Practical signal conditioning practice that protects data integrity from strain, force, pressure, and acceleration sensors. Covers excitation, isolation, amplification, filtering, and the deployment patterns that prevent the most common forms of unrecoverable signal corruption.
What’s Covered
Why signal conditioning is the highest-leverage decision in your chain
Signal conditioning sits between the sensor and the ADC. Because everything downstream of conditioning operates on the conditioner’s output, errors introduced here propagate unfixed through the rest of the system. That makes conditioning the single highest-leverage decision in a measurement chain, and the most common cause of unrecoverable data quality problems.
Specifying the right conditioner is rarely about brand preference. It’s about matching four functions, excitation, isolation, amplification, filtering, to the sensor type and the environment.
The four functions of signal conditioning
| Function | What it does | What goes wrong if mis-specified |
|---|---|---|
| Excitation | Provides power to passive sensors (bridge, IEPE, LVDT, RVDT) | Sensor outputs zero, drifts, or behaves nonlinearly |
| Isolation | Breaks galvanic ground paths between sensor and DAQ | Ground loops add hum and DC bias; equipment damage in fault conditions |
| Amplification | Matches signal level to ADC input range | Lost dynamic range or clipping; raised noise floor |
| Filtering | Limits bandwidth to band of interest, prevents aliasing | Out-of-band content folds into your signal |
Isolation: stop ground loops before they start
Industrial environments are electrically noisy. Drives, switchgear, and high-current loads inject noise onto every available conductor, including ground. When sensor and DAQ are grounded at different potentials, current flows through the signal cable shield or the sensor wiring, producing the classic 60 Hz hum, drift, and erratic noise that is impossible to fix in software.
- Use galvanically isolated conditioners on sensors mounted to grounded equipment
- Tie cable shields at one end only, typically the conditioner end, unless the cable specification calls for both
- For high-noise environments, isolation per channel may be necessary; for benign benches, a common isolated input stage is sufficient
- Confirm the conditioner’s isolation voltage is high enough to survive worst-case fault potentials
Durham Instruments stocks isolated conditioners (ISG and similar) and JM Concept conditioners specifically designed for high-noise environments. Browse the full signal conditioning catalog and isolation & signal conditioning catalog.
Amplification: gain structure and dynamic range
Dynamic range is the ratio of the largest signal a system can record without clipping to the smallest signal that rises above its noise floor. Where you place gain in the chain affects dynamic range almost as much as ADC bit depth.
Gain near the sensor
Amplifying close to the sensor, at the conditioner, raises signal above the noise floor before any cable noise can be injected. This is the right answer for most low-level signals (strain, accelerometer, microphone).
Gain near the ADC
Late amplification raises both signal and noise equally. It’s appropriate when the sensor signal is already at a usable level and you only need to match the ADC’s input range.
Set gain so the largest expected signal reaches roughly 80% of full-scale ADC input. Lower than that wastes resolution; higher risks clipping on transients. Verify with a known signal, a bridge calibration shunt for strain, a calibrator for acoustic, a known load for force.
Filtering and aliasing
Anti-aliasing filters reject content above the Nyquist frequency, preventing it from folding into the band you care about. Without a properly specified anti-aliasing filter, fast transients from a switching power supply can appear as DC drift, and high-frequency mechanical noise can appear as low-frequency content that looks indistinguishable from real data.
- Anti-aliasing filter cutoff must be set well below Nyquist, typically 0.4× the sample rate
- Filter order must be steep enough to attenuate out-of-band content to below the system noise floor
- For programmable conditioners, document the filter setting actually applied
- For dynamic measurements, prefer modules with built-in matched anti-aliasing filtering rather than relying on post-acquisition software filters
Conditioning product types and where they fit
| Product class | Sensor types | Where it fits |
|---|---|---|
| Bridge amplifiers (BSC4A multi-channel) | Strain, load, pressure | Test rigs, durability stands |
| Industrial digital amplifiers (BX series) | Strain, load with CANbus / USB / Bluetooth output | In-line measurement, mobile rigs |
| Modular industrial systems (PME / PMX) | Mixed industrial sensors with fieldbus | Process and production environments |
| Portable data loggers (BX6-BT) | Strain, load with Bluetooth | Service and commissioning, mobile data capture |
| LVDT/RVDT conditioners (LVM-110, SCM100, LDM-1000) | LVDT, RVDT, half-bridge | Position measurement, aerospace test |
| Isolated conditioners (ISG) | Any sensor in high-noise environment | Industrial floor, near drives and switchgear |
| In-line conditioners (Model 500, CSC, LCSC) | Force and load with integrated conditioning | Embedded measurement in production lines |
| Distributed I/O (ioPro) | Mixed analog signals with networked output | Plant-wide signal aggregation |
FAQ
What is the difference between bridge excitation and IEPE excitation?
Bridge excitation provides a stable DC voltage to a Wheatstone bridge sensor (strain gauge, load cell). IEPE excitation provides a constant current (typically 2–20 mA) to power the integrated electronics inside an IEPE accelerometer or microphone. They are not interchangeable, using one in place of the other will at best produce no signal and at worst damage the sensor.
Do I need isolation on every channel?
Channel-to-channel isolation matters when sensors are mounted on different equipment with potentially different ground potentials. For sensors all referenced to the same ground (e.g., a single test article on an isolated rig), a common isolated input stage is usually sufficient. Industrial floors with significant ground noise typically benefit from per-channel isolation.
What is shunt calibration and when do I use it?
Shunt calibration places a known resistance in parallel with one arm of a strain bridge, producing a predictable bridge output that simulates a known strain. It’s used to verify the bridge wiring and conditioning end-to-end without applying physical load. Most professional bridge conditioners include a built-in shunt calibration mode.
How do I prevent aliasing in my measurements?
Use an anti-aliasing filter with cutoff well below Nyquist (typically 0.4× sample rate) and steep enough roll-off to attenuate out-of-band content below the system noise floor. Verify the filter is actually applied, programmable conditioners can have filter bypassed during initial setup.
Does Durham Instruments help size conditioning for a complete system?
Yes. Durham Instruments supplies sensors, conditioning, and DAQ as integrated systems and supports specification, installation, and validation. Contact our team for application-specific recommendations.
Specifying signal conditioning?
Send your sensor list and operating environment, Durham Instruments will recommend matched conditioning, isolation, and filter settings.