Power supply ripple acceptance criteria depend on both the connected load and the measurement method. Set one criterion for the supply output and another for the receiving circuit, then compare each result with its own limit under defined conditions. A source can pass its specification while the installed load fails, as the worked 24 V example below shows.
Table of Contents
Set separate limits for the supply and the load
Before putting a ripple limit into an OEM purchase specification, define the quantity being measured: peak-to-peak voltage, RMS noise, or noise within a particular frequency range. A number without its metric and method leaves room for different interpretations.
The referenced ripple-and-noise training note distinguishes peak-to-peak and RMS reporting and explains why consistent measurement conditions matter. Apply that distinction to the requirement, rather than selecting whichever result looks smaller.
The supplier’s output-terminal limit applies to its documented test configuration. Your load requirement applies to the receiving circuit with the project’s harness, fitted components and operating states. Record both. For bandwidth, probe connections and output-loading setup, use WEHO’s ripple measurement guide. This article turns those measurements into separate source-terminal and receiving-circuit pass/fail decisions.
Check the receiving device’s steady-state voltage window alongside its ripple requirement so the supply and distribution path satisfy both.
Build the pass/fail matrix before testing
In this hypothetical acceptance example, a project uses a 24 V supply at 230 V AC input, a steady 5 A total load, and 25 °C ambient. The source requirement permits 120 mV peak-to-peak at test point TP1; the receiving circuit permits 80 mV peak-to-peak at TP2.
The example compares the readings directly with their limits. For an actual test, agree in advance how measurement uncertainty and near-limit results will be handled.
| Requirement and owner | Point and method | Assumed limit | Example reading | Decision |
|---|---|---|---|---|
| Supplier conformity; supplier quality | TP1 supply output, agreed method S1 | ≤120 mVp-p | 96 mVp-p | Pass this defined requirement |
| Equipment suitability; equipment engineering | TP2 receiving circuit, agreed method E1 | ≤80 mVp-p | 104 mVp-p | Fail this operating-state requirement |
| Additional diagnostic capture; test engineer | TP1, different bandwidth from S1 | No acceptance limit assigned | 150 mVp-p | Diagnostic only; cannot judge S1 conformity |
Give methods S1 and E1 a setup sheet recording bandwidth, coupling, probe connection, acquisition settings, capture duration, measurement algorithm, output components and exact test point. Both example methods use a 20 MHz upper bandwidth limit. Select the actual bandwidth from the applicable requirement and method.
S1 uses the agreed source fixture. E1 uses the actual project harness and fitted load components. Identify any capacitors and filters in each configuration; do not silently add a measurement capacitor to make a result pass.
Worked result: source test passes, load test fails
At TP1, 96 mVp-p is below the assumed 120 mVp-p source limit by 24 mV: the source requirement passes. At TP2, 104 mVp-p exceeds the assumed 80 mVp-p equipment limit by 24 mV: the equipment requirement fails.
Next decision: Keep the source pass in the supplier record and resolve the TP2 failure before equipment release. Confirm method E1 and the receiving-device requirement, then compare supply, harness or load changes against the same acceptance test.
Do not subtract the two peak-to-peak readings and call the difference “cable noise.” Separate captures and measurement networks do not establish the origin or phase relationship of their waveform components. Investigate that question with an appropriate diagnostic plan.
Compare supplier quotes on the same measurement basis
In a hypothetical quote comparison, Supplier A lists 100 mVp-p with a stated fixture, while Supplier B lists 80 mVp-p without a method. Obtain B’s method before ranking the two figures, and ask both suppliers which operating states their data covers.
The external application note AN1105 explains that reported ripple and noise depend on setup and measurement point, and that load and temperature matter when comparing specifications.
For initial candidate screening, the public WEHO NDR-240 specification lists a 150 mVp-p ripple-and-noise entry (labeled “Wave and Noise” on that one-page sheet). When preparing a purchase specification, request the applicable bandwidth, probe connection, external capacitor network and document revision for the exact model and operating conditions. Compare the response with your source and receiving-circuit requirements before sample acceptance.
Follow a clear sequence when results disagree
Before changing the power supply or adding a filter, check whether the results answer the same question:
- Match the method. Check the point, bandwidth, probing, acquisition and fitted components. A changed setup calls for a repeat of the agreed method or an agreed method revision.
- Match the operating state. Confirm the input, load, output setting, temperature and stabilization conditions against the requirement.
- Apply the relevant limit. Compare TP1 with its source criterion and TP2 with its equipment criterion. Keep a wider-bandwidth diagnostic trace separate.
The referenced probing application note describes how lead geometry and probe sensitivity can affect low-level measurements. Keep a setup photograph with the record so a second laboratory can reproduce the connection.
For a hypothetical reading of 119 mVp-p against a 120 mVp-p limit, the agreed decision rule matters. Record measurement uncertainty and repeatability, then apply the rule consistently. Changing averaging until the display drops below the limit changes the test rather than resolving the result.
If a supply change is still needed, use the replacement compatibility checklist to review startup, status interfaces and recovery behavior alongside ripple performance.
Save a test record that another lab can reproduce
Download the ripple acceptance matrix with the source and equipment examples. Replace the assumed values and complete the method fields for your project.
Retain one row per requirement and operating state, with these fields:
- Identity: project, sample serial or lot, complete model, revision, and document version.
- Requirement: owner, measurement point, metric, limit, and decision rule.
- Condition: input, output setting, load state, ambient, stabilization criterion, harness, and fitted components.
- Evidence: method ID, instruments, settings, setup photograph, waveform file, observation, and disposition.
Keep each result attached to its tested configuration and operating state. Extend the plan to the input, load and temperature conditions the equipment must cover. Define startup and load-step tests separately from the steady ripple measurement. Keep the continuous-current budget from the 24 V DIN rail power supply sizing guide alongside this record; sufficient current capacity alone does not establish ripple acceptance.
Questions about ripple acceptance tests
Can RMS ripple be converted directly to peak-to-peak?
Not with a universal multiplier for an unknown waveform. Request the specified metric and method; narrow spikes and waveform shape affect the relationship.
Does adding a capacitor invalidate the test?
A capacitor belongs in the test if the agreed method or installed design includes it. An undocumented addition changes the configuration and cannot silently replace the original acceptance test.
Does a ripple pass establish EMC compliance?
No. It closes only the defined ripple requirement. Keep equipment EMC evidence and its applicable test requirements as a separate assessment.
