Using Bode Plots
Using Bode Plots
1/13
H ( s)
1
1 RC
1 RCs s 1 RC
H ( )
1
1 jRC
2/13
RC Highpass Filter
H ( s)
RCs
s
1 RCs s 1 RC
c= 1/RC
2n s
s 2 2n s n2
H ( s)
2n s
( s a )2
H ( s)
10
172
5828
0
|H( )| (dB)
2n s
( s a )( s b)
-10
-20
-30
-40
-50
-60 1
10
10
10
CT Frequency (rad/sec)
10
100+j995 5
100-j995 10
0.1
1000
1000
1
3/13
A Better Bandpass Filter? Suppose you want a BPF. with faster rolloff!
1
2
K1s
K2s
H ( s)
( s 1 )( s 2 ) ( s 1 )( s 2 )
BPF RLC w/
distinct poles
K1s K 2 s
H ( s)
2
2
( s 1 ) ( s 2 )
BPF RLC w/
repeated poles
K1s 2 K 2
H ( s)
2
2
( s 1 ) ( s 2 )
HPF RLC w/
Rep. poles
LPF RLC w/
Rep. poles
Looks like we could just cascade two of our RLC circuits Here we cascade BPFs.
Our cascade theory only
holds when attaching the
2nd system does not change
how the first one behaves!
Although TF Theory says this will work the problem is that the second
circuit Loads the first one!!!
So one approach would be to re-analyze this cascade and see if it will still
work but with some tweeks on the component choices.
Another approach is to use an op amp as a buffer between the stages!
Bandpass RLC
Bandpass RLC
It may be desirable
to add another
buffer here
5/13
Although these ideas lead to workable circuits they are not necessarily the best
For one thing they need inductors (which are big and cant be made in an IC!)
There are other forms See this link for the form used below.
http://pdfserv.maxim-ic.com/en/an/AN1795.pdf
6/13
20dB
0dB
f (Hz)
100
1000
From our Bode Plot Insight we know we can get this from a single real pole,
single real zero system with the zero first, then the pole:
H ( s)
(1 s / 1 )
(1 s / 2 )
H ( )
(1 j / 1 )
(1 j / 2 )
Z1 ( s ) R1 1 / C1s C2 sR1C1 1
Z 2 ( s ) R2 1 / C2 s C1 sR2C2 1
Aha!!! What we want!
Let : 1 1 / R1C1
2 1 / R2C2
Z1 ( ) C2 1 j / 1
Z 2 ( ) C1 1 j / 2
8/13
Gain
vin (t )
vo (t )
Rf
Ratio of
resistances
Ri
Z f (s )
Z i (s )
vin (t )
H ( s)
Z f ( s)
Zi ( s)
vo (t )
Wont affect our
magnitude: |H()|
9/13
vin (t )
vo (t )
But wait!! You then remember that op amps must always have negative
feedback at DC so putting Cf here is not a good idea
So we have to continue
We also might not like this circuit because it might not give us a very
large input impedance and that might excessively load the circuit
that you plug into this (e.g., the guitar)
Back to the drawing board!!!
10/13
Gain 1
vin (t )
Rf
Ri
vo (t )
We avoid the no DC feedback
issue but were not yet sure well
get what we want!
Oh Cool!! We Get
What We want!
Rf
Ri 1 / Ci s
Set:
( Ri 1 / Ci s ) R f
Ri 1 / Ci s
1
R R C
i
( Ri R f ) 1 / Ci s
628 rad/s
1
6283 rad/s
RiCi
Ri 1 / Ci s
H (s)
( R f Ri )Ci s 1
Ri Ci s 1
Choose:
R f 15k
Ri 1.5k
Ci 0.1F
11/13
|H( f )| (dB)
Discrepancy
due to use of
standard
values
f (Hz)
12/13
2.
3.
4.
5.
We used further insight into op-amps to realize that we could fix the
input impedance issue using a non-inverting form of the op-amp
circuit
6.
We didnt give up at first sign that the inverting form might not give
us the form we want
a. Through mathematical analysis we showed that we did in fact
get what we wanted!!!!!!
13/13