home *** CD-ROM | disk | FTP | other *** search
- ond changed very little. This serves to demonstrate the use-
- fulness of sensitivity analysis. The filter now has the fol-
- lowing characteristics:
-
- FREQ S11 S21 S12 S22
- in GHz in dB in dB in dB in dB
- 0.3 -32.71 0 0 -32.71
- 0.6 -17.3 -0.08 -0.08 -17.3
- 0.9 -13.18 -0.21 -0.21 -13.18
- 1.4 -0.04 -19.99 -19.99 -0.04
- 1.7 0 -29.98 -29.98 0
-
- We see here that the filter behaves almost exactly as we
- had requested with very minor exceptions. Now, this becomes a
- good starting point for realizing the filter in hardware.
-
- b. FEEDBACK AMPLIFIER
-
- Now, let's try something a little more difficult. Let's
- take the feedback amplifier on page 21 and smooth out the re-
- sponse. This is going to require that we make certain judge-
- ments in the optimization elements this time, as we have eight
- that we can select from. We choose the following objectives:
-
- FREQ OBJECTIVES
- in GHz S11 WGT S22 WGT S21 WGT
- 0.05 -15 10 -15 10 25 10
- 0.0505 -15 10 -15 10 25 10
- 0.051 -15 10 -15 10 25 10
-
- These objectives are derived from the observations made
- from the previous response analysis. 25 dB of gain was readily
- attainable, but the return losses left quite a bit to be de-
- sired. Therefore, an idealistic -15 dB for the return losses
- was chosen in an attempt to improve on these two parameters.
- In choosing optimization elements, it is important that we
- first note that the value of the neutralization inductor, INS 8,
- was chosen so as to enhance the stability of the amplifier, as
- was the parallel resistor at the gate (the first parameter of
- CRP 3). This leaves us with six candidates for optimization.
- An initial stability analysis shows:
-
- ELEMENT PARAMETER GRADIENT
- TYPE NUM NUM VALUE
- 1 INP 1 1 177 -2.143
- 2 INS 2 1 462 1.818
- 3 CRP 3 2 14 10.482
- 4 CPP 5 1 8.59 1031.067
- 5 INS 6 1 1114 7.457
- 6 INP 7 1 177 -0.731
-
- ITERATION #0 ERROR = 540.8
-
-
-
- - 32 -