Signal Conditioning & Data Acquisition

The measurement chain

Every bridge measurement passes through the same chain:

A sensor converts a physical quantity (strain, displacement, acceleration) into a weak electrical signal. Signal conditioning makes it strong and clean. The analogue-to-digital converter (ADC) turns it into numbers. The data acquisition system (DAQ) records, time-stamps, stores and sends the data. A weakness at any stage limits the quality of the whole result.

Signal conditioning

Excitation

Many sensors are passive and need a supply: strain-gauge bridges and load cells need a stable excitation voltage (typically 2.5–10 V) or a constant current; LVDTs need an AC carrier; accelerometers (IEPE) need a constant current. The stability of the excitation directly affects the accuracy, since the output is proportional to it (ratiometric measurement removes the effect: the ADC reference is taken from the excitation).

Amplification

Sensor outputs are often in millivolts, while the ADC input range is volts. An instrumentation amplifier provides gain (×10 to ×1000) with a high input impedance and a high common-mode rejection ratio (CMRR) — the ability to reject the noise voltage that appears equally on both wires. Gain should use most of the ADC range for resolution, without saturating on peaks.

Filtering

• Low-pass filter — removes high-frequency noise and, importantly, the frequencies above half the sampling rate that would cause aliasing (an anti-aliasing filter).
• High-pass filter — removes DC drift and slow trends when only vibration is of interest.
• Band-pass and notch filters — isolate a band or remove mains hum at 50 Hz.
• Order and type (Butterworth, Bessel) affect the sharpness of the cut-off and the phase distortion; Bessel filters preserve the waveform shape, Butterworth filters give flat pass-band response.

Isolation, grounding and shielding

Noise in bridge sites — electrified railways, power lines, radio transmitters, lightning — is a serious problem.

Lead-wire effects

Long cables add resistance and pick up noise: three-wire or four-wire (remote sense) connections compensate for lead resistance in strain-gauge and RTD circuits.

Analogue-to-digital conversion

The ADC samples the analogue voltage at fixed instants and converts each sample into a binary number. Key parameters:

• Resolution — the number of bits $N$. The voltage represented by one bit (LSB) is

where $V_{FS}$ is the full-scale range. Quantisation error is up to ±½ LSB.

• Sampling rate ($f_s$) — number of samples per second.
• Number of channels and simultaneous vs multiplexed sampling — multiplexed systems scan the channels one after another (small time skew between channels), while simultaneous-sampling ADCs capture all channels at the same instant — important for modal analysis where phase between sensors matters.
• Input range and gain settings.

Sampling theorem and aliasing

The Nyquist–Shannon theorem states that to reproduce a signal without distortion, the sampling rate must be at least twice the highest frequency present:

If the signal contains frequencies above $f_s/2$ (the Nyquist frequency) they are folded back into the lower band as false frequencies: aliasing. It cannot be removed after sampling, so an analogue anti-aliasing low-pass filter is placed before the ADC. In practice, $f_s$ is chosen at 5–10 times the highest frequency of interest so that the waveform of the peaks (for example a train passage) is captured.