Analog Circuit Design

Analog design is where the precision work lives. Digital circuits deal in ones and zeros; analog circuits must get a signal to behave exactly right across temperature, supply voltage, component aging, and manufacturing tolerance. The goal is usually to amplify, filter, or measure something to a specified accuracy — and then verify it actually achieves that in the worst case. Small systematic errors compound through gain stages and become large errors at the output.

Transistor Amplifier Configurations

Beyond op-amps, discrete transistor stages are still essential for high-frequency work, driving large currents or voltages, or situations needing specific impedance characteristics:

ConfigurationVoltage gainInput impedanceOutput impedancePhase shiftPrimary use
Common Emitter (BJT)High (−RC/RE)Moderate (β×RE)Moderate (RC)180° (inverting)Voltage amplification, general purpose
Common Collector (Emitter Follower)~1High (β×RE)Low (RE/β)0° (non-inverting)Impedance buffer, current driving
Common Base (BJT)High (RC/RE)Very low (~26mV/Ic)High0° (non-inverting)RF amplifiers, current buffer
Common Source (FET)High (−gm×RD)Very high (GΩ)Moderate (RD)180° (inverting)FET voltage amplification
Common Drain (Source Follower)~1Very highLow (1/gm)Impedance buffer with FET input

Common Emitter with Emitter Degeneration

Common Emitter amplifier with degeneration A NPN transistor with a collector resistor RC to VCC, an emitter resistor RE to ground, and output taken from the collector. VCC RC (sets gain) Vout B C E NPN RE (degeneration)
Common-emitter NPN amplifier with collector resistor RC and emitter-degeneration resistor RE; the output is taken at the collector.
Voltage gain Av ≈ −RC / RE   (with degeneration, approximate)
Input impedance ≈ β × RE     (at low frequencies)
Quiescent collector current: Ic ≈ (VCC × R2/(R1+R2) − 0.7V) / RE

The emitter resistor RE provides local negative feedback that stabilises the operating point against β variation and temperature drift. Without it (bypassed with a capacitor), Av = −gm×RC = −RC×Ic/26mV — much higher gain but very sensitive to temperature and transistor β.

Instrumentation Amplifiers

When amplifying small differential signals in the presence of large common-mode noise — thermocouple outputs, Wheatstone bridge measurements, biosignals — an instrumentation amplifier is the correct tool. A simple difference amplifier has limited CMRR because resistor mismatch degrades rejection; the INA topology uses three op-amps with laser-trimmed resistors to achieve very high CMRR:

Three-op-amp Instrumentation Amplifier Two input op-amps buffering V1 and V2, connected by a shared gain-setting resistor RG, feeding a third difference op-amp. + V1 U1 + V2 U2 RG Gain setting Outputs feed difference amplifier U3
Three-op-amp instrumentation amplifier: input buffers share gain-setting resistor RG and feed a difference amplifier.
Gain = 1 + 2R/RG   (R is internal, RG is external)

INA128 with RG=10kΩ: Gain = 1 + 100k/10k = 11
INA128 with RG=1kΩ:  Gain = 1 + 100k/1k = 101

CMRR: 80–120dB (rejects common-mode, amplifies only difference)
PartGain rangeVin (cm) rangeVos (max)CMRR (min)Notes
INA128 / INA1291–10,000±10V25 µV90 dBClassic precision INA. ±15V supply.
INA3331–1,000−0.1 to VCC+0.1V25 µV100 dBSingle supply, low power — good for battery applications
AD82211–1,000−Vs+2.1V to +Vs−1.1V25 µV80 dBLow noise, rail-to-rail output. High-performance medical/industrial.
INA8261–1,000−0.1 to VCC+0.1V100 µV80 dBTiny, cheap. Good for data acquisition.

Noise in Analog Circuits

Every resistor and active component adds noise. Understanding noise sources lets you design for a required signal-to-noise ratio:

Noise typeSourceFormulaSpectrum
Thermal (Johnson) noiseAny resistanceV_n = √(4kTRB)White (flat)
Shot noiseDC current through junctionI_n = √(2qI_DC×B)White
Flicker (1/f) noiseSemiconductor surfaces, contactsIncreases as 1/fPink (rises at low freq)
Popcorn (burst) noiseDefects in ICsRandom step changesIrregular
Thermal noise example: R=10kΩ, T=300K, B=20kHz (audio bandwidth)
V_n = √(4 × 1.38×10⁻²³ × 300 × 10,000 × 20,000)
V_n = √(3.31×10⁻¹²) ≈ 1.82 µV RMS

Op-amp noise: input-referred voltage noise density en (nV/√Hz)
Total input noise in bandwidth B:
  V_noise = en × √B

TL071: en = 18 nV/√Hz → in 20kHz audio BW: V_noise = 18 × √20000 = 2.55 µV
NE5532: en = 5 nV/√Hz → V_noise = 5 × √20000 = 0.71 µV

Noise Budget Strategy

Total system SNR requirement → input-referred noise budget → op-amp selection

Example: 12-bit ADC, 3.3V reference, required SNR = 72dB (12-bit ideal)
LSB = 3.3V / 4096 ≈ 806 µV
Required noise < LSB/2 ≈ 400 µV RMS at the ADC input

With gain of 10 before the ADC:
Required input-referred noise < 40 µV RMS

Available bandwidth = 1kHz:
Required en < 40µV / √1000 Hz ≈ 1.26 µV/√Hz = 1260 nV/√Hz
Any op-amp with en << 1260 nV/√Hz is adequate → even LM358 (en ≈ 40nV/√Hz) is fine here

Transimpedance Amplifier (TIA)

Converts current to voltage. Essential for photodiode circuits, current-mode sensors, and output of charge-based sensors like accelerometers:

Transimpedance Amplifier A photodiode connected to the inverting input of an op-amp, with a feedback resistor Rf connecting the output to the inverting input. The non-inverting input is grounded. + PD Rf Vout = I_diode × Rf
Transimpedance amplifier converts photodiode current to output voltage through feedback resistor Rf.
Bandwidth: f_-3dB ≈ 1/(2π × Rf × Cf)   (Cf in parallel with Rf to prevent oscillation)
Stability: Cf ≥ √(Cd / (2π × Rf × GBW))  (Cd = photodiode capacitance)

Example: Rf=1MΩ, Cd=20pF, GBW=10MHz (op-amp):
Cf ≥ √(20×10⁻¹² / (2π×10⁶×10⁷)) ≈ 0.56pF → use 1pF
f_-3dB ≈ 1/(2π×10⁶×10⁻¹²) ≈ 160kHz

ADC Input Conditioning

A complete ADC input chain must address multiple issues:

ProblemSolution
Signal range doesn't match ADC input rangeScaling amplifier — gain or attenuation to fill 0 to Vref
Signal has DC offsetDifferential input or subtraction of offset before scaling
Frequencies above Nyquist presentAnti-aliasing low-pass filter before ADC
Source impedance too highVoltage follower or low-impedance buffer before filter
Impedance mismatch causes RC errorSource impedance must be ≪ ADC sample-and-hold input impedance
Ground loops / common-mode interferenceDifferential sensing + instrumentation amplifier

PCB Layout for Analog

Analog circuits are much more sensitive to layout than digital. These rules matter:

RuleReasoning
Solid ground plane on one layerLow-impedance return path; reduces ground bounce and electromagnetic pickup
Decoupling caps within 2mm of each IC VCC pinInductance in series with a cap renders it ineffective at high frequencies
Keep high-gain signal traces short and directLong traces act as antennas — pick up interference from adjacent signals
Guard ring around high-impedance nodesDriven ground ring around op-amp + input, TIA input prevents leakage from adjacent net biasing the node
Route digital clock signals away from analog pathsSwitching edges radiate EMI that couples into sensitive nodes
Separate analog and digital sections of the PCBCommon return current flows under traces — keep digital return currents from passing through analog section
Place filter caps at output, not input of amplifierFiltering at the output removes noise; filtering at the input sources are already at the op-amp input — mismatch biases can worsen noise

Precision Voltage References

A good ADC is only as accurate as its reference. Zener-based shunt references and bandgap references are the two families:

PartVoltageInitial accuracyTempcoNotes
LM385-2.52.5V±1.5%20 ppm/°CMicropower shunt reference. Good enough for 8-bit work.
TL4312.5–36V adj.±0.5%25 ppm/°CProgrammable with two resistors. Ubiquitous in power supply feedback.
LM4040-5.05V±0.1%50 ppm/°CPrecision shunt. Good for 10-bit ADC reference.
REF33252.5V±0.2%10 ppm/°CTI series reference. Good 12-bit performance.
MAX63505V±0.02%1 ppm/°CPrecision for 16-bit ADC. Low noise, laser-trimmed.

References