Supplier Checklist: Questions to Ask Before Buying a Triaxial Test Data Acquisition System

Choosing a DAQ for triaxial testing is about more than channel count. The right system must match your sensors, sample cleanly and in sync, fit your software workflow, and scale with your lab. Use this checklist to interrogate suppliers and avoid costly rework.

Is the System Compatible with Your Current Sensors?

Questions to ask

• Which sensor interfaces¹ are natively supported?
  • mV/V bridges (load cells, pressure transducers) with stable bridge excitation (2.5–10 V)?
  • LVDT inputs (AC excitation + demodulation) or guidance for using clean analogue inputs?
  • ±10 V, 0–5 V, 4–20 mA² for conditioned transmitters?
  • Digital: encoder inputs, RS-485/Modbus, CAN, Ethernet?
• What connectors are provided (BNC, 5-pin LVDT, DIN)? Are shielding/strain-relief options standard?
• Does the software support multi-point, up/down calibration (linearity, hysteresis) with certificate storage and expiry alerts?
• Can I store units/scale factors per channel and lock them in a versioned template?

What good looks like

• Native bridge conditioning with low-noise amps and 24-bit ADC for mV/V channels.
• Dedicated LVDT modules or proven AC excitation/demod approach.
• Mixed analogue + digital I/O on the same chassis; clear pinout docs.
• Calibration workflows with shunt-cal, zero/tare, and audit trail.

Evidence to request

• Datasheets showing excitation stability/noise; example calibration reports; a wiring guide for your exact sensors.

What Sampling Speed Does the System Offer?

Questions to ask

• What sampling modes³ are available: simultaneous vs. multiplexed? Is there one hardware clock for all channels?
• What max rates per channel and aggregate? What anti-alias filters⁴ are configurable (cut-off choices, filter type)?
• Typical settings for monotonic CU/CD and cyclic tests?

What good looks like

• Published noise (rms) / ENOB specs; per-channel low-pass filters; shared clock; simultaneous sampling option.

Evidence to request

• Raw data snippets (CSV) from a load cell at rest; filter configuration screenshots; clock/sync topology.

How Well Does the Software Integrate with Lab Workflow?

Questions to ask

• Are calculated channels⁵ built in? ( (q), (p’), ( \varepsilon_a ), ( \varepsilon_v ), B-value, stiffness windows )
• Can I script test stages⁶ (saturation → consolidation → shearing) with interlocks (over-pressure/over-stroke) and stop rules (strain limit, post-peak %, u-excess)?
• Does it support auto-save, checkpoint files, checksums, role-based access, and audit logs (ISO/IEC 17025-friendly)?
• Are exports open (CSV/TDMS/Parquet) with full metadata (operator, calibration IDs, firmware, method version)? LIMS/API available?
• Can I manage versioned templates for channels and methods to prevent unit/scale drift?

What good looks like

• Real-time plots and calculated channels; scriptable stages; robust data integrity; export that drops straight into your reports/LIMS.

Evidence to request

• A demo project file: method + channel template + sample dataset; proof of API calls or LIMS connector.

Can the System Scale with Future Needs?

Questions to ask

• How do I add more triaxial cells, a cyclic module, or extra sensors—modular slots⁷ or networked DAQs?
• Product lifecycle: parts/firmware support for 5–10 years? Field-replaceable modules?
• Can the system run multiple rigs in parallel without bottlenecks (CPU/network/disk)?
• What are the service, calibration, and warranty terms and pricing?

What good looks like

• Modular chassis or stackable Ethernet DAQs; spare analogue/digital channels; clear upgrade path.
• Storage plan (local + mirrored + NAS archive) and configuration backup tools.

Evidence to request

• A scalability diagram, firmware roadmap, example multi-rig deployment, and service menu with SLAs.

Supplier Q&A Table (fill during calls)

Red Flags (walk away if you see these)

• No native bridge excitation for load cells; only generic ±10 V inputs.
• Channels on different clocks with no hardware sync option.
• Closed, proprietary data formats with limited export.
• No audit logs or checkpointed saving for long tests.
• Vague lifecycle/support answers; unclear calibration workflow.

One-Page Printable Checklist

• [ ] mV/V, LVDT, ±10 V/4–20 mA, and digital interfaces supported natively
• [ ] Stable bridge excitation; 18–24-bit ADC; published noise/ENOB
• [ ] Single hardware clock; simultaneous sampling where needed; configurable low-pass filters
• [ ] Calculated channels ( (q, p’, \varepsilon_a, \varepsilon_v, B) ); scripted stages; interlocks and stop rules
• [ ] Auto-save, checkpoints, checksums, audit trails; role-based access
• [ ] Open exports (CSV/TDMS/Parquet) with full metadata; LIMS/API integration
• [ ] Multi-point calibration; certificate storage; expiry reminders
• [ ] Modular/network expansion; multi-rig performance; 5–10 yr support plan
• [ ] Clear service/warranty pricing; configuration backup/restore tools

Bottom Line

Ask suppliers to prove fit across four pillars: sensor compatibility, clean/synchronised sampling, workflow-ready software, and real scalability. If they can show raw noise data, method templates, and a growth path you trust, you’ll buy once, test faster, and produce triaxial results that stand up to audits and design reviews.