An IP67 label on a datasheet answers a narrow question: can dust enter, and does water enter when the enclosure sits one metre deep for thirty minutes. It says nothing about whether the product survives three years on a pole in a coastal climate, where the enclosure cycles through day and night temperatures twice a day and the air inside it expands and contracts with every cycle.
That gap is where most outdoor electronics fail. The water that kills the board usually did not get in past the gasket; it condensed out of the air that was already inside, or it was drawn in through the cable gland by the same thermal pumping that moves air in and out of every sealed box. Buyers who specify sealing without specifying pressure equalisation pay for the same leak twice: once in the failed unit and once in the requalification.
The Two Numbers, Decoded
An IP code has two digits under IEC 60529. The first is protection against solid objects, from 0 (none) to 6 (dust-tight). The second is protection against water, from 0 (none) to 9K (high-pressure, high-temperature jets). Both digits describe a test that was actually performed, not an intention, and a rating without a test report is marketing.
| Code | What is actually tested | Realistic use |
|---|---|---|
| IP5X | Dust protected: some dust may enter but not enough to interfere | Indoor industrial, sheltered outdoor |
| IP6X | Dust-tight: talc chamber with vacuum, no ingress after 8 hours | Any outdoor or dusty indoor product |
| IPX4 | Splashing water from any direction, 10 minutes, 10 litres/minute | Outdoor rain exposure on a sheltered product |
| IPX5 | Water jets 6.3 mm nozzle, 12.5 litres/minute, 3 minutes | Hose-down cleaning |
| IPX6 | Powerful water jets 12.5 mm nozzle, 100 litres/minute, 3 minutes | Deck wash, heavy spray |
| IPX7 | Immersion at 1 m for 30 minutes | Occasional flooding, shallow submersion |
| IPX8 | Continuous immersion at depth and duration agreed with the customer | Subsea, permanently submerged sensors |
| IPX9K | High-pressure jets at 80 degrees Celsius, 80-100 bar, four angles, 30 seconds each | Food processing, washdown, vehicle underbody |
Two things follow from that table. First, IP6X is usually easier to achieve than IPX7, which is why an "IP67" claim is almost always a claim about the water test. Second, IPX8 has no fixed test condition. It is defined by agreement between customer and supplier, so an IPX8 claim without a stated depth and duration is not a specification you can rely on.
What the Dust Test Actually Involves
The IP6X test places the enclosure in a chamber filled with circulating talc and applies a partial vacuum inside it, so that any leak is drawn inward rather than waiting for dust to diffuse in. It runs for eight hours. A product that passes IP6X has a gasket that seals against a pressure differential, which is a much more searching condition than dust settling on a closed box.
The consequence for design is that IP6X failure almost always points at the same places as IPX7 failure: a gasket that is not continuously compressed, a screw boss that bottoms out before it loads the gasket, or a cable entry that relies on a friction fit. If a product passes IP6X, most of the sealing architecture is already right.
IPX7, IPX8 and IPX9K: Three Tests, One Spec Sheet
These three are often quoted interchangeably in marketing and are nothing alike in what they demand from a design.
| Test | Load on the design | Typical failure mode | What it proves |
|---|---|---|---|
| IPX7 | Static pressure, roughly 0.1 bar at 1 m | Gasket compression set, bond-line void, gland not tightened | The seal holds at low static pressure |
| IPX8 | Static pressure at agreed depth, can be many bar | Enclosure wall deflection, window bond failure | The seal holds under real hydrostatic load |
| IPX9K | Kinetic energy: 80-100 bar at close range, 80 degrees Celsius | Gland seal extrusion, gasket thermal softening, label lift | The seal survives a directed jet and thermal shock |
A product can pass IPX7 and fail IPX9K on the same day, because IPX7 is a pressure test and IPX9K is an energy test. A jet at 100 bar does not just raise the pressure at the surface, it pushes water along any interface at speed, which finds the gap between a gasket and its groove that static pressure never exploited.
Why IPX4 Is the Specification Most Outdoor Products Should Actually Use
Most outdoor electronics see rain at an angle, splashing from the ground, and washdown from a hose. That is IPX4 or IPX5 territory. Specifying IP67 because it sounds safer has three costs: a heavier gasket, a fully sealed enclosure that cannot breathe, and a test that proves nothing about the product's actual exposure.
The exception is anything that can be submerged by accident or by design: pit-mounted sensors, flood-prone installations, handheld instruments that get dropped in water. For those, IP67 or IPX8 is the right specification, and the breather question becomes urgent rather than optional.
Where Sealing Actually Fails: Six Mechanisms
Failure analysis on sealed enclosures returns the same six mechanisms repeatedly. Five are mechanical, one is thermodynamic, and the thermodynamic one is the one that is not in the IP code at all.
- Gasket compression set — the gasket takes a permanent deformation under continuous load at temperature. After two or three years it no longer springs back, the compression ratio falls, and the seal that passed at the factory no longer exists.
- Cable gland interface — the gland body seals to the enclosure wall and to the cable jacket by two different mechanisms. A gland tightened onto the wall but not onto the jacket passes a factory test and leaks in service as the jacket relaxes.
- Screw-boss creep — plastic bosses under continuous gasket load creep, the clamping force decays, and the compression ratio drops without any visible change.
- Moulded-in insert porosity — metal inserts moulded into plastic enclosures can leak along the interface between metal and polymer, a path that only opens under pressure.
- Window and display bond line — the adhesive bead around a transparent window is a seal in its own right, and it is the first thing to fail under thermal cycling because the two materials expand at different rates.
- Absence of pressure equalisation — covered next, and the largest single cause of water inside a sealed box.
Condensation: The Failure IP67 Does Not Cover
Every enclosure is assembled in air, and that air carries water vapour. When the product warms in sunlight, the air inside expands and a small amount leaves through whatever path it can find. When the product cools at night, the air contracts and draws that moisture back in, along with more from outside. Over a few hundred cycles, a sealed box acts as a pump that concentrates humidity inside itself.
Once the inside air reaches dew point against a cool metal wall or a cool board surface, it condenses. The water is then inside a product that passed IP67, and no amount of gasket improvement prevents it, because it did not cross the gasket as liquid.
There are three practical answers, and they differ in cost and in serviceability:
| Approach | How it works | Limit | Relative cost |
|---|---|---|---|
| ePTFE breather membrane | Vent equalises pressure by diffusion; the membrane passes vapour but blocks liquid water | Needs a mounting point and a protected location; not a solution for full immersion cycling | Low, plus one hole and one part |
| Desiccant | Absorbs moisture inside the sealed volume | Finite capacity; needs replacement at service, or it is a countdown | Low, but a service item |
| Potting / encapsulation | Removes the air volume entirely | Not repairable; thermal and CTE implications; mass increase | High |
For an outdoor enclosure that will see daily thermal cycling over a five-year service life, the breather membrane is usually the cheapest effective answer and the one most often omitted, because it adds a hole to a design whose entire purpose is to not have holes. The membrane is the reason the hole is acceptable: it passes water vapour by diffusion while holding liquid water out.
What IP69K Costs
IP69K is a different class of test from IP67 and the price reflects it. The test alone requires a specialised rig, four orientations, 30 seconds each at 80 to 100 bar and 80 degrees Celsius. Passing it changes the design, not just the test plan.
| Change | Why IP69K forces it | Relative cost impact |
|---|---|---|
| Gasket material to silicone or EPDM with low compression set | 80 degrees Celsius jets soften nitrile and accelerate set | +1.2 to 1.8x gasket cost |
| Captive or retained gasket | A jet can displace a loose gasket during cleaning | Tooling change |
| Stainless or plated glands | Repeated caustic wash corrodes zinc-plated parts | +1.5 to 2.5x gland cost |
| Vent relocated to a shielded face | A direct jet will force water through a membrane | Design iteration |
| Retest budget | Each design change after a failure requires a new test cycle | One to three test cycles typical |
The retest line is the one that surprises programmes. An IP69K failure typically means a gasket or gland change, which invalidates the previous test result, which means another cycle. Budgeting one test and expecting one pass is optimistic on a first design.
Gasket Material Selection
| Material | Temperature range | Compression set | Chemical resistance | Choose it when |
|---|---|---|---|---|
| Silicone | -55 to +200 degrees Celsius | Excellent | Poor against oils and fuels | Wide temperature, outdoor, washdown |
| EPDM | -40 to +120 degrees Celsius | Very good | Good against water and steam, poor against oils | Outdoor, UV, water and steam exposure |
| Nitrile (NBR) | -30 to +100 degrees Celsius | Good | Excellent against oils and fuels | Oil-contact industrial, not washdown |
| TPU / TPE | -40 to +90 degrees Celsius | Fair | Moderate | Consumer products with a soft-touch seal |
Compression set is the property that decides whether a seal still works at year three, and it is the property least often quoted on a datasheet. Ask for the set value at the service temperature after the specified time, not the room-temperature figure.
Pre-Test Checklist
Before booking a lab, eight items decide whether the test is a measurement or a coin flip:
- Gasket durometer, cross-section and groove fill ratio confirmed against the drawing, not assumed.
- Screw torque specified and applied with a calibrated driver on every fastener.
- Cable glands torqued to the maker's value, with the jacket diameter inside the gland's clamping range.
- Every port either populated with a rated gland or fitted with a rated blanking plug. An open hole with a label over it is a hole.
- Window and display bond line cured per the adhesive datasheet, including the full cure time.
- Vent fitted, oriented to a shielded face, and confirmed to pass air.
- Sealing surfaces free of mould release agent, which is the most common cause of a bond-line leak.
- Test unit built from production tooling, not a machined prototype, because the prototype usually seals better than the mould.
What to Ask a Box-Build Supplier
Ask which IP rating the enclosure was designed to and which test report exists for that exact configuration. Ask what gasket material and compression set figure is specified. Ask whether the design includes pressure equalisation, and if not, what the condensation plan is. Ask how glands and blanking plugs are specified and torqued. And ask what happens to the rating when a connector or a label changes, because a configuration change invalidates the report.
The sealing decision and the harness decision are the same programme seen from two ends: a sealed enclosure is only as sealed as its cable entries, and the crimp that fails inside it fails in a box nobody can open without breaking the seal. Our guide to IPC-A-620 harness inspection covers the crimp acceptance side that determines whether the entry point holds. Where a design goes further and fills the volume, the potting and encapsulation guide covers what potting does to thermal path and repairability.
For boards that stay unsealed but exposed, the relevant control is coating thickness rather than gasket compression, and the measurement method is in our conformal coating thickness control guide. The ESD packaging standards guide covers what happens to a sealed product between the factory and the customer, where moisture barrier bags and desiccant do the same job as an internal breather.
Have Your Enclosure Reviewed Before Tooling
Send the enclosure drawing, the gasket specification and the cable entry list. We will identify the sealing path that is most likely to fail first, flag any port that cannot be rated to the target IP level as drawn, and tell you whether pressure equalisation is needed for the service environment. Reviewing it before tooling is cut is the difference between a gasket change and a mould change.
