HOW-TO GUIDE
How to Measure Ripple Voltage With an Oscilloscope
By the OhmSavvy Research Team ·
Safety first: the probe ground clip goes to mains earth
- On a mains-powered bench scope, the probe's ground clip is connected to mains earth through the power cord. Only clip it to points that are already at earth or low-voltage ground, such as the output ground of an isolated supply. Clipping it to the primary side of a mains-powered supply shorts that point to earth and can destroy the probe, the scope or the circuit, or injure you.
- Never float the scope by removing its earth connection. To probe the mains side, use an isolation transformer on the device under test or a differential probe rated for the job.
- Large filter capacitors can stay charged after power is removed. Measure ripple on the low-voltage output with the enclosure closed where possible.
Why ripple is hard to see on a DC rail
Ripple is the small AC variation left on a DC supply after rectification or switching. On a 5 V rail it might be a few tens of millivolts, which is about 1% of the DC level. If you probe the rail with DC coupling at a range that fits 5 V on screen, the ripple is far too small to see:
To look at the switching waveform that creates this ripple, see how to measure a switching power supply waveform.
Turning the V/div knob down to millivolts would show the ripple, but the 5 V level would then be pushed hundreds of divisions off the top of the screen. The steps below fix that. If reading divisions is new to you, start with how to read an oscilloscope screen.
1Set up the probe and use a short ground spring
Ripple measurements fail at the probe more often than at the scope. The long ground lead with its crocodile clip forms a loop several centimeters across, and that loop picks up magnetic fields from switching currents, inductors and transformers nearby. On a switching supply the pickup can be as large as the ripple itself.
- Remove the long ground lead. Pull the probe's hood off and fit the short ground spring that comes with many probes, so the tip and the ground barrel touch the capacitor's two terminals directly.
- Probe at the output capacitor. Measure across the supply's output capacitor or at the load, not at the end of a long cable.
- Choose 1x or 10x deliberately. A 1x probe gives ten times the sensitivity but usually much less bandwidth; a 10x probe keeps bandwidth but makes the most sensitive range ten times coarser. Set the channel's probe ratio to match.
- Compensate a 10x probe first. The steps are in how to compensate an oscilloscope probe.
A quick check: touch the probe tip to its own ground spring on the same ground point. Whatever appears on screen now is pickup and scope noise, the floor below which you cannot measure.
2How to use oscilloscope AC coupling to measure ripple
AC coupling puts a capacitor in series with the channel input, blocking the DC part of the signal so only the variation reaches the screen, centered on the ground line. Open the channel menu, change coupling from DC to AC, and the 5 V trace drops to the center line. Now you can turn the sensitivity right up:
AC coupling has a low-frequency cutoff, usually a few hertz, and the exact figure varies by model, so check your manual. It passes 100 or 120Hz rectifier ripple and switching-frequency ripple without trouble, but it also hides slow drift and the DC level itself. If you need the rail's actual voltage, measure that separately with DC coupling or a multimeter. The voltage measurement guide covers DC levels in detail.
3Zoom in with V/div and set the timebase
Reduce V/div until the ripple fills as much of the screen height as possible without clipping off the top or bottom. Then pick a timebase that shows two or three cycles of the ripple. Converting the reading is the same arithmetic as any voltage: divisions × V/div.
| V/div setting | Ripple height | Ripple (peak-to-peak) |
|---|---|---|
| 20 mV/div | 3 div | 3 × 20 mV = 60 mV |
| 10 mV/div | 4.4 div | 4.4 × 10 mV = 44 mV |
| 50 mV/div | 2.4 div | 2.4 × 50 mV = 120 mV |
| 5 mV/div | 6 div | 6 × 5 mV = 30 mV |
For the timebase, start from the frequency you expect. Rectifier ripple at 100Hz has a 10 ms period, so 2 ms/div shows two cycles across the 10-division screen (2.4 cycles at 120Hz). Ripple from a 100kHz switching regulator has a 10 µs period, so 2 µs/div does the same job. If you don't know the frequency, measure it with the method in how to measure frequency with an oscilloscope.
4Turn on the bandwidth limit
Many scopes have a bandwidth-limit option in the channel menu, commonly 20MHz. It filters out high-frequency noise that is not part of the supply's ripple, which otherwise thickens the trace and inflates the peak-to-peak number. Many power supply datasheets state that their ripple and noise figure was measured with a 20MHz bandwidth limit, so using the same setting lets you compare your reading with the spec.
Ripple and switching spikes are different measurements
5Trigger and measure ripple peak-to-peak
Set an edge trigger on the ripple channel with the level near the center line, so the waveform stops drifting; see how to trigger an oscilloscope if it still rolls. On a switching supply, triggering on the switch node with a second channel often gives a steadier picture. Then take the reading in one of three ways:
- Count divisions from the highest to the lowest point of the trace and multiply by V/div, as in the table above.
- Use voltage cursors on the top and bottom of the ripple; the scope shows the difference as ΔV.
- Use the automatic Vpp measurement. It reports the highest sample minus the lowest, so it includes every spike on screen.
Leave the acquisition mode on normal sampling for the ripple figure. Averaging mode smooths away noise and makes ripple look smaller than it is, so use it only to study the ripple's shape. Also measure under a realistic load: many supplies show more ripple at higher output current.
What ripple looks like on typical supply outputs
The shape and frequency of ripple tell you what kind of supply you are looking at, and that decides the timebase. The rows below are typical patterns, not limits; for the acceptable amount, check the supply's or regulator's datasheet.
| Supply type | Typical ripple pattern | Timebase to start with |
|---|---|---|
| Unregulated linear (transformer, rectifier, capacitor) | Sawtooth at twice mains frequency: 100Hz on 50Hz mains, 120Hz on 60Hz | 2-5 ms/div |
| Linear regulator output (3.3 V, 5 V, 12 V rails) | The same 100/120Hz component, much reduced; may be close to the noise floor | 2-5 ms/div |
| Switching regulator (buck or boost) | Roughly triangular ripple at the switching frequency, plus narrow spikes at each switching edge | µs/div range |
| Mains wall adapter or USB charger (switching) | Switching-frequency ripple, sometimes with a 100/120Hz component as well | Check both ms/div and µs/div |
Switching frequencies vary widely between designs, commonly from tens of kilohertz to a few megahertz, so read the regulator's datasheet or measure it rather than assuming.
Ripple measurement checklist and common mistakes
The whole procedure in order:
01
Short ground
Spring at the output capacitor
02
AC coupling
Removes the DC level
03
Zoom V/div
Fill the screen height
04
BW limit
Often 20MHz, as in datasheets
05
Read Vpp
Divisions, cursors or auto
- Long ground lead left on: pickup is added to the ripple and changes when you move the lead.
- Probe ratio mismatch: a 10x probe on a 1x channel setting reads ten times too low.
- Averaging left on: spikes and noise disappear and the reading drops.
- No load: an unloaded supply often shows much less ripple than it will in use.
- Trusting AC coupling for slow signals: below its cutoff, it attenuates the signal, so slow drift reads low.
Which scope do you need for ripple measurements?
Rectifier ripple at 100 or 120Hz is within reach of almost any scope, including a pocket unit like the FNIRSI DSO152 (200kHz bandwidth). Switching supplies are different: their ripple sits at tens of kilohertz to megahertz and the edge spikes go much higher, so a bench scope such as the Rigol DS1054Z (50MHz) or the Siglent SDS1104X-E (100MHz) is the practical choice. Whichever you use, check its manual for the bandwidth-limit and AC-coupling options before you start. Compare all four models in our best beginner oscilloscopes roundup, or see what oscilloscope bandwidth means before you buy.
Bench scopes in our reviews
50MHz, 4 channels, 1 GSa/s, with a large 7-inch display and strong community documentation.
100MHz, 4 channels, 1 GSa/s: more bandwidth headroom than entry-level scopes.
As an Amazon Associate I earn from qualifying purchases. Prices and availability are shown on Amazon.
Frequently asked questions
Can I measure ripple voltage with a multimeter?
Only roughly. A multimeter on its AC voltage range can show that some ripple is present, but it reports an RMS figure, may not respond well at switching frequencies, and cannot show spikes or the shape of the ripple. Power supply ripple is normally specified in millivolts peak-to-peak, which needs an oscilloscope to read.
Should I use a 1x or 10x probe to measure ripple?
Each has a trade-off. A 1x probe gives ten times the sensitivity, which helps with ripple of a few millivolts, but its bandwidth is usually far lower. A 10x probe keeps more bandwidth but makes the scope's most sensitive range ten times coarser. Many people use 1x for low-frequency ripple and 10x for switching supplies.
Why does my ripple reading change when I move the probe ground lead?
Because the long ground lead is acting as an antenna. Its loop picks up magnetic fields from nearby switching currents and transformers, and that pickup adds to the reading. If the number changes as you move or reshape the lead, the extra is pickup, not ripple. Replace the long lead with a short ground spring at the output capacitor.
What is the difference between ripple and noise on a power supply?
Ripple is the repeating variation linked to how the supply works: the mains rectifier frequency on a linear supply or the switching frequency on a switcher. Noise is the less regular, higher-frequency content such as the narrow spikes at each switching edge. Many power supply datasheets quote a combined ripple and noise figure in millivolts peak-to-peak.
Why does ripple get worse as a power supply gets older?
The usual culprit is the output filter capacitors. Electrolytic capacitors slowly dry out with heat and age, so their capacitance falls and their internal resistance (ESR) rises, and both let more ripple through. A supply that once measured a few tens of millivolts can creep far higher. Bulging or leaking capacitors confirm it, but rising ripple often shows up before any visible damage.
