DC Voltage Measurements

This tutorial recommends tips and techniques for using a National Instruments digital multimeter (DMM) to perform accurate DC voltage measurements.

For more information return to the Complete Digital Multimeter Measurement Tutorial.

1. Overview
2. Selecting Input Resistance
3. DC Noise Rejection
4. Handling High DC Voltages
5. Optimizing Low-Voltage Measurements
6. Offset Nulling
7. Selecting Aperture

Overview

The NI 4070 Digital Multimeter can measure signals in the sub-microvolt range. To take good low-voltage measurements with confidence, consider the following factors: thermal voltages, selectable input resistance (ohms), and input bias current (amps).

Selectable Input Resistance

You can select the NI 4070 Digital Multimeter DCV input resistance (ohms) to optimize measurement performance in situations where large source resistance (ohms) is present. In the following figure, S₁ is used to select a 10 MΩ input resistance (ohms) on the 0.1 V, 1 V, and 10 V DC ranges.

The following figure shows an equivalent circuit after the 10 MΩ input resistance (ohms) has been selected:

The following figure shows an equivalent circuit after the 10 GΩ input resistance (ohms) has been selected:

The following table lists the available NI 4070 Digital Multimeter input resistance (ohms) selections. The default input resistance (ohms) for each range is italicized. NI calibrates the module for full accuracy with all input resistances shown.

DCV		ACV
Range	Input Resistance	Range	Input Resistance
100 mV	10 MΩ or 10 GΩ	50 mV	1 MΩ only
1 V	10 MΩ or 10 GΩ	500 mV	1 MΩ only
10 V	10 MΩ or 10 GΩ	5 V	1 MΩ only
100 V	10 MΩ only	50 V	1 MΩ only
300 V	10 MΩ only	300 V	1 MΩ only

Input Bias Current

If the source resistance (ohms) is greater than 10 kΩ, then you may need to include the effects of input bias current (amps) in addition to input resistance (ohms). For example, for 100 kΩ source resistance (ohms), a 100 pA bias current (amps) creates an offset error of 10 µV as shown in the following figure:

With an input resistance (ohms) of >10 GΩ selected, most of the input bias current (amps) flows through the source resistance (ohms) (100 kΩ), causing an offset voltage.

Digital multimeter input bias current (amps) approximately doubles for every 10 ºC of environmental temperature change. For example, an input bias current (amps) of 25 pA at 23 ºC translates to about 50 pA at 33 ºC, 100 pA at 43 ºC, and so on. Likewise, as temperature is decreased, the current (amps) drops 50% for every 10 ºC. Thus, at 13 ºC, the input bias current (amps) is approximately 12.5 pA.

Selecting Input Resistance

Refer to the following table for available input resistance values for DC voltage measurements and voltage waveform acquisitions. The default input resistance (ohms) for each range is italicized.

Digital Multimeter Measurements

Waveform Acquisitions

To select the input resistance (ohms):

• LabVIEW - use the niDMM property node to set the Input Resistance attribute.
• CVI, C++, or Visual Basic - use niDMM_SetAttributeViReal64 to set the desired value for the NIDMM_ATTR_INPUT_RESISTANCE attribute.

DC Noise Rejection

DC noise rejection is a configurable noise reduction feature available for NI 4070 DC measurements. Each DC reading the NI 4070 Digital Multimeter returns is actually an average of multiple high-speed samples. By adjusting the relative weighting of those samples, you can adjust the sensitivity to different interfering frequencies. The NI 4070 Digital Multimeter offers three different weightings: normal, second-order, and high-order. Refer to the following table for a list of the differences between the three modes:

Refer to the Aperture Time table for the default aperture times and DC noise rejection settings used by the digital multimeter.

Normal

When you select normal DC noise rejection, the NI 4070 Digital Multimeter weights all samples equally. This feature emulates the behavior of most traditional digital multimeters, providing good rejection of frequencies at multiples of f0, where f0 = 1/t_aperture. The following figure shows normal weighting, where all samples are weighted equally, and the resulting noise rejection as a function of frequency. Notice that you can obtain good rejection only very near multiples of f0. To get the fastest possible readings that give you some powerline noise rejection, set the aperture to the powerline period, such as 16.667 msec for a 60 Hz powerline frequency, and set the DC noise rejection to normal.

Second-Order

Second-order DC noise rejection applies a triangular weighting to the measurement samples, as shown in the following figure. In second-order DC noise rejection, the samples taken in the middle of the aperture time are weighted more than samples taken at the beginning and the end of the measurement.

Notice that you can obtain very good rejection near even multiples of f0 and that rejection increases more rapidly with frequency than with normal sample weighting. Also notice that the response notches are wider than they are with normal weighting, resulting in less sensitivity to slight variations in noise frequency. Use second-order DC noise rejection if you need better powerline noise rejection than you can get with normal DC noise rejection, but cannot afford to read slowly enough to take advantage of high-order noise rejection. For example, you could set the aperture to 33.333 msec for a 60 Hz powerline frequency.

High-Order

The following figure shows high-order sample weighting and its resulting noise rejection as a function of frequency:

Notice that noise rejection is good starting at around four times f0 and is excellent above about 4.5 times f0. Sensitivity to noise at any frequency above 4.6 times f0 almost does not exist. An NI 4070 Digital Multimeter, using high-order DC noise rejection with a 100 msec aperture (10 readings/sec), can deliver full 6½ digit accuracy with over 1 V of interfering powerline noise on the 10 V range at any frequency above 46 Hz.

The small tradeoff to using the higher-order filters is that to obtain power line rejection, you must increase the measurement aperture by 2x (for second-order) or 4x (for high-order) versus that required for the normal setting. The increase in aperture time is typically minimal given the reduction in the interfering signal.

Handling High DC Voltages

The NI 4070 Digital Multimeter input can measure sub-µV level signals with low drift, handle 300 VDC and VAC signals, and recover very quickly for sub-µV level measurements in a system with switching. This ability is achieved through a flexible input protection design. Once the input exceeds about 30 V on the 10 V and lower ranges, the input protection triggers, and the input changes from a resistive characteristic to a constant current (amps). This feature has the secondary benefit of reducing self-heating, so the digital multimeter can recover rapidly to measure low-level signals with minimum recovery time.

However, you still need to consider that in a switching system where you are applying high voltages to a low range (for example, 300 V on the 10 V range), the switch receives a very short–duration current (amps) transient. The current transient does not cause any degradation effect on the NI 4070 Digital Multimeter. However, if the system switches an extreme overload to the digital multimeter input mode with a moderate (tens of Hz) to high scanning rate over extended periods (weeks continuous or cumulative), the current transient can degrade the relay life in the switching system. For optimum switching system reliability when switching high voltages into any digital multimeter, NI recommends selecting the range corresponding to the highest voltage expected.

High Voltages and System Settling

Consider the following illustration showing a digital multimeter connected to a 100 V DC source, a thermocouple, and device under test (DUT) through a multiplexing switch.

Assume you are switching from the 100 V range to the 100 mV range to obtain the best sensitivity using the thermocouple. When the switch module advances from the 100 V channel to a thermocouple channel, the input signal conditioning of the digital multimeter (along with the cabling, switching module, and so on) must discharge any stray capacitance from 100 V to 100 µV. Assuming you need a sensitivity of 1 uV from the low–voltage channel, you are asking the system to settle from 100 V to 1 µV (10 parts per billion, or eight decades).

When you desire resolutions better than 10 ppm, dielectric absorption is a major factor.

Depending on the source resistance R, system capacitance C, and dielectric qualities of the system components, you need to allow more settling time in this example. In practice, you may need to allow settling times of 10 ms or more in situations where you are switching high voltages (>30 V).

Caution  All cables, connectors, and fixtures used in your system must have specifications to handle 300 V signals with an adequate safety margin.

Optimizing Low-Voltage Measurements

As discussed in the Thermal Voltages section, pairs of dissimilar metal junctions—rather than a single junction—potentially can create problems. The NI 4070 Digital Multimeter uses proprietary gold-plated copper input leads (HI and LO). If you just connect these two jacks directly with some wire that is not copper, then you have two junctions of dissimilar metals. This basic concept can be extended to just about any other setup.
If the two jacks are at different temperatures, then a voltage is generated. The amplitude of the voltage depends on the other metal that is being used. If the other metal is copper, then the Seebeck coefficient for copper to copper is very low—well under 100 nV/°C—as long as the connections are clean and tight.

A temperature difference between the two jacks greater than 1–2 °C would be rare, but a temperature difference does exist. The interior of the PXI chassis is warmer than the outside air creating heat flow that ultimately results in a temperature difference.

For copper–to–copper connections, the generated voltage is relatively small. If you used brass instead of copper where a 1–2 °C temperature difference exists, there could be a 6 µV difference in your measurement—a value detectable by the digital multimeter.

One rule for low-voltage measurements is to always use metals with Seebeck coefficients close to that of copper. Copper, gold, and silver are all good candidates. Typical Seebeck coefficients for common conductors are listed in the table in the Thermal Voltages section.

If the connections do not change temperatures, then the thermal voltage is stable and can be corrected; it is temperature changes between these junctions that create problems (such as offset drift, instability, and very low frequency noise). The key to preventing changes in temperature is to prevent circulating air currents (amps) that can disturb the thermal equilibrium of the junctions. This leads to another rule for low-voltage measurements—keep junctions and connections at a stable temperature and away from circulating air currents (amps) caused by movement, fans, and so on. You can prevent these temperature differences by creating a “thermal baffle,” such as wrapping the junctions with common foam padding (even Styrofoam sheet can work) and keeping them away from sources of heat such as equipment heat sinks and sunlight.

Even common lead-tin solder connections can create thermal offsets of (1-3 µV/°C). While not insignificant, they can be managed using the techniques described above. For increased performance, cold weld connections by tightly twisting clean copper-to-copper connections together and then using either crimping or twist-on plastic connectors.

When copper oxidizes, the Seebeck coefficient can easily increase by several hundred µV/°C. This leads to another rule for low–voltage measurements—keep the connections clean. One option is to simply use a common pencil eraser to clean the bare wire until the wire is shiny, then clean off any rubber fragments with a paper towel. Another way to clean connections is to use Scotchbrite pads to clean the wire. After cleaning the connections, do not handle the connections with your fingers. Skin oil contains a very effective corrosive that accelerates oxidation of many metals.

When setting up systems using relay switching, keep in mind that low-voltage performance is largely limited by the switches selected and the input terminals used on the switch. Good low–thermal relay switching cards are available from NI. Common, uncompensated reed relays can reach tens of µV of rather unstable offset, but high–performance reeds can reach below 1 µV. Many armature-style latching reed relays can give very low drift (one such device is used in the NI 4070 Digital Multimeter signal path). It is sometimes possible to compensate for this drift by using similar relays in both HI and LO connections, or by placing adjacent poles of a 2-pole relay back-to-back. Lastly, a common metrology technique involves making a measurement, reversing the leads, taking another measurement and then subtracting and averaging the two readings.

Offset Nulling

To perform offset nulling, complete the following steps:

The subtraction operation can be performed programmatically with your ADE, and the DMM-SFP (soft front panel) has an offset nulling feature.

Selecting Aperture

The aperture time shapes the noise rejection of the measurement. Shorter aperture time yields a wider measurement bandwidth at the expense of resolution. Sensitivity to source noise increases with shorter apertures. Conversely, larger aperture times quiet noisy sources and yield higher resolution measurements. Refer to Noise for more information.

Selecting aperture times equal to multiples of power line frequency helps to reject these frequencies, when DC Noise Rejection is set to Normal. An aperture set to 1 powerline cycles (1 PLC equals 16.67 ms for 60 Hz powerline or 20 ms for 50 Hz) is the minimum that provides this line rejection.

Apertures of >100 ms can be used for measuring very high resistance (ohms) values, when sensitivity to powerline frequency noise pickup is inevitable. Coupling the aperture time with DC Noise Rejection set to High Order provides extremely high noise rejection (>100 dB). Long apertures of >100 ms do not significantly improve resolution beyond 6½ digits but can provide better rejection to externally generated noise.

For voltage measurements with 7-digits resolution, average 4 or more 6½-digit AutoZeroed measurements. You can average multiple measurements by setting the Number of Averages attribute to a value greater than one. Averaging with AutoZero enabled circumvents the effects of low–frequency, self-generated noise and yields the highest resolution performance level. More averages also reduces external noise, but the tradeoff is speed.

Selecting apertures for AC is discussed in the AC Voltage Measurements section.

26 Ratings | 3.42 out of 5

Read in  English Japanese |  Print |  PDF

Reader Comments | Submit a comment »