Understanding

specifications for
precision multimeters
Application Note

Digital multimeters are the work- you must be confident that your From this formal definition we
horses of the electronics industry. readings accurately reflect reality. can draw the characteristics of
Almost all of the electronic This application note discusses good specifications:
products we use in our personal some of the thinking behind DMM • Completeness – any factor
and professional lives are built or specifications and spec sheets. It that impacts uncertainty is cov-
serviced using multimeters. defines the various elements of ered, including operating limits
Different DMM applications DMM specs and gives tips on how such as humidity, altitude, or
require different degrees of to apply them. vibration
attention to specs. A technician • Clarity – all efforts should be
checking a logic power supply Specifications and the made to make the specifica-
can use his trusty bench DMM spec sheet tions straightforward
confidently to verify that the • Objectivity – does not attempt
Whenever we take a measure-
supply is within a few percent to mislead for the sake of
ment with any meter, we’re
of 5 V. However, when the job promotion
gambling. We’re gambling that
requires testing critical circuits, A well-written specification
the instrument is going to give
checking precision components, should maintain the same level
us the “real” reading. Fortunately,
making fine adjustments in of integrity as a medical chart or
it’s a very safe bet that a quality
production, verifying compli- bank statement. Manufacturers
multimeter will deliver readings
ance with industry standards, must stand firmly behind their
that coincide with reality. Specifi-
or taking measurements outside specs and you should fully expect
cations quantify the confidence of
the controlled environment of the that the information you are get-
getting accurate readings and the
lab, then you’ll need to evaluate ting is accurate and complete.
risk of seeing inaccurate readings.
specifications carefully. On the spec sheet, you should
A specifications document is
A solid understanding of see measurement uncertainty
a clearly written description of
specifications is critical when specs and modifiers that affect
an instrument’s performance.
you’re evaluating the suitability of the uncertainty. You will also see
It should quantify an instru-
DMMs for an application, or when operating limits that constrain
ment’s capabilities objectively
under well-defined operating the environment in which the
conditions. uncertainty specifications will
hold true. These are stated in
numerical values (e.g. humidity)
or with reference to international
standards (shock and vibration).
First let’s take a closer look at
how we quantify measurement
uncertainty.

Where do uncertainty
specifications come from?
The main job of a DMM
specification is to establish the
measurement uncertainty for any
input in the instrument’s range.
The spec answers the question,
“How close is the value on the
meter display likely to be to the
actual input to the meter?”

F ro m t h e F l u k e D i g i t a l L i b r a r y @ w w w. f l u k e . c o m / l i b r a r y
 Meter manufacturers bet their normal distribution curve cen-
reputations on how a large popu- tered at 10 V.
lation of instruments is going to Using experimentation and
behave for the duration of calibra- experience, instrument designers
tion cycle. (A typical calibration set specifications by assuming a
cycle is one year.) Instrument normal distribution and find-
engineers and metrologists use ing the standard deviation for
laboratory testing and carefully a significant number of design
applied statistics to set the specs. samples. Adopting a normal
DMM specifications apply to a distribution allows us to relate
particular model (i.e. design), not standard deviation to the per-
to any individual instrument. centage of readings that occur, 9.9200 9.9400 9.9600 9.9800 10.0000 10.0200 10.0400 10.0600 10.0800
Any single instrument of a par- by measuring the area under the
ticular design should perform well curve. Figure 1: A normal distribution with a mean of 10 volts and standard
within the specification, espe- deviation 0.02 volts.
68 % of the readings will be
cially toward the beginning of its within 1 standard deviation of
calibration cycle. A model’s specs the mean
are based on testing a significant
sample of products and analyz- 95 % of the readings will fall discussed how we make sure
ing the collected data from the within 2 standard deviations of we’re all talking about the same
instruments. the mean volt, ohm or amp. DMMs must
If we take measurements of trace their measurement perfor-
99.7 % of the readings will fall
a nominal input from, say, 50 mance back to national laboratory
within 3 standard deviations of
instruments of the same design, standards.
the mean
we are going to get a range of DMMs are usually calibrated
readings. Many of the instruments Statisticians refer to these per- using multifunction calibrators
will have the same readings, but centages as confidence intervals. like the Fluke 5700A or Fluke
we would expect some varia- They might say, “We are 95 % 9100. But there are usually a
tion due to normal uncertainty. confident that a reading will not number of links between the
For example, we can record the be more than 2 standard devia- DMM and national standards,
readings from 50 Fluke Model tions of the actual value.” including calibrators and transfer
xyz DMM’s hooked up to the same In the simple example above standards. As you move through
precision calibrator outputting the chain between your DMM
10 volts. We will record a narrow 1 standard deviation corresponds and the national standards lab,
spread of readings around 10 to ± 0.02 V the calibration standards become
volts. We can calculate the mean 2 standard deviations corresponds increasingly accurate. Each
(average) of all the measure- to ± 0.04 V calibration standard must be
ments, which we would expect to traceable to national standards
3 standard deviations corresponds through an through an unbroken
be 10 V. We can also calculate the to ± 0.06 V
standard deviation of the read- chain of comparisons, all having
ings (Equation 1). So the questions for the manu- stated uncertainties.
facturer become, “How many So the uncertainty of a DMM
standard deviations do we use depends on the uncertainty of the
for our spec?” “What confidence calibrator used to calibrate it. Most
interval do we use to build our DMM specs are written assuming
specs?” The higher the number two things:
of standard deviations, the lower • The DMM has been calibrated
Equation 1. N = sample size the probability that an instru- using a particular model of
X = measurement ment will fall out of spec between calibrator, usually specified in
The standard deviation is a calibrations. The manufacturer’s the DMM service manual.
measure of the “spread” of the internal engineering standards • The calibrator was within its
sample of measurements, outward will determine how many stan- operating limits and traceable
from the mean. This measure of dard deviations are used to set to national standards.
spread is the basis of uncertainty the spec. Fluke uses a confidence This allows a DMM manufacturer
specifications. of 99 %, which corresponds to to include the uncertainty of the
If we plot the number of times 2.6 standard deviations on a calibrator in the DMM uncertainty
each reading occurs, we should normal distribution. specs. If you see an uncertainty
see a bell-shaped normal distri- listed as “relative” this means
bution. (Almost all measurements Traceability and the uncertainty in the calibrator
follow a normal distribution, specifications output has not been considered
including those made with simple So far we have described how and it must be added to the DMM
instruments like rulers and mea- much uncertainty we can expect uncertainty.
suring cups.) Figure 1 shows a from a DMM, but we have not
2 Fluke Corporation Understanding specifications for precision multimeters
 Temperature affects the per- thermal effects. Table 1 shows the
National Standards (e.g. NIST)
formance of every component in elements of a DMM specification
an instrument—from the simplest and gives examples for each.
Secondary
Standards resistor to the most elegant
integrated circuit. DMM design- Signal
Input
AC PWR IN CAL CHARGE LOW BAT AC PWR IN CAL CHARGE LOW BAT AC PWR IN CAL CHARGE LOW BAT

ers are good at building circuits Input

Divider Analog
Switching to Digital
that compensate for temperature
10V 1.018V CHASSIS 10V 1.018V CHASSIS 10V 1.018V CHASSIS

10V COM 1.018V COM GUARD 10V COM 1.018V COM GUARD 10V COM 1.018V COM GUARD
Protection Convertor
variation. This ability to operate at
In-house various temperatures is captured Amplifiers and
Standards in a specified operating range Signal Conditioning

and is often accompanied by a Figure 3: Simplified DMM analog signal path for dc volts measurement.
temperature coefficient. (More on
this later.) Example
DMM
Multimeter uncertainties cannot Baseline uncertainty

be given as simple percent- Input ± (0.001 % of reading + 3 digits)

Scale term + floor ± (0.001 % of reading + 3 digits)
ages, although it is tempting to Uncertainty modifiers
Figure 2: Traceability is the path from your over simplify. You might see a Temperature coefficient ± (0.003 % of reading) per °C from 0 °C
DMM back to national standards.
sales brochure that touts, “Basic to 18 °C and 28 °C to 50 °C
accuracy to 0.002 %”. This is only Time 1 year

Elements of digital giving a small part of the picture,

Qualifiers
Warm-up time Specs are valid after 1 hour warm-up
multimeter specifications and it’s usually an optimistic view Operating temperature 23 ± 5 °C
of the data. RH 80 % RH from 0 to 35 °C, 70 % to 50 °C
Among the many standards that
The reasons for complexity in Storage temperature 40 ºC to 60 ºC
govern instrumentation, there
the specifications have to do with Vibration Meets requirements of MIL-T-28800E for
is no standard for writing DMM Type III, Class 3,
the multimeter’s ability to perform Style E equipment
specs. Over the years, though,
many different measurements, EMI susceptibility Complies with EN 50082-1
manufacturers have converged
over many different ranges, using Altitude 2000 meters
on similar formats, making it a bit Power line regulation 100 V/ 120 V/ 220 V/ 240 V ± 10 %
several different internal signal
easier to compare multimeters. Overvoltage protection 600 V overvoltage category III
paths.
This application note covers the Table 1: Key elements of the DMM specification.
Consider the diagram in Figure
most common conventions for
3. It shows the analog signal path
specifications. Baseline uncertainty
for a dc voltage measurement,
As described above, uncertainty specifications
also known as “the front end”.
specifications define a range Baseline specifications are usually
Each block contributes uncer-
around a nominal value. When given as:
tainty in the form of nonlinearity,
taking a measurement within the
offset, noise and thermal effects. ± (percent of reading + number
specified limits of time, tempera-
The front end contributes most of of digits)
ture, humidity, etc., you can be
the uncertainty of the instrument. or
confident that you won’t get a
Depending on the design, ± (percent of reading + number
reading outside that range.
changing ranges affects the of counts)
Time and temperature are
divider performance or the
crucial for determining uncer- “Digits” or “counts” are used
amplifier performance or both.
tainty. Electronic components interchangeably and they indicate
Internal noise, for example, has a
experience small changes (or the value of the least significant
greater relative impact on lower
“drift”) over time. Because of drift, digits for a particular range. They
ranges and at the low ends of
DMM uncertainties are valid only represent the resolution of the
ranges. Changing functions alters
for a specified period of time. This DMM for that range. If the range
the signal path. For example, a
period usually coincides with is 40.0000 then one digit, one
resistance measurement requires
the recommended calibration count, is worth 0.0001.
the addition of a current source to
cycle and is typically one year. At Let’s say you want to measure
the analog path. So each function
calibration, the clock starts over 10 V on a 20 V range in which
and range must be specified in a
again and the uncertainties are the least significant digit repre-
way that considers the effects of
valid for another period. sents 0.0001 V. If the uncertainty
non-linearities, offsets, noise and
for the 20 V range is given as
± (0.003 % + 2 counts) we can
calculate the uncertainty in mea-
Top three pitfalls of using DMM specs surement units as:
1. Using only the percent of reading as a shortcut ± ((0.003 % x 10 V + 2 x 0.0001
2. Applying specs to DMM’s that fall outside of V) = ± (0.0003 V + 0.0002 V) =
their calibration cycle ± (0.0005 V) or ± 0.5 mV
3. Using a DMM outside its temperature range Some spec sheets use the form:
without de-rating uncertainties ± (percent of reading + percent
of range)
3 Fluke Corporation Understanding specifications for precision multimeters
 In this case you simply multiply modified uncertainty is more Comparing different
the maximum reading for the than three times larger than digital multimeters
range by a percentage to get the the baseline! When you are evaluating the
second term. In both cases the Qualifier specifications suitability of several digital
second term is called “the floor”. DMM uncertainties depend on multimeters, the best approach is
The floor considers the effects of other conditions besides time to choose a set of measurements
offsets and noise associated with and temperature. Environmental and conditions that approximate
a single range as well as those factors such as storage tempera- your application. First, make sure
common to all ranges. Ignoring ture, humidity, air density, and the qualifiers of each DMM are
this term can have significant electromagnetic radiation can compatible with the application
consequences, especially for affect uncertainty. The DMM must environment. Then consider all of
measurements near the bottom of receive reasonably clean power if the functions (dc volts, ac volts,
a range. its precision internal power sup- dc amps, ohms and so forth) and
Uncertainty modifiers plies are to function properly. ranges you are likely to use. Are
Modifiers can be applied to the Some qualifiers can be easily you going to be making a lot of
uncertainty specs to account for specified by numerical values, like measurements near the bottoms
common environmental or time power line regulation, altitude of ranges? Then make sure you
factors. Some specifications will and relative humidity. DMMs are compare with low readings to
give not only one-year specs, but not hermetically sealed, so air check the contribution of the floor.
also specs that apply for, say, becomes a component of their For each measurement, convert
90 days after calibration. The circuitry. The electrical charac- the uncertainty into measure-
90-day spec will be tighter than teristics of air are affected by ment units like volts, ohms or
the 1-year spec. This allows a density (altitude) and humidity, so amps. Compare the uncertainties
DMM to be used in more demand- designers set boundaries on these in measurement units to decide
ing applications by calibrating parameters. Excessive storage which DMM is better suited to the
more frequently. temperatures can irreversibly alter task at hand.
For reasons already covered, the operating characteristics of The ability to work with DMM
uncertainty specs are valid over electronic components. uncertainty specifications is a
a specified temperature range. More complex qualifiers like fundamental engineering skill.
Commonly the range encom- overvoltage protection, shock When you get two different
passes “room temperature”, from and vibration, or electromagnetic answers from two different DMMs,
18 °C to 28 °C (64.4 °F to 82.4 °F) compatibility are given by noting measurement uncertainty might
when calibrated at 23 °C. Over a compliance with standard mea- explain the difference. When you
wider range the uncertainty can surement techniques and limits. need to compare the performance
be modified to account for the International standards docu- of DMMs to decide the right tool
temperature. ments for these characteristics for the job, being proficient with
Say you need to take the same typically require a series of test uncertainty specs will help you
10 V measurement performed procedures along with applicable compare apples to apples. And
above, at a field location where limits. Adding all of the limits whenever you are depending on
the temperature is 41 °C (106 °F). would render the DMM specifica- an important measurement, you
The temperature coefficient of tions too cumbersome, so DMM can be comfortable understand-
your DMM is given as: designers just list the standards ing how well your instrument will
with which the DMM complies. really perform.
± (0.001 % of reading) per °C
from 0 °C to 18 °C and 28 °C
to 50 °C
The temperature is 13 °C above Fluke Calibration. Precision, performance, confidence.™
the 28 °C boundary for using
unmodified baseline uncertainty.
For each degree above the
boundary, we have to add
0.001 % x 10 V = 0.1 mV/°C to Fluke Calibration Fluke Europe B.V.
the baseline uncertainty. The PO Box 9090, PO Box 1186, 5602 BD
added uncertainty at 41 °C is Everett, WA 98206 U.S.A. Eindhoven, The Netherlands
13 °C x 0.1 mV/°C = 1.3 mV. So For more information call:
the total voltage uncertainty, com- In the U.S.A. (877) 355-3225 or Fax (425) 446-5116
In Europe/M-East/Africa +31 (0) 40 2675 200 or Fax +31 (0) 40 2675 222
bining the baseline uncertainty In Canada (800)-36-FLUKE or Fax (905) 890-6866
calculated in the example above From other countries +1 (425) 446-5500 or Fax +1 (425) 446-5116
and the temperature modifier Web access: http://www.fluke.com
would be ± (0.5 mV + 1.3 mV) = ©2006-2011 Fluke Corporation. Specifications subject to change without notice.
Printed in U.S.A. 1/2011 2547797B A-EN-N Pub_ID: 11066-eng Rev 02
± 1.8 mV. Notice that the
Modification of this document is not permitted without written permission from Fluke Corporation.

4 Fluke Corporation Understanding specifications for precision multimeters