Choosing the best USB DAQ under $1,000

LabJack, MCC, NI and Digilent sell USB DAQ devices from the $204 MCC USB-1208FS-Plus to the NI USB-6003 (from $928), and they differ most in input ranges and OS support.

Britt Espinosa··6 min read

This comparison covers four vendors found on many lab benches: LabJack, Measurement Computing (MCC, now sold through Digilent), NI and Digilent itself. Their devices differ less in headline specs than in input ranges, driver support across operating systems and the software you are expected to use. Prices are US list prices from each vendor’s own store at the time of writing, and they change, so check before ordering.

Specs and list prices

DeviceAnalog inputsResolutionMax AI rateInput rangesAnalog outList price
LabJack T44 high-voltage + 8 flexible low-voltage12-bit50 kS/s shared±10 V (AIN0–3), 0–2.5 V (flex)2, 0–5 V$305
LabJack T71416-bit (T7-Pro adds a 24-bit low-speed ADC)100 kS/s total±10, ±1, ±0.1, ±0.01 V2, 12-bit, 0–5 V$570 (T7-Pro $850)
MCC USB-1208FS-Plus8 SE / 4 DIFF12-bit50 kS/s±10 V SE; ±20 V to ±1 V DIFF2, 12-bit$204
MCC USB-1408FS-Plus8 SE / 4 DIFF14-bit48 kS/s±10 V SE; ±20 V to ±1 V DIFF2, 12-bitSee Digilent store
MCC USB-1608G16 SE / 8 DIFF16-bit250 kS/s±10, ±5, ±2, ±1 VNone (GX-2AO: 2)$479
MCC USB-2058 SE12-bit500 kS/s±10 V2See Digilent store
NI USB-60018 SE / 4 DIFF14-bit20 kS/s±10 V2, 16-bit, 5 kS/s/chFrom $353
NI USB-60028 SE / 4 DIFF16-bit50 kS/s±10 V2, 5 kS/s/ch$715
NI USB-60038 SE / 4 DIFF16-bit100 kS/s±10 V2, 16-bit, 5 kS/s/chFrom $928
Digilent Analog Discovery 32 differential scope channels14-bit125 MS/s per channel±25 V2 AWG, 14-bit, ±5 V$379

Specs come from the vendor pages: LabJack, Digilent’s MCC listings for the USB-1208FS-Plus series and USB-1608G series, NI’s USB-6001, USB-6002 and USB-6003 pages, and Digilent’s Analog Discovery 3 page. NI’s “from” prices cover more than one part number, including OEM board-only kits. Use the device finder to filter these and other models by channel count, rate and supported software.

Specs the table leaves out

Aggregate versus per-channel rate. Every device here except the Analog Discovery 3 multiplexes one ADC across its inputs, so the headline rate is shared. A USB-1608G at 250 kS/s gives 31.25 kS/s per channel with all 8 differential channels scanning; a LabJack T7 at 100 kS/s total gives 10 kS/s each across 10 channels.

Resolution versus accuracy. Bits tell you the step size, not how close the reading is to the truth. NI lists an absolute accuracy of 26 mV on the ±10 V range for the 14-bit USB-6001 and also for the 16-bit USB-6003, even though the USB-6003’s code width is about 0.3 mV. For small signals, an input range close to the signal size helps more than extra bits, so weigh the T7’s ±0.1 V and ±0.01 V ranges and the MCC differential ranges down to ±1 V. The details are in ADC resolution vs accuracy.

Single-ended versus differential. The MCC USB-200 series is single-ended only. Single-ended inputs are fine for signals that share ground with the DAQ. Sensors with long leads, separate supplies or small outputs read better differentially. See single-ended vs differential inputs for how to tell which you need.

Current loop sensors. None of these devices reads 4–20 mA directly. A precision shunt resistor turns the loop current into a voltage, for example 250 Ω gives 1–5 V, which any of them can measure. The 4–20 mA calculator converts between current, voltage and engineering units, and reading 4–20 mA loops with a DAQ covers wiring and resistor choice.

Discontinued and retired models

NI lists the USB-6008 and USB-6009 as Mature, not recommended for new designs, with the USB-6001 and USB-6002 as their recommended replacements, and at least some USB-6008 part numbers are no longer available to buy. Digilent marks the MCC USB-201 as retired and suggests the USB-202, and the 500 kS/s USB-1608GX is retired as well, leaving the USB-1608G and USB-1608GX-2AO in that series. NI still offers support, repair and calibration for the USB-6008, so an inherited unit remains usable.

Driver and OS support by vendor

If you don’t work on Windows, driver support narrows the list more than any spec in the table.

  • LabJack: the LJM library has installers for Windows, macOS (Intel and Apple Silicon) and Linux, and the same library is used from Python, C, MATLAB and LabVIEW. LabJack’s simple loggers, LJLogM and LJStreamM, are Windows only.
  • MCC: the Universal Library and InstaCal cover Windows. On Linux and macOS, MCC’s open-source uldaq library supports every MCC device in the table, plus the USB-TC and USB-TEMP. DAQami, MCC’s logging app, is Windows only.
  • NI: NI-DAQmx runs on Windows and on supported 64-bit Linux distributions, without the MAX configuration utility on Linux. Check NI’s compatibility page for your exact device before planning a Linux setup. There is no current macOS driver; see NI DAQ on a Mac for workarounds.
  • Digilent: WaveForms and its SDK run on Windows, macOS and Linux.

Recommendations by use case

Temperature logging

Thermocouples produce tens of microvolts per degree, so the useful specs are a small input range, cold junction compensation and noise, not sample rate. On a device with only a ±10 V range, such as the USB-6001, the input spans 20 V with a 14-bit ADC, about 1.2 mV per code, which is tens of degrees for a type K thermocouple. Use the ADC resolution calculator to run that math for any device.

A dedicated module does better, such as the MCC USB-TC with 8 differential channels, a 24-bit converter and types J, K, R, S, T, N, E and B at 2 S/s per channel, or a LabJack T7 on its ±0.01 V range, where the T7-Pro’s 24-bit ADC helps further. The thermocouple measurement guide covers cold junction compensation and wiring, and the thermocouple calculator converts between millivolts and temperature.

Vibration and audio-band signals

To capture content up to a few kilohertz without aliasing, you need a per-channel rate well above twice the highest frequency, plus an anti-alias plan. The sample rate and aliasing guide explains why, and the sample rate calculator gives numbers. In this price range the USB-1608G has the most headroom, with 250 kS/s aggregate at 16-bit, enough for four channels at 50 kS/s each. The LabJack T7’s 100 kS/s total is shared across channels, which suits one or two vibration channels. For one or two fast channels, an Analog Discovery 3 samples far faster, at 14-bit.

None of these devices power IEPE accelerometers or microphones directly. You need an external signal conditioner, or a dedicated IEPE device such as the NI 9234, a C Series module that also needs a cDAQ chassis.

General bench multifunction work

For a mix of voltages, a couple of control outputs and some digital lines, the LabJack T7 covers the most ground, with 14 inputs on four ranges, Ethernet as well as USB, onboard Lua scripting and drivers for Windows, macOS and Linux. On a tighter budget, the USB-1208FS-Plus at $204 gives you 8 inputs, 2 outputs and 16 digital lines with differential ranges down to ±1 V. If your lab standardizes on NI software, the USB-6001 or USB-6002 keeps you in NI-DAQmx, LabVIEW and FlexLogger.

Education and teaching labs

Teaching labs need low prices and software students can install on their own laptops. The Analog Discovery 3 at $379 covers the scope, function generator and logic analyzer for circuits courses, and WaveForms runs on every student OS. The LabJack T4 at $305 suits measurement and control labs, with ±10 V inputs and Python support on all three platforms. NI’s USB-6000 (from $289, 12-bit, 10 kS/s, no analog outputs) or the USB-6001 fits courses already built on LabVIEW.

Software to run them with

Each vendor ships free software for basic logging, mostly Windows only, and every device here also has a Python library for scripted tests. Mixing vendors on one bench, or using these devices from a Mac, is where the options thin out. DAQAtlas, which is in development, records MCC, LabJack T4/T7/T8 and Digilent Analog Discovery devices on macOS, Windows and Linux, and NI-DAQmx hardware through a bridge on a Windows or Linux PC. For the software side in detail, see data logging software compared and Python data acquisition.