Short answer: work four causes in order — the fan itself, the filter and airflow path, the heat budget, and last a false alarm from dirty board contacts. Measure the stopped fan's supply first: 24 V DC at the connector with the blade still means a new fan, and no voltage means the supply or its fuse. Run the heat budget before you trust the cabinet again: a 20 x 16 x 8 in stainless box sheds about 2.9 W per kelvin of rise, so 150 W inside at 35 C ambient reaches 87 C sealed. And when your overtemp flag lands with a communication fault on the same controller card, pull that card, clean its edge connector and reseat it — the alarm can clear while the cooling problem survives.
What each symptom points at
- One fan stopped, the others running. The compartment that fan serves has lost its cooling while its neighbors still look healthy. Expect that fan's bearings first, then its branch supply.
- Overtemp plus a communication fault on one card. The card reporting its own temperature is also losing its bus, and one oxidized edge connector can raise both flags. Jump to the contact check early.
- Temperature creeping over days, alarm late in the shift. That is a filter loading up: the cabinet recovers overnight and fails again during the warm hours.
- Alarm on the season's first warm day. The cabinet ran marginal against 20 C ambient all winter and crossed its limit at 35 C. The load grew or the fan weakened; the weather only revealed it.
- Alarm text with no temperature value. Treat the message as a switch, not a thermometer, until you measure the air at the exhaust outlet yourself.
Likeliest cause: the fan and its supply
Live measurement on an energized cabinet is qualified-person work with rated probes. Lock out and prove zero voltage before you touch any wiring.
With the fan called to run, put your meter across its connector. A typical 4 in [120 mm] 24 V DC cabinet fan draws 0.2-0.4 A, so expect its rated voltage within 10%. Full voltage on a still blade convicts the fan: enclosure fans are brushless DC or EC units with no brushes to wear, so a stall with supply present means bearings or the internal driver. Replace the unit. To reverse one, turn it around in its mount — swapping the two leads can damage a brushless fan instead of reversing it.
Low or missing voltage points upstream: an unseated connector, the branch fuse, or the 24 V supply itself. Check that fuse with power isolated, and fit the specified type and rating only after you know why it opened, because a seized rotor opens the next one too. When several fans stop together, read shared-fault diagnosis for control cabinet fans first — one supply feeds them all, and it is the common culprit.
Next: the filter and the airflow path
Pull the filter mat and hold it up to the shop light. A mat you can barely see light through has been starving the fan for weeks, and on typical fan curves a clean mat and guard already cost 20-30% of free-air flow; a loaded mat can take half of what is left. Work the whole path while the cabinet is open: the exhaust screen, shipping caps left in a fitting, and cable bundles laced across the inlet all cut delivered flow. For the numbers behind that, airflow loss through fan guards and what a fan grille's pressure drop changes carry the detail.
Then: the heat budget, with the arithmetic
A sealed stainless box cools only through its walls, at 3.7 W per square metre per kelvin of inside-to-outside difference (this site's reference figure for stainless; painted steel manages 5.5). Take the 20 x 16 x 8 in 316L wall-mount box: its six faces total 0.78 m², from 2 x (0.508 x 0.406 + 0.508 x 0.203 + 0.406 x 0.203) with the inches converted to metres. Sealed, it sheds 3.7 x 0.78 = 2.9 W for every kelvin of rise.
Drives, power supplies and most industrial electronics are rated for 40 C [104 F] ambient at the device, unless the datasheet says otherwise. At 35 C ambient your allowed rise is 5 K, and 2.9 x 5 = 14 W is everything that box can hold sealed. Now total a realistic load: a 5 hp drive at half load gives up about 60 W of heat at 3% losses, a 24 V supply 20 W, an industrial PC 40 W, relays and I/O 30 W — 150 W. Sealed, the rise is 150 / 2.9 = 52 K, and the inside sits at 35 + 52 = 87 C. The alarm is telling the truth, and cleaning a card changes none of it.
Moving air fixes it. For air at 1.2 kg/m³, the working rule is 3 x watts / allowed rise, in m³/h. The walls still shed their 14 W, so the fan carries 136 W: 3 x 136 / 5 is about 82 m³/h, and you round to 90 m³/h [53 CFM] to cover the filter loss. Fan airflow calculation near ambient temperature walks the full method. One rule bites here: a filter fan lowers the whole assembly to the fan's rating, IP54 on a typical unit, so keep a washdown box sealed and cool it with a heat exchanger or an air conditioner sized to the watts.
A bigger box helps in proportion to wall area. The 24 x 24 x 8 in 316L box carries 1.24 m² of surface: 3.7 x 1.24 = 4.6 W per kelvin, or 23 W sealed at the same 5 K rise. Double the area, double the sealed watts; everything above that needs moving air. Stainless wall-mount enclosures run from junction-box sizes to 60 x 36 x 16 in if the budget points you up a size.
Last: the false alarm on the card's contacts
When the overtemp flag and a communication fault land on the same card, that card is both the reporter and a suspect. Its edge connector carries power and signals on contact fingers working at milliamps, and years of heat cycles oxidize the plating until the card misreads its temperature line while dropping the bus. The repair is mechanical, and it works. With the cabinet locked out and proven dead, photograph the cabling and note the slot positions, pull the card, wipe the contacts with the board maker's approved cleaner on a lint-free cloth, inspect for worn plating, then seat it fully home with its retainer secured. Skip abrasive erasers — they take plating off along with the oxide.
Then make the decision this repair invites you to skip. A card that cleans up and talks again has fixed the alarm, and it has fixed nothing about the heat; the dead fan beside it is still a finding. Restore cooling before the machine runs a full production shift, and re-run the heat budget above.
How you prove it fixed
- Run the cabinet at full load for one hour at the warmest point of the day, and measure the air at the top of the compartment and at the exhaust outlet.
- Target 40 C [104 F] inside against 35 C ambient, and record the rise next to the 5 K design figure.
- Confirm every fan turns, the exhaust flow is steady, and the alarm stays quiet through a full shift including its warmest hour.
- Log the temperatures. A baseline turns the next overtemp into a five-minute diagnosis.
For pinning the documented limit of every device in the box, how to check control cabinet temperature limits is the companion page.
Keep it from coming back
Put the filter on the PM schedule and log the internal rise at every visit; when the rise climbs between visits, shorten the interval. Replace fans at the first speed drop or new bearing noise rather than at failure — a fan turning 20% slow still spins, and it has already spent several of your kelvins. Set the alarm to trip 5 K below the lowest component limit in the cabinet, so you hear about the heat while margin remains. When the same cabinet alarms twice in one quarter, spend an afternoon on the first why instead of the symptom; a fishbone cause that becomes the answer to Why #1 is a sound pattern for that session.
When to re-specify the enclosure
| Your situation | What to order |
|---|---|
| Clean indoor air, 50-200 W | Filter fan and exhaust filter sized at 3 x watts / allowed rise — 90 m³/h for the example load; the cabinet then runs IP54 |
| Dusty plant air | Coarse prefilter on the fan with a short change interval, or a sealed box with a heat exchanger |
| Washdown, outdoor, chlorides | Sealed cooling only: heat exchanger or air conditioner sized to the watts, with heat exchanger cleaning when airflow drops on the PM list |
| Load keeps growing | Order by wall area at 3.7 W per m² per K sealed, then add air movement — the 24 x 24 x 8 in box holds 23 W at a 5 K rise |
Material follows the room: 304 serves indoor panels, and 316L earns its keep wherever washdown chemicals, salt or chlorides reach the box. The enclosures here are rated NEMA 4X / IP66 by the manufacturer and not UL listed, and the UL 50E and NEMA 250 listing guide explains when a listing matters to your inspector. Check the space rules before you order up a size: NEC 312.6 sets the wire bending space at terminals by conductor size, and NEC 110.26(A) keeps at least 3 ft [900 mm] of clear working depth in front at 0-150 V to ground, with room for a 90-degree door swing. The electrical enclosure sizing guide walks both tables.
Decision in one line: clean air with a modest load takes a filter fan on a short PM interval; washdown or outdoor duty takes a sealed box with a heat exchanger sized to the watts; a load that keeps growing takes wall area first and air movement second.
Enclosures in this article
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 20 × 16 × 8 in | 316L stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 20 × 16 × 8 in | 304 stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 24 × 24 × 8 in | 316L stainless | 24 x 24 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
Can I bypass the overtemp switch to finish the shift?
No. The switch protects electronics rated around 40 C [104 F] inside, and heat ages them fast: on electrolytic capacitor life curves, every 10 K above rating roughly halves the hours. Bypassing trades a fan repair for an early drive or power-supply failure. Silence it only long enough to measure, and restore it the same shift.
Why would cleaning an edge connector clear an overtemp alarm?
Because the alarm comes from the controller card, and an oxidized edge connector disturbs the very lines the card reads. A contact film adds resistance to a sensor or bus line carrying milliamps, so the card reports a temperature it never measured while its communication drops out. Cleaning and full reseating restore the contact. Treat the dead fan found beside it as a separate repair.
How often should the cabinet filter be changed to prevent this alarm?
On typical panel maintenance schedules, quarterly in dusty plants and every six months in clean areas. Log the internal temperature rise at each change: when the rise starts climbing between visits, shorten the interval. A mat that comes out grey and stiff has been costing you airflow for weeks before any alarm fired.
How long do enclosure fans last before they cause an overtemp alarm?
Typical ball-bearing DC fans carry about 50,000 hours at 40 C on manufacturer life curves — roughly six years of continuous running — and a hot cabinet cuts that figure. Plan replacement at four to five years in a warm enclosure, or at the first speed drop or bearing noise. A fan turning 20% slow still spins, and it has already spent several of your kelvins.
