially when the higher Q
eters: f0, the filter sections center frequency; H0, the pass- is greater than 0.707. This is due to the higher relative gain
band gain; and the filters Q. These are determined by the at the center frequency of a higher-Q stage. Placing a stage
characteristics required of the filter being designed. with lower Q ahead of a higher-Q stage will provide some
As an example, lets assume that a system requires a attenuation at the center frequency and thus help avoid clip-
fourth-order Chebyshev low-pass filter with 1 dB ripple, unity ping of signals near this frequency. For this example, stage
gain at DC, and 1000 Hz cutoff frequency. As the system A has the lower Q (0.785) so it will be placed ahead of the
order is four, it is realizable using both second-order sec- other stage.
tions of an MF10. Many filter design texts include tables that For the first section, we begin the design by choosing a
list the characteristics (fO and Q) of each of the second-or- convenient value for the input resistance: R1A e 20k. The
der filter sections needed to synthesize a given higher-order absolute value of the passband gain HOLPA is made equal

13
3.0 Applications Information (Continued)
to 1 by choosing R4A such that: R4A e bHOLPA R1A e The resistors for the second section are found in a similar
R1A e 20k. If the 50/100/CL pin is connected to mid-sup- fashion:
ply for nominal 100:1 clock-to-center-frequency ratio, we R1B e 20k
find R2A by:
R4B e R1B e 20k
f0A2 (529)2
R2A e R4A e 2 c 104 c e 5.6k and f0B2 (993)2
(fCLK/100)2 (1000)2 R2B e R4B e 20k e 19.7k
(fCLK/100)2 (1000)2
R3A e QA 0R2AR4A e 0.785 05.6 c 103 c 2 c 104 e 8.3k
R3B e QB 0R2BR4B e 3.55901.97 c 104 c 2 c 104 e 70.6k
The complete circuit is shown in Figure 16 for split g 5V
power supplies. Supply bypass capacitors are highly recom-
mended.

TL/H/10399 25
FIGURE 16. Fourth-Order Chebyshev Low-Pass Filter from Example in 3.1.
g 5V Power Supply. 0V5V TTL or b 5V g 5V CMOS Logic Levels.

TL/H/10399 26
FIGURE 17. Fourth-Order Chebyshev Low-Pass Filter from Example in 3.1.
Single a 10V Power Supply. 0V 5V TTL Logic Levels. Input Signals
Should be Referred to Half-Supply or Applied through a Coupling Capacitor.

14
3.0 Applications Information (Continued)

TL/H/10399 28
TL/H/1039927 (b) Voltage Regulator
TL/H/10399 29
(a) Resistive Divider with
Decoupling Capacitor (c) Operational Amplifier
with Divider
Va
FIGURE 18. Three Ways of Generating for Single-Supply Operation
2

3.2 SINGLE SUPPLY OPERATION 10 will have a 20 dB peak in its amplitude response at fO. If
The MF10 can also operate with a single-ended power sup- the nominal gain of the filter HOLP is equal to 1, the gain at
ply. Figure 17 shows the example filter with a single-ended fO will be 10. The maximum input signal at fO must therefore
power supply. VA a and VD a are again connected to the be less than 800 mVp p when the circuit is operated on
positive power supply (8V to 14V), and VAb and VDb are g 5V supplies.
connected to ground. The AGND pin must be tied to V a /2 Also note that one output can have a reasonable small volt-
for single supply operation. This half-supply point should be age on it while another is saturated. This is most likely for a
very clean, as any noise appearing on it will be treated as circuit such as the notch in Mode 1 (Figure 7) . The notch
an input to the filter. It can be derived from the supply volt- output will be very small at fO, so it might appear safe to
age with a pair of resistors and a bypass capacitor (Figure apply a large signal to the input. However, the bandpass will
18a) , or a low-impedance half-supply voltage can be made have its maximum gain at fO and can clip if overdriven. If
using a three-terminal voltage regulator or an operational one output clips, the performance at the other outputs will
amplifier (Figures 18b and 18c) . The passive resistor divider be degraded, so avoid overdriving any filter section, even
with a bypass capacitor is sufficient for many applications, ones whose outputs are not being directly used. Accompa-
provided that the time constant is long enough to reject any nying Figures 7 through 15 are equations labeled circuit
power supply noise. It is also important that the half-supply dynamics, which relate the Q and the gains at the various
reference present a low impedance to the clock frequency, outputs. These should be consulted to determine peak cir-
so at very low clock frequencies the regulator or op-amp cuit gains and maximum allowable signals for a given appli-
approaches may be preferable because they will require cation.
smaller capacitors to filter the clock frequency. The main
3.4 OFFSET VOLTAGE
power supply voltage should be clean (preferably regulated)
and bypassed with 0.1 mF. The MF10s switched capacitor integrators have a higher
equivalent input offset voltage than would be found in a
3.3 DYNAMIC CONSIDERATIONS typical continuous-time active filter integrator. Figure 19
The maximum signal handling capability of the MF10, like shows an equivalent circuit of the MF10 from which the out-
that of any active filter, is limited by the power supply volt- put DC offsets can be calculated. Typical values for these
ages used. The amplifiers in the MF10 are able to swing to offsets with SA/B tied to V a are:
within about 1V of the supplies, so the input signals must be Vos1 e opamp offset e g 5 mV
kept small enough that none of the outputs will exceed
Vos2 e b150 mV @ 50:1 b 300 mV @ 100:1
these limits. If the MF10 is operating on g 5V, for example,
Vos3 e b70 mV @ 50:1 b 140 mV @ 100:1
the outputs will clip at about 8 Vp p. The maximum input
voltage multiplied by the filter gain should therefore be less When SA/B is tied to Vb, Vos2 will approximately halve. The
than 8 Vp p. DC offset at the BP output is equal to the input offset of the
lowpass integrator (Vos3). The offsets at the other outputs
Note that if the filter Q is high, the gain at the lowpass or
depend on the mode of operation and the resistor ratios, as
highpass outputs will be much greater than the nominal filter
described in the following expressions.
gain (Figure 6) . As an example, a lowpass filter with a Q of

15
3.0 Applications Information (Continued)
Mode 1 and Mode 4 Mode 2 and Mode 5

# J
1 V
VOS(N) e VOS1 a 1 HOLP b OS3
#R JV
R2 1
Q Q VOS(N) e a1 OS1 c
p 1 a R2/R4
VOS(BP) e VOS3
1 VOS3
VOS(LP) e VOS(N) b VOS2 a VOS2 b :
1 a R4/R2 Q01 a R2/R4
Mode 1a Rp e R1//R3//R4
#1 QJ V
1 VOS3
VOS(N.INV.BP) e a OS1 b VOS(BP) e VOS3
Q
VOS(LP) e VOS(N) b VOS2
VOS(INV.BP) e VOS3
Mode 3
VOS(LP) e VOS(N.INV.BP) b VOS2
VOS(HP) e VOS2
VOS(BP) e VOS3

( # R2 J
R4 R4
VOS(LP) e VOS1 1 a b VOS2
Rp

# R3 J
R4
b VOS3

Rp e R1//R2//R3

TL/H/10399 30
FIGURE 19. MF10 Offset Voltage Sources

TL/H/10399 31
FIGURE 20. Method for Trimming VOS

16
3.0 Applications Information (Continued)
For most applications, the outputs are AC coupled and DC was fs/2 b 100 Hz. This phenomenon is known as alias-
offsets are not bothersome unless large signals are applied ing, and can be reduced or eliminated by limiting the input
to the filter input. However, larger offset voltages will cause signal spectrum to less than fs/2. This may in some cases
clipping to occur at lower AC signal levels, and clipping at require the use of a bandwidth-limiting filter ahead of the
any of the outputs will cause gain nonlinearities and will MF10 to limit the input spectrum. However, since the clock
change fO and Q. When operating in Mode 3, offsets can frequency is much higher than the center frequency, this will
become excessively large if R2 and R4 are used to make often not be necessary.
fCLK/fO significantly higher than the nominal value, espe- Another characteristic of sampled-data circuits is that the
cially if Q is also high. An extreme example is a bandpass output signal changes amplitude once every sampling peri-
filter having unity gain, a Q of 20, and fCLK/fO e 250 with od, resulting in steps in the output voltage which occur at
pin 12 tied to ground (100:1 nominal). R4/R2 will therefore the clock rate (Figure 21) . If necessary, these can be
be equal to 6.25 and the offset voltage at the lowpass out- smoothed with a simple R C low-pass filter at the MF10
put will be about a 1V. Where necessary , the offset voltage output.
can be adjusted by using the circuit of Figure 20 . This allows
The ratio of fCLK to fC (normally either 50:1 or 100:1) will
adjustment of VOS1, which will have varying effects on the
also affect performance. A ratio of 100:1 will reduce any
different outputs as described in the above equations. Some
aliasing problems and is usually recommended for wide-
outputs cannot be adjusted this way in some modes, how-
band input signals. In noise sensitive applications, however,
ever (VOS(BP) in modes 1a and 3, for example).
a ratio of 50:1 may be better as it will result in 3 dB lower
3.5 SAMPLED DATA SYSTEM CONSIDERATIONS output noise. The 50:1 ratio also results in lower DC offset
The MF10 is a sampled data filter, and as such, differs in voltages, as discussed in Section 3.4.
many ways from conventional continuous-time filters. An im- The accuracy of the fCLK/fO ratio is dependent on the value
portant characteristic of sampled-data systems is their ef- of Q. This is illustrated in the curves under the heading
fect on signals at frequencies greater than one-half the Typical Performance Characteristics. As Q is changed,
sampling frequency. (The MF10s sampling frequency is the the true value of the ratio changes as well. Unless the Q is
same as its clock frequency.) If a signal with a frequency low, the error in fCLK/fO will be small. If the error is too large
greater than one-half the sampling frequency is applied to for a specific application, use a mode that allows adjustment
the input of a sampled data system, it will be reflected to of the ratio with external resistors.
a frequency less than one-half the sampling frequency. It should also be noted that the product of Q and fO should
Thus, an input signal whose frequency is fs/2 a 100 Hz will be limited to 300 kHz when fO k 5 kHz, and to 200 kHz for
cause the system to respond as though the input frequency fO l 5 kHz.

TL/H/10399 32
FIGURE 21. The Sampled-Data Output Waveform

17
18
Physical Dimensions inches (millimeters)

20-Lead Ceramic Dual-In-Line Package (J)

Order Number MF10AJ or MF10CCJ
NS Package Number J20A

Molded Package (Small Outline) (M)

Order Number MF10ACWM or MF10CCWM
NS Package Number M20B

19
MF10 Universal Monolithic Dual Switched Capacitor Filter
Physical Dimensions inches (millimeters) (Continued)

20-Lead Molded Dual-In-Line Package (N)

Order Number MF10ACN or MF10CCN
NS Package Number N20A

LIFE SUPPORT POLICY

NATIONALS 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 2. A critical component is any component of a life
systems which, (a) are intended for surgical implant support device or system whose failure to perform can
into the body, or (b) support or sustain life, and whose be reasonably expected to cause the failure of the life
failure to perform, when properly used in accordance support device or system, or to affect its safety or
with instructions for use provided in the labeling, can effectiveness.
be reasonably expected to result in a significant injury
to the user.

National Semiconductor National Semiconductor National Semiconductor National Semiconductor

Corporation Europe Hong Kong Ltd. Japan Ltd.
1111 West Bardin Road Fax: (a49) 0-180-530 85 86 13th Floor, Straight Block, Tel: 81-043-299-2309
Arlington, TX 76017 Email: cnjwge @ tevm2.nsc.com Ocean Centre, 5 Canton Rd. Fax: 81-043-299-2408
Tel: 1(800) 272-9959 Deutsch Tel: (a49) 0-180-530 85 85 Tsimshatsui, Kowloon
Fax: 1(800) 737-7018 English Tel: (a49) 0-180-532 78 32 Hong Kong
Fran3ais Tel: (a49) 0-180-532 93 58 Tel: (852) 2737-1600
Italiano Tel: (a49) 0-180-534 16 80 Fax: (852) 2736-9960

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.