{"id":135530,"date":"2026-11-06T09:19:40","date_gmt":"2026-11-06T09:19:40","guid":{"rendered":"https:\/\/www.seeedstudio.com\/blog\/?p=135530"},"modified":"2026-11-06T09:21:51","modified_gmt":"2026-11-06T09:21:51","slug":"ip66-vs-ip67-vs-ip69k-rugged-j4012","status":"publish","type":"post","link":"https:\/\/www.seeedstudio.com\/blog\/2026\/11\/06\/ip66-vs-ip67-vs-ip69k-rugged-j4012\/","title":{"rendered":"IP66 vs IP67 vs IP69K: Which Rating Does Your Edge AI Deployment Actually Need?"},"content":{"rendered":"\n

IP ratings are easy to look up and harder to choose between. The codes are public and the test methods are standardised, yet selection meetings still stall on two assumptions: that a higher number is always safer, and that a higher rating automatically includes the lower one. Neither is true.<\/p>\n\n\n\n

This guide is for automation and robotics integrators specifying an edge AI computer, equipment engineers reviewing an enclosure, and field teams responsible for installations after commissioning. It explains what each rating tests, where the distinctions matter, and what to check beyond the rating printed on the enclosure.<\/p>\n\n\n\n

The examples in this guide use the reComputer Rugged J4012, an IP66-rated rugged edge AI computer built on the NVIDIA Jetson Orin NX 16GB module. The same selection logic applies to any enclosure you are evaluating.<\/p>\n\n\n\n

Definition.<\/strong> An Ingress Protection (IP) rating is a code that classifies how an enclosure protects against access to hazardous parts and the ingress of solid objects and water. The familiar two-digit code is defined by IEC 60529, and the rating is established by specified laboratory tests on a new, clean sample in the configuration that will be installed, including its cable entries and connectors.<\/p>\n\n\n\n

Key takeaway.<\/strong> Choose a rating by the way water reaches the enclosure, not by the number. IP66 covers dust-tight protection and powerful water jets. IP67 adds a defined temporary-immersion test. IP69K, under ISO 20653, addresses high-pressure, high-temperature water jets for road-vehicle electrical equipment. None of these ratings replaces separate corrosion, chemical-compatibility or durability testing.<\/p>\n\n\n\n

\"reComputer<\/figure>\n\n\n\n

The rating covers the installed enclosure and its cable entries, not the site around it. Shown: reComputer Rugged J4012.<\/em><\/p>\n\n\n\n

\n\n\n\n

The two digits answer two different questions<\/h2>\n\n\n\n

The first and second digits describe separate protections, tested independently. A product can be dust tight and only splash resistant.<\/p>\n\n\n\n

First digit: solids.<\/strong> Only two values matter for industrial computers.<\/p>\n\n\n\n

Code<\/th>Test<\/th>What it means<\/th>Where it shows up<\/th><\/tr><\/thead>
IP5X<\/td>Talcum powder chamber, 8 hours, partial vacuum for category 1 enclosures<\/td>Dust protected: dust may enter, but not in a quantity that interferes with operation<\/td>Indoor cabinets, warehouse floors, most general factory areas<\/td><\/tr>
IP6X<\/strong><\/td>Same chamber and duration, no dust<\/strong> ingress<\/strong> permitted<\/strong><\/td>Dust tight: no dust enters at all<\/strong><\/td>Foundries, mines, grain and feed handling, carbon or metal dust<\/strong><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

The gap between 5 and 6 matters in one situation: when the dust conducts. Settled non-conductive dust is a cleaning problem. Metal or carbon dust bridging two pins is a failure.<\/p>\n\n\n\n

Second digit: water.<\/strong> This is where selection mistakes happen, because each code uses different equipment and simulates a different exposure.<\/p>\n\n\n\n

Code<\/th>Test conditions<\/th>Where it shows up<\/th><\/tr><\/thead>
IPX4<\/td>Splashing water, two methods: spray nozzle at 10 L\/min for at least 5 minutes, or oscillating tube for 10 minutes<\/td>Rain blown onto an enclosure, spill from any direction<\/td><\/tr>
IPX5<\/td>6.3 mm nozzle, 12.5 L\/min, 2.5 to 3 m, at least 3 minutes<\/td>A low-pressure hose aimed at the unit<\/td><\/tr>
IPX6<\/strong><\/td>12.5 mm<\/strong> nozzle<\/strong>, 100 L\/min, 2.5 to 3 m, at least 3 minutes<\/strong><\/td>Heavy seas, powerful jets, direct hose-down<\/strong><\/td><\/tr>
IPX7<\/strong><\/td>Immersion at 1 m for 30 minutes<\/strong><\/td>Temporary<\/strong> flooding<\/strong>, standing water, a dropped unit<\/strong><\/td><\/tr>
IPX8<\/td>Continuous immersion, depth and duration declared by the manufacturer<\/td>Long-term submersion, custom to the product<\/td><\/tr>
IPX9K<\/strong><\/td>80 \u00b1 5\u00b0C water, 8,000 to 10,000 kPa (80 to 100 bar), 14 to 16 L\/min, four angles, 30 s each, 100 to 150 mm<\/strong><\/td>Hot high-pressure washdown, industrial and vehicle<\/strong><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Three notes on this table.<\/p>\n\n\n\n

IEC 60529 defines each jet test by nozzle diameter, flow rate, distance and duration. Nozzle pressure is not an independent setting: it is whatever pressure produces the specified flow, roughly 30 kPa for IPX5 and 100 kPa for IPX6. That is why a pressure figure taken off a test rig, or off a pressure washer, cannot be compared directly. The tests are defined by geometry and flow, not by a pressure number.<\/p>\n\n\n\n

IPX9K is not part of IEC 60529. It comes from ISO 20653, the road-vehicle ingress standard, and the figures in that row are its spray conditions. IEC 60529 does define an IPX9, without the K. When a supplier claims IP69K, ask which standard the claim rests on, and whether the test was run on the configuration you are buying.<\/p>\n\n\n\n

The third column describes the field situation each code is meant for, not what the test applies. Industrial washdown often involves detergents, but the IPX9K test sprays hot water only. Chemical compatibility with cleaning agents is a separate question the rating does not answer.<\/p>\n\n\n\n

\n

The short version.<\/strong> IP66, IP67 and IP69K are different water-ingress tests, not grades on one scale. IPX6 is a 12.5 mm nozzle at 100 L\/min from 2.5 to 3 m for at least 3 minutes. IPX7 immerses the sample at 1 m for 30 minutes. IPX9K sprays 80\u00b0C water at 80 to 100 bar from 100 to 150 mm under ISO 20653.<\/p>\n<\/blockquote>\n\n\n\n

\n\n\n\n

IP66, IP67 and IP69K are not a ladder<\/h2>\n\n\n\n

Each claim has to be supported by its own test. A higher number never includes the lower one.<\/p>\n\n\n\n

IP66<\/strong> does not include IP67.<\/strong> An IP66 enclosure has passed the dust and powerful water-jet tests. It has not been immersed. If the equipment can sit in standing water, check for a separate immersion rating.<\/p>\n\n\n\n

IP67 does not include<\/strong> IP66<\/strong>.<\/strong> IP67 establishes protection under its specified temporary-immersion test. It does not, by itself, establish resistance to the IPX6 water jet. If the equipment must survive both, look for an explicit IP66\/IP67<\/code> marking rather than inferring one rating from the other.<\/p>\n\n\n\n

IP69K does not include IP67.<\/strong> The IPX9K test never submerges the sample, so a unit can pass IPX9K and still leak in a puddle.<\/p>\n\n\n\n

IP69K is not<\/strong> IP66<\/strong> with a bigger number.<\/strong> The 80\u00b0C jet at 80 to 100 bar from 100 to 150 mm loads the enclosure differently from the IPX6 jet at 2.5 to 3 m: a different nozzle, a different distance, a different water temperature. It is not a harder setting of the same test.<\/p>\n\n\n\n

What to<\/strong> look for<\/strong> on the datasheet.<\/strong> Dual marking such as IP66\/IP67<\/code> or IP66\/IP69K<\/code>, with the applicable standard named on the datasheet. Do not infer an unlisted rating from a higher-looking number.<\/p>\n\n\n\n

\n\n\n\n

What an IP rating does not test<\/h2>\n\n\n\n

An IP rating is an ingress test on a clean, new sample in a laboratory. It is not a durability grade and it is not a system-level guarantee.<\/p>\n\n\n\n

Not covered<\/th>Why it matters in the field<\/th><\/tr><\/thead>
Pressure washing<\/strong><\/td>A washer changes pressure, flow, distance, temperature and impact at once<\/td><\/tr>
Salt spray and corrosion<\/strong><\/td>IP ratings say nothing about salt. Coastal sites need ISO 9227 or ASTM B117 data<\/td><\/tr>
UV<\/strong>, ozone and ageing<\/strong><\/td>Gaskets harden and take a compression set. The rating describes day one, not year three<\/td><\/tr>
Vibration and shock<\/strong><\/td>An IEC 60068 claim, not an IP one. A vibrating cable can work a connector loose<\/td><\/tr>
Temperature<\/strong><\/td>A separate claim. Seals behave differently at \u221220\u00b0C and at 60\u00b0C<\/td><\/tr>
Condensation<\/strong><\/td>Sealed enclosures breathe under thermal cycling. Moisture enters as vapour and condenses inside<\/td><\/tr>
Cleaning chemicals<\/strong><\/td>Detergents and solvents attack different elastomers at different rates<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Installation is the gap that causes most warranty disputes, and it needs its own section.<\/p>\n\n\n\n

\n\n\n\n

The weakest link is often outside the enclosure<\/h2>\n\n\n\n

An IP rating applies to the enclosure as tested, with the cable entries and connectors named in the test report. Change any of them without validating the new configuration, and the original rating may no longer describe what is installed on site. In practice, failures cluster in five places.<\/p>\n\n\n\n

\"Two<\/figure>\n\n\n\n

An unmated M12 port is a hole in the enclosure. The rating holds only while every entry is mated or capped.<\/em> Shown: reComputer Rugged J4012 M12 panel.<\/em><\/p>\n\n\n\n

    \n
  1. An unmated connector.<\/strong> The rating belongs to the mated and torqued pair, not to the receptacle alone. With the cap off, the port is a hole in the box. The cap belongs on the bill of materials, not in someone’s pocket.<\/li>\n\n\n\n
  2. Connectors tightened by feel.<\/strong> Some assemblies rely on a defined mating torque to compress the O-ring. Finger-tight is a feeling, not a specification. Field teams need the value and the tool.<\/li>\n\n\n\n
  3. A reused or damaged gasket.<\/strong> Compression, ageing and maintenance all affect performance, and a gasket that looks intact may no longer seal to specification. Follow the service interval, not a reuse rule.<\/li>\n\n\n\n
  4. The wrong gland for the cable.<\/strong> A gland seals against a specific diameter range. A 6 mm cable in an 8 to 12 mm gland leaves the seal under-compressed, and that is one of the most common leak paths on site.<\/li>\n\n\n\n
  5. A pressure washer at close range.<\/strong> IPX6 is a 12.5 mm nozzle at 100 L\/min from 2.5 to 3 m. A pressure-washing procedure can involve much higher pressure, shorter distances and heated water. That is a different loading case, and it can expose weak points around joints, gaskets and other interfaces that the jet test never loaded.<\/li>\n<\/ol>\n\n\n\n

    If the site routinely applies close-range, high-pressure, hot-water washdown, IP66 was never tested for that duty. That is an IPX9 or IPX9K conversation, usually with different materials and a different price. The other four are connector, sealing and installation issues. No higher IP rating fixes an open port or an incorrectly sized cable gland.<\/p>\n\n\n\n

    An IP66 enclosure with twelve cable entries has twelve places to get wrong. A gland is hand-assembled, its seal depending on cable diameter and torque. An M12 port is sealed at the factory, so field work is a mate and a torque value. On the J4012, Ethernet, CAN, serial, I\/O and power are all M12, and four PSE ports power four cameras without a second cable each.<\/p>\n\n\n\n

    \n\n\n\n

    Four questions, and the answers stack<\/h2>\n\n\n\n
    \"Four<\/figure>\n\n\n\n

    Figure 1. Four separate questions. Every yes adds a requirement, and the answers combine.<\/em> From the reComputer Rugged selection guide.<\/em><\/p>\n\n\n\n

      \n
    1. Is dust<\/strong> ingress<\/strong> a concern?<\/strong> If so, specify IP5X or IP6X by the dust load on site.<\/li>\n\n\n\n
    2. Will the enclosure face powerful water jets?<\/strong> If so, evaluate IPX6 against the actual cleaning procedure, not against a pressure number.<\/li>\n\n\n\n
    3. Could the equipment be immersed?<\/strong> If so, specify IPX7 or IPX8 by the expected depth and duration.<\/li>\n\n\n\n
    4. Will it face high-pressure, hot-water cleaning?<\/strong> If so, identify whether the applicable standard is IPX9 under IEC 60529 or IPX9K under ISO 20653, and ask what the claim was tested on.<\/li>\n<\/ol>\n\n\n\n

      Treat these as four parallel checks, not as a path with one exit. A washdown station beside a vehicle bay can need IP66, IP67 and IP69K at once, which is why datasheets list combined codes, each one backed by its own test. Teams also tend to stop at question 2 because IP66 is the familiar answer, then discover at commissioning that the cleaning crew uses a 90 bar washer. That is question 4, and it never gets asked unless someone asks it.<\/p>\n\n\n\n

      \n\n\n\n

      Three checks you can run on the unit you already have<\/h2>\n\n\n\n

      Specification work usually happens with no hardware in the room. These three checks take about 30 seconds each and work on anything already installed, including a unit from another vendor.<\/p>\n\n\n\n

        \n
      1. Read the second digit.<\/strong> A 6<\/code> means a jet test, not an immersion test. If the site can flood and the code says 6<\/code>, that gap belongs in the risk register today.<\/li>\n\n\n\n
      2. Count the ports with nothing plugged in.<\/strong> Every unused entry needs its sealed cap, and the cap has to be on the bill of materials. An open port is a hole in the box.<\/li>\n\n\n\n
      3. Document the washdown.<\/strong> Nozzle, flow, distance, water temperature and chemicals. Pressure alone does not settle it, and an IPX6 rating does not automatically cover every hose-down routine.<\/li>\n<\/ol>\n\n\n\n
        \"M12<\/figure>\n\n\n\n

        A cap without its O-ring closes the hole but does not seal it. Shown: reComputer Rugged M12 sealing cap.<\/em><\/p>\n\n\n\n

        \n\n\n\n

        Where this decision framework stops<\/h2>\n\n\n\n

        The four questions cover the common industrial cases. Four situations need a different framework altogether.<\/p>\n\n\n\n