Written and maintained by CASRAI Editorial Board
Last updated
This page describes products and general use; it is not medical advice.
A data acquisition system (DAQ) is the combination of sensors, measurement hardware and software that turns a physical or electrical signal into numbers a computer can store and analyse. This page is for lab staff, students and research managers who need to understand what a DAQ is made of, what the main architectures are, and which settings belong in the lab record.
What a data acquisition system is
National Instruments (NI) defines data acquisition as “the process of measuring an electrical or physical phenomenon”, with voltage, current, temperature, pressure and sound given as examples. The same page describes a DAQ system as consisting of “sensors, DAQ measurement hardware, and a computer with programmable software”. In a research lab the purpose is to record a signal reproducibly, so it can be analysed, shared and audited later.
Who uses it: physiology, engineering, chemistry and teaching labs that record signals such as force, voltage or temperature. ADInstruments, for example, describes its PowerLab hardware as built to record data for research and education, including analog physiological signals such as ECG, EEG, EMG, EOG, force, galvanic skin response (GSR), non-invasive blood pressure (NIBP) and dP/dt.
How it is built
The signal path has three parts, following the NI description.
- Sensor or transducer: converts the phenomenon (temperature, force, a bioelectric potential) into an electrical signal.
- DAQ measurement hardware: the device that digitises the signal. ADInstruments notes that PowerLab works with compatible signal conditioners and transducers and has independent stimulator outputs, so one unit can both record and drive a stimulus.
- Computer and software: programmable software such as LabVIEW (named by NI) or the vendor’s own recording application controls acquisition, display and storage.
Generic engineering references also describe a signal-conditioning stage (amplification, filtering and similar) between sensor and digitiser, and an anti-aliasing filter ahead of sampling. Which conditioning is needed depends on the sensor, so it is specified in the sensor and hardware documentation for your setup.
Types and architectures
| Architecture | What the vendor describes | Key difference | Source |
|---|---|---|---|
| Multifunction DAQ device | Analog I/O (±10 V), 5 V/TTL digital lines and counters in one device | One device covers analog, digital and counter tasks | NI data acquisition overview |
| CompactDAQ | Modular chassis connecting by USB or Ethernet | Modules are chosen per measurement type | NI data acquisition overview |
| FieldDAQ | Connects by Ethernet | Network-attached, suited to equipment that is not beside the computer | NI data acquisition overview |
| Physiological recording unit (PowerLab) | Records analog physiological signals; works with compatible signal conditioners and transducers; independent stimulator outputs | Built around bioelectric and biomechanical signals; PowerLab C also offers digital acquisition | ADInstruments PowerLab page |
NI lists USB, Ethernet, PCI Express and PXI as the bus options across its DAQ range. The bus affects how the hardware attaches to the computer, not what is being measured.
Specifications that matter: sampling rate, resolution, bandwidth
Sampling rate
NI states that the sample rate fs must be greater than twice the highest frequency component of interest; half of fs is the Nyquist frequency. Sampling below that risks distorting the signal (aliasing).
Resolution
NI explains that bits of resolution is the number of unique vertical levels the digitiser can represent, and that more bits means a more precise measurement. By arithmetic, an n-bit converter has 2n levels: 8 bits gives 256 and 16 bits gives 65,536.
Bandwidth
NI defines bandwidth as the frequency at which a sine-wave input is attenuated to 70.7 percent of its amplitude (the −3 dB point). Signals near or above that frequency are measured with reduced amplitude.
The specific channel counts, sample rates and bit depths of a given product are in its datasheet; the vendor pages cited here do not publish a single set of numbers for the whole range, so none are stated on this page.
How to choose
Only criteria stated in the cited sources are listed here.
- Match the measurement to the hardware: NI’s modular systems are organised by measurement type, and ADInstruments specifies compatible transducers and signal conditioners for PowerLab. Check the sensor is on the vendor’s compatibility list.
- Check sampling rate against your fastest signal component using the greater-than-twice rule above.
- Check resolution against the smallest change you must detect, using the 2n level count.
- Check bandwidth so the −3 dB point sits above the frequencies you care about.
- Check the interface (USB, Ethernet, PCI Express, PXI) is supported by the computer and software you will use.
Safe use and handling basics
Use only the sensors, transducers and signal conditioners the manufacturer lists as compatible, and stay within the input ranges in the hardware documentation (NI describes multifunction devices as ±10 V analog and 5 V/TTL digital). Follow your institution’s electrical and human-subjects rules when recording from people or animals; those rules sit outside this page.
Common mistakes
- Sampling at or below twice the highest frequency of interest, which makes the Nyquist condition fail.
- Choosing by bit depth alone while ignoring bandwidth and conditioning.
- Changing channel, rate or filter settings between sessions without recording the change.
- Saving only processed output and discarding the raw recording and its settings.
What to record in the lab record
Editorial guidance, not a vendor requirement: for each recording, note the channels and sensors, sampling rate, input range, filter settings, hardware and software versions, and the file format. These are the technical details that make a dataset interpretable later, and they belong in the protocol and in the data management plan as technical metadata. They also support reproducibility and the data-integrity expectations summarised in ALCOA+. For the general concept, see metadata.
FAQ
What is a data acquisition system?
It is sensors, measurement hardware and a computer with software that together measure an electrical or physical phenomenon and record it digitally (NI, read 2026-11-08).
What does DAQ stand for?
DAQ is the abbreviation for data acquisition.
What is the minimum sampling rate?
It must be greater than twice the highest frequency component of interest (NI, read 2026-11-08).
How many levels does a 16-bit system resolve?
65,536, since an n-bit converter has 2n levels.
Sources
- NI, Data Acquisition, read 2026-11-08.
- NI, Acquiring an Analog Signal: Bandwidth, Nyquist Sampling Theorem, read 2026-11-08.
- ADInstruments, PowerLab, read 2026-11-08.
