IP65 is a test, not a promise

Ask what a floor terminal needs to survive and the answer comes back as a rating. IP65, usually, sometimes IP66, occasionally a NEMA Type from someone who has been doing this longer. It is a satisfying answer because it is a number, it appears on every data sheet, and it lets a purchasing conversation end. It is also the least informative sentence in the specification, and terminals fail in service for reasons the number never covered.

This is not an argument that ratings are meaningless. They are the product of real procedures, and an unrated enclosure in a wet area is a straightforward mistake. It is an argument that a rating is a test result, that a test has a defined scope, and that most of what destroys hardware on a plant floor is outside that scope by design.

What the code actually classifies¶

The IP code is IEC 60529 (opens in a new tab), in its consolidated second edition. Its scope is stated in three parts: protection against access to hazardous parts, protection against ingress of solid foreign objects, and protection against ingress of water. Two digits, two independent scales. The first digit runs to 6, which is dust-tight; the second describes water, and the levels are not a ladder in the way buyers read them.

That second point is the one that costs money. The water digit describes different tests, not increasing severity along one axis. A device rated for powerful jets from any direction has not been shown to survive immersion, and a device tested for immersion has not necessarily been tested against a jet, which is why the higher-sounding number can be the wrong specification for a washdown area. If the process involves a hose, the specification has to name the jet test, not a larger digit.

The two systems are not interchangeable, and the asymmetry has a direction¶

North American estates mix the vocabularies constantly, with an IP number on one quotation and a NEMA Type on the next, and somebody in the middle producing a conversion chart. NEMA answers this directly in its own published FAQ, and the answer is worth having verbatim because it is routinely used backwards.

The same answer supplies the direction: a NEMA Type meets or exceeds a corresponding IP degree, but an IP degree cannot be stated as equivalent to a NEMA Type. So a conversion chart is safe read one way and misleading read the other, and the way a purchasing conversation reads it is the misleading one, because the substitution people want to make is an IP-rated unit into a place where a Type was specified.

NEMA adds a second distinction that is easy to skim past and matters on site: its Type ratings require that the enclosures are fully installed and ready for use, whereas IP degrees may apply to partially completed states of installation. The rating on the box is a claim about a condition, and the condition includes what the installer did with the entries.

The three things that actually take a terminal out of service¶

Across floor deployments, the failures that generate calls are consistently these, and it is worth noticing how little the ingress rating has to do with any of them.

• Condensation. A sealed enclosure that heats during a shift and cools overnight moves air past whatever it can, and moisture ends up inside a box that no water entered. The better the seal, the fewer chances it has to leave. This is a thermal design question, and it is why heaters and breather vents exist on hardware that already passed its ingress test.
• Chemistry. Washdown is not water; it is a hot caustic or acidic solution applied under pressure, and it attacks gaskets, membranes, screen bonding and any exposed metal that is not the right alloy. A unit that passes a jet test with clean water at ambient temperature has been tested against something that does not exist in a food plant.
• Readability and reach. A terminal that cannot be read under the actual lighting, or cannot be operated by somebody wearing the gloves the job requires, is out of service in every sense that matters even while it is powered on. This is where deployments quietly fail: the hardware works and the work moves back to paper.

Touch, gloves and the input nobody specifies¶

A projected-capacitive screen senses a change in an electric field, so a glove is not one problem but several: thickness, material and whether it is wet all change the answer, and a screen tuned to be sensitive enough for a thick glove becomes a screen that registers a splash of water as a touch. Vendors expose this as a sensitivity mode for exactly that reason, and it is a trade rather than a setting with a right value.

This is why the useful test is not a specification review, it is ten minutes on the floor with the gloves the operators actually wear, which are frequently not the gloves the specification named. Where the trade cannot be resolved, the honest design answer is to stop asking the screen to carry the interaction: physical buttons for the two or three actions performed constantly, and the screen for everything else. A terminal that requires a glove to come off is one that either gets a bare hand in a place that needed the glove, or gets ignored.

How we would write the requirement instead¶

1. 1.State the cleaning regime first: chemistry, temperature, pressure, frequency, and whether the terminal is in the path or adjacent to it.
2. 2.State the thermal cycle: the range across a day and whether the space is heated, because that decides condensation, not the seal.
3. 3.State what is on the operator’s hands, and what the room’s lighting does at the worst hour of the day.
4. 4.Then state the ingress requirement, naming the test rather than the digit, and add the corrosion requirement separately if the estate speaks in IP degrees, since the IP code does not carry it.
5. 5.Ask for the installation condition to be part of the acceptance: entries fitted and sealed to at least the rating claimed, checked on a real unit rather than assumed from the enclosure’s certificate.

Where this argument stops¶

Reading an IP rating as a test rather than a promise can be taken too far. There are plenty of environments where a commodity tablet in a decent case, replaced when it breaks, beats a specified industrial terminal on every axis including uptime, because the replacement is a cupboard away and costs a fraction of the service call. Specifying for the worst imaginable environment when the actual one is an office corridor is a way to spend a fleet’s budget on protection nobody needed, and to end up with a slower, dimmer screen for the privilege.

The rating is not the enemy. It is a genuine, useful, narrowly scoped test result that answers one question well. The mistake is treating it as a summary of survivability, and the correction is cheap: write down what the room does to things, and let the ratings follow from that rather than stand in for it.