Technical Insights > Benchtop

How to Reduce DAQ Measurement Errors

2019-01-17  |  7 min read 

A data acquisition (DAQ) system is normally used to automate the characterization, monitoring, and control of your test process. DAQ systems have come a long way with years of improvements. For example,

  • DAQ systems can measure multiple types of signal inputs such as temperature, voltage, current, resistance, capacitance, diode, signal frequency and more.
  • DAQ systems can take periodic scan measurements of a large number of channels. It can also provide real-time monitoring of scanned data on instrument display or on PC.
  • DAQ systems can input and output control signals to control device under test or to control other instruments or environmental chambers.

To really make the most out of your DAQ system, you will want to reduce DAQ measurement errors during your wiring and configurations. The following are the steps in reducing DAQ measurement errors.

Steps in reducing measurement errors

1) Group all similar measurement functions together

It is a good practice to group all similar measurement functions together. For example, group all the thermocouple measurements together, group all the DC voltage inputs together, group all the resistance inputs together and so on. By doing so, when the DAQ system is scanning, this minimizes function changes. Organization by measurement functions generally helps reduce configuration errors.

It also helps to reduce mistakes while doing configuration setups.

2) Match and grouping all similar input levels together

While grouping all similar measurement functions together, try to also group measurements with the same input levels or the same range together. For example, grouping all DC voltage inputs with +/-10 V range together, this helps minimize time-consuming range changes.  Grouping voltages also reduce parasitic measurement errors. Low voltage measurements can be affected by large voltages charging parasitic capacitance.

3) Grouping of industrial sensors together

Industrial process control sensors have a standard analog current loop output of 4-20 mA representing 0-100% of the range of measurement or control. You can standardize by terminating all the 4-20 mA current loops with 250 Ω shunt resistor, converting the 4-20 mA current loop into a voltage measurement range of 1-5 V. In the DAQ configuration, you will only need to select DC voltage measurement with +/- 10 V range. See Figure 1.

Sensor output conversion
Figure 1. Process control sensor – current loop measurement

4) Need for isolated inputs

Mixed-signals sharing the same ground or cables bunch together can create ground noise and cross-talk in measurements. Some DAQ systems provide isolation. What does isolation do? Isolation breaks ground loops, reduce noise and delivers better measurement accuracy. It is a good practice to separate digital and analog grounds. Some instrument will indicate the type and extent of built-in isolation.  Isolation topologies include channel-to-channel or channel-to-earth.

For example, Keysight’s DAQ 34970A/34972A and DAQ970A have a channel-to-channel and channel-to-earth isolation of 300 V.  This means that the “low” terminals of the differential inputs can float up to 300 V from earth ground.

Isolation circuit
Figure 2. Isolation inputs that float up to 300 V

Isolation also provides safety, which is very important.  It provides a barrier between the user and the equipment power supply.

5) Need for a shielded and twisted pair of cables

When bundling mixed-signal wires together, cross-talk between cables will happen. It is worse if the wrong cables are used and the connections are not made properly. It is especially critical if we have noisy digital lines bundled together with very sensitive low voltage analog signal lines.

One way to reduce noise from our system is always to use a twisted pair shielded cable as shown in Figure 3.

Shielded and twisted pair cable
Figure 3. Example of shielded and twisted pair cables

6) Need for Normal Mode Rejection (NMR)

Another key factor causing measurement errors and prevalent while doing mixed-signal measurements is the normal mode noise. This noise is caused by power-line and its harmonics. Not all DAQ systems readily have normal mode rejection capability. It requires built-in integrating analog to digital (A/D) converter that averages out the unwanted AC components, including its harmonics to zero.

Power Line Noise
Figure 4. Power-line noise creeps into measurement path

Summary

DAQ systems can be very versatile for the characterization, monitoring, and control of your test process. To make the most out of your DAQ system, you can reduce measurement errors by taking simple steps during your setup and configuration. Follow them, and it will benefit you for a long time to come.

Besides this blog’s tips to reduce measurement errors, Keysight has just introduced the next generation DAQ970A data acquisition system. It has capabilities many have become accustomed to with the 34970A/34972A such as built-in 6 ½ digit DMM and signal conditioning (no external box required), capable of measuring multiple types of signals (temperature, voltage, current, resistance, diode, capacitance, frequency) and universal inputs allowing you to wire any signal to any channel.

The DAQ970A has speed and accuracy enhancements over the 34970A/34972A in:

  • Get up to 2X scan rate speed improvement (up to 450 channel/second)
  • Get up to 10X reading rates improvement to I/O interface & memory
  • Get up to 90% accuracy improvements across voltage, current, resistance and temperature measurements

DAQ970A data acquisition system
Figure 5. The next generation DAQ970A with its interface modules

For more information about the Keysight’s DAQ970A system, please go to http://www.keysight.com/find/daq970a

To learn more about DAQ systems, please read our eBook on Four Things to Consider when using a DAQ as a Data Logger.