Guide · Meters, supplies and scopes
Do you need an oscilloscope?
A multimeter tells you the voltage; a scope shows what a signal does over time. When a repair needs one, how much bandwidth to buy, and why a meter comes first.
We may earn a commission if you buy through links on this page, at no extra cost to you. Unless a page says otherwise, our picks come from published specifications and listing research, not hands-on testing. As an Amazon Associate I earn from qualifying purchases. How we make money · How we research
Short answer
You need an oscilloscope when the fault is about timing or shape rather than a steady value: a clock that may not be running, ripple on a supply rail, a PWM signal, a power-up sequence or serial data. For continuity, shorts and steady voltages, a multimeter is the right tool and the better first purchase. If you do buy a scope, choose bandwidth with headroom: Keysight suggests at least five times the highest clock rate for digital signals.[1]
What a scope shows that a meter cannot
A multimeter gives you one number: a steady voltage, a resistance or a current. An oscilloscope draws voltage against time, so you can see whether a signal is there, what shape it has and when it happens.
| Repair question | Multimeter | Oscilloscope |
|---|---|---|
| Is there power on this rail? | Yes | Yes, but a meter is quicker |
| Is something shorted? | Yes (continuity or resistance) | No |
| Is the clock or crystal running? | Not reliably | Yes |
| Is there ripple or noise on a supply rail? | Not reliably | Yes |
| What is this PWM signal doing? | Partly, if the meter measures duty cycle | Yes |
| Do the rails come up in the right order? | No | Yes, with two or more channels |
| Is there activity on a data line? | No | Yes; some scopes also decode serial protocols |
Buy a meter first
Most first-line diagnosis is a meter's job: is power present, is something shorted, does a fuse or diode read correctly. A scope comes next, when you have ruled those out and need to see what a signal is doing. If you do not yet own a decent multimeter, start there, and read multimeter CAT ratings explained before it goes near mains.
Two meters cover the two cases. Fluke rates its 107 at 600 V CAT III under IEC 61010-1,[2] which suits a meter that may also touch outlet-connected equipment. The UNI-T UT61E+ listing shows no CAT rating, so it is a bench meter for low-voltage boards until you have read the marking on the meter itself.
How much bandwidth?
Tektronix defines a scope's bandwidth as the frequency at which a sine wave is shown at 70.7% of its real amplitude, the -3 dB point.[3] Near and above that frequency, signals look smaller than they are and fast edges look slower. So you need headroom:
- Tektronix's "5 times rule": bandwidth at least five times the highest frequency component keeps amplitude error under about ±2%.[3]
- Keysight: at least three times the highest sine frequency for analog signals, and at least five times the highest clock rate for digital systems.[1]
- Rise time: bandwidth = K / rise time, where Tektronix gives K as 0.35 to 0.45.[3]
Worked examples using those rules:
| Signal | Rule applied | Bandwidth to look for |
|---|---|---|
| 20 kHz sine wave | Keysight, 3x[1] | At least 60 kHz |
| 1 MHz digital clock | Keysight, 5x[1] | At least 5 MHz |
| 20 MHz digital clock | Keysight, 5x[1] | At least 100 MHz |
| Edge with a 3.5 ns rise time | Tektronix, 0.35 / rise time[3] | About 100 MHz |
Sample rate and memory
Bandwidth is only half the story; the scope also has to take samples fast enough.
- Tektronix: sample at least 2.5 times the highest frequency component with sin(x)/x interpolation, or 10 times with linear interpolation.[3]
- Keysight: a maximum sample rate of at least four to five times the scope's bandwidth.[1]
- Memory depth decides how long a stretch of signal you can capture at full sample rate, which matters for catching a fault that happens once during start-up.
A current benchtop example: Rigol's DHO800 series comes in 70 MHz (DHO802, DHO804) and 100 MHz (DHO812, DHO814) models, with 1.25 GSa/s, 25 Mpts of memory and 12-bit vertical resolution.[4] On the 100 MHz models, 1.25 GSa/s is 12.5 times the bandwidth, well over Keysight's four-to-five-times guide.
Headline specs are claims
Treat a scope's headline numbers as the maker's claim until someone measures them. FNIRSI, for example, lists 100 MHz, 1 GS/s and two channels for its 1014D, and 10 MHz with 50 MS/s, two channels and a built-in signal generator for its 2C23T.[5][6] Members of the EEVblog forum dispute FNIRSI's headline figures; one estimated about 30 MHz of real bandwidth for a FNIRSI scope.[7] That is a forum estimate, not a controlled measurement, but it is a good reason to look for independent measurements before you trust any headline figure.
When a scope earns its place
- A board powers up but does not boot, and you need to check clocks, resets and data lines.
- A power supply misbehaves: ripple, oscillation, or a rail that comes up and then collapses.
- Audio and analog work, where the shape of the waveform matters.
- Microcontroller and hobby projects with PWM outputs and serial buses.
If none of those describe your repairs yet, a meter and a bench power supply will take you further for now.
Next
Questions
Should I buy an oscilloscope or a multimeter first?
A multimeter. It answers the first questions in most repairs, such as whether power is present and whether something is shorted; a scope comes next, when you need to see a signal over time.
How much oscilloscope bandwidth do I need?
Use a rule with headroom: Keysight suggests at least three times the highest sine frequency for analog signals and five times the highest clock rate for digital ones, and Tektronix's five-times rule keeps amplitude error within about 2%.[1][3]
Is 1 GSa/s enough sample rate?
It depends on the bandwidth. Keysight suggests a maximum sample rate of at least four to five times the scope's bandwidth, so on that rule 1 GSa/s suits a scope of up to about 200 to 250 MHz.[1]
Are budget handheld oscilloscopes accurate?
Treat their headline figures as claims. FNIRSI lists 100 MHz for its 1014D, while EEVblog forum users dispute FNIRSI's headline specs, one estimating about 30 MHz of real bandwidth for a FNIRSI model.[5][7]
Can I measure mains with an oscilloscope?
Not with an ordinary setup. Only use a scope on mains-powered circuits if the scope, its probes and your method are rated for it; Fluke notes that higher measurement categories carry higher-energy transients and that test tools should be independently certified.[8]