# Calculations and Simulations with Pads and Feedback Amplifiers

### Further Thoughts and Conclusions

**1.** Pads or attenuators can be extremely useful in RF systems for establishing a stable, wideband impedance. The worst case match looking into a pad (with the output termination being a short or open circuit) is a return loss equal to twice the attenuation. Hence, a 6 dB pad will establish a return loss looking into the pad of 12 dB or better. This is intuitively reasonable. Envision a pulse of energy launched toward the 6 dB pad. The signal will undergo a 6 dB attenuation. It then encounters a discontinuity at the open or short circuit. All energy is then reflected with the reflected wave going back toward the pad. It then returns to the source where it is again attenuated by 6 dB. The wave returns 12 dB weaker than it was when launched. (Another term we might used is “out and back” loss.)

**2.** Amplifiers are usually desired in our communications systems. An bipolar transistor amplifier should ideally have modest to high standing current, for the high power capability means that it can amplify small signals without excess distortion. But high current also means high gain. This gain can be reduced with negative feedback while maintaining the high current needed for low distortion.

**3.** Emitter degeneration is a form of negative feedback that increases input impedance. This is why followers have increased input resistance.

**4.** Parallel feedback in the form of resistance between collector and base reduces both input and output impedance.

**5.** Combinations of the two feedback forms can be combined to establish a desired gain and low S11 and S22. A good starting point for feedback elements is ( R-fb)(R-degen) = (R-source)(R-load) In our example amplifier, the source is 50 Ohms while the load (at the collector) is 200 Ohms. Hence, with a feedback resistor of 1300 Ohms, we would start with a degeneration R of 50*200/1300=7.7 Ohms. We used a value of 6 in parallel with the bias R of 68 Ohms in series with the r-e of the biased transistor (26/19=1.37) for a net of 6.9 Ohms, which is close.

**6.** Decreasing parallel feedback R while increasing degeneration move toward lower gain.

**7.** The feedback amplifier shown is a low power design. This is limited to much less than a Watt output. (Recall the equation for output load resistance for a given power: R-load = (Vcc-Ve)^2/(2*Pout) . ) If higher output power is desired, the transformer should be eliminated.

**8.** Any form of narrow band output termination will be reflected back to an amplifier input. This occurs to a greater extent with negative feedback amplifiers. (Problem: Simulate a lower gain amplifier (more feedback) than the one we have used and examine the effect of narrow band terminations.) Transformer feedback amplifiers (EMRFD Fig 2.81) are useful for low noise and low distortion with excellent impedance matches. However, they can be especially bad in reflecting an output load back to the input.

**9.** The problems of amplifiers reflecting impedances from one port to the other can often be fixed to some extent with a pad. In the example we used, we showed that a pad at the output improved the amplifier input match. It also improves the output match seen looking back from a point “to the right” (in our examples) of the pad. When we used a 6 dB pad, the worst case mismatch seen when looking back is 12 dB return loss. Hence, a filter just after the pad is going to see a reasonable termination.

**10.** Just one pad in a multiple stage amplifier can serve to preserve match in more than one stage.

**11.** It is often convenient to merge biasing and feedback resistors in one design, realizing fewer components in an amplifier. It is just as possible to build circuits where the bias elements and feedback components are well isolated. See EMRFD, Fig 2.66.

**12.** Although not a “universal truth,” negative feedback tends to enhance amplifier stability with regard to parasitic oscillation.

## Last Comments