The practical-circuits pool moves quickly from counters to power supplies, mixers, software-defined radio, op-amps, and oscillators. The common skill is not memorizing a catalogue of circuits; it is tracing one conserved or transformed quantity through the stage before attaching a circuit name.
Use the right trace for the circuit
First classify the stem as state, energy, spectrum, sampling, or feedback. Write the smallest model that preserves that quantity, then test the options against it. A choice may describe a real circuit yet belong to the wrong model: voltage regulation is not frequency selection, an RF filter does not create gain, and a timing label does not describe a logic operation.
| Question family | Minimal model |
|---|---|
| Counters and multivibrators | Count stable states and ask what triggers the transition; cascaded toggles give a division of 2n, so solve for n rather than counting answer choices |
| Amplifiers and regulators | Mark cutoff, saturation and the linear region, then account for power: input becomes useful output plus heat and unwanted spectral energy |
| Filters and matching | Separate impedance transformation from frequency rejection; for a filter, inspect passband, transition and stop band independently |
| Mixers and detectors | List the frequencies entering a nonlinear stage and the information retained by the modulation before choosing the following detector |
| Converters and DSP | Keep sample rate, bit depth and algorithm length on separate axes: bandwidth, quantization step and response sharpness are not interchangeable |
| Op-amps and oscillators | Distinguish the ideal model from finite gain-bandwidth, then identify the feedback element and the physical disturbance it can admit |
Three compact calculations cover several stems. A chain of n toggling stages changes frequency by 2n. A converter with n bits has 2n available codes, so compare that count with full-scale range divided by required step size. In a series pass regulator, compute loss as input power minus load power; with nearly the same current on both sides, the algebra reduces to the voltage removed by the pass element times that current. When the sample rate is reduced, recompute the new Nyquist boundary before discarding samples.
The traps
- Collapsing amplifier classes into one conduction angle. Class AB is the overlap region: each push-pull device conducts longer than a half-cycle but not for the entire cycle.
- Explaining switching efficiency by high gain or a special load. The loss falls because the device spends most of its time fully off or fully on, avoiding simultaneous large voltage and current.
- Trying to cancel the resistive part with negative resistance. A reactive matching network instead cancels the load's reactance and transforms the remaining resistance to the required value.
- Describing a switchmode regulator like a linear pass device. The switch changes pulse duty cycle and the following filter averages those pulses; continuous conduction control belongs to the linear regulator.
- Expecting mixer outputs only at two and four times an input frequency. The principal components are both inputs and their sum and difference; overload is a separate trap that adds spurious products rather than automatic limiting or AGC action.
- Confusing decimation with deleting digits or merely attenuating a signal. It lowers the effective sample rate by removing samples, which is why the spectrum must be limited first.
Try it
How many flip-flops are required to divide a signal frequency by 16?
- 4
- 6
- 8
- 16
Which of the following calculates power dissipated by a series linear voltage regulator?
- Input voltage multiplied by input current
- Input voltage divided by output current
- Voltage difference from input to output multiplied by output current
- Output voltage multiplied by output current
Why is an anti-aliasing filter required in a decimator?
- It removes high-frequency signal components that would otherwise be reproduced as lower frequency components
- It peaks the response of the decimator, improving bandwidth
- It removes low-frequency signal components to eliminate the need for DC restoration
- It notches out the sampling frequency to avoid sampling errors