Voltage-divider bias is preferred over fixed-base bias chiefly because

the operating point becomes nearly independent of beta

it eliminates the need for an emitter resistor

it increases the transistor's breakdown voltage

In an ideal op-amp under negative feedback, the 'virtual short' means

the two inputs sit at the same voltage without current flowing between them

the two input terminals are physically connected

the two inputs sit at the same voltage without current flowing between them

the output is shorted to ground

the feedback resistor carries zero current

The instrumentation amplifier's high CMRR is set mainly by

resistor matching in the three-op-amp topology

a single op-amp's own open-loop CMRR

resistor matching in the three-op-amp topology

A practical differentiator circuit adds a small series resistor at the input chiefly to

limit the high-frequency gain and reduce noise amplification

protect the capacitor from reverse voltage

A Schmitt trigger differs from an ordinary comparator in that it

switches at two different thresholds depending on direction, from added positive feedback

uses negative feedback to force a virtual short

switches at two different thresholds depending on direction, from added positive feedback

has no output saturation levels

A monostable multivibrator, on receiving a trigger,

produces one output pulse of fixed width and returns to its stable state

free-runs indefinitely between two states

produces one output pulse of fixed width and returns to its stable state

toggles permanently and needs a second trigger to return

Which statements about the Barkhausen criterion for sinusoidal oscillators are correct?

It requires loop gain magnitude of exactly 1 and loop phase of exactly 0 (or a multiple of 360 degrees).

A real oscillator is typically built with slightly more than unity loop gain at startup.

Non-linearity is typically what settles the gain back to exactly 1 once amplitude is established.

It requires loop gain magnitude of exactly 1 and loop phase of exactly 0 (or a multiple of 360 degrees).

A real oscillator is typically built with slightly more than unity loop gain at startup.

The condition alone guarantees that oscillation will start from noise.

Non-linearity is typically what settles the gain back to exactly 1 once amplitude is established.

A phase-locked loop (PLL) locks a VCO's frequency to a reference by

feeding back the phase difference between VCO output and reference through a loop filter

directly shorting the VCO output to the reference

feeding back the phase difference between VCO output and reference through a loop filter

increasing the VCO's supply voltage until frequencies match

replacing the VCO with a crystal oscillator

Shot noise arises specifically from

the discreteness of charge carriers crossing a potential barrier under DC current

random thermal motion of carriers in a resistor

the discreteness of charge carriers crossing a potential barrier under DC current

imperfections and traps dominant at low frequency

mutual inductance between two current loops

Inductive coupling between two circuits is best reduced by

twisted-pair wiring, which cancels the enclosed loop area

a grounded conductive shield around the sensitive wire

twisted-pair wiring, which cancels the enclosed loop area

adding a series resistor at the input

Diodes, Transistor Biasing and Op-Amp Circuits

Section 6 of the GATE Instrumentation Engineering (IN) paper: the analog electronics that turns a sensor's microvolt or microamp signal into something the digital and communication sections downstream can use. Diodes and their small-signal models first, then BJT and MOSFET biasing — the DC operating point that must be set correctly before any small-signal gain formula means anything — then the op-amp, both as an ideal building block for linear signal shaping (adders, integrators, instrumentation amplifiers) and as a nonlinear one for waveform generation (comparators, multivibrators, oscillators, phase-locked loops), and finally the noise and grounding practice that decides whether a beautifully designed circuit actually reads a clean signal on the bench.

1. Diode applications and transistor biasing

A diode's applications follow directly from its one-way conduction: rectification (converting AC to pulsating DC), clipping (removing a signal above or below a reference), clamping (shifting a waveform's DC level without changing its shape), and voltage reference (the Zener, operated deliberately in reverse breakdown, where voltage stays nearly constant over a wide current range). For a transistor, biasing means fixing the DC operating point — the quiescent collector current and voltage for a BJT, the quiescent drain current and voltage for a MOSFET — independent of the device parameter spread between two units of the same part number.

🎯 Why voltage-divider bias is the standard choice

Fixed-base bias sets base current directly, so the operating point moves proportionally with beta, which varies widely between transistors and with temperature. Voltage-divider bias fixes the base VOLTAGE instead (with a stiff divider and an emitter resistor), so the emitter current is set almost entirely by that voltage and the emitter resistor, and barely at all by beta — this is what makes the operating point stable enough to mass-produce a circuit.

Small-signal analysis is done AFTER the DC bias point is fixed: replace every DC source by its AC equivalent (a voltage source by a short, a current source by an open) and replace the transistor by its small-signal model (h-parameters or the hybrid-pi model for a BJT, the transconductance gm and output resistance for a MOSFET) to find gain, input and output resistance. The two analyses answer different questions and neither can substitute for the other — a transistor correctly biased but analysed with the wrong small-signal model gives a plausible-looking wrong gain.

2. The op-amp as an ideal linear building block

The ideal op-amp has infinite open-loop gain, infinite input impedance, zero output impedance and infinite bandwidth; under negative feedback this collapses to two working rules — the virtual short (both inputs sit at the same voltage) and zero input current (nothing flows into either input terminal). Every standard linear application follows from applying just these two rules with Kirchhoff's current law at the summing node: the inverting amplifier, the adder (summing several inputs with individual weighting resistors), the subtractor (a differential configuration), the integrator (a capacitor in the feedback path, output proportional to the running integral of the input) and the differentiator (a capacitor at the input, output proportional to the input's rate of change).

Two application families worth telling apart
Circuit	Reactive element	What it does
Integrator	C in feedback path	output = running integral of input, scaled by 1/RC
Differentiator	C at the input	output proportional to d(input)/dt, scaled by RC
Instrumentation amplifier	none (three op-amps, resistors)	very high input impedance, high CMRR, gain set by one resistor

⚠️ Why the plain differentiator is avoided in practice

An ideal differentiator's gain rises without limit as frequency increases, which means it amplifies high-frequency noise more than it amplifies the signal — a syllabus fact worth remembering as the reason a practical differentiator always adds a small series resistor to cap the high-frequency gain, unlike the integrator, whose gain FALLS with frequency and is inherently well-behaved.

The instrumentation amplifier is worth its own entry rather than being read as three ordinary op-amps: two input buffers give it very high input impedance without loading the (often high-impedance) sensor, and the three-op-amp topology gives a high common-mode rejection ratio that is set almost entirely by resistor matching rather than by any single op-amp's own CMRR. A precision rectifier places a diode inside an op-amp's feedback loop so that the circuit rectifies signals well below a silicon diode's own 0.6 V forward-drop threshold — the op-amp's open-loop gain, not the diode alone, is what does the rectifying.

3. Comparators, multivibrators, oscillators and PLL

A comparator runs the op-amp open-loop (or with positive feedback) so the output slams to one supply rail or the other depending on which input is higher — no virtual short applies here, because there is no negative feedback forcing the two inputs together. A Schmitt trigger is a comparator with positive feedback added deliberately: it switches at two different thresholds depending on direction, and that gap — the hysteresis — is exactly what rejects noise near a single threshold from causing multiple spurious transitions.

A monostable multivibrator has one stable state and produces a single output pulse of fixed width per trigger — used for pulse-width standardisation and simple timing.
An astable multivibrator has no stable state and free-runs, continuously switching — the basic square-wave generator.
A bistable multivibrator has two stable states and needs a trigger to move between them — the basic flip-flop behaviour, and the building block of memory.

Sinusoidal oscillators need two conditions simultaneously, the Barkhausen criterion: loop gain magnitude exactly 1, and loop phase exactly 0 (or a multiple of 360 degrees) at the frequency of oscillation. A Wien-bridge oscillator sets this with an RC network giving zero phase shift at one frequency; a phase-shift oscillator uses three RC stages, each contributing 60 degrees, to reach the needed 180 degrees together with the amplifier's own inversion. A voltage-controlled oscillator (VCO) is different in kind: its output frequency is set by an input voltage rather than fixed by RC or LC values, which is exactly what a phase-locked loop needs at its output stage — a PLL locks a VCO's frequency to a reference by feeding back the phase difference between the two through a loop filter, and this is the standard way to generate a clean, frequency-agile clock or to demodulate an FM signal.

⚠️ Barkhausen is necessary, not sufficient

Loop gain magnitude of exactly 1 and phase of exactly 0 describes a SUSTAINED oscillation once it exists; a real oscillator is deliberately built with slightly MORE than unity loop gain at startup (so noise grows into oscillation) and relies on some non-linearity — amplifier saturation, an AGC element — to settle the gain back to exactly 1 once the amplitude is established. Designing for gain exactly 1 from the start is the standard exam trap: with component tolerances, the circuit as likely as not simply fails to start.

4. Noise, interference, shielding and component selection

Noise INTERNAL to a component has three named sources: thermal noise (from random carrier motion, present in any resistor above absolute zero, flat with frequency, and proportional to resistance, bandwidth and absolute temperature), shot noise (from the discreteness of charge carriers crossing a junction, present wherever DC current flows through a potential barrier such as a diode or transistor junction) and flicker or 1/f noise (dominant at low frequency, tied to imperfections and traps in the device, and the reason precision DC and low-frequency instrumentation circuits are the hardest to make quiet).

Interference EXTERNAL to the circuit couples in three ways: conductively (through a shared power or ground path — a common-impedance coupling), capacitively (through stray capacitance between a noisy conductor and a sensitive one, worse at high frequency and high impedance), and inductively (through mutual inductance between current loops, worse with high di/dt and large loop area). Each coupling path suggests its own cure: conductive coupling is broken by separate return paths or isolation, capacitive coupling by shielding (a grounded conductive enclosure) and physical separation, and inductive coupling by twisted-pair wiring, which cancels the loop area two wires would otherwise enclose.

ℹ️ Grounding is a design choice, not an afterthought

A single-point (star) ground avoids the common-impedance coupling that a chain of ground connections creates, where one stage's return current develops a voltage across shared wiring resistance and injects it into another stage's reference. This is why instrumentation-grade layouts route every sensitive return separately back to one point rather than daisy-chaining grounds along a rail.

Component selection against a specification is a matter of matching the right parameter to the right requirement, not choosing the cheapest or most precise part available: a resistor's tolerance and temperature coefficient matter where a bridge's balance depends on ratio stability, an op-amp's input bias current matters where source impedance is high (a picoammeter front end wants FET or CMOS input, not bipolar), and a capacitor's dielectric absorption and leakage matter in an integrator, where either directly corrupts the stored value.

Key takeaways

Voltage-divider bias is preferred because the operating point is set mainly by the divider and emitter resistor, barely by beta, which varies widely between transistors.
The virtual short and zero input current are the only two rules needed to derive every standard op-amp linear circuit — adder, subtractor, integrator, differentiator.
A practical differentiator adds a series resistor because its ideal gain rises without limit with frequency, amplifying noise; the integrator has no such problem because its gain falls with frequency.
A Schmitt trigger's hysteresis, from deliberate positive feedback, is what stops noise near a single threshold from causing multiple false transitions.
Barkhausen's loop gain of 1 and phase of 0 describes sustained oscillation; real oscillators start with slightly more than unity gain and rely on non-linearity to settle it.
Conductive, capacitive and inductive coupling each have a distinct cure — separate returns, shielding, and twisted-pair wiring respectively.

Practice questions (10)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

Voltage-divider bias is preferred over fixed-base bias chiefly because
1. it requires fewer resistors
2. the operating point becomes nearly independent of beta
3. it eliminates the need for an emitter resistor
4. it increases the transistor's breakdown voltage
Show answer
Answer: B — the operating point becomes nearly independent of beta
By fixing the base voltage with a stiff divider and adding an emitter resistor, the emitter (and hence collector) current is set almost entirely by that voltage and resistor, not by the widely-varying beta. It needs an extra resistor pair, not fewer, keeps the emitter resistor as an essential part of the scheme, and has nothing to do with breakdown voltage.
In an ideal op-amp under negative feedback, the 'virtual short' means
1. the two input terminals are physically connected
2. the two inputs sit at the same voltage without current flowing between them
3. the output is shorted to ground
4. the feedback resistor carries zero current
Show answer
Answer: B — the two inputs sit at the same voltage without current flowing between them
The virtual short is that the two input terminals are forced to the same voltage by the feedback action, with no actual current flowing between them (they are not physically joined) — that is exactly why it is called 'virtual'. The output is not grounded, and the feedback resistor generally does carry current in these circuits.
The instrumentation amplifier's high CMRR is set mainly by
1. a single op-amp's own open-loop CMRR
2. resistor matching in the three-op-amp topology
3. the supply voltage magnitude
4. the input bias current alone
Show answer
Answer: B — resistor matching in the three-op-amp topology
The classic three-op-amp instrumentation amplifier achieves its high CMRR largely through matched resistor ratios in the difference stage, which is why laser-trimmed resistor networks are used in precision parts — a mismatch there degrades CMRR far more than any single op-amp's own figure does.
A practical differentiator circuit adds a small series resistor at the input chiefly to
1. increase the DC gain
2. limit the high-frequency gain and reduce noise amplification
3. protect the capacitor from reverse voltage
4. convert it into an integrator
Show answer
Answer: B — limit the high-frequency gain and reduce noise amplification
An ideal differentiator's gain rises without bound as frequency rises, which amplifies high-frequency noise severely; a small series resistor caps this gain at high frequency while leaving the intended differentiating action largely intact at the frequencies of interest.
A Schmitt trigger differs from an ordinary comparator in that it
1. uses negative feedback to force a virtual short
2. switches at two different thresholds depending on direction, from added positive feedback
3. has no output saturation levels
4. requires a dual power supply
Show answer
Answer: B — switches at two different thresholds depending on direction, from added positive feedback
The defining feature is hysteresis from deliberately added positive feedback: the switching threshold going up is different from the threshold coming down, and that gap rejects noise near a single level. It is not a negative-feedback (virtual-short) circuit, still saturates to the rails, and does not specifically require a dual supply.
A monostable multivibrator, on receiving a trigger,
1. free-runs indefinitely between two states
2. produces one output pulse of fixed width and returns to its stable state
3. toggles permanently and needs a second trigger to return
4. produces a continuous sinusoid
Show answer
Answer: B — produces one output pulse of fixed width and returns to its stable state
'Mono' names the one stable state: a trigger pushes it to its single temporary state for a fixed, RC-set duration, after which it returns on its own — no second trigger needed. Free-running indefinitely describes astable, permanent toggling describes bistable, and a continuous sinusoid describes an oscillator, not a multivibrator at all.
Which statements about the Barkhausen criterion for sinusoidal oscillators are correct?
1. It requires loop gain magnitude of exactly 1 and loop phase of exactly 0 (or a multiple of 360 degrees).
2. A real oscillator is typically built with slightly more than unity loop gain at startup.
3. The condition alone guarantees that oscillation will start from noise.
4. Non-linearity is typically what settles the gain back to exactly 1 once amplitude is established.
Show answer
Answer: A — It requires loop gain magnitude of exactly 1 and loop phase of exactly 0 (or a multiple of 360 degrees).; B — A real oscillator is typically built with slightly more than unity loop gain at startup.; D — Non-linearity is typically what settles the gain back to exactly 1 once amplitude is established.
The criterion states the SUSTAINED-oscillation condition correctly, and it is true that practical designs start with excess gain and rely on non-linearity to settle it. But the criterion by itself is necessary, not sufficient, for startup — exact unity gain from a cold start would not reliably grow from noise, which is precisely why designs deliberately exceed it.
A phase-locked loop (PLL) locks a VCO's frequency to a reference by
1. directly shorting the VCO output to the reference
2. feeding back the phase difference between VCO output and reference through a loop filter
3. increasing the VCO's supply voltage until frequencies match
4. replacing the VCO with a crystal oscillator
Show answer
Answer: B — feeding back the phase difference between VCO output and reference through a loop filter
A PLL's phase detector compares the VCO output's phase to the reference's, and that error, after the loop filter, drives the VCO's control voltage until the phase (and hence frequency) difference is driven to zero — a closed feedback loop, not a direct connection or a component swap.
Shot noise arises specifically from
1. random thermal motion of carriers in a resistor
2. the discreteness of charge carriers crossing a potential barrier under DC current
3. imperfections and traps dominant at low frequency
4. mutual inductance between two current loops
Show answer
Answer: B — the discreteness of charge carriers crossing a potential barrier under DC current
Shot noise is specifically tied to DC current crossing a junction or barrier in discrete charge packets. Thermal noise is the random-motion source in any resistor; flicker (1/f) noise is the low-frequency, trap-related one; mutual inductance between loops is an external inductive-coupling mechanism, not an internal noise source at all.
Inductive coupling between two circuits is best reduced by
1. a grounded conductive shield around the sensitive wire
2. twisted-pair wiring, which cancels the enclosed loop area
3. increasing the supply voltage
4. adding a series resistor at the input
Show answer
Answer: B — twisted-pair wiring, which cancels the enclosed loop area
Inductive coupling scales with the loop area enclosing current flow, and twisting the go-and-return conductors together cancels that area almost completely. A shield is the standard cure for CAPACITIVE (electric-field) coupling instead; the other two options address neither mechanism.