Menu

Enclosures That Survive the Real World

Enclosures That Survive the Real World: IP Ratings, Heat, and Repair

Notes · Mechanical designNovember 19, 202614 min readSenova Systems Research

IP67 is a laboratory test on a new unit, for thirty minutes. It says nothing about year two, when the gasket has taken a compression set and the box has baked in the sun. A sealed 100 × 60 × 40 mm box with 5 W inside runs 25 °C hotter than the air around it. The enclosure is not a box, it is a promise: keep the water out, let the heat out, and open again when something breaks.

First, a simple picture

A house, not a box.

A box keeps the rain off. A house keeps the rain off, and it also lets the heat out, and it has a door you can open, and the roof can be fixed without tearing the house down. Nobody calls a sealed crate a house, because a house is a place where things live, and things that live need air, and repair, and a way in.

An enclosure is the house the electronics live in. It is not a box. It is a promise with three parts: keep the water out, let the heat out, and open again when something breaks. Most enclosures keep one promise and break the other two, and the breakage shows up in year two, when the gasket has died, the battery has baked, and the unit can't be opened without breaking it.

This article is about the three promises, the numbers behind them, and the demo that shows what each one costs.

The rating: what the IP code actually proves

An IP rating is a short code, IP65, IP66, IP67, that tells you how well an enclosure keeps out solid objects and water. It comes from one international standard, IEC 60529, published in the UK as BS EN 60529. The letters stand for International Protection in the standard itself, though almost everyone says Ingress Protection, and either is understood. The code is not a quality grade or a marketing term. Each digit stands for a specific laboratory test, with a defined probe, nozzle, flow rate, depth, or duration.

The first digit (0–6) is solids. Digits 1–4 are tested with rigid probes of decreasing size, pushed against every opening. Digits 5 and 6 are tested in a chamber of circulating talcum powder, for up to 8 hours: 5 means dust-protected, some dust may get in, but not enough to stop the equipment working; 6 means dust-tight, no dust may get in at all. Almost every outdoor product you will look at scores 6. On the dust axis, IP65, IP66, and IP67 are identical.

The second digit (0–9K) is water. The tests change character as they go: drips, then spray, then jets, then immersion.

IP ratings are separate tests, not a ladder (IEC 60529; IP69K per ISO 20653)

RatingSolidsWater test
IP65Dust-tight6.3 mm nozzle, 12.5 L/min, about 30 kPa, 2.5 to 3 m, at least 3 min
IP66Dust-tight12.5 mm nozzle, 100 L/min, about 100 kPa, 2.5 to 3 m, at least 3 min
IP67Dust-tightImmersion to 1 m for 30 minutes
IP68Dust-tightContinuous immersion, depth set by the maker
IP69KDust-tight80 °C water at 80 to 100 bar, 14 to 16 L/min, 10 to 15 cm, 30 s at 0°, 30°, 60° and 90°

The number is not a ladder. This is the fact most charts leave out: above IPX6, each water digit is a different test, not a stronger version of the one before. IP69K covers hot, high-pressure jets, but not immersion. IP67 covers immersion, but not jets. A datasheet that says IP66/IP67 means the enclosure passed both tests, which is what jet-washed equipment that may also be submerged needs. A datasheet that says only IP67 does not promise jet resistance, and a datasheet that says only IP69K does not promise immersion. IP66 is roughly eight times the water volume and more than three times the pressure of IP65 (100 L/min against 12.5, 100 kPa against 30 kPa), and the minimum duration of either jet test is three minutes, conducted on a new unit.

The test allows some water in. For every water digit, the standard allows some water to get in. The test passes if whatever enters doesn't stop the equipment working, doesn't reach live parts, and doesn't reach insulation that isn't designed to be wet. The rating is a promise that the water does no harm, not a promise that no water enters. And IP68's conditions are agreed between maker and user and must be stated on the datasheet: how deep, and for how long. An IP68 with no depth and no duration on the datasheet is a number without a test.

The rating is day one. The jet tests run for three minutes, on a new unit, with a new gasket. The rating says nothing about the gasket in year two, the sun on the lid, or the hundred temperature cycles the box has lived through. That is the rest of this article.

The heat: a sealed box cooks its own electronics

Sealing and heat rejection are in direct conflict, and this is the core of the enclosure problem. A fully sealed housing has no convective path to the outside air, so all the heat generated inside must leave by conduction through the housing walls or through a dedicated thermal path. Most of the electrical power consumed by power supplies, processors, converters, and radios eventually becomes heat, and whether the sealed box can get rid of it depends on the power level, not on intent.

A sealed 100 × 60 × 40 mm box with 5 W inside can only lose heat through its walls (about 0.025 m² at about 8 W/m²K), so on a 40 °C day the inside reaches about 65 °C:

ΔT=Ph⋅A=5 W8 W/m2K×0.025 m2=25 K\Delta T = \dfrac{P}{h \cdot A} = \dfrac{5\ \text{W}}{8\ \text{W/m}^2\text{K} \times 0.025\ \text{m}^2} = 25\ \text{K}

The arithmetic is first-order, and it is the article's own, with its assumption stated: a 100 × 60 × 40 mm box has a surface area of about 0.025 m². With a combined natural-convection-plus-radiation coefficient of roughly 8 W/m²K, a standard first-order estimate for a small enclosure in still air, a 5 W load inside it raises the inside temperature by 5 ÷ (8 × 0.025) ≈ 25 °C. On a hot day, with the air at 40 °C, the inside of the box sits at 65 °C.

At 25 °C ambient, it sits at 50 °C. The battery from Article #6 bakes at 65 °C, cells age faster and deliver less usable energy when they run hot, and the BMS ends up derating the pack to protect the cells. That failure looks like an undersized battery.

It is actually an undersized thermal design.

And the sun adds more: a dark box in direct sun absorbs solar heat on top of its own 5 W, and a design rated for 25 °C laboratory conditions may not survive a 50 °C factory floor or direct solar exposure. The permitted internal component temperature matters too, semiconductor junction limits may be much lower than the enclosure material's temperature limit. The box that is fine on the bench is not fine on the roof.

The fixes are the heat paths, not the wishes. A conduction path to the host chassis, mount the pack so its hottest surfaces press against a metal surface, with a thermal interface material in the joint. An internal heat spreader, a metal plate that carries heat from the hot parts to the walls. A metal lid instead of a plastic one, the lid becomes the radiator. And a vent membrane, an ePTFE membrane that lets air and pressure through and blocks water, which also solves the next problem: pressure.

The pressure problem, and the pump. A sealed box breathes. When the temperature changes, the air inside expands and contracts, and the pressure swings with it. On a sealed IP67 box, that pressure cycling can pump moisture past the gaskets, the box inhales, a little, on every cool night, and the moisture it inhales condenses inside, where it cannot get out. The vent membrane equalises the pressure without letting water in, and it stops the pump. The membrane is the cheapest part of the enclosure, and the one most often left out.

The gasket: a service item with a life

The gasket is the soft seal between the lid and the body, the rubber that makes the box watertight. It is also the part with a life, and the life is the thing nobody budgets for.

Compression set is the gasket's memory loss: a gasket is squeezed to a percentage of its thickness (its compression) to seal, and over time, heat, UV, ozone, repeated temperature cycling, it stops springing back. The set is measured by ASTM D395. Weatherstrips with poor compression set lose 20% of their sealing ability after 2 years of temperature cycling, failing to maintain their required IP ratings. And industry consensus from 2026 telecom infrastructure reports puts 32% of outdoor edge-node failures directly tied to faulty enclosure sealing, a vendor-guide figure, labelled as such, but the direction is the point: the seal is the failure, not the electronics.

The materials, with their lives. EPDM rubber (ethylene propylene diene monomer, the black rubber in car door seals): −40 °C to 120 °C, excellent UV and ozone resistance, supports IP67, about $0.85 per linear metre, 8–10 years field service life. High-temperature silicone: −60 °C to 200 °C, superior UV and ozone resistance, supports IP68, about $1.72 per linear metre, 10–15 years field service life. Cellular silicone foam (HT-800 class) holds roughly −55 °C to 200 °C with a UL 94 V-0 flame class. Replacement is needed every 7–10 years for EPDM and 10–15 years for silicone, earlier if visible hardening or cracking is detected. Utility PV plants are designed for a 25-year life outdoors, and the controlling numbers are compression set, UV/ozone durability, and wear rate over decades, not day-one performance. A gasket that is perfect on day one and dead in year eight is a gasket that was never specified for the product's life.

The rating is day one; the gasket is year two (FMSeal 2026; H-O Products; battery enclosure guides)

FindingFigure
Sealing lost after two years of temperature cycling (poor compression set)20%
Outdoor edge-node failures tied to enclosure sealing32%
EPDM gasket life (−40 to 120 °C, IP67)8 to 10 years
Silicone gasket life (−60 to 200 °C, IP68)10 to 15 years

The gasket rules, in plain words. The gasket must hold sealing contact across the whole channel, a relaxed run or a gap is where the field enclosure leaks. The compression must be set by the design, not by how hard someone tightens the screws: a channel with a defined depth, a gasket with a defined thickness, and captive screws that bottom out at the right compression. The material must match the environment: EPDM for UV and ozone, silicone for desert heat and coastal salt, and a flame class where the product needs one. And the gasket must be replaceable, because a gasket with a 10-year life in a product with a 25-year life is a service item, and a service item needs a door.

The repair: recalibrated, not replaced

The site's own rule is "made to be kept, instruments designed to be recalibrated, not replaced." The enclosure is where that rule is won or lost.

Cells wear out and connectors wear out while housings do not. A design that allows the pack to be opened, inspected, and rebuilt keeps the host equipment alive for longer. But every opening, fastener, and cable gland is a potential leak path, so the serviceability has to be designed, not wished for: a lid that opens on captive screws, a gasket in a channel that survives opening and closing, a cable gland that seals the cable where it enters, and a battery that can be replaced without cutting the box open. The sealed box that cannot be opened without breaking it is a disposable product wearing a waterproof rating.

The cable is the leak nobody publishes. The rating that matters more than any of the digits is the one nobody publishes: how well the installer sealed the cable. A cable gland, the fitting that seals the cable where it enters the box, is a leak path if it is the wrong size, the wrong material, or tightened by hand instead of to a torque. The box can be IP67 and the cable entry can be a hole, and the hole is the rating.

DEMO

Example Labs is a made-up company we use for worked examples.

This product is made up. Call it the Example Labs field sensor hub: the fence-post unit from Article #6, the weather sensor, the event camera's cousin, the Kalman filter, the PID loop, now with a 5 W average load (camera, processor, radio, solar-charged battery), in a 100 × 60 × 40 mm box on a fence post, in the sun. The demo's thermal arithmetic is the article's own first-order calculation, verified before writing; the design targets are the article's own recommendations, labelled as such.

Step 1, The sealed box (IP67, proud of the number). The box is sealed, gasketed, screwed shut, and rated IP67, immersion to 1 m for 30 minutes, on a new unit, in a laboratory. On a hot day, the air is 40 °C and the inside sits at 65 °C, the 25 °C rise from the 5 W load, through the walls, with no other way out. The battery bakes; the BMS derates the pack; the gasket bakes with it. The box passes its IP67 test on day one. The box is not a house. It is an oven with a rating.

Step 2, Year two. The gasket has taken a compression set, 20% of its sealing ability gone after two years of temperature cycling, because the material was chosen for price, not for set. The box breathes on every cool night, pumping a little moisture past the gasket, and the moisture condenses inside, where it cannot get out. The battery is dead, aged by the heat, and the box cannot be opened without cutting it, because the lid is glued and the screws are buried. The field failure is one scrapped unit, one truck roll, and one customer who will not buy the next one.

Step 3, The serviceable box (IP65, the right test for the job). The rating is lower, IP65, low-pressure jets, not immersion, because the job is rain, not submersion, and the lower number buys the things that matter: an ePTFE vent membrane that equalises pressure and stops the pump; a metal lid with a heat spreader that carries the 5 W to the outside, with a design target of a 10 °C rise instead of 25 °C (a target, not a measured result, the mechanism is the spreader and the lid, and the measurement comes from the thermal test); an EPDM gasket in a channel, compressed to a defined depth by captive screws, with an 8–10 year service life; and a battery that unclips, a lid that opens, and a cable gland torqued to a number. The box is IP65. The box is a house.

Step 4, What each one costs. The sealed box costs one scrapped unit, one truck roll, and one lost customer, and the next hundred units carry the same design. The serviceable box costs a gasket swap at year eight and a battery swap at year five, both done in the field, both done by the customer, both done without a truck roll. The difference between a repairable product and a disposable one is the lid, and the lid costs nothing at design time, only the discipline to put it there.


The decision test

Four questions. They sort the enclosures that are honest from the products that need a mechanical review first:

  1. Does the rating match the job, jets for rain, immersion for submersion, and both tests if both are needed? No, the number is a trophy, not a test, and the wrong test is worse than no test. Yes, the rating is a specification, not a slogan.
  2. Has the heat been calculated, the watts inside, the surface area, the rise, at the hottest ambient the product will see? No, the box is an oven, and the first hot day will prove it. Yes, the thermal design is a number, not a hope.
  3. Is the gasket a service item, a named material, a defined compression, a replacement interval, and a lid that opens? No, the seal will die in year two, and the product will die with it. Yes, the seal has a life, and the life is budgeted.
  4. Can a technician open the product, replace the battery, and close it again to the same rating? No, the product is disposable, and the site's own rule says instruments are made to be recalibrated, not replaced. Yes, the product is a product, not a consumable.

If the answers are yes, yes, yes, yes, the enclosure is honest, and the review is a final check, not a rescue. If any answer is no, the job is a mechanical review, and the honest thing is to do it before the tooling is cut, not after the first hundred units come back from the field. And if the answer to question 1 is "we picked IP67 because it is the biggest number", the problem is the rating, not the design, and the review starts by asking what the product actually has to survive.

"But IP67 is the best rating", and two more fair doubts

"IP67 is the best rating." It is the best rating for submersion. It is not the best rating for rain, and it is not a rating for heat, and it is not a rating for year two. The number is two tests, not one score, and the test that matters is the one the product will actually face. A box that passes immersion and cooks its battery is a worse product than a box that passes jets and runs cool. The rating is a laboratory result. The product is a field result.

"A sealed box is simpler." A sealed box is simpler to draw and harder to live with. The heat has nowhere to go, the pressure has nowhere to equalise, the moisture has nowhere to leave, and the battery has nowhere to be replaced from. The vent membrane, the heat spreader, and the opening lid are not complications, they are the three promises, kept. Simplicity that breaks the promises is not simplicity. It is a delay.

"Our supplier certifies the enclosure." Then ask for the test report: which tests, on how many units, at what depth and duration, and whether the rating was tested with the cable gland fitted and the lid opened and closed first. A certified enclosure with an uncertified cable entry is a certified box with a hole in it. The review's job is to read the report, not the datasheet.

Sources

  1. Things Embedded, "IP Rating Chart: IP65, IP66, IP67, IP68 and IP69K Explained" (IEC 60529)
    What we used

    First digit 0–6 solids, second 0–9K water; IPX5 6.3 mm nozzle, 12.5 L/min, ~30 kPa, 2.5–3 m; IPX6 12.5 mm nozzle, 100 L/min, ~100 kPa; IPX7 immersion to 1 m for 30 min; IPX8 continuous immersion, maker-specified; IP69K 80 °C water at 8–10 MPa from 100–150 mm; above IPX6 the water digits are separate tests; IP66/IP67 dual ratings mean both tests passed

  2. Enclosure Company, "IP Ratings Explained" (IEC 60529 / BS EN 60529)
    What we used

    First digit tested with probes (1–4) or talcum powder up to 8 hours (5–6); IPX5 and IPX6 jet tests at least 3 minutes; IPX7 30 minutes under water (lowest point 1 m down, or top 150 mm down for enclosures 850 mm or taller); IPX8 conditions agreed between maker and user, more severe than 7; IP69K 80 °C, 80–100 bar, 14–16 L/min, 10–15 cm, 30 s at 0°, 30°, 60°, 90°; for every water digit, the standard allows some water in, the test passes if it does no harm

  3. FIRGELLI, "Complete Guide to IP Ratings"
    What we used

    IP67 = dust-tight + temporary immersion to 1 m for 30 minutes; manufacturers cannot self-assign IP ratings; an independent certified laboratory must test and verify

  4. SafeBode, "IP65 vs IP66 vs IP67: What a Camera's Weather Rating Really Promises"
    What we used

    IPX5 and IPX6 test parameters from an ISO 17025 accredited laboratory's published IEC 60529 procedure; IP66 is roughly 8× the water volume and more than 3× the pressure of IP65; the rating that matters more than any digit is how well the installer sealed the cable

  5. HRLI Battery, "Battery Pack Enclosure Design: Sealing, Heat and Serviceability Trade-offs"
    What we used

    Four jobs (environmental protection, thermal management, serviceability, plus the pack itself); sealed housing has no convective path, all heat leaves by conduction; BMS derates the pack when the walls can't shed the heat, a failure that looks like an undersized battery; IP67 pressure cycling can pump moisture past gaskets; cells and connectors wear out while housings do not

  6. Jinhui Precision, "Electrical Enclosure Design Guide: Cooling, EMI, and Sealing"
    What we used

    Most electrical power consumed eventually becomes heat; define the continuous and peak power dissipated, the permitted internal component temperature, and the ambient operating range; a design rated for 25 °C laboratory conditions may not survive a 50 °C factory floor or direct solar exposure; set serviceability requirements (door opening angle, filter replacement time, connector access)

  7. FMSeal, "Outdoor Telecom Enclosure Weatherstrip" (2026)
    What we used

    32% of outdoor edge-node failures tied to faulty enclosure sealing (industry-consensus claim from 2026 telecom infrastructure reports, labelled as such); weatherstrips with poor compression set lose 20% of sealing ability after 2 years of temperature cycling; EPDM −40 °C to 120 °C, IP67, $0.85/m, 8–10 year field life; high-temperature silicone −60 °C to 200 °C, IP68, $1.72/m, 10–15 year field life; replacement every 7–10 years (EPDM) and 10–15 years (silicone)

  8. H-O Products, "PV Inverter & Combiner Box Gaskets"
    What we used

    Utility PV plants designed for a 25-year life outdoors; controlling numbers are compression set, UV/ozone durability, and wear rate over decades; cellular silicone (HT-800 class) −55 °C to 200 °C, UL 94 V-0; standards cited: IEC 60529, NEMA 250, UL 50E, ASTM D395 (compression set), D1056, D2240, D1149, G154, B117, D5470, UL 94

More in design notes

Injection Moulding DFMInjection molding design determines 70–80% of final part cost and quality before production begins. The five rules are simple: a uniform 1.5–3.0 mm wall, 1–3° of draft per side, ribs at half the wall, bosses at six-tenths, and shrinkage built into the cavity, PP shrinks 1.0–2.5%, so a 100 mm part needs a cavity of 101–102.5 mm. Roughly 70% of first-submission designs contain at least one feature that would cause a defect or force an expensive mould action, and one post-tooling modification typically costs $2,000–$20,000 and 2–6 weeks. The fix is a CAD edit, not a tool edit.19 minSheet Metal DFMYour enclosure came back 7 millimetres too wide, with a cracked bend and a hole stretched into a slot. Sheet metal is not machining, the flat pattern is the part, and the bend is where the size lives. Twelve rules, in the order you meet them on a model.17 minThe Clip That Doesn't BreakThe lid has to open for battery service and calibration, and every time it opens, the clip bends past its limit, because the clip is a spring, and the spring was sized by feel instead of by strain.21 minThe Screw That Doesn't StripThe housing has to open for battery service and calibration, and every time it opens, the joint gets weaker. A self-tapping screw holds 380 N on the first pass and is stripped by the eleventh; a brass insert still holds 1,056 N at pass 25. The strength does not come from the brass, it comes from plastic locked into the knurls.21 minThe Seal That Actually SealsYour enclosure passed IP65 on the bench and leaks in the field, because the gasket took a set after two years of temperature cycling and the joint designed at 30% compression is now working at 15%. The seal is a set of numbers, not a material.26 min