Power supply trips when enclosure hits 60 C: find the cause

Short answer: the trips are derating, and the arithmetic is checkable. A 24 V, 5 A DIN-rail supply holds its full 5 A only to 40 C [104 F]. At 60 C [140 F] the curve most makers publish lets it deliver about 70% of that, 3.5 A. Above that line the output folds back, the bus sags, and a system sized at room temperature faults intermittently. Log the peak temperature inside the box, read the supply's curve at that peak, then either size the supply for the hot point or take watts out of the box. The checks below run in the order you would do them on the bench, numbers included.

Why does heat make a power supply fail?

Heat cuts the current a supply can deliver: most 24 V DIN-rail units hold full current to 40 C and fall to about 70% by 60 C on the curve the maker publishes, so a room-temperature load outruns the supply when the box runs hottest.

Two facts sit behind that curve. Makers set the 100% point at 40 C because a convection-cooled unit is already close to its thermal design limit there. They draw the curve from a maximum internal temperature, usually specified below 125 C, which is why it bends through your working range instead of stopping at it. Treat 40 C as the full-load point until your own datasheet says otherwise.

The failure mode explains the intermittency. A derating supply does not switch off: it folds the output back, the bus sags, and the most sensitive load resets or raises a fault. When the box cools, the full 5 A returns and a reset clears everything. Hours of heat, a fault, then a clean reset is the signature, and it repeats for as long as the heat does.

What do the symptoms tell you before you measure?

The pattern of the trips narrows the cause before you touch a meter: trips that follow hours of hot running point at derating, while trips that also happen cold point at connections, the supply itself or its upstream wiring. Sort your case against the list below.

• Trips only after hours of heat, resets clean when cool — thermal derating, in the supply or in a neighbor on the same rail.
• Everything alarms at once but the supply's LED stays green — the output folded back and the bus sagged; the supply never died.
• The backup or battery module drops out first — backup modules often carry a 60 C ceiling on their datasheets, lower than the supply's curve end.
• Peak logged inside sits above the curve's last printed point — the box carries more watts than the supply alone explains.
• Trips also happen on a cool morning — probably not heat: work the terminal torque, the unit and its upstream wiring.

For a whole cabinet running hot rather than one supply, the same causes-in-order method behind the backplate and DIN-rail layout guide still applies: measure first, then work the list.

How much current is left at 60 C?

Read the derating curve, not the nameplate: a 24 V 5 A unit that holds 5 A to 40 C gives about 3.5 A at 60 C on the curve most makers publish, and your measured load has to sit under it with margin.

The curve lives on the datasheet next to the environmental data. Its x-axis is ambient temperature at the unit's intake, and its y-axis is percent of rated current or watts. Most curves hold 100% to 40 C, bend to about 70% at 60 C and end near 50% at 70 C. The temperature you read against is the logged peak at the intake, not the room outside the box.

Measure the load with the door closed. A DC clamp meter around the positive output lead gives the running amps; one with min/max hold catches the peak overnight. A cheap logger at the output bus tells you what the voltage did at the moment of the trip.

Worked example. These are example inputs; put in your curve's percentage and your own measured amps before you order anything.

• Rated output: 24 V × 5 A = 120 W, good to 40 C.
• Curve at the logged 60 C peak: 70%, so 5 A × 0.7 = 3.5 A available.
• Measured load: 4.5 A — 1.0 A past the derated output, so every hot cycle folds the supply back.
• Size to order: 4.5 A ÷ 0.7 = 6.4 A, which puts you in the 10 A class. At 60 C that unit still holds 10 A × 0.7 = 7 A, a 2.5 A margin over the load.

The bigger unit does not cook the box. At half output its losses track the watts delivered, so its waste heat stays close to the 5 A unit's.

How hot is the inside of the box?

Log the peak with a max/hold thermometer or a small logger set within 50 mm of the supply's intake, closed up through a full hot cycle: that peak, not the room and not the wall, is the point the curve is read at.

Set the sensor, close the door, and let it ride one full hot cycle. Read the room ambient at the same time. The gap between the two is the box's own temperature rise, and the rise tells you whether the problem lives in the box or in the room. Log the other points the same way you would when you size an electrical enclosure: room, wall and intake together.

Then balance the watts against the walls. The wall transfer values used across this site's sizing guides are 3.7 W/m²K for stainless, 5.5 W/m²K for painted steel and 3.5 W/m²K for plastic.

Worked example. Same box, same cycle: a 16 × 12 × 8 in [400 × 300 × 200 mm] plastic box peaks at 60 C [140 F] with the room at 38 C [100 F].

• Wall area: 2 × (0.4 m × 0.3 m) + 2 × (0.4 m × 0.2 m) + 2 × (0.3 m × 0.2 m) = 0.52 m².
• Rise over the room: 60 C − 38 C = 22 K.
• Heat leaving through the walls: 3.5 W/m²K × 0.52 m² × 22 K = 40 W.
• The supply's waste heat at 120 W out and 88% efficiency: 120 W ÷ 0.88 − 120 W = 16 W.
• Unaccounted heat: 40 W − 16 W = 24 W. Something else is burning watts in there, so find it before you buy hardware.

What is heating the box beyond the room?

Add up the watts: in the worked box, holding 60 C [140 F] at a 38 C [100 F] room pushes 40 W through the walls while a 120 W supply at 88% efficiency wastes about 16 W, so the remaining 24 W comes from another heat source.

Work the causes in this order; the first two cover most hot boxes.

1. A crowded rail. Convection needs air. Manuals commonly ask for 25 to 50 mm of free space above and below the housing and empty rail at the sides. Check your manual, then check the rail.
2. A second heat source nobody counted. A battery module, a second supply, an HMI or a transformer adds its own watts. An infrared thermometer sweep at peak heat finds them in minutes.
3. Wall gain from outside. Sun on the box or hot plant against the wall adds to the rise. Hold the infrared gun on the outside wall at peak: a wall above shade temperature is heat load, not scenery. The same physics drives outdoor electrical enclosures in summer.
4. Terminals and wiring. A resistive lug is a small heater; the same infrared sweep shows it as a spot running hot against its neighbors.
5. Not heat at all. Anything that trips with the box cool lives in connections, the unit or the wiring feeding it.

One test splits box heat from the rest: run the next hot cycle with the door open and a fan moving air across the contents. Trips stop, and the problem is heat inside the box. Trips continue at a cool box temperature, and you are chasing connections or a failing unit. Keep the test short, attended and behind a barrier — an open, energized panel is a shock hazard.

What fixes the trips without giving up the seal?

Fix the load first: the example needs 4.5 A at 60 C, so a 10 A supply that still gives 7 A there covers it with margin, and running a 24 V supply well below its rating is safe.

• Size the supply to the curve, as the example shows. The 6.4 A result lands you in the 10 A class, and oversizing a 24 V PLC supply carries no reliability penalty of its own.
• Give back the convection. One empty module each side of the supply, 25 to 50 mm clear above and below, and the hottest unit placed so its rising plume passes other devices rather than into them.
• Move the batteries to the coolest spot. Sealed lead-acid float life halves for roughly every 10 K above 25 C, so a battery rated 5 years at 25 C lasts months at 60 C.
• Cool the sealed box without opening it. A plate heat exchanger or an enclosure cooling unit moves watts out and holds the box's manufacturer rating. An open fan-filter or a row of drilled holes does the opposite: the assembly then carries the lowest-rated part's rating, and washdown fails there first. A clogged exchanger quietly undoes the fix, so give the exchanger core the same airflow check you would give a NEMA 4X versus IP66 gasket: verify the weakest part, not the box.
• Keep the open door as a test. It is a one-hour diagnostic, never the permanent fix.

When do you re-specify the enclosure instead?

Re-specify when the room itself is the problem: at a steady 38 to 45 C [100 to 113 F], spacing and a bigger supply trim the peak but cannot restore curve margin, so the answer is a bigger sealed box with active cooling and the right wall.

That last row carries its number. Plastic walls pass heat at 3.5 W/m²K and stainless at 3.7 W/m²K, so a stainless wall lowers the internal rise by 1 − 3.5 ÷ 3.7 = 5%. Wall metal is rarely the fix; the watts and the air are.

When the decision is a new box, ours is built your way. You set the material — 304 or 316L stainless, carbon steel, aluminum, polycarbonate or fiberglass — plus the width, height, depth and every cutout, from one piece. Each box is manufacturer-rated to the Type you specify and delivered in 10 business days from your approval of the drawing. Size the box with the enclosure sizing guide, shortlist it in the enclosure selector, and send the finished layout from the custom enclosures page. It goes into the queue the same day you approve the drawing.

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