Derating a power supply inside a plastic enclosure

Short answer: load the supply to 60-70% of its nameplate (12-14 A on a 5 V 20 A unit), turn that load into a heat number, and size the walls or a small fan so the air inside stays at or below 40 °C [104 °F]. The worked example assumes an 80% efficient open-frame supply, a continuous load at the bottom of that band and a plastic box. Swap in your datasheet's efficiency, your measured load and your wall area before you order anything. The three sums take minutes and decide between a bigger box, a 2.6 CFM fan, or no fan at all.

How much current can a closed-box supply feed?

Keep the continuous draw at 60-70% of the nameplate, so a 20 A supply feeds 12-14 A. The label assumes free air at 40 °C [104 °F], and a closed box gives the supply none of that margin.

The nameplate figure assumes free air at 40 °C [104 °F], the ambient most electronics are rated for. Seal the supply in and its own losses heat that air, so the usable figure falls. Inexpensive open-frame units also run thin internal margins, which is where the 60-70% working range comes from. A unit with a published derating curve and convection fins takes 80-90% comfortably. Duty matters too: a load that runs for hours sits at the conservative end, while a few minutes per hour can ride higher. The wiring and clearance side is in mounting a 5 V power supply inside an enclosure.

How many watts of heat does the supply leave in the box?

Heat equals output watts times (1 ÷ efficiency - 1), so a 60 W load on an assumed 80% efficient supply leaves 15 W: 60 ÷ 0.80 = 75 W in, and 75 - 60 = 15 W warms the box.

Treat the 80% as the example's assumption, in the usual band for this supply class near half to full load; the datasheet curve at your load is the number that counts. Efficiency often dips at light load, so a 4 A draw can leave proportionally more heat than the table shows.

Every figure after this point uses that example heat, so when your datasheet says 88% instead, divide each value by 0.88/0.80 before you size anything. The same sum for a whole panel, supply plus controller plus relays, is in calculating enclosure heat load from component losses.

How hot does a sealed box get?

A sealed box rises by heat ÷ (k × wall area), so 15 W in a 12 x 10 x 6 in [305 x 254 x 152 mm] plastic box with 0.325 m² of wall climbs 15 ÷ (3.5 × 0.325) ≈ 13 °C [23 °F].

That 13 °C sits on top of the room: a 25 °C [77 °F] bench leaves the internal air near 38 °C, just under the full-rating point. The same heat load in the 10 x 8 x 4 in 304 stainless enclosure rises 21 °C. Wall material barely moves the answer, because the guides' wall figures are 3.5 W/m²K for plastic and 3.7 W/m²K for stainless; area and airflow decide.

For flame grade, bonding, stiffness and cost, see plastic vs metal electrical enclosures. The case for a metal box over a plastic one lays out the grounding argument in detail. The full sizing method lives in the guide to sizing an electrical enclosure. Our 12 x 10 x 6 in 304 wall-mount enclosure carries the example supply on its mounting panel, with room left for the terminals. The same box in 316L suits washdown and coastal air, and the 16 x 12 x 6 in 304 box drops this load to a 9 °C rise.

When does the derating curve cut the available current?

Most open-frame ratings hold full output to 40 °C [104 °F] and then fall linearly. A curve that loses 50% by 70 °C [158 °F] gives up about 1.7% per °C, so air at 48 °C leaves roughly 17 A of the rating.

Read the curve rather than assuming it: some units hold full output to 50 °C [122 °F], and the slope flattens near the end. The example's 50-point loss over 30 °C works out to about 1.7% per °C. At 48 °C inside, the 35 °C room plus the 13 °C rise, the supply still offers about 17 A, comfortably above the 12 A load. A 17 A continuous load at that temperature fails the same check. The curves also bend for altitude and mounting orientation, and the datasheet shows both.

When does the box need a fan, and how big?

Add airflow when the sealed rise would push the internal air past the full-rating point. Size it by the standard rule CFM ≈ 1.76 × watts ÷ allowed °C rise, which turns the example's 15 W at 10 °C into a 2.6 CFM [4.4 m³/h] fan.

The rule comes from the specific heat of air; in °F form it reads CFM ≈ 3.17 × watts ÷ °F of rise, the same figure. Any small fan covers 2.6 CFM, so the air path matters more than the volume. Run intake low on one face, exhaust high on the other, fins lined up across the flow, and keep at least 5 mm [0.2 in] between the supply's case and every wall. A vented hobby box has no rating to lose, but a stainless box manufacturer-rated NEMA 4X / IP66 (not UL listed) keeps its rating only while every penetration stays sealed. Fit a fan-and-filter with a documented rating to match, or keep the box sealed and let the walls carry the heat. The fan airflow calculation runs the same sum for larger loads.

How do you build it, step by step?

Mount the supply on standoffs with its fins across the air path, wire the AC side shrouded and bonded, fuse the DC output, and settle the box or fan choice from the sums above before you cut anything.

1. Read the datasheet: note the full-rating temperature (40 or 50 °C), the derate slope, and the efficiency at your load.
2. Add up the continuous load current, or clamp-meter the 5 V run; aim for the 60-70% band from the first section.
3. Compute the heat: output watts × (1 ÷ efficiency - 1); the table above gives the figure for each load.
4. Compute the wall area and the sealed rise from the table above, then add the site's warmest ambient.
5. If the total lands at or below the full-rating point, go sealed; if it misses, step up a box size or fit the 2.6 CFM fan.
6. Mount the supply on standoffs, fins across the airflow, 5 mm [0.2 in] clear of every wall, and keep any plastic flame-rated UL 94 V-0, the grade whose specimens self-extinguish within about 10 seconds of each ignition.
7. Set the input selector to the mains you have before the first plug-in, land the protective earth on the supply's PE terminal, and bond a metal box so fault current has a metal path; NEC 250.96(A) calls for metal parts serving as grounding conductors to be bonded together (grounding and bonding).
8. Fuse the 5 V output above its running current: the 15 A fuse from the table, at the supply terminal.

How do you check the finished box?

Run the full continuous load for an hour at the warmest temperature the site sees. The air next to the supply should read at or below 40 °C [104 °F], with the rise close to the calculated 13 °C.

• Air beside the supply reads 40 °C [104 °F] or less after an hour at full continuous load.
• The measured rise above room air lands within a few degrees of the calculated 13 °C; a big miss means blocked airflow or a heavier real load.
• On a fanned box, the exhaust-to-intake air difference sits near the designed 10 °C.
• Output voltage at the load holds 5.0 V with everything running, with no droop at peak.

Measure at the supply, not at the far wall: where to measure ambient temperature inside an enclosure shows the spots that mislead.

Which mistakes cost the rating?

The costly mistakes are loading to the nameplate inside a tight sealed box, blocking the fin-to-fin air path, and treating the label's full rating as a continuous enclosed figure. Any of them turns the example's 15 W design into the 25 W of the table's last row.

• Loading to the nameplate: full output leaves 25 W against the example's 15 W, two-thirds more heat. Keep the continuous draw in the 60-70% band.
• Supply flat against a wall or lid: standoff it 5 mm [0.2 in] and line the fins with the flow.
• Unrated plastic: choose UL 94 V-0 plastic or a stainless box.
• Slots cut into a rated box: the assembly drops to the vent's rating, so seal it or fit a documented fan-and-filter.
• No output fuse: fit the 15 A fuse at the terminal so a short clears at the source.
• Bench check at 20 °C only: repeat it near the site's warmest ambient, because 10 °C of room temperature eats most of a 13 °C margin.

Re-run the three sums whenever an input changes. The box that stays under 40 °C at the example load needs a fan once the draw reaches 17 A.

Catalog dimensions to compare

Compare these catalog dimensions with your component layout. Open each enclosure for its rating, material and usable space, then check those specifications against your application.

Questions