Analog vs. DSP Lock-In Amplifiers: Trade-Offs in Weak Signal Recovery
Filed under Amplifiers
Lock-in amplifiers are indispensable instruments for recovering weak AC signals obscured by noise sources that are thousands of times larger. Historically, these instruments relied on analog phase-sensitive detection architectures. Modern designs have transitioned to digital signal processing (DSP) to perform demodulation, filtering, and phase shifting mathematically. This guide details the core technical trade-offs between analog and DSP lock-in topologies. It is intended for experimental physicists, test engineers, and technical buyers evaluating instruments for low-level signal recovery. By comparing standard design parameters, such as dynamic reserve, phase stability, and operational frequency limits, engineers can determine which architecture is best suited for their specific measurement constraints.
- Analog Operating Frequency (SR510)
- 0.5 Hz to 100 kHz
- DSP Operating Frequency (SR830)
- 1 mHz to 102.4 kHz
- Analog Dynamic Reserve (SR510)
- Up to 60 dB
- DSP Dynamic Reserve (SR830)
- >100 dB
- DSP Phase Resolution (SR850)
- 0.001°
Dynamic Reserve and Pre-filtering Limits
Dynamic reserve is a critical metric in phase-sensitive detection, representing the ratio of the largest tolerable noise signal to the full-scale input signal. In analog lock-in amplifiers, the dynamic reserve is constrained by the linearity and headroom of the analog mixer and output amplifiers. For instance, the analog Stanford Research Systems SR510 provides a dynamic reserve of up to 60 dB. Exceeding this limit causes saturation in the analog circuitry, leading to measurement errors. To manage large noise signals, analog systems often require manual gain adjustments or tracking band-pass filters.
In contrast, DSP lock-in amplifiers digitize the input signal early in the signal path using high-precision analog-to-digital converters (ADCs). Because the phase-sensitive detection is performed numerically, DSP instruments like the Stanford Research Systems SR830 and Stanford Research Systems SR850 achieve a dynamic reserve of >100 dB without requiring pre-filtering. This allows DSP systems to resolve signals buried under noise levels that would instantly saturate an analog mixer.
Phase Stability and Orthogonality
Phase-sensitive detection relies on multiplying the input signal with a reference signal. In analog lock-ins, temperature-induced drift in the analog multiplier and phase-shifting components can introduce gain and phase errors. The analog SR510 exhibits a gain stability of ±200 ppm/°C.
DSP architectures resolve these drift issues by executing the multiplication and phase-shifting steps using digital logic. The reference channel phase is controlled mathematically, providing stable performance. For example, the DSP-based SR850 features a reference phase drift of less than 0.01°/°C below 10 kHz and less than 0.1°/°C above 10 kHz. It also maintains channel orthogonality within 0.001° and offers a precise phase resolution of 0.001°. This degree of orthogonality and stability is physically difficult to achieve or maintain over time in purely analog demodulators.
Frequency Range and Signal Input Constraints
The operating frequency range of a lock-in amplifier depends heavily on its design. Analog instruments face low-frequency limits due to component drift and 1/f noise in the analog phase detectors. The SR510 operates down to a lower limit of 0.5 Hz and extends up to 100 kHz.
DSP lock-ins can operate much closer to DC because digital processing does not suffer from 1/f noise or drift in the demodulation stage. Both the SR830 and SR850 support an operating frequency range starting at 1 mHz and extending up to 102.4 kHz.
Input configurations also differ between these generations. The analog SR510 has a voltage input impedance of 100 MΩ in parallel with 45 pF and a full-scale sensitivity range starting at 100 nV. The DSP-based SR830 features a voltage input impedance of 10 MΩ in parallel with 25 pF, with a more sensitive full-scale range down to 2 nV.
Example instruments
Frequently asked questions
- Why does a DSP lock-in amplifier have a higher dynamic reserve than an analog lock-in?
- Analog lock-in amplifiers are limited by the physical dynamic range of their analog multipliers, which saturate when noise levels are high. DSP instruments digitize the input signal first and perform multiplication mathematically, preventing the digital mixer from saturating and allowing dynamic reserves of more than 100 dB.
- Does a DSP lock-in amplifier have lower input noise than an analog one?
- Not necessarily. The noise floor of both types is determined by the analog front-end preamplifier. For example, the analog SR510 has a typical voltage input noise of 7 nV/√Hz at 997 Hz, which is comparable to the DSP-based SR830, which has a voltage input noise of 6 nV/√Hz at 1 kHz.
- What are the low-frequency limitations of analog lock-in amplifiers?
- Analog lock-in amplifiers suffer from increased component drift, DC offset issues, and 1/f noise at very low frequencies. Consequently, models like the SR510 are limited to a minimum frequency of 0.5 Hz, whereas DSP models like the SR830 can operate down to 1 mHz.