Microprocessor Programmable
Universal Active Filters
10
-180
-10
GAIN
-90
-30
0
-50
PHASE
90
-70
100 200
500 1K 2K
FREQUENCY (Hz)
VIN
5
INA
1
BPA
23
INB
180
5K 10K
VOUT
21
BPB
0
fO = 15.6kHz
fCLK = 1MHz
-10
-20
-30
-40
-50
1K 2K
VIN
5
fO = 31.3kHz
fCLK = 2MHz
fO = 47kHz
fCLK = 3MHz
5K 10K 20K 50K 100K
FREQUENCY (Hz)
VOUT
1
23
21
INA
BPA
INB
BPB
CLKA
11
MAX260
CLKB
12
WR, AX, DX
PROGRAM
CLK
CLKA,B MODEA MODEB fOA fOB QA
QB
120kHz 1
2 N = 56 N = 5 N = 96 N = 83
Figure 24. Wide Passband Chebyshev Bandpass Filter
Fourth-Order Butterworth Lowpass
Figure 26 shows a fourth-order Butterworth lowpass
with a cutoff frequency of 3kHz. Sections A and B of a
MAX260 are cascaded. The f0 and Q parameters for
each section are:
f0A = 3kHz
f0B = 3kHz
QA = 1.307
QB = 0.541
Mode 1 and a 400kHz clock are used. Because of low
Q values, the sampling errors of Figure 20 begin to look
significant in this case. From the graphs, using fCLK/f0
ratio near 133, f0A is about 4% high, f0B is 1.5% high,
QA is -1.2% low, and QB is -0.5% low. If these errors
are not a problem, the corrections can be ignored.
They are included here for best possible accuracy:
CLKA
11
MAX262
CLKB
12
WR, AX, DX
PROGRAM
CLK
CLKA,B MODE fOA
fOB
1 to 3MHz
1
N = 38 N = 0
QA
QB
N = 96 N = 96
Figure 25. High-Frequency Chebyshev Bandpass Filter
CLKA = CLKB = 400kHz
fCLK/f0A = 135.08 (N = 22), f0B = 2961Hz
(-1.3% correction)
fCLK/f0B = 139.80 (N = 25), f0A = 2861Hz
(-4.6% correction)
QA = 1.306 (N = 79, Q resolution prevents +0.5%
correction)
QB = 0.547 (N = 11 +1.1% correction)
Measured wideband noise for this filter is 123µV RMS.
If mode 2 were used, the noise would be 87µV RMS.
For lower noise with either mode, the first section
should have the highest Q (section A in this example).
______________________________________________________________________________________ 25