Microphone & Sensor Preamps#

The first active stage in a signal chain has an outsized influence on the entire system’s noise performance. A preamplifier takes a weak signal — microvolts from a thermocouple, millivolts from a microphone, or nanoamps from a photodiode — and amplifies it to a level where subsequent stages can process it without degrading the SNR. Getting the preamp wrong means no amount of downstream processing can recover the lost signal quality.

Why the First Stage Dominates: Friis Intuition#

The Friis formula for cascaded noise figures shows that the first stage’s noise figure dominates the system:

NF_total ≈ NF₁ + (NF₂ - 1)/G₁ + (NF₃ - 1)/(G₁ × G₂) + …

The key insight: each subsequent stage’s noise contribution is divided by the gain that precedes it. If the first stage has 20 dB of gain and a noise figure of 3 dB, the second stage’s noise is attenuated by 20 dB when referred back to the input. This is why low-noise design effort concentrates on the preamp.

Practical consequence: A cheap noisy ADC following a good preamp can outperform an expensive low-noise ADC with a mediocre preamp, as long as the preamp has enough gain to lift the signal well above the ADC’s noise floor.

Impedance: Matching vs Bridging#

How the source connects to the preamp determines both signal transfer and noise behavior:

Impedance matching (Z_load = Z_source) — Maximizes power transfer. Used in RF systems (50 Ω), some professional audio (600 Ω legacy), and transmission line termination. The signal voltage at the load is half the open-circuit source voltage.

Impedance bridging (Z_load » Z_source) — Maximizes voltage transfer. Used in most audio and instrumentation applications. The preamp input impedance is typically 10× or more the source impedance. Nearly all the source voltage appears across the load.

Noise implications: The optimal source impedance for minimum noise depends on the amplifier’s input noise voltage (e_n) and noise current (i_n). There’s an optimum source impedance R_opt = e_n / i_n where total input-referred noise is minimized. For BJT-input op-amps, this favors low-impedance sources; for JFET/CMOS inputs, it favors higher impedances.

Preamp Topologies#

Inverting amplifier — Simple, predictable, with virtual ground at the input. Input impedance equals the input resistor (can be low). Good for current-output sources and when a defined input impedance is needed.

Non-inverting amplifier — High input impedance (set by the op-amp, not external resistors). Signal appears directly at the non-inverting input. Gain is 1 + Rf/Rg, so minimum gain is 1 (unity). The default choice for voltage-output sensors.

Instrumentation amplifier — Three op-amps (or a dedicated IC like INA128) providing differential input with high CMRR, high input impedance on both inputs, and gain set by a single resistor. Essential for bridge sensors, thermocouples, and any source where the signal rides on a common-mode voltage. CMRR > 80 dB is typical; precision instrumentation amps exceed 120 dB.

Charge amplifier (transimpedance) — Converts charge or current to voltage. A capacitor (for charge) or resistor (for current) in the feedback path of an op-amp. Used for piezoelectric sensors, photodiodes, and capacitive transducers. The feedback element sets the sensitivity; bandwidth is limited by the feedback network and op-amp GBW.

Source-Specific Considerations#

Dynamic microphones — Low impedance (150-600 Ω), low output (~1-5 mV), balanced. Need 40-60 dB of clean gain. BJT-input preamps often have lower noise at these source impedances. Phantom power (48 V) is not needed and should be blocked.

Condenser/electret microphones — Higher output, require bias voltage or phantom power. Electret capsules with built-in JFET buffer present a moderate source impedance. MEMS microphones often have digital (PDM) or analog outputs with built-in amplification.

Piezoelectric sensors — Very high source impedance (capacitive), charge output. Require either a charge amplifier or a buffer with extremely high input impedance (>10 MΩ, often >1 GΩ). JFET-input op-amps or dedicated charge amplifiers are necessary. Cable capacitance affects sensitivity if not using a charge amp.

Resistive bridges (strain gauges, load cells) — Differential millivolt output on a common-mode voltage. Instrumentation amplifier is the standard approach. Bridge excitation voltage, lead resistance, and CMRR all affect accuracy.

Tips#

  • Put as much gain as possible in the first stage, consistent with not clipping expected signals
  • Match the op-amp input type to the source impedance — BJT inputs for low-Z sources, JFET/CMOS for high-Z sources
  • Use an instrumentation amplifier for any differential or bridge-type sensor

Caveats#

  • Gain bandwidth product limits high-gain stages — An op-amp with 10 MHz GBW at a gain of 1000 (60 dB) only has 10 kHz bandwidth. High-gain preamps may need faster op-amps or cascaded lower-gain stages
  • Input bias current creates offset with high-impedance sources — A JFET-input op-amp has pA bias current; a BJT-input op-amp has nA to µA. With a 1 MΩ source, 100 nA of bias current creates 100 mV of offset — which may rail the output or reduce headroom
  • Phantom power can damage equipment — 48 V phantom power applied to an unbalanced input or a ribbon microphone can cause damage. Preamp design must account for phantom power routing
  • Cable capacitance acts as a low-pass filter — Long cables from high-impedance sources (guitar pickups, piezo sensors) roll off high frequencies. Moving the preamp close to the source (or using a buffer) solves this
  • EMI pickup scales with impedance — Higher-impedance nodes pick up more electromagnetic interference. High-impedance preamp inputs need careful shielding, short traces, and guard rings on PCBs

In Practice#

  • A preamp that produces excessive noise with a high-impedance source but not with a low-impedance source suggests input bias current or EMI pickup issues
  • An instrumentation amplifier that shows poor CMRR in practice may have mismatched gain resistors or layout-induced imbalance
  • A charge amplifier whose sensitivity changes with cable length is operating in voltage mode rather than true charge mode
  • High-frequency rolloff on a piezo or high-impedance source indicates cable capacitance loading the source
Backward Analog Front Ends Analog Filtering Forward