Design Principles of RF and Microwave Filters and Amplifiers

Solution to Assignment-3: Amplifier Design

Q1. A silicon transitor has following S-matrix at 1 GHz with a 50Ω reference impedance,
 
0.386 −1580 0.016 340
3.506 800 0.406 −430

The source and load impedances are 25Ω and 40Ω respectively. Compute the unilateral power
gain, unilateral available power gain and unilateral transducer power gain.
Solution:
From the S-matrix, it is obvious that, the transistor has S12 = 0.016 340 which is comparatively
small and closes to zero. Hence, it can be assumed that the transistor is unilateral.

The value of source reflection coefficient Γs and load reflection coefficient Γl can be calculated as

Zs − Z0
Γs =
Zs + Z0
25 − 50
=
25 + 50
1
=−
3
1
= 6 1800
3
Zl − Z0
Γl =
Zl − Z0
40 − 50
=
40 + 50
1
=−
9
1
= 6 1800
9

From the generic formula we can calculate the power gain of the transistor,

|S21 |2 (1 − |Γl |2 )
G=
(1 − |Γin |2 )|1 − S22 Γl |2

Now, as the transistor is approximated as unilateral, therefore, Γout ≈ S22 and Γin ≈ S11 . So,

|S21 |2 (1 − |Γl |2 )
G = GU =
(1 − |S11 |2 )|1 − S22 Γl |2
(3.50)2 × 0.99
=
0.86 × [1 − 0.056 1370]2
3.52 × 0.99
=
0.86 × 1.08
= 13.05
= 11.16dB

Similarly by applying the generic formula of available power gain as given below, the required

2
 parameter can be calculated.
|S21 |2 (1 − |Γs |2 )
GA = GAU =
|1 − S11 Γs |2 (1 − |S22 |2 )
3.52 × 0.89
=
0.84 × |1 − 0.136 270 |2
3.52 × 0.89
=
0.84 × 0.79
= 16.49
= 12.15dB

In similar fashion,
|S21 |2 [1 − |Γs |2 ][1 − Γl ]2
GT = GT U =
|1 − S11 Γs |2 |1 − S22 Γl |2
3.52 × 0.99 × 0.89
=
|1 − 0.136 270 |2 |1 − 0.056 1370 |2
3.52 × 0.99 × 0.89
=
0.79 × 1.08
= 12.65
= 11.02dB

Q2. A transistor has following S matrix at 1 GHz (50Ω) and connected to the network given below.

 
0.916 −440 0.066 680
3.926 1490 0.936 −170

Determine the value of Γs and Γl to maximize the transducer power gain.

Solution:
We need to determine the values Γs and Γl such that the transducer gain is maximum. Therefore,
at first we need to check the stability of the transistor.
By applying the Rollet’s stability criteria,

1 − |S11 |2 − |S22 |2 + |△|2

K=
2|S12 S21 |
where △ = S11 S22 − S12 S21 and |△| = |0.916 −440 0.936 −170 − 3.926 1490 0.066 680 | = |0.6 − j0.6| =
0.84 < 1. Hence,
1 − 0.912 − 0.932 + 0.842
K=
2 × 3.92 × 0.06
= 0.026
<1

As the transistor fails the Rollet’s stability test, the transistor network will be conditionally stable
for some values of Zs and Zl . Therefore, we need to determine the range of values Γl and Γs for

3
 Figure 1: The stable and unstable regions for a transistor with specific S-matrix

which the transistor network will be stable. We need to determine the centers and radii of input
and output stability circles.
∗ ∗
(S22 − △S11 )
Cl = 2
|S22 | − |△| 2

0.936 −170 − (0.846 −450 × 0.916 440 )

=
0.932 − 0.842
= 0.78 + j1.62
= 1.86 64.30
|S12 S21 |
Rl = 2
|S22 | − |△|2
0.235
=
0.932 − 0.842
= 1.48
∗ ∗
(S11 − △S22 )
Cs = 2
|S11 | − |△| 2

0.916 −440 − (0.846 −450 × 0.936 170 )

=
0.912 − 0.842
= −0.29 + j2.17
= 2.196 97.60
|S12 S21 |
Rs = 2
|S11 | − |△|2
0.235
=
0.912 − 0.842
= 1.92

The input stability circle divides the smith chart into two parts. The region of stability can be
determined by considering the valued of |S11 |. If |S11 | < 1, then the region containing the center
of the smith chart will be stable region for both input stability circle case (i.e. load reflection

4
 coefficient) and output stability circle case (i.e. source reflection coefficient).
Same recipe can be applied for identifying the stable region for output side but the decision has
to be taken based on the value |S22 |. In the present case the |S11 | < 1 and |S22 | < 1, thus, the
portion of smith chart containing the center will be stable region. The stable and unstable regions
are determined using Smith chart and shown in Figure-1.
√ √
B1 ± B12 −4|C1 |2 B2 ± B22 −4|C2 |2
Now, Γs = 2C1
and Γl = 2C2
. Where,
B1 = 1 + |S11 |2 − |S22 |2 − |△|2
= 0.257
B2 = 1 + |S22 |2 − |S11 |2 − |△|2
= 0.331
∗
C1 = S11 − △S22
= 0.916 −440 − (0.846 −450 × 0.936 170 )
= −0.03 − 0.27j
= 0.276 −96.30
∗
C2 = S22 − △S11
= 0.936 −170 − (0.846 −450 × 0.936 440 )
= 0.12 − 0.26j
= 0.296 −65.220
In order to have a solution of Γl and Γs for conjugate matching, the quantities, B12 − 4|C1 |2 and
B12 − 4|C1 |2 must have positive values. In this case, none of the quantities gives a positive value.
Therefore, the maximum gain can not be obtained and the transistor network can not be matched
to achieve the maximum gain.
Q3. An amplifier uses a GaAs HBT device having the following scattering parameters (Z0 = 50Ω).
The input of the transistor is connected to a source with Vs = 2V (peak) and Zs = 25Ω and the
output is connected to Zl = 100Ω load. Compute the available power from the source and power
delivered to the load.  
0.616 −1700 0.066 700
2.36 800 0.726 −250
Solution:
The reflection coefficients Γs and Γl can be determined as
Zs − Z0
Γs =
Zs + Z0
25 − 50
=
25 + 50
1
=−
3
1
= 6 1800
3
Zl − Z0
Γl =
Zl + Z0
100 − 50
=
100 + 50
1
=
3

5
 From the signal flow graph, the value of input and output reflection coefficients can be determined
as
S12 S21 Γl
Γin = S11 +
1 − S22 Γl
1
2.36 800 × 0.66 700 ×
= 0.616 −1700 + 3
1 − 0.726 −25 × 31
= −0.65 − j0.07
= 0.656 −173.770
S12 S21 Γs
Γout = S22 +
1 − S11 Γs
0 2.36 800 × 0.66 700 × 31 6 1800
= 0.72 −25 +
6
1 − 0.616 −1700 × 13 6 1800
= 0.703 − j0.33
= 0.7766 −250

Now, the average power supplied by the source to the transistor network is
1
Ps = |V + |2
2Z0
|Vs |2 |1 − Γs |2
=
8Z0 | − Γs Γin |2
2 |1 − 31 6 1800|2
= ×
8 × 50 |1 − ( 13 6 1800 × 0.656 −1730 )|2
= 0.014W

In similar fashion, the power delivered to the load can be determined as

|Vs |2 |S21 |2 (1 − |Γl |2 )|1 − Γs |2
Pl =
8Z0 |1 − S22 Γl |2 |1 − Γs Γin |2
2 2.32 × (1 − 91 ) × |1 − 31 6 1800 |2
= ×
8 × 50 |1 − 0.726 −250 |2 × |1 − (136 1800 × 0.656 −173.770 )|2
= 0.108W

Q4. Determine the length of short-circuited shunt stubs (l1 and l4 ) & the length of the transmis-
sion line (l2 and l3 ) at 3 GHz for the network shown below. The s-matrix of the network is.

 
0.666 −1420 0.036 620
2.396 760 0.726 −680

Solution:
The transmission lines used in the matching networks are assumed to be lossless. In order to

6
design the matching network, wee need to check the stability conditions of the transistor from the
supplied S−matrix values.
Applying Rollet’s stability criteria we get

1 − |S11 |2 − |S22 |2 + |△|2

K=
2|S12 S21 |

where △ = S11 S22 − S12 S21 and |△| = |0.666 −1420 0.726 −680 − 2.396 540 0.036 620 | = | − 0.38 +
j.173| = 0.42 < 1. Hence,

1 − 0.662 − 0.722 + 0.422

K=
2 × 2.39 × 0.03
= 1.55
>1

Therefor, the transistor network passes the Rollet’s stability criteria. So, the network is absolutely
stable for any values of Γs and Γl .
√ √
B1 ± B12 −4|C1 |2 B2 ± B22 −4|C2 |2
Hence, Γs = 2C1
and Γ l = 2C2
. Where,

B1 = 1 + |S11 |2 − |S22 |2 − |△|2

= 0.74
B2 = 1 + |S22 |2 − |S11 |2 − |△|2
= 0.91
∗
C1 = S11 − △S22
= −0.3 − 0.198j
∗
C2 = S22 − △S11
= 1.98 × 10−4 − j0.73

So, Γs = 0.78 + 0.62j = 0.796 33.10 and Γl = 0.78 + 0.62j = 0.996 38.480 . The length of the short
circuited shunt stubs and the length of the transmission line can be easily determined by using
Smith chart. Smith chart based solutions for the same are shown below.

Smith Chart based solution for input matching Smith Chart based solution for output matching
network network
7
It is mentioned in the problem that all the values are taken at a frequency of 3 GHz. Hence, the
10
value of wavelength λ = 3×10
3×109
= 10 cm.
So, the length of the short circuited stubs l1 = 0.06λ = 0.6cm, l4 = 0.182λ = 1.82cm, and the
length of the lossless transsion lines are l2 = 0.24λ = 2.4 cm and l3 = 0.482λ = 4.82cm. The entire
transistor network after conjugate matching is shown in Figure-2

Figure 2: A transistor network with conjugate matching sections