How to Read a Power Adapter Derating Curve Correctly

A power adapter derating curve shows how much output load the adapter may deliver as a condition such as ambient temperature, input voltage, airflow, or mounting orientation changes. The nameplate wattage is therefore not automatically available at every temperature in the operating range. OEM teams should read the curve for the exact model and configuration, then verify that the adapter stays inside the permitted region in the finished product.

The curve is a boundary, not a prediction of the temperature your application will reach. You still need the real load profile, local air temperature, enclosure conditions, cable losses, input range, orientation, and operating time. Those inputs turn a supplier graph into a usable design and acceptance decision.

What is a power adapter derating curve?

Illustrative power adapter derating curve showing full-load and reduced-load regions

A derating curve plots an allowed output quantity against an operating condition. For external AC/DC adapters, the vertical axis may show output power, output current, or percent load. The horizontal axis commonly shows ambient temperature, but a separate curve may address input voltage, altitude, orientation, or airflow.

The typical temperature curve has two regions:

A full-load region where the model can supply its rated output under the stated test conditions.
A derated region where allowable continuous output falls as temperature rises.

The endpoint may mean zero permitted output, the maximum specified ambient, or another limit defined by the manufacturer. Do not extend a plotted line beyond its stated range or assume that operation outside the curve is acceptable for a short time.

The curve must stay attached to its conditions. A graph for one output voltage, enclosure, cable, plug, or regional model does not automatically apply to another member of the family. Check the model number, suffix, output rating, input range, test orientation, and cooling statement on the same controlled document.

COSEL’s derating guidance shows that allowable output can depend on ambient temperature, mounting method, and input voltage. It also illustrates a supplier-specific method for combining two factors. Use that calculation only when the manufacturer defines the factors and how they interact.

Which temperature belongs on the derating curve?

Engineer measuring local air temperature beside an operating desktop power adapter

Use the temperature definition in the adapter specification. “Ambient” usually refers to air around the unit, but the measurement location, distance, stabilization time, and airflow condition can change the result. Room temperature, enclosure temperature, and adapter surface temperature are different measurements.

An adapter can sit in a 25 °C laboratory while the air trapped behind a display, inside a kiosk, or under a bench reaches a much higher value. Nearby processors, motors, batteries, lighting drivers, and sealed housings can warm the air entering the adapter. Sunlight, a heated wall, or restricted clearance can add another local source.

Case temperature is not a substitute for ambient temperature unless the supplier defines a case-temperature limit and measurement point. Surface readings also depend on emissivity, sensor attachment, location, and whether the reading comes from an infrared camera or a contact sensor. Record the method rather than comparing an unqualified case reading with an ambient-temperature curve.

For an OEM test, place calibrated sensors at the locations named in the supplier method and at any application hot spots identified during design review. Keep sensor wires from blocking vents or changing the adapter’s contact with the mounting surface. Wait for the required thermal condition or documented stabilization criterion before accepting the result.

How do input voltage, orientation, and airflow change allowable load?

Thermal lab comparing power adapters at different orientations with controlled airflow

Heat generation and heat removal depend on more than room temperature. Input voltage can change current and internal losses. Orientation can change natural-convection paths. Airflow can improve cooling, while an enclosure, cable bundle, foam insert, dust filter, or nearby surface can restrict it.

Treat each published curve as a defined configuration. Look for notes covering:

minimum and maximum AC input used for the curve;
natural convection or forced airflow, including airspeed and direction;
wall-mount, desktop, vertical, horizontal, or other orientation;
clearance around the enclosure and whether it rests on a surface;
output cable, connector, and load connection;
altitude or atmospheric-pressure limitations;
single-output versus combined-load rules;
warm-up time, duty cycle, and continuous-operation definition.

Do not assume that airflow from the finished device cools an external adapter. The adapter may be outside the product, behind furniture, inside an accessory compartment, or connected at a wall outlet with little useful air movement. If the design relies on forced air, specify how that airflow reaches the adapter and what happens when a fan slows, a filter loads with dust, or an installation blocks the path.

Advanced Energy’s external-adapter application note tells users to keep an adapter exposed to surrounding air and follow the model’s derating curve at elevated temperature. That is a model-family example, not a universal temperature limit for other adapters.

How do you calculate the allowable output from a derating curve?

Engineer calculating allowable adapter load from a model-specific derating graph

Start with the exact rated quantity used by the graph. If the vertical axis is percent load, multiply the model’s rated output by the permitted fraction at the required condition. If the graph gives watts, use the watt limit directly. If it gives amperes, do not convert through a different nominal voltage unless the supplier’s specification permits that interpretation.

For an illustrative 24 V, 90 W adapter rated at 3.75 A, a model-specific curve that allows 60% load at the required condition would set these screening values:

Permitted output power: 90 W × 0.60 = 54 W
Equivalent current at 24 V: 54 W ÷ 24 V = 2.25 A

This arithmetic does not establish a limit for any Merryking model or another supplier’s product. The actual curve may use a different axis, breakpoint, slope, input condition, cooling condition, or output-voltage rule.

When more than one derating factor applies, follow the manufacturer’s combination method. Some suppliers multiply independent factors; others provide a combined curve, a table, or a more restrictive rule. Choosing the lowest percentage, multiplying percentages, or adding reductions can produce different answers. The datasheet or a controlled engineering response must state which method is valid.

Compare the result with the complete device load profile. Continuous load, startup demand, short peaks, charging intervals, and repeated duty cycles do not have the same thermal effect. If the supplier permits an overload or short-duration excursion, record its duration, repetition rate, recovery condition, and temperature boundary separately from the continuous derating limit.

Is the operating-temperature range the same as the full-load range?

Power adapter operating in an environmental chamber during a temperature and load test

No. An operating-temperature range may include conditions where the adapter can run only at reduced load. A label or datasheet line such as “operating temperature: −20 °C to 70 °C” does not by itself state that full rated power is available across that span.

Separate these specifications:

Storage temperature: the non-operating range allowed under stated storage conditions.
Startup temperature: the range in which the adapter is expected to start, sometimes with a warm-up or performance qualification.
Operating temperature: the range in which operation is permitted under the accompanying load and cooling conditions.
Full-load temperature range: the region where rated continuous output is available.
Derated region: the region where output must be reduced.
Case or component limit: a temperature measured at a defined point that may constrain the application.

Low-temperature operation also needs model evidence. Capacitor characteristics, startup timing, ripple, and regulation can change before the unit warms. Do not infer cold-start performance from a high-temperature derating graph.

Thermal protection is not a replacement for staying inside the curve. A protection circuit may shut down, reduce output, latch, or restart in a way that disrupts the device. Its threshold and recovery behavior can vary with design and tolerance. Normal production operation should meet the specified thermal boundary without depending on protection activation.

How should an OEM validate derating in the finished product?

OEM engineer validating adapter and device temperatures with thermocouples under full system load

Build a thermal test around the worst credible installation, not an uncovered adapter on an open bench. Use production-intent hardware, the specified cable and connector, representative mains input, device operating modes, and the orientation and clearances users will have.

A useful validation matrix records:

Adapter model, suffix, sample ID, hardware revision, label, and cable.
AC input voltage and frequency, including the condition that produces the relevant thermal stress.
Device load modes, continuous load, peak load, duty cycle, and test duration.
Chamber or local ambient temperature, sensor position, stabilization rule, and uncertainty.
Adapter orientation, supporting surface, clearances, enclosure, airflow, and nearby heat sources.
Output voltage at the adapter and at the device when cable drop matters.
Adapter case points or internal points only when the supplier defines them.
Functional behavior, temperature results, protection events, and post-test inspection.

Run enough time to reach the specified steady condition or the worst point in a documented duty cycle. A short functional check can miss slow thermal rise. Conversely, a chamber setpoint is not proof that the local air around the adapter has stabilized at that value.

The site’s power adapter qualification test plan can tie the thermal matrix to sample control, failures, retests, and release status. If the same project also has millivolt-level output concerns, keep the thermal run linked to the separate ripple and noise measurement method because temperature and load can affect the captured waveform.

What belongs in the RFQ and acceptance record?

Buyer and engineer reviewing adapter thermal curves and an OEM acceptance checklist

Ask for more than a maximum operating temperature. The RFQ should give the supplier enough application detail to identify the correct model and thermal boundary.

Include:

target countries, AC input range, frequency, and installation altitude;
required output voltage, continuous current, peak profile, and duty cycle;
maximum local ambient temperature at the adapter;
wall-mount or desktop format, orientation, clearance, and surrounding materials;
natural-convection or forced-air condition and any credible blocked-airflow case;
enclosure or compartment details if the adapter is not fully exposed;
cable length, conductor requirement, connector, and allowable voltage at the device;
required derating curve, temperature definition, model identity, and revision;
test matrix, sensor locations, stabilization rule, acceptance limits, and report format;
change-control triggers for enclosure, transformer, semiconductor, capacitor, PCB, cable, or thermal-material changes.

The acceptance record should bind the curve to the ordered SKU. Match the model and suffix across the quotation, datasheet, drawing, label artwork, sample, test report, and purchase order. Record whether the evidence is typical characterization or a guaranteed specification and whether the limit applies to every unit, a family, or one evaluated configuration.

Avoid adding an arbitrary “20% safety margin” to every project. Engineering margin can be appropriate, but its size should follow the device load uncertainty, ambient uncertainty, installation variation, production spread, reliability target, and supplier evidence. A lightly loaded adapter is not automatically the best answer either; size, efficiency at the actual load, no-load behavior, cost, and device startup requirements remain part of the selection.

For a model-level review, send Merryking the target output, load profile, local ambient, input range, orientation, airflow, enclosure, cable, and required acceptance method. The response should identify the applicable adapter model, its controlled thermal data, and any sample testing needed before production approval.