LOW COST HIGH ACCURACY STM32 FFT LCR METER

STM32 LCR Meter Combines Analog Measurement With DSP

This low-cost LCR meter uses a mixed-signal architecture that combines a configurable analog front end with STM32-based digital signal processing. Instead of estimating only signal amplitude, the design measures both magnitude and phase so it can calculate complex impedance.

e Important Trade-Offs

A traditional I-V measurement places a resistor in series with the device under test and calculates impedance from the measured current and DUT voltage.

The difficulty is selecting the shunt resistance. A very small resistor minimizes its effect on the DUT but produces a correspondingly small measurement voltage. Increasing the shunt improves sensitivity but also increases burden voltage and changes the measurement conditions.

Switching multiple passive shunts can extend measurement range, but relay size or semiconductor-switch resistance introduces additional design complications.

Differential Measurements Improve Measurement Accuracy

The project modifies the basic auto-balancing approach by making differential measurements across both the DUT-related voltage path and the transimpedance feedback path.

Differential sensing can help reject unwanted common-mode signals and reduce errors caused by measuring relatively small signals in the presence of larger voltages.

The signal paths are also arranged so voltage and current measurements share much of the same downstream circuitry, helping reduce mismatch between the two measurement channels.

Four-Wire Kelvin Connections Reduce Lead Resistance Error

The DUT connects through a four-wire Kelvin interface using SMA connectors.

Kelvin sensing separates the current-carrying connections from the voltage-sensing connections. This is particularly important when measuring low resistance because ordinary probe and cable resistance can otherwise become a significant part of the measured value.

The technique helps explain how the project can meaningfully attempt measurements toward the low-ohm end of its stated range.

Configurable TIA Provides Automatic Current Ranging

A four-to-one analog multiplexer selects between four precision feedback resistors in the transimpedance amplifier.

Changing the feedback resistor changes the current-to-voltage conversion gain. The firmware can therefore select a resistance appropriate to the DUT impedance and keep the measured signal within a useful range.

The design then uses a programmable gain amplifier for additional adjustment, giving the current path two stages of controllable gain.

Shared Signal Chain Reduces Channel Mismatch

Buffered voltage and current signals are switched into a common precision instrumentation amplifier and programmable gain amplifier.

Using much of the same signal-processing chain for both measurements can reduce gain and phase differences that would otherwise appear if two completely separate analog channels were used.

This is especially important because impedance is calculated from the relationship between the voltage and current vectors. Any mismatch between their measurement paths can directly become impedance error.

Anti-Aliasing Protects the Digital Measurements

Before reaching the ADC, the selected signal passes through scaling circuitry and an anti-aliasing filter.

An ADC cannot distinguish a genuine low-frequency component from a higher-frequency signal that aliases into the sampled bandwidth. Filtering therefore helps suppress unwanted frequencies before digitization.

The design also uses oversampling, which can improve digital signal quality when implemented appropriately, although it does not eliminate the need for adequate analog filtering.

DAC Generates Controlled Excitation Signals

The test stimulus is generated digitally using the microcontroller’s DAC and then passed through filtering and buffering circuitry.

Sallen-Key filters remove unwanted spectral components from the generated waveform, helping provide a cleaner excitation signal to the DUT.

An external buffering arrangement is necessary because the internal DAC output stage may not provide the bandwidth, drive capability, or settling performance required for accurate operation at the highest measurement frequencies.

DSP Extracts Magnitude and Phase Information

After voltage and current signals are sampled, the firmware determines their in-phase and quadrature components.

These components form complex voltage and current vectors. Complex division of the voltage vector by the current vector then produces the complex impedance:

Z = V / I

The real and imaginary components of that impedance can subsequently be converted into resistance, capacitance, inductance, magnitude, and phase information.

Single-Bin Frequency Analysis Reduces Processing Work

The article describes the processing as a reduced one-bin FFT. In practical terms, the firmware is interested primarily in the known excitation frequency rather than calculating an entire spectrum.

Processing only the required frequency component reduces computational work while still extracting the real and imaginary components needed for impedance calculations.

This is particularly appropriate because the meter generates its own excitation signal and therefore already knows which frequency should contain the useful measurement information.

Interleaved ADCs Increase Effective Sampling Rate

The STM32’s three ADCs are used in an interleaved arrangement, producing stated sampling rates between approximately 1.4MSPS and 7.2MSPS.

A higher sampling rate provides many samples across each 1kHz, 10kHz, or 100kHz measurement waveform and gives the DSP more information from which to estimate magnitude and phase.

Interleaved ADC architectures must still be handled carefully because timing, offset, and gain mismatch between converters can introduce artifacts if they are significant.

Input Protection Helps Protect Sensitive Analog Stages

The DUT interface includes series resistance and bidirectional low-leakage diode clamps.

These components help prevent excessive voltages from reaching sensitive operational-amplifier inputs when an unexpected signal or charged component is connected.

Protection design is particularly important in test equipment because the instrument must tolerate a wider range of external conditions than circuitry connected permanently to a known signal source.

Analog Stability Matters at Higher Frequencies

The transimpedance amplifier uses an op-amp with approximately 20MHz gain-bandwidth product to reduce phase error, but its bandwidth must also be restricted sufficiently to maintain feedback stability.

This highlights an important LCR-meter design trade-off. Wide bandwidth helps preserve magnitude and phase accuracy at higher test frequencies, but excessive bandwidth or unsuitable feedback components can cause instability or ringing.

The analog front end therefore has to be optimized together with the DSP rather than considering either section independently.

Digital Control Enables Extensive Autoranging

The combination of selectable TIA feedback resistors, programmable gain, analog multiplexers, and MCU-controlled routing creates a large number of possible measurement configurations.

The project describes roughly 100 theoretical voltage and current ranges, although only a subset is used to keep calibration manageable.

This is a sensible engineering compromise because additional ranges are useful only when their improvement in measurement coverage justifies the extra calibration and firmware complexity.

Display Presents Primary and Secondary Information

A 128 × 64 LCD provides the local user interface.

The primary area shows the selected measurement quantity, while secondary information indicates details such as the active range and impedance representation.

A compact interface of this type is appropriate for bench equipment because the most important result remains immediately visible while diagnostic information is still available to the operator.

Calibration Is Essential Across a Wide Impedance Range

The project reports testing from approximately 0.1Ω to 10MΩ after calibration, with measurements performed at 1kHz, 10kHz, and 100kHz.

Such a wide range places very different demands on the analog circuitry. Low-impedance measurements are sensitive to connection and series resistance, while very high impedance measurements become increasingly sensitive to leakage, parasitic capacitance, input bias current, and electrical noise.

The stated performance should therefore be understood as results reported for the author’s prototype rather than a guaranteed specification for independently reproduced builds.

£55 Prototype Demonstrates Capable Low-Cost Instrumentation

The reported project cost was approximately £55, which is notable given the combination of programmable test frequencies, four-wire sensing, automatic ranging, digital phase detection, and impedance calculations.

Its strongest aspect is not simply low component cost but the use of measurement techniques found in more sophisticated instruments, including an auto-balancing bridge, programmable analog gain, Kelvin connections, and complex digital signal analysis.

The project demonstrates how careful analog design and microcontroller DSP can be combined to create genuinely useful laboratory instrumentation without relying on specialized impedance-measurement ICs.

Read more: LOW COST HIGH ACCURACY STM32 FFT LCR METER