Choosing the right DMM for testing applications requires a closer look at its measurement capabilities and characteristics.


click to enlarge
Figure 1. 10 steps to selecting a DMM. Source: Keithley Instruments Inc.   

Selecting the most appropriate digital multimeter (DMM) for a specific application requires analyzing both the device under test (DUT) and the test environment, then answering ten key questions on measurement functions and performance requirements (Figure 1).

The device under test
First, analyze the DUT and required measurements. A typical five-function DMM measures direct current (DC) and alternating current (AC) volts, DC and AC currents, and two-wire resistances (ohms). However, many available instruments include four-wire resistance, temperature, frequency, period, and capacitance, as well as diode and continuity test functions, without a higher price tag.

The DMM's measurement range must span the expected minimum and maximum test signals. At the low end, the DC noise specification is critical. If the noise floor is higher than the signal to be measured, it will be necessary either to slow down measurements to employ noise reduction techniques or choose a DMM with a lower range and noise. At the high end, the maximum allowable input might include specifications for a volt-hertz product limitation or a time limit on AC signals.

Three factors define DMM performance more than any others. Accuracy is a measurement's degree of conformity to a standard or true value, usually specified as a percent of reading plus a percent of range (or a number of counts of the least significant digit). It may also be specified in parts per million (ppm). Accuracy is also specified over a period of time—for example: 24 hours, 90 days, or 1 year—and within a temperature window—typically 1 C or 5 C around room temperature, with deratings for temperatures outside those ranges.

Resolution is the smallest portion of a signal that can be measured or displayed, for example, one digit out of 20,000 on a 4-1/2-digit display. The resolution of the display is the ratio of the smallest count to the maximum count.

Sensitivity is the smallest change of the measured signal the instrument can detect. It depends on both resolution and the lowest measurement range of the instrument. For example, the sensitivity of a 5-1/2-digit DMM on the 200 mV range is 1 mV.

For testing high-impedance devices, a DMM might not be the optimum solution. DMMs generally have 10 mega-ohms to 10 giga-ohms input resistance. If the DUT is susceptible to device self-heating when measuring its resistance, the ability to control the level of test current on each range is important. For contact-resistance measurements, dry circuit testing may be necessary, which limits test voltage to prevent disturbing the oxidation on the contacts.

The test environment
Although DMMs are still widely used in benchtop applications, a variety of interconnect options are available for system integration. For example, Keithley's Model 2110 DMM (Figure 2) offers engineers a USB 2.0 interface and an optional general purpose interface bus (GPIB) for external personal computer control via Standard Commands for Programmable Instruments (SCPI) test commands. LabVIEW (National Instruments, Austin, Texas), IVI-C, and IVI-COM drivers enable integration of the DMM into test systems. Front panel thermocouple inputs facilitate connections of popular temperature measurement sensors.

A DMM's specifications for normal mode rejection ratio (NMRR) and common mode rejection ratio (CMRR) indicate its noise susceptibility. Typical values for NMRR and CMRR are greater than 60 dB and greater than 120 dB, respectively. The amount of noise that emanates for the DMM itself is especially critical when measuring low signal levels. Common EMI standards, including the European Union Directive 89/336/EEC, FCC part 15 class B, and IEC 801-2 will specify compliance levels.

Measurement speed is critical in production environments because it's a major determinant of throughput. It's usually stated as readings per second for a specific level of resolution. A variety of other factors can also affect throughput.

Many DMMs today have programmable resolution levels, which simplifies tradeoffs between speed and accuracy. The integration period—the window of time in which a signal is sampled by the A/D converter—is often expressed in number of power line cycles (NPLCs). NPLCs that are integer multiples (1, 5, 10, etc.) can reduce the most common type of normal mode noise, that of 50/60 Hz line pickup. The larger the value of N, the greater the line noise reduction, but the longer the measurement takes to complete.

General-purpose DMMs, like Keithley's Model 2110 Dual-Display Digital Multimeter, typically offer the flexibility needed to handle a broad variety of test and measurement applications. Image: Keithley Instruments Inc.Digital filtering—or the number of A/D conversions averaged for each reading—stabilizes noisy readings, but slows measurements.

When autozero is activated, the DMM measures internal voltages to maintain stability and accuracy as temperature changes over time; but doing so too frequently will affect the reading rate. It may be possible to disable this function, just perform it periodically, or program the DMM to do so only during the test fixture's unload/load cycle to increase throughput.

Other speed specifications that can affect system throughput include the range changing speed, autorange time, and function changing rate. Trigger latency can also affect system throughput. Hardware triggers are typically faster than software triggers. Some DMMs have a special microprocessor just for triggers that can shorten trigger latency considerably. Settling time (or response time) is another speed specification to consider, especially when testing high-impedance devices.

Fast, accurate switching is particularly crucial in a production test environment where hundreds or thousands of devices must be tested each work shift. A growing number of system builders are turning to DMMs that offer the simplicity of a built-in switch mainframe. Newer models can support hundreds of multiplexer channels or thousands of matrix crosspoints.

Other useful features
Some functions and features are helpful when building larger systems:

  • Faster and simpler hardware triggering to reduce testing time.
  • An internal data buffer.
  • Measurement limit testing and digital I/O for binning DUTs without controller intervention.
  • Additional display capabilities. Some new DMMs can show the results from two measurements simultaneously, such as DC voltage and temperature.
  • Built-in mathematical functions such as percentage, average, min/max, null, limits, mX+B, dB, and dBm.

Total cost of ownership
A DMM's overall cost of ownership can be much higher than the purchase price. Startup costs, including programming, system integration, and training can be significant. Downtime and shipping costs for recalibration processes, availability of spare units, and warranties should be calculated in the total costs.

Modern DMMs offer a wide range of accuracy, resolution, speed, and special features, and can often prove to be the most cost-effective solutions. However, for low-level signals or high-source resistances, more specialized instruments such as nanovoltmeters, picoammeters, electrometers, or source measurement units may be better choices.