When we take a measurement with a multimeter, caliper, thermometer, or any other instrument, it is common to look at the displayed number and assume that it is simply the value of the quantity we are measuring.
But every measurement has limitations.
If a multimeter displays:
5.02 V
this does not necessarily mean that the voltage is exactly 5.020000… V.
The result depends on the instrument, the selected range, the test leads, environmental conditions, and even how we perform the measurement.
In the video featured in this article, I share some tips on how to obtain better measurement results. This topic also helps us understand a very important distinction between resolution, precision, and accuracy.
Measuring Is Not Simply Reading a Number
A measurement is a comparison between a quantity and a reference.
NIST defines measurement as an experimental process that produces a value that can reasonably be attributed to the quantity we want to determine.
Therefore, when we obtain:
R = 1.003 Ω
we should ask ourselves:
Which instrument was used? On which range? Under what conditions? With what uncertainty?
The more demanding the application, the more important these questions become.
Precision, Accuracy, and Resolution Are Not the Same Thing
These concepts are frequently confused.
Imagine that we take several measurements of the same quantity.
Precision
Precision is related to how close repeated measurement results are to one another.
For example:
5.01 V 5.01 V 5.02 V 5.01 V 5.01 V
The results are closely grouped.
NIST characterizes repeatability by the agreement between successive measurements performed under the same conditions.
Accuracy
Accuracy is related to how close a measurement result is to the reference value.
We might have an instrument that consistently displays:
5.20 V 5.20 V 5.20 V 5.20 V
when the reference value is:
5.00 V
It has excellent repeatability, but there is a significant deviation from the reference value.
Fluke also distinguishes between these concepts: greater precision is associated with repeatability, while greater accuracy means a reading closer to the value considered true or to the reference value.
What Is Resolution?
Resolution is the smallest change that an instrument can indicate.
Imagine two multimeters.
The first displays:
5.0 V
The second:
5.023 V
The second has higher resolution.
But there is a catch:
more decimal places do not automatically mean greater accuracy.
An instrument can display many digits and still have considerable measurement uncertainty.
The instrument’s specifications tell us what those digits actually represent.
Choose the Correct Multimeter Range
This is one of the simplest ways to obtain a reading with better resolution.
Suppose we want to measure approximately:
1.25 V
If we use a range much higher than the measured value, we may lose resolution.
In an example provided by Fluke, different ranges offer different resolutions:
| Range | Resolution |
|---|---|
| 300.0 mV | 0.1 mV |
| 3.000 V | 1 mV |
| 30.00 V | 10 mV |
| 300.0 V | 100 mV |
Therefore, when an instrument allows manual range selection, we normally choose the lowest range that can safely accommodate the measured value. With autoranging multimeters, the instrument itself attempts to select an appropriate range.
Naturally, when we do not know the approximate value of the quantity being measured, we begin in a safe manner appropriate to the instrument and the application.
More Digits Do Not Solve Everything Either
There are multimeters with:
3½ digits 4½ digits 5½ digits 6½ digits …
as well as specifications expressed in counts.
Bench multimeters can achieve much higher resolutions than ordinary handheld multimeters. Keysight, for example, currently offers bench instruments ranging from 5½ to 8½ digits, depending on the equipment class.
This is important for certain applications.
But there is no point in buying an instrument with extremely high resolution and then using it incorrectly.
Read the Accuracy Specification
With digital multimeters, it is common to find a specification such as:
±(0.5% of reading + 2 digits)
This means that the permitted error is not simply:
±0.5%
There is also a contribution associated with the least significant digits.
Fluke presents, for example, a hypothetical specification of:
±(2% + 2)
and shows that a displayed reading of 100.0 V could correspond, within that specification, to a range of approximately 97.8 to 102.2 V.
Therefore, we should never interpret every digit on the display as equally reliable without knowing the instrument’s specifications.
Test Probe Contact Makes a Difference
An extremely common source of error is not inside the multimeter.
It is here:
MULTIMETER
│
↓
CABLES
│
↓
PROBES
│
↓
CONTACT
│
↓
CIRCUITOxidized probes, loose connectors, unstable contacts, or varying contact pressure can affect the measurement.
This becomes particularly evident when measuring very low resistances.
Perform a Simple Test with the Probes
Set the multimeter to resistance measurement and touch the two test probes together.
Ideally, we would like to see:
0 Ω
but we might obtain:
0.2 Ω
for example.
This resistance may include contributions from the:
cables + connectors + probes + contacts.
Now imagine trying to measure a resistor with a resistance of:
0.5 Ω
The resistance of the measurement system is no longer negligible.
Some instruments offer compensation functions or specific techniques for low-resistance measurements.
Four Wires Can Be Better Than Two
When measuring very low resistances, laboratories and bench instruments may use the four-wire technique, also known as Kelvin measurement.
In simplified form:
current
↓
┌───────────┐
───────┤ ├───────
│ RESISTOR │
───────┤ ├───────
└───────────┘
↑
voltage
One pair carries the current, while the other measures the voltage.
This can significantly reduce the influence of cable resistance on the result.
For ordinary workbench measurements, this is not always necessary, but it demonstrates an important principle:
the measurement system itself can affect the result.
The Multimeter Is Also Part of the Circuit
When we connect an instrument to a circuit, it is not completely invisible from an electrical standpoint.
When measuring voltage, we want the voltmeter to have a sufficiently high input impedance so that it has little effect on the circuit.
When measuring current, the ammeter is connected in series and introduces some internal resistance.
Therefore:
ORIGINAL CIRCUIT
↓
add instrument
↓
CIRCUIT + INSTRUMENTIn ordinary situations, this influence may be small.
In high-impedance circuits or more delicate measurements, it can become significant.
Do Not Hold Certain Parts with Your Fingers
When measuring high resistances, our own bodies can create an additional electrical path.
Imagine:
probe ── RESISTOR ── probe
↑ ↑
finger fingerYou may end up placing a resistance associated with your body in parallel with the component.
In certain measurement ranges, this can change the reading.
Therefore, especially when measuring high resistances, avoid simultaneously touching the metal parts of the probes and terminals.
Temperature Also Changes Values
Electronic components do not necessarily have the same value at every temperature.
A resistor has a temperature coefficient.
Sensors, semiconductors, and other devices may exhibit even more significant variations.
The instrument itself also has specified environmental conditions under which it achieves its stated accuracy. Keysight’s documentation emphasizes that accuracy specifications only make sense when accompanied by the conditions under which they are valid, including factors such as temperature, humidity, and the time elapsed since calibration.
This explains why metrological measurements require much more controlled conditions than ordinary workbench measurements.
Allow the Reading to Stabilize
Not every quantity reaches its final value instantly.
We might observe something like this:
5.17 5.11 5.08 5.06 5.05 5.05 5.05
If we immediately record:
5.17
we may be measuring a transient rather than the condition we actually wanted to analyze.
That is why we need to understand:
what we are measuring and when we should take the reading.
This applies to electrical instruments, temperature, mass, dimensions, and many other quantities.
Take More Than One Measurement
Another useful practice is to repeat the measurement.
Instead of:
MEASURE → RECORD → FINISH
we can do:
MEASURE
↓
REPEAT
↓
COMPARE
↓
CHECK CONSISTENCYIf we obtain:
10.01 10.02 10.01 10.02
we have a very different situation from:
9.72 10.34 9.89 10.28
Repeating measurements helps identify variability in the process. NIST considers the dispersion of a series of measurements to be one of the pieces of information that can contribute to evaluating uncertainty.
But Averaging Does Not Eliminate Every Error
Imagine an instrument with a systematic error.
It repeatedly measures:
10.20 10.21 10.20 10.20 10.21
when the reference value is close to:
10.00
Taking the average of a thousand measurements will not necessarily eliminate this deviation.
This helps us understand the difference between:
random errors and systematic effects.
NIST notes that measurement uncertainty can include contributions from both random and systematic effects.
Compare Instruments When It Makes Sense
A simple workbench experiment is to measure the same power source with two multimeters.
For example:
MULTIMETER A → 5.01 V
MULTIMETER B → 5.07 V
Which one is correct?
This comparison alone cannot tell us.
Perhaps A is closer to the reference value.
Perhaps B is.
Perhaps both are within their respective specifications.
But the comparison is useful because it reminds us that the number displayed by an instrument is not an absolute truth.
To properly determine accuracy, we need a known reference and a calibration process.
Calibration Is Different from Simply Comparing
In professional applications, instruments can be calibrated using standards with known traceability.
This creates a chain of comparisons:
INSTRUMENT
↓
REFERENCE STANDARD
↓
HIGHER-LEVEL STANDARDS
↓
METROLOGICAL REFERENCESWe do not need to turn every home workbench into a metrology laboratory.
But understanding this concept changes the way we interpret a reading.
How Many Decimal Places Should I Record?
An important practical rule is:
do not attribute more precision to a result than your measurement process can support.
If the instrument and method do not justify:
12.347891 V
there is no scientific advantage in recording six decimal places simply because a calculation produced that number.
The reported digits must be consistent with the measurement’s resolution and uncertainty.
This Applies to Much More Than Multimeters
The same principles apply when we use:
calipers, micrometers, scales, thermometers, oscilloscopes, frequency counters, or any other instrument.
For example, when using calipers:
correct positioning
+
appropriate pressure
+
clean surfaces
+
suitable instrument
↓
better measurementWith a thermometer:
sensor positioning
+
stabilization time
+
surface characteristics
+
environmental conditions
↓
better measurementMeasurement technique matters just as much as the instrument.
Watch the Tips in the Video
In the video below, I share some tips on how to obtain more precise measurements with a multimeter or other instruments.
The goal is not to turn a simple measurement into a laboratory procedure.
It is to develop an important habit:
before trusting a number, understand how it was obtained.
A Better Instrument Helps, but Technique Is Still Important
Buying a more expensive multimeter may offer:
better accuracy, higher resolution, better stability, and more features.
But that does not automatically correct a poorly performed measurement.
We can summarize a good measurement as follows:
SUITABLE INSTRUMENT
+
APPROPRIATE RANGE
+
GOOD PROBES AND CONTACTS
+
CORRECT TECHNIQUE
+
KNOWN CONDITIONS
+
REPETITION WHEN NECESSARY
↓
MORE RELIABLE RESULTThis is perhaps the most important point.
Measuring well does not simply mean owning a good instrument. It means knowing how to use it and understanding the limitations of the measurement we are performing.
To learn more about using instruments, testing components, and diagnostic techniques at the workbench, explore the Workbench, Instrumentation, and Maintenance section.