Power Adapter Hold-Up Time: A Practical OEM Test Guide

Power adapter hold-up time is the interval that the DC output remains within a defined acceptable range after the AC input is removed or falls below the level needed for normal conversion. It is not a universal promise that an adapter will keep a device running through every mains disturbance. The result depends on the exact model, input voltage, output load, voltage limit, temperature, unit history, and measurement method.

For an OEM project, the useful question is not simply “How many milliseconds does it have?” It is whether the adapter and the powered device remain in an acceptable state during the interruption that the finished system is expected to encounter.

What does power adapter hold-up time actually measure?

Power adapter hold-up time diagram showing the input removal and output-voltage limit

Hold-up begins at a defined input event and ends when the output crosses a defined limit. In an AC/DC adapter, energy stored mainly in the input-side bulk capacitor and other reactive components can continue feeding the converter briefly after the mains input disappears. The available interval is finite, and the output eventually falls as that stored energy is used.

Three points must travel with any result:

Start point: for example, the commanded removal of AC or a specified point on the input waveform.
End point: the minimum acceptable DC output voltage, not an arbitrary point chosen after viewing the trace.
Operating condition: input voltage, output load, temperature, cable, and the adapter’s stabilized state.

Hold-up time is different from startup time, output rise time, load-transient recovery, and battery backup duration. A UPS or an energy-storage subsystem may bridge a long interruption; an adapter’s hold-up interval is normally a short ride-through characteristic. Keep those requirements separate in the product specification.

The exact duration is model-specific. A supplier datasheet may state a typical value under one line and load condition. That value is useful only when the project’s conditions and acceptance boundary are compatible with the stated test.

Why do input voltage and output load change the result?

Engineer comparing hold-up behavior at different input and output load conditions

The energy available before the interruption and the rate at which the load consumes it both matter. A lower AC input can leave the storage stage with less usable energy before input removal. A heavier DC load draws stored energy faster. For that reason, a result measured at high line and light load should not be treated as a guaranteed result at low line and full load.

A defensible test matrix records at least:

Variable	Conditions to define	Why it matters
AC input	Required low, nominal, and high values; frequency	Changes pre-interruption operating state and stored energy
DC load	Minimum, typical, maximum continuous, and relevant dynamic state	Changes the energy drain and device tolerance
Output limit	Minimum voltage at the adapter connector or device input	Establishes the actual end point
Temperature	Ambient, warm-up time, and hot or cold state	Can change component behavior and available margin
Cable and connector	Production length, conductor, mating interface, measurement point	Voltage drop can move the device below its limit before the adapter terminal does
Interruption	Removal method, phase relationship, duration, and repetition	Makes captures reproducible

The output power relationship also matters. A device drawing constant power can demand more current as its input voltage falls, accelerating the decline. A resistive electronic load may therefore fail to reproduce the finished device. Where the load behavior changes with voltage, use the real device or a programmed profile that represents it.

Do not infer hold-up performance from capacitor size alone. Converter topology, control thresholds, efficiency, input range, output capacitance, protection behavior, and load characteristics all affect the usable interval.

How should an OEM measure hold-up time safely?

Qualified engineer measuring power adapter hold-up time with a controlled test bench

Testing combines hazardous mains voltage with oscilloscope measurements. Use qualified personnel, an appropriately rated and isolated test arrangement, correct differential probes, and protective equipment. Do not clip an earth-referenced probe ground onto a live mains node.

A repeatable procedure is:

Identify the exact adapter model, rating, hardware revision, sample, cable, and load.
Stabilize the unit at the specified input, load, orientation, and temperature.
Monitor input voltage, DC output voltage, and—when useful—output current on a common timebase.
Remove or interrupt the AC input using a controlled method suitable for the test voltage and current.
Measure from the defined input event to the point where the monitored output crosses the stated limit.
Repeat at each required line, load, and thermal condition and across enough samples for the project’s risk.
Save the waveform and instrument settings, not only a typed millisecond value.

Probe at the location used by the acceptance requirement. An adapter-terminal capture can pass while a long cable and connector cause the voltage at the device to cross its limit earlier. If the device input is the system boundary, measure there or record both points.

The AC source or switch must produce a known interruption. Contact bounce, source ramp-down, residual coupling, and unsynchronized switching can make two nominally identical tests different. Capture the input waveform so the test event is visible beside the output response.

The site’s power adapter qualification test plan provides the broader framework for sample identity, test conditions, failures, retests, and release decisions.

How do you turn a waveform into an acceptance decision?

Engineer reviewing a power adapter hold-up waveform against device requirements

A waveform is evidence, but the end product defines what counts as acceptable. Establish the device’s minimum operating voltage and the longest interruption it must tolerate. Also decide whether the requirement is uninterrupted operation, controlled shutdown, data retention, communication recovery, or simply no unsafe behavior.

Review these outcomes:

output voltage stays above the device limit for the required interval;
the device does not reset, corrupt data, chatter a relay, or enter an undefined state;
output recovery after AC returns is orderly and does not create unacceptable overshoot;
repeated interruptions do not expose an unexpected restart or protection interaction;
the adapter remains within its own specified operating conditions.

Avoid using a single cursor measurement as the pass rule when the output contains ripple or switching noise. Define how the threshold crossing is determined and whether brief excursions count. The same discipline used for power supply ripple and noise measurement helps prevent probe and bandwidth choices from deciding the result accidentally.

If the required ride-through time exceeds what a practical adapter can provide, the system solution may belong elsewhere: local capacitance evaluated with the load, controlled power-fail signaling, a battery or supercapacitor subsystem, or a UPS. Adding capacitance without analysis can change inrush current, startup, protection, discharge time, and safety behavior.

What belongs in the RFQ and production record?

OEM team reviewing hold-up evidence for a power adapter and powered device

“Long hold-up time” is not a testable purchasing requirement. Give the supplier the exact scenario and ask for conditioned evidence.

Include in the RFQ:

input-voltage and frequency range, including the worst required line condition;
nominal output, continuous load, and the real device behavior during voltage decay;
minimum permitted voltage and the measurement location;
interruption duration, repetition, and required device outcome;
cable, connector, temperature, orientation, and warm-up conditions;
sample quantity, waveform channels, timing definition, and report format;
recovery behavior after the input returns;
exact-model binding and change-control expectations.

The acceptance record should connect the waveform to the tested model, sample, revision, test date, instruments, setup, and limits. A family datasheet or a report for a nearby wattage cannot prove the production configuration. If a supplier changes the input capacitor, transformer, control IC, output capacitor, protection thresholds, or relevant firmware, require engineering review when the change can affect hold-up behavior.

Hold-up time also trades against size, cost, inrush, efficiency, and thermal design. The best choice is not automatically the adapter with the largest number. It is the model that meets the finished product’s defined interruption requirement while satisfying the complete electrical, thermal, mechanical, and compliance specification.

For a model review, send Merryking the target markets, input range, output rating, device load profile, minimum device voltage, cable details, interruption scenario, and required evidence. The engineering review can then determine whether an existing platform should be tested or whether the requirement calls for a different system approach.