Skip to content
AkerminMeasure. Monitor. Improve. Compare Products
Learning Center

Oscilloscope or digital multimeter? One gives a number, the other gives time

The comparison is usually framed as cheap against expensive, which misses the point. These instruments trade accuracy against time resolution in opposite directions, and the trade is published in their own specifications if you know which numbers to read against each other.

By the Akermin Editorial Team · 7 min read

As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.

A digital multimeter collapses a signal into a single number and refreshes it a few times a second. An oscilloscope plots voltage against time and can sample millions of times a second. If what you are chasing lasts less than a quarter of a second, no meter will show it and almost any scope will. If what you need is a direct voltage you can trust to four or five digits, the meter is the more accurate instrument by a wide margin and the scope is not even close. Neither is an upgrade on the other; they answer different questions.

One number per quarter second against a quarter billion samples

Put the two update rates side by side and the difference stops being abstract. The Fluke 177 publishes a digital display updating four times a second and a 33 segment analogue bar graph updating forty times a second. The FNIRSI 2C53T publishes 50 MHz of bandwidth and a 250 megasample per second sampling rate.

Four readings a second against 250 million samples a second is a factor of about 62 million. Anything that happens and finishes inside a quarter of a second is, to the meter, an event that occurred between two readings. To the scope, a single microsecond of activity is 250 individual samples, which is enough to draw its shape, measure its amplitude and time its edges.

Event Duration On a display updating four times a second On a 250 megasample per second scope
Mains half cycle at 50 Hz 10 milliseconds Folded into one averaged or RMS number 2.5 million samples
Motor starting surge About 100 milliseconds Missed unless a dedicated capture function exists Fully resolved, shape and peak
Relay contact bounce 1 to 10 milliseconds Invisible 250,000 to 2.5 million samples
Switching supply ripple at 100 kHz 10 microseconds per cycle Invisible, averaged into the direct reading 2,500 samples per cycle
Commutation or discharge spike 100 nanoseconds Invisible 25 samples

The middle row is the interesting one, because it is where meter designers fought back. A starting surge lasts long enough that a purpose built capture function can catch it, and that is exactly what the inrush mode on a modern clamp does: the Neoteck 1000A clamp publishes true RMS measurement, a variable frequency drive mode, inrush current capture and low impedance operation, with a measurement category of three at 1000 volts and four at 600 volts, a 9999 count colour display and an analogue bar. An inrush function is a time windowed measurement wearing a meter’s clothes.

What a steady reading hides

A meter displaying 12.01 volts direct on a supply rail that carries 400 millivolts of switching ripple will still display about 12.01 volts, because direct coupling averages the ripple away. The reading is correct and the rail is not healthy. Three families of fault behave this way.

  • Ripple and noise riding on a steady level. Averaged out by definition. A scope shows the amplitude, the frequency and whether it is a clean sawtooth from a linear supply or the sharp spikes of a switching one.
  • Intermittent dropouts shorter than the update interval. A connection that opens for five milliseconds every few seconds will reset a microcontroller and never register on a meter. On a scope with a trigger set below the nominal level it is captured on the first occurrence.
  • Waveform shape. A variable frequency drive output is a train of pulses whose average has little to do with its effective value. True RMS measurement fixes the number, which is covered in what true RMS means on a multimeter, but a correct number still tells you nothing about why the waveform looks the way it does or whether the switching is misbehaving.

The honest way to state the limitation is that a meter answers how much and a scope answers what is it doing. A great many frustrating faults are shape faults, and no amount of digits resolves them.

Where the meter is the more accurate instrument

Run the comparison the other way and the scope looks poor. The Fluke 8808A publishes a five and a half digit dual display, a basic direct voltage accuracy of 0.015 percent, a dedicated direct current leakage measurement and a four wire resistance technique. Five and a half digits is roughly 200,000 counts of resolution, and 0.015 percent on a ten volt reading is about 1.5 millivolts of uncertainty.

A general purpose scope is nowhere near that, for two structural reasons. The vertical path is an analogue amplifier chain ahead of a fast converter, and fast converters are shallow: most instruments in this class use an eight bit converter, which divides the whole screen height into 256 levels, so the smallest distinguishable step is roughly four tenths of a percent of the displayed span before any amplifier error is considered. The FNIRSI listing publishes vertical sensitivity from 10 millivolts to 10 volts per division and a maximum input of 400 volts, but a vertical accuracy figure and a converter depth are both Not stated in the listing bullets, which is normal for the category and is itself informative: accuracy is not what the instrument is sold on.

The resolution gap is visible inside the same product. The 2C53T publishes a four and a half digit 19999 count multimeter mode covering 0 to 750 volts alternating, 0 to 999.9 volts direct, up to 9.999 amps, 19.99 megohms and 99.99 millifarads. The manufacturer put a separate meter section in the instrument rather than asking the scope channels to do that job, which is the clearest possible admission of what each half is for. Our notes on resolution in a digital meter and choosing a digital multimeter go further into how resolution and accuracy come apart.

Record depth and the width of the window you can see

This is the specification nobody mentions and almost everybody runs into. The 2C53T publishes a record depth of 1 kilopoint, meaning one capture stores about a thousand samples.

Do the arithmetic. At the full 250 megasample per second rate, a thousand samples spans four microseconds of time. That is ample for a single fast edge and useless for watching a mains cycle. To see ten milliseconds in one sweep with a thousand points, the instrument must sample at around 100 kilosamples per second, one sample every ten microseconds, at which rate a one microsecond glitch lands between samples and simply does not exist in the capture.

So sample rate and time window compete for the same memory, and record depth is the referee. This is the real difference between a basic scope and a serious one, far more than the headline bandwidth: deep memory lets you capture a long window at a high sample rate and then zoom in afterwards, which is precisely the workflow you want when hunting an intermittent fault. With a shallow buffer you must choose in advance between seeing the context and seeing the detail, and if you choose wrongly the event is gone.

Combination instruments and what they give up

Combined instruments are genuinely useful and the compromises are specific rather than vague. The 2C53T bundles a two channel 50 MHz scope, a four and a half digit meter and a signal generator reaching 50 kilohertz with thirteen waveform shapes and one hertz steps, which is a complete small bench in one hand.

What it gives up is visible in its own numbers: a 400 volt ceiling on the scope inputs, a thousand point record depth, and no published measurement category rating for the scope channels, which matters because category ratings describe survival during a transient rather than behaviour during a normal reading.

The traffic runs both ways. Meters have been absorbing time domain functions for years: inrush capture, peak hold on a millisecond timescale, minimum and maximum recording, frequency and duty cycle, and drive filtering modes. Each of those is a scope measurement reduced to a single number by a dedicated circuit, which is a good trade when you know in advance which number you want and no help at all when you do not. Our handheld meter notes and mid range meter notes cover which of those functions the listings in that class publish.

Choosing by the shape of the question

Let the question pick the instrument.

  • How much voltage, current or resistance is there? Meter, and the more digits the better if the decision is tight.
  • Is this rail clean, and what is riding on it? Scope. A meter cannot see ripple at all.
  • Why does this breaker trip only on starting? A clamp with an inrush function answers it in one reading. A scope with a current probe answers it in more detail.
  • Why does this device reset occasionally? Scope with a trigger armed below the nominal level, and record depth is the specification that decides whether you catch it.
  • Is this circuit safe to touch? Neither of these. A properly rated tester, every time.
  • What shape is this signal? Scope, by definition. This is the one question a meter cannot approach.

If you can own only one, own the meter, because the questions it answers come up a hundred times more often and it is the instrument you can trust with a number. Add a scope when you find yourself asking why rather than how much, and when that happens look at record depth at least as hard as you look at bandwidth.

Frequently Asked Questions

Can an oscilloscope measure direct voltage?

Yes, with the input set to direct coupling the trace sits at a vertical offset proportional to the voltage, and almost every scope will put a number on screen for you. The accuracy is the problem rather than the capability. Vertical accuracy on general purpose scopes is typically quoted in whole percentage points of full scale, against hundredths of a percent on a decent bench meter, so a scope is the right instrument for seeing whether a direct rail is clean and the wrong one for deciding whether it is 4.98 or 5.02 volts.

What does 50 MHz of bandwidth actually let me see?

Less than the number suggests, because bandwidth is the frequency at which the response has already fallen by three decibels, meaning an amplitude error around thirty percent at that point. The working convention is to want bandwidth of at least five times the highest frequency you need to measure accurately, so 50 MHz is comfortable up to roughly 10 MHz. The other useful form is risetime: dividing 0.35 by the bandwidth gives about 7 nanoseconds, which is the fastest edge such an instrument can represent at all.

Do I need an oscilloscope for household electrical work?

Almost never. Mains work is about presence of voltage, continuity, insulation and current, and all of those are meter or tester questions. A scope becomes useful at the boundary where electrical work meets electronics: inverter outputs, variable speed drive waveforms, switch mode supply ripple, dimmer behaviour and anything involving a signal rather than a supply. If the question is how much, buy the meter first.

What is record depth and why does 1 Kpts matter?

Record depth is how many samples the instrument stores for one captured sweep, and it sets the hard trade between how fast you sample and how long a window you can see. A thousand points at a 250 megasample per second rate covers four microseconds of time. To view ten milliseconds with only a thousand points the instrument has to drop to about 100 kilosamples per second, at which rate a one microsecond glitch falls between two samples and is simply absent from the capture. Deep memory is what lets you have both at once.

Is a combined scope and meter a sensible first instrument?

It is a reasonable way to see waveforms without buying two instruments, provided you treat the meter half as a convenience rather than as a reference. The compromises are in the specifications: limited record depth, modest input voltage ceilings on the scope channels, and measurement category ratings that are often unpublished. If your work is mostly electrical rather than electronic, a properly rated handheld meter and a separate scope bought later is the better order.