frequency response
TRANSCRIPT
ELC331T Amplifier Frequency Response
1
Electronics III Topic: Amplifier Frequency Response Compiled by: J. Pretorius January 2005 Introduction In Electronics I and Electronics II component- and circuit behavior were investigated involving conditions of only DC voltages and/or small signals with a constant frequency. Capacitors approached open circuit behavior at low frequencies and DC, and approached short circuit behavior at high frequencies. Which frequencies can be considered to be low and which as high? How will capacitors behave at frequencies that are neither low nor high, i.e. in the intermediate frequency range? How will other components behave at over a range of frequencies? These are all questions to be answered in Electronics III. To be able to investigate frequency response of components and circuits, some mathematical concepts need to be understood first. Frequency response is normally expressed graphically, with the frequency range indicated on the horizontal axis and the magnitude of the transfer function (gain, power or phase) on the vertical axis. Analysis will be done over a very wide range of frequencies. To be able to indicate a complete frequency range on one singular graph, the frequencies have to be expressed on a LOGARITHMIC scale. The transfer characteristic can also vary from very small values to very high values. Here the DECIBEL concept is used to express the values on the vertical axis. This graphical presentation is known as the BODE representation and referred to as the Bode Plot (gain and phase). The transfer function is used to describe the frequency-domain relationship between input and output signals appearing in different parts of a circuit. The phase plot or phase response shows the relationship between the output signal to input signal. Logarithmic scale (Frequency representation) The relationship between the variables of a logarithmic function: xba= and ax blog= for instance if b = 10 and x = 2, 100)10( 2 === xba and 2100loglog 10 === ax b Example 1.1 Find the value of x so that 100 000 = 10x. Solution
5)100000(log10 ==x Table 1.1 indicates how the logarithm of a number increases only as the exponent of the
ELC331T Amplifier Frequency Response
2
number.
)10(log 010 0
)10(log 110 1
)10(log 210 2
)10(log 310 3
)10(log 410 4
)10(log 510 5
)10(log 610 6
)10(log 710 7
and so on Table 1.1 Frequency can range from DC (0Hz) up to several Giga Hertz (GHz). Table 1.2 express frequencies in logarithms. 1Hz )1(log10 0 2Hz )2(log10 0.301 3Hz )3(log10 0.477 5Hz )5(log10 0.699 7Hz )7(log10 0.845 9Hz )9(log10 0.954 10Hz )10(log10 1 50Hz )50(log10 1.699 100Hz )100(log10 2 500Hz )500(log10 2.699 1000Hz (1kHz) )1000(log10 3 1.5kHz )1500(log10 3.699 10kHz )10(log 4
10 4 100kHz )10(log 5
10 5 1000kHz or 1MHz )10(log 6
10 6 and so on
Table 1.2 Graph 1.1 is a semi-logarithmic graph showing only one cycle. The graph should be viewed sideways for the log-sale to be on horizontal axis. Graph 1.2 is a semi-logarithmic graph showing four cycles.
ELC331T Amplifier Frequency Response
3
1
2
3
4
5
6
7
8
9
10
Graph 1.1. One cycle semi-logarithmic
ELC331T Amplifier Frequency Response
4
1
2
3
4
5
67891
2
3
4
5
67891
2
3
4
5
67891
2
3
4
5
678910
Graph 1.2. Semi-logarithmic four cycle
ELC331T Amplifier Frequency Response
5
Figure 1.1 shows the magnitude and phase plots for the transfer function
)2(1)(+
=s
sG
Figure 1.1. Frequency Response. Example of a magnitude and phase plot Decibel The magnitude of a frequency response graph (Bode plot) is indicated on the vertical axis. This magnitude can be in the form of relationship between input and output power or input and output voltage (signal). The range of this magnitude can spread over a wide range of values and is there for expressed in decibels.
Decibel to relate power levels is calculated with )(log101
210 PPGdB = dB and the
decibel to relate voltage levels is calculated with )(log201
210 VVGdB = dB
Table 1.3 below shows voltage gain in
out
VV expressed in dB.
Voltage gain dB level 0.5 -6 0.707 -3 1 0 2 6 10 20 100 40 1000 60 10000 80 1000000 120 and so on
ELC331T Amplifier Frequency Response
6
Another advantage working in dB is the logarithmic relationship of adding (summing) logs.
nT AAAAA ......321 ∗∗∗= then )(....)()()()( 321 dBAdBAdBAdBAdBA nT +++= This allows cascaded system’s gains to be added when working in dB. Figure 1.1 shows an example with the magnitude-expressed dB on the vertical axis. Plotting a frequency response (Bode Plot) The log-magnitude and phase frequency response curves as function of log ω are called Bode plots or Bode diagrams. The transfer function of a circuit in the frequency domain can take on the following general form:
))....()(())....()((
)(21
21
nm
k
pspspsszszszsK
sG++++++
=
with K a constant, z1,z2, to zk zeros, p1, p2, to pn poles and s the frequency. Simplify by working with the logarithm of the magnitude:
ω
ω
jsnm
k
pspspss
zszszsKjG
→⋅⋅+−−+−+−−
+++++++=
)(log20....)(log20)(log20log20
)(log20....)(log20)(log20log20)(log20
21
21
Magnitude plot Thus if the response of each term is known, the algebraic sum would yield the total response in dB. Further, if an approximation of each term could be made, which will only consists of straight lines, graphic addition of terms would be greatly simplified. To plot the magnitude plot a simple set of rules can be applied: Zeros have 0dB value from start freq. and then break upwards along an asymptote with 20dB/decade slope from its break point. Poles have 0dB value from start freq. and then break downwards along an asymptote with 20dB/decade slope from its break point. A zero at jω = 0 has a continues slope of +20dB/dec and a value of 0dB at ω = 1 rad/sec. A pole at jω = 0 has a continues slope of –20dB/dec and a value of 0dB at ω = 1 rad/sec. These graphs continue to infinity frequency. One decade is the start frequency multiplied with 10, and two decades is the start frequency multiplied by 100, and so on. A break point is a specific zero or pole value. Figure 1.2 below shows examples of magnitude plots.
ELC331T Amplifier Frequency Response
7
Figure1.2. Example of magnitude and phase plots. Example 1.2 Draw the magnitude Bode plot for the system transfer function
)1021)(
25001)(
251(
10)(5×
+++= ssssH
Solution
Let s = jω: )
1021)(
25001)(
251(
10)(5×
+++= ωωωω jjjjH
The system has three poles and no zeros. Poles are at: Pole Frequency p1 25 rad/sec p2 2500 rad/sec P3 2 x 105 rad/sec The graph can be subdivided into four magnitude graphs: Straight line gain at dBM 20)10log(20 ==
ELC331T Amplifier Frequency Response
8
A straight line at 0dB until p1 and then breaking downwards with a slope of -20dB/decade. A straight line at 0dB until p2 and then breaking downwards with a slope of -20dB/decade. A straight line at 0dB until p3 and then breaking downwards with a slope of -20dB/decade.
ELC331T Amplifier Frequency Response
9
The complete magnitude graph is the sum of above four graphs:
ELC331T Amplifier Frequency Response
10
Phase Plot The relation between the phase of the output signal to input signal can also be plotted on a log frequency graph with the phase relation indicated in degrees on the vertical scale. To plot the magnitude plot a simple set of rules can be applied: A zero at ω = 0 has a phase of 900 at all frequencies.
• A pole at ω = 0 has a phase of -900 at all frequencies. • A constant gain term has a 00 phase at all frequencies. • The phase plot for a zero is 00 until 1 decade before the zero
break point. It then breaks upwards with 450 per decade slope until 1 decade after the zero break point. The phase stays constant for the rest of the frequencies.
• The phase plot for a pole is 00 until 1 decade before the pole break point. It then breaks downwards with 450 per decade slope until 1 decade after the pole break point. The phase stays constant for the rest of the frequencies.
See Figure 1.2 for examples of phase plots. Example 1.3 Draw the phase Bode plot for the system transfer function in Example 1.2. Solution: The solution is shown in example 1.2. Example 1.4 Draw the Bode magnitude and phase plots for the following function:
)50)(7)(1(
)20()(+++
+=
sssssG
ELC331T Amplifier Frequency Response
11
Solution:
First normalize the function: )
501)(
71)(1(
)20
1()50)(7(
20
)( sss
s
sG+++
+=
• Zero break point at 20 • Pole break points at 1,7 and 50 • Constant gain factor of - 24.9dB
1.1. Problems 1.1.1. For each of the following functions plot the magnitude and phase Bode
plots.
1.1.1.1. )4)(2(
1)(++
=sss
sG
1.1.1.2. )4)(2(
)5()(++
+=
ssssG
1.1.1.3. )4)(2()5)(3()(
++++
=ssssssG
1.1.2. For each of the following functions plot the magnitude and phase Bode plots.
1.1.2.1. )
101)(
50001(
)90
1)((900)(
5ωω
ωωω jj
jjjH
++
+=
ELC331T Amplifier Frequency Response
12
1.1.2.2. )
1021)(
25001)(
251(
10)(5×
+++= ωωωω jjjjH
1.1.2.3. 2
2
)5000
1(
)90
1(200)( ω
ω
ω j
j
jH+
+=
1.2. Sources of Capacitance and Inductance in electronic circuits
The most common origins of capacitance include the discrete capacitors used in single element design, the stray capacitances contributed by interconnections such as wires or printed circuit board paths, and the internal capacitances that originate within electronic devices (ex. transistors).
1.2.1. Internal capacitance in Bipolar Junction Transistors
Figure 1.3. Small signal model of BJT including internal junction capacitors. On data sheets of discrete BJT’s , Cµ is also referred to as Cob. The lateral base resistance rx, which interacts with Cπ and Cµ at high frequencies is also shown in the device model in Fig 1.3. With Cπ and Cµ included, the dependent current source can still be considered a function of the total base current Ib if β0 is represented as a function of frequency β(ω).
Figure 1.4. Small signal BJT model with frequency dependent β(ω)
ELC331T Amplifier Frequency Response
13
µωπ CgCT
m −= With ωT the unity gain frequency.
µππππ
βωCC
gCCr
mT +
=+
=)(
0
On manufacturers’ data sheets, β(ω) is sometimes labeled with the symbol hfe which is short hand notation for hfe(ω).
1.2.2. Internal capacitance in Field Effect Transistors.
Figure 1.5.
1.3. Frequency response of electronic circuits and devices
Figure 1.6. RC circuit with capacitor as shunt element.
RCjCjRCj
VV
in
out
ωω
ω+
=+
=1
11
1
ELC331T Amplifier Frequency Response
14
Figure 1.7. Bode plots for RC circuit in Fig. 1.6.
Figure 1.8. RC circuit with C as series element.
RCjRCj
CjRR
VV
in
out
ωω
ω +=
+=
11
ELC331T Amplifier Frequency Response
15
Figure 1.9. Bode plots for RC circuit in Fig. 1.8. 1.3.1. Bode Plot of systems with complex transfer functions.
The task of constructing the Bode plot of any circuit, no matter how complex, is greatly simplified if its system function is expressed in the
general form )....1)(1()....1)(1()(
31
42
ωωωωωωωωω
ωjjjjjAjH
++++
=
with the numbered frequencies ω1….ωn the break points of the system function and A constant. The solitary factor jω is not present for all circuits. The super position of poles approximation can be applied to any system function that can be put in the form a midband-gain multiplied by separate low-frequency and high-frequency system functions.
+++
×
+++
=
⋅⋅=
)1)...(1()1(1
)1()(
...)1()(
)1()/)(
)(
210
0
nm
m
b
b
a
a
HL
jjjjj
jj
jj
A
HHAjH
ωωωωωωωωωω
ωωωω
ωωωω
ω
The breakpoints ωa….ωm of HL jointly define the low-frequency limit of the midband, and the breakpoints ω1….ωn of HH jointly define the high-frequency limit of the midband. The high-frequency –3dB point ωH of the system function which constitutes the upper limit of the midband region, is given by
ELC331T Amplifier Frequency Response
16
nn
H ωωωωωω
ω ....1....1121
1
21
=
+++=
−
This is known as the superposition-of-poles approximation at the high-frequency end of the midband. The high-frequency poles with the lowest frequency will make the most contribution to ωH. If one pole is significantly lower than the others, it will dominate ωH. Poles located near each other will make nearly equal contributions to ωH. Poles well above ωH will make little contribution to the value of ωH. The low-frequency –3dB point ωL of the system function which constitutes the lower limit of the midband region, is given by
mbaL ωωωω +++= .... This is known as the superposition-of-poles approximation at the low-frequency end of the midband. The low-frequency poles with the highest frequency will make the most contribution to ωL. If one pole is significantly higher than the others, it will dominate ωL. Poles located near each other will make nearly equal contributions to ωL. Poles well below ωL will make little contribution to the value of ωL.
1.3.2. High- and Low-frequency Capacitors
Capacitors only influence frequency response below or above the midband region. Conversely, the midband represents the frequency range over which circuit behavior is unaffected by circuit capacitors. Certain capacitors influence the region below the midband range, called low-frequency capacitors and others the region above the midband range, and called high-frequency capacitors. High-frequency capacitors degrade the gain above the midband and low-frequency capacitors degrade the gain below the midband. Low-frequency capacitors behave like short circuit in midband and the high-frequency capacitors like open circuit in the midband. General rules:
• Capacitors will function as low-frequency capacitors if they are in series with a circuit’s input or output terminal.
• Capacitors will function as high-frequency capacitors if they shunt input or output terminals to small-signal ground.
• All internal device and stray lead capacitors are high-frequency capacitors.
• All external series capacitors act as low-frequency capacitors. • A bypass capacitor that shunts a device terminal to ground acts
as a low-frequency capacitor. (Example CE)
ELC331T Amplifier Frequency Response
17
+
-gmVpi CL
+
-
Vs1
Cu
Cpi
Ce
CcCs
RLRc
Re
RpiRb
Rs
Cpi
Figure 1.10. Small-signal model showing all capacitors.
In Fig1.10 above all capacitors are shown. This circuit can be split into a low-frequency model showing only the low-frequency capacitors and a high-frequency model showing only the high frequency capacitors. Low-frequency capacitors: Cs, Cc, and CE High-frequency capacitors: CL, Cπ, and Cµ
Each capacitor is important to only one end of the frequency spectrum.
• At frequencies well below high-frequency end of midband: o All high-frequency capacitors can be treated as open
circuits. • At frequencies well above low-frequency end of midband:
o All low-frequency capacitors can be treated as short circuits.
• A model where low-frequency capacitors are short circuits and high-frequency capacitors are open circuits is called the midband model.
Example 1.5
Identify the high- and low-frequency capacitors in the following circuit and compute the midband gain and separately examine the high- and low-frequency ends of the Bode plot. Assume the following conditions for the BJT: VT = 25mV Cµ = 0 rx =10Ω fT = 450MHz βF = βo = 100 Vf = 0.7V ro = ∞
ELC331T Amplifier Frequency Response
18
+ Vcc12V
+
-
Vs110V
CL14pF
Cc10uF
Cs10uF RL
1M
Rc5.1k
RB1M
Rs1k
Q1
Figure1.11. Common emitter amplifier Solution
• Identify the low- and high-frequency capacitors. Low-frequency caps: CS and CC High-frequency caps: Cπ and CL
• Compute bias current through Q1 so that gm, rπ and Cπ can be determined.
B
fCCB R
VVI
−= and mAII BFC 13.1≈= β
VAmVI
gT
Cm 45≈=
η and Ω≈= k
gr
m
o 2.2β
π
pFCfg
CT
m 162
≈−= µπ π
• Evaluate the low-frequency model of the circuit
VoutVb
+
-gmVpi
+
-
Vs1
CcCs
RLRcRpi
Rb
Rs
Figure 1.12. Low-frequency small-signal model With voltage division Vb can be found:
sss
ss
sb
SSSxB
xBb
VRrCjRrCj
Rrr
VV
VRCjrrR
rrRV
)(1)(
)1()()(
+++
+==
+++
+=
π
π
π
ππ
π
π
ωω
ω
with (rπ + rx) ≈ rπ and RB (rπ + rx) ≈ rπ
ELC331T Amplifier Frequency Response
19
π
π
ωω
ω
VRRCjRRCj
RRRR
gV
RVgCjRR
RV
RIV
CLC
CLC
LC
LCmout
LmCLC
Cout
Loutout
)(1)(
)(1
+++
+−=
−++
=
⋅=
Substituting the equation for Vπ into the Vout equation will yield
)(1)(
)(1)(
)(1)(
)(1)(
CLC
CLC
Ss
SS
LC
L
SCm
S
out
SCLC
CLC
Ss
SS
LC
LC
SCmout
RRCjRRCj
RrCjRrCj
RRR
Rrr
RgVV
VRRCjRRCj
RrCjRrCj
RRRR
Rrr
RgV
+++
+++
++−=
+++
+++
++−=
ωω
ωω
ωω
ωω
π
π
π
π
π
π
π
π
The above expression is in the form H(jω)=A0HL.
Hzf
radRRC
Hzf
radRrC
CLC
SS
016.02
sec1.0)(
1
52
sec3.31)(
1
22
2
11
1
==
=+
=
==
=+
=
πω
ω
πω
ωπ
• Evaluate the high-frequency model of the circuit
rx
+
-gmVpi CL
+
-
Vs1 CpiRLRc
RpiRb
Rs
Figure 1.13. High-frequency small-signal model
VoutVpi
gmVpiCL
+
-
VTh CpiRLRc
Rp
gmVpi
Figure 1.14. Simplified high-frequency small-signal model
ELC331T Amplifier Frequency Response
20
SS
xSBSB
BSTh
xSBp
Rrr
VrRRr
rRR
RVV
rrRRR
+=
+++=
+=
π
π
π
π
π
)(
)(
SPS
ThP
VRCjRr
rV
RCjCj
Vππ
π
π
ππ ωω
ω++
=+
=1
111
[ ]
)(1)(
)()1(
LCL
LCmout
mLLCout
RRCjRR
VgV
VgCjRRV
ω
ω
π
π
+−=
−=
Substituting Vπ into Vout equation:
MHzf
RC
MHzf
RRC
RRCjRCjRrr
RRgVV
P
LCL
LCLPSLCm
S
out
5.142
1
24.22
)(1
)(11
11)(
44
4
33
3
==
=
==
=
+++−=
πω
ω
πω
ω
ωω
π
ππ
π
• Evaluate the midband-frequency modal for the circuit.
(Determination of midband gain) The midband gain can be determined from the low-frequency model equation by letting the frequency be much higher than ω1 or the high-frequency model equation by letting the frequency be much lower than ω3.
dBRR
RRrr
RgaLC
L
SCmv 44157≡−=
++−=
π
π
with CmRg− the basic midband gain
and SRr
r+π
π the input-loading factor
and LC
L
RRR+
the output-loading factor.
ELC331T Amplifier Frequency Response
21
Students have to be able to do above example with a FET transistor instead of a BJT. This is assumed to be self-study.
1.4. Problems (Frequency response of Transistor amplifiers) 1.4.1. Find the low- and high-frequency –3dB endpoints of the circuit. The
MOSFET has internal capacitance Cgs = 8pF, and negligible Cgd and Cds. The output resistance is ro = 100kΩ and K = 1mA/V2 and VTR = 4V so that ID = 1mA and gm = 2(KID)½ = 2mA/V.
C210uF
1kHz
V1-1/1V
+ VDD10V
CL1pF
C110uF
RL10k
RD5k
R21M
R11M
Rs10k Q1
Figure 1.15 Circuit for problem 1.8.1
1.4.2. Find the –3dB midband of the endpoints of the BJT amplifier using the superposition-of-poles technique. Assume the BJT to be at room temperature with parameters Cµ = 0, rx = 10Ω, fT = 450MHz, βF = βo = 100, Vf = 0.7V, η = 1, and ro = ∞.
ELC331T Amplifier Frequency Response
22
CM4pF
RE3.3k
+ Vee6V
+ Vcc6V
Vout
+
C2100uF
+
C12.2uF
1kHz
V1-10m/10mV
Q1
Rs1k RM
1M
Rc5k
R210k
R120k
Figure 1.16. Circuit for problem 1.8.2
1.5. Effect of Transverse Capacitance on amplifier response
Up to now the transverse capacitance Cµ was assumed to be negligible. This capacitance is almost always present in a real transistor. Two methods can be used to determine the influence of this capacitance on the frequency response of a circuit. One method uses the Thevenin-resistance to find the contribution of Cµ to the high-frequency midband point, and the other method is called the Miller’s Theorem. The later will be discussed and used. 1.5.1. Miller’s Theorem and Miller Multiplication
Miller’s Theorem is valid for all resistive, capacitive, inductive and circuits containing general linear complex impedance elements. A transverse impedance like Zµ can be modeled by the equivalent parallel impedance ZA.
Input
Va
Output
Vb
Input
Va
CA
Zu
Figure 1.17. Miller’s Theorem )1( abA VVCC −= µ Example1.6.
ELC331T Amplifier Frequency Response
23
Find the –3dB high-frequency endpoint using Miller’s theorem for Figure 1.16. Omit the load elements RM an CM and assume the BJT to be biased at IC = 1mA with parameters Cµ = 2pF, fT = 450MHz, ro = ∞, rx = 10Ω, βo = 100, gm = 40mA/V, and rπ = 2.5kΩ. Solution
gmVpi+
-
Vs1
Cu
Cpi
rx
RcRpi
Rb
Rs
↓
CA
gmVpi+
-
Vs1Cpi
rx
RcRpi
Rb
Rs
CA
Figure 1.18. Replacing Cµ with CA
with
1.5.2. High-Frequency poles with feedback resister
When a device is connected in the follower configuration, or when a resistor is shared between the input and output loops of an inverter, Miller’s Theorem must be used with care.
[ ] Ω=+=
=+
=
=−=−=
+=
650)(
591)(2
1402)1(
)(
21 RRRrrr
kHzCCr
f
pFVVCCRVgVCCCC
Sxth
AthA
outA
Cmout
AA
ππ
ππ
πµ
π
ππ
π
ELC331T Amplifier Frequency Response
24
Vout
gmVpi+
-
Vs1
Cu
Cpi
rx
Rc
Re
RpiRb
Rs
Figure 1.19. High-frequency model of BJT follower. Using Miller’s theorem will result in CA being in parallel with Cπ. CA appears between the base-emitter terminals and not between the base and ground. This will result in a tedious calculation to find the Miller’s capacitance CA. The high-frequency response of an amplifier with feedback resistor is best derived using the Thevenin-resistance method and the superposition of poles. Example 1.7 (Feedback resistor and Thevenin-resistance method) Find the upper –3dB endpoint of the circuit for which Cµ = 2pF, fT = 400MHz and rx = 10Ω. Suppose βo = 100 and IC = 2.5mA so that gm = 100mV/A and rπ = 1kΩ at room temperature. Remember CS behaves like a short circuit at the frequencies of interest. Note that the poles determined using Thevenin-resistance method apply whether the amplifier is used as a follower or an inverter.
RE1k
+
C110uF
1kHz
V1-10m/10mV
Q1
Rg1k
Rc5k
R26.8k
R127k
1kHz
V1-10m/10mV
Figure 1.20 The small-signal model of Fig1.19 can represent this amplifier.
ELC331T Amplifier Frequency Response
25
Let
pFCg
C
VRRR
RRV
RRR
v
RRRRRR
T
m
gS
Sb
bin
gbSs
38
845
121
21
21
≈−=
+=
+=
Ω===
µπ ω
The high-frequency –3dB point can be found by evaluating the small-signal Thevenin-resistance seen by Cπ and Cµ.
• Thevenin-resistance seen by Cπ
Vin = 0
e
c
ie
i(test)xb
+
-Vtest
gmVpi
rx
Rc
Re
Rpi
Rs
Figure 1.21 Applying KCL to nodes x and e:
)(1
)1(11
)(
)()(
xsE
EmE
xstest
xs
Etesttest
testmtesttest
test
Etesttest
testmEtestmEee
xs
etesttesttest
rRRrRgR
rRrv
rR
Rirvvgv
rvi
RirvvgRiivgRiv
rRvv
rvi
++
++
++
=
+
−+++=∴
−+=−+==
++
+=
ππ
π
π
πππ
π
The small-signal Thevenin-resistance seen by Cπ is given by the expression
Ω≈++
++
++== 18
)1(11)(1
ππ
π
rRgR
rRr
rRRiv
rE
mExs
xsE
test
testth
MHzCr
fth
23321
==ππ
π π
ELC331T Amplifier Frequency Response
26
• Thevenin-resistance seen by Cµ
ic
Vin = 0
e
c
ie
xb
+ -
Vtest
gmVpi
rx
Rc
Re
Rpi
Rs
+ -
Vtest
Figure 1.22 Applying KCL to nodes c and x yields
)()(
))((
)(
Cmxs
Cxstesttest
Ccxstesttest
Ccxsscxtets
testtests
mtestc
RgrrR
vRrRiv
RirRrv
iv
RirRivvvrv
iiii
vgii
++
−+++=∴
++−=∴
++=−=
−=−=
+=
ππ
π
π
π
ππ
π
Apply KVL around input loop:
MHzCr
f
kRgrrR
rRRrgrrrR
RrRiv
r
RgrrR
rRRrgrrrRi
RrRiv
rRRrgrrrRi
v
Rvgrv
vrRrv
i
Rvgrv
vrRi
th
Cmxs
xsEm
xsCxs
test
testth
Cmxs
xsEm
xstestCxstesttest
xsEm
xstest
Emxstest
Emxss
96.721
10))()()1(
)(()(
))()()1(
)(()(
)()1()(
)())((
)()(
≈=
Ω≈++
−++++
++++==∴
++
−++++
++++=∴
+++++
=
++=+−∴
++=+
µµµ
πππ
πµ
πππ
π
ππ
ππ
ππ
ππ
π
π
ππ
ππ
π Using the superposition of poles to find the upper –3dB point:
MHzMHzMHz
fffH 7.7)96.71
2331( 1 ≈+== −
µπ
ELC331T Amplifier Frequency Response
27
Important notice:
The equations for rthπ and rthµ show that the Thevenin-resistance seen by Cπ and Cµ can be reduced by increasing the value of RE which will increase the bandwidth of the amplifier. This increase comes with the expense of midband gain that will decrease as RE increase.
1.6. Frequency Response with Bypass Capacitor
The simultaneous goals of amplifier bias stability and large gain are often not compatible. A large resister like RE shared by the input and output terminals produces bias stability but reduce gain. The gain can be improved by bypassing (shunting) RE down to signal ground with a large capacitor that will act as a short circuit at frequencies of interest where high gain is required.
gmVpiCL
+
-
Vs1
Cu
Cpi
Ce
CcCs rx
RLRc
Re
RpiRb
Rs
Figure 1.23 The small-signal gain without CE of the above circuit was determined in Electronics II.
E
c
EbS
c
in
outv R
RRrRR
Rvv
A−
≈+++
−==
)1( 0
0
ββ
π
With CE is present and the frequency high that CE function as a incremental short circuit to signals, the gain is given by:
π
βrRR
Rvv
AbS
c
in
outv +
−== 0
The frequency where CE undergoes the transition from open- to short-circuit is of importance. The problem can be addressed by treating RE and CE in parallel as a complex impedance ZE.
EE
E
EEE CRj
RCj
RZωω +
==1
1
Substitution of ZE in the place of RE in the equation for Av will yield:
ELC331T Amplifier Frequency Response
28
++
=
=
++
+++−
=
1
1
111
)1(
0
2
1
2
1
0
0
β
ω
ω
ωωωω
ββ
π
π
rRRC
CR
jj
RrRRR
vv
SEE
EE
EbS
C
in
out
The zero ω1 represents the frequency at which CE begins to bypass the emitter node to ground and increasing the magnitude of the gain. The pole ω2 represents the frequency at which complete bypassing is achieved.
1.7. Problems (Bypass capacitor) 1.7.1. The following amplifier must provide a midband gain magnitude of at
least 120, beginning at a frequency no higher than 100Hz. Resisters R1, R2, and RE have been chosen to set the BJT bias current to about 1.4mA. Choose appropriate values for RC, CE and CS such that the gain objectives are met. The pole of CE should dominate the low-frequency response.
RE1k
+Vcc10V
Vout
+
CE
+
Cs
1kHz
V1-10m/10mV
Q1
Rs
Rc
R227k
R1100k
+Vcc10V
Figure 1.24
1.7.2. Consider the circuit in 1.7.1 with β0 = 90. Draw the low-frequency end of the Bode plot of Vout/Vs if RC = 2k2Ω, CS = 150µF and CE = 10µF.
ELC331T Amplifier Frequency Response
29
1.8. Active Filter Frequency Response (First Order) 1.8.1. Introduction
Filters of some sort are essential to the operation of most electronic circuits. It is therefore in the interest of anyone involved in electronic circuit design to have the ability to develop filter circuits capable of meeting a given set of specifications. Unfortunately, many in the electronics field are uncomfortable with the subject, whether due to a lack of familiarity with it, or a reluctance to grapple with the mathematics involved in a complex filter design. In circuit theory, a filter is an electrical network that alters the amplitude and/or phase characteristics of a signal with respect to frequency. Ideally, a filter will not add new frequencies to the input signal, nor will it change the component frequencies of that signal, but it will change the relative amplitudes of the various frequency components and/or their phase relationships. Filters are often used in electronic systems to emphasize signals in certain frequency ranges and reject signals in other frequency ranges. Such a filter has a gain which is dependent on signal frequency. As an example, consider a situation where a useful signal at frequency f1 has been contaminated with an unwanted signal at f2. If the contaminated signal is passed through a circuit (Figure 1.25) that has very low gain at f2 compared to f1, the undesired signal can be removed, and the useful signal will remain. Note that in the case of this simple example, we are not concerned with the gain of the filter at any frequency other than f1 and f2. As long as f2 is sufficiently attenuated relative to f1, the performance of this filter will be satisfactory.
In general, however, a filter's gain may be specified at several different frequencies, or over a band of frequencies. Since filters are defined by their frequency-domain effects on signals, it makes sense that the most useful analytical and graphical descriptions of filters also fall into the frequency domain. Thus, curves of gain vs frequency and phase vs frequency are commonly used to illustrate filter characteristics, and the most widely-used mathematical tools are based in the frequency domain.
Figure 1.25. Using a Filter to Reduce the Effect of an Undesired Signal at Frequency f2, while Retaining Desired Signal at Frequency f1.
ELC331T Amplifier Frequency Response
30
1.8.2. The Basic Filter Types There are five basic filter types (band-pass, notch, low-pass, high-pass, and all-pass).
1.8.3. Band-pass Filter
Figure 1.26. Examples of band-pass filter responses
1.8.4. Band-reject Filter (Notch)
Figure 1.27. Examples of band-reject filter responses.
ELC331T Amplifier Frequency Response
31
1.8.5. Low-pass Filter
Figure 1.28. Examples of Low-pass filter responses.
VoutVin
Z2
Z1
Figure 1.29. Inverting amplifier topology with feedback. In the inverting amplifier in Fig. 1.29. let Z1 be purely resistive R1. Let Z2 be a resistor R2 and capacitor C in parallel.
CRjRR
ZZ
VV
CRjR
CjR
CjR
CjRXRZ
RZZZ
VV
in
out
C
in
out
21
2
1
2
2
2
2
2
222
11
1
2
11
11
11
ω
ωω
ωω
+−=−=
+=
+===
=
−=
ELC331T Amplifier Frequency Response
32
Figure 1.30. Frequency response of an active low-pass filter.
1.8.6. High-pass Filter
Figure 1.31. Examples of high-pass filter responses.
VoutVin
Z2
Z1
Figure 1.32. Inverting amplifier. For the inverting amplifier in Fig. 1.32., let Z1 be R1 and C in series, and Z2 be R2. It is left to the reader to proof that:
CRjCRj
VV
in
out
1
2
1 ωω+
−=
and to plot the frequency response to indicate that it is a high-pass filter.