An IP rating is a two-digit code for how well an enclosure keeps out solid objects and water, defined in IEC 60529. The first digit covers solids, the second covers water. IP67 and IP68 share the same first digit — full protection against dust — and differ in the second, which is where the misunderstanding starts: IP67 is a defined test, IP68 is a defined category whose test conditions the manufacturer states.
That single sentence decides more sourcing arguments than any other line in this article.
Reading the two digits
The difference that actually matters
So the answer to "is IP68 better than IP67" is: it is stricter only if the stated depth and duration are stricter. An IP68 part tested at 1 m for 30 minutes is an IP67 part with a different label. Ask for the condition, in writing, before treating the higher number as the safer choice.
The test conditions matter more than the number
Two suppliers can both claim IP67 and produce very different parts, because the number does not describe how the test was run:
- Sample state. Tested as a loose connector, or as a terminated harness with the correct wire seals, the right cable diameter and the retainer locked? Most real failures happen at the wire-to-seal interface, which a bare connector test never exercises.
- Number of specimens. A single sample that passed is not a rating; a rating is a set of specimens with a defined pass criterion.
- Whether the part was mated. An unmated connector tests the housing. A mated pair tests the interface, which is the part that leaks.
- Ageing before the test. Seals harden and take a set. A part that passes when new and leaks after two winters is a thermal cycling question, not an IP question.
A supplier who can answer all four has done the test properly and knows exactly what it does and does not prove. A supplier who answers only with the number has read the datasheet.
IP69K, and where IP stops helping
IP69K adds protection against close-range high-pressure, high-temperature water jets and is the rating to ask for on anything cleaned with a pressure washer — construction machinery, food processing equipment, some agricultural applications. It is not "higher than IP68"; it tests a different exposure. A part can be IP69K and fail an immersion test, and vice versa.
More importantly, the IP code says nothing about three conditions that decide whether a harness survives in the field:
- Temperature cycling. Repeated expansion and contraction is what eventually breaks a seal or a crimp. It is tested as a thermal shock or thermal cycling qualification, not as an IP rating.
- Chemical and salt exposure. Road salt, coolant, hydraulic fluid and cleaning agents attack different materials differently. That is a materials question, covered by corrosion testing rather than by any IP digit.
- Flex and vibration. A sealed assembly that is also a moving cable needs a cycle life qualification in addition to an ingress rating; passing one says nothing about the other.
Low temperature: the other half of the specification
Cold climate does not show up in an IP rating at all, and it changes the material selection more than a supplier's datasheet usually admits. Three effects matter:
- Insulation stiffens. PVC becomes noticeably less flexible well before its rated temperature, so the minimum bend radius that was fine at 20 ℃ can crack a jacket at −30 ℃. A harness routed in a door loom or a moving assembly needs the gauge and the bend radius checked against the cold end of the range, not the ambient temperature of the test bay.
- Seals take a set. A seal compressed cold and then warmed repeatedly loses elastic recovery, which is why a part can pass an ingress test new and leak in the second winter. Ask for a thermal cycling result rather than an IP test alone.
- Materials diverge. Some jacket materials keep flexibility at low temperature far better than others, and the choice is not the same as the choice for high temperature. If your operating range is −40 ℃, say so explicitly — it is a material decision, not a note on the drawing.
Applying it to a harness
An ingress rating is a property of the whole assembly, not of the connector alone. Four things have to be right at the same time:
- The connector is sealed and the right wire seals are fitted for the actual cable diameter — the most common leak path is a seal sized for a nominal cable that is undersized in production.
- The cavity is plugged where a circuit is unused, rather than left open.
- Terminations are sealed or overmoulded where the assembly is exposed, and the overmould adhesion is verified rather than assumed.
- The finished assembly is leak-tested, if the programme requires it — a rating claimed from a component datasheet is a component rating, not an assembly rating.
That is why an IP requirement belongs in the specification rather than in the part number: it constrains the seals, the moulding, the test and often the choice of connector family, and it needs to be on the drawing before the part is quoted. For the families where this comes up most:
- On an EV high-voltage assembly, sealing is normally part of the requirement, and the shield termination and the seal have to be designed together — an overmould that solves the ingress problem can compromise the shield path.
- On a two-wheeler harness, the exposure is continuous weather rather than occasional spray, so the wire-to-seal interface and the connector-to-connector interface both matter.
- On a construction machinery harness, wash-down usually means IP69K rather than IP68, plus the corrosion resistance that the IP code says nothing about.
- On an industrial automation harness, the requirement is often coolant and cutting-fluid resistance in a control cabinet rather than an immersion rating at all.
Frequently asked questions
Is IP68 better than IP67 for a wire harness?
Only if the stated conditions are more severe. IP67 is a fixed test — one metre for thirty minutes — while IP68 is a category in which the manufacturer states the depth and the duration. IP68 tested at one metre for thirty minutes is the same performance as IP67 under a different label. Before treating the higher figure as the safer choice, ask for the depth and duration in writing; if the supplier cannot state them, the IP68 claim is a marketing position rather than a test result.
What does an IP rating not tell me about a harness?
It says nothing about temperature cycling, chemical or salt exposure, vibration, flex life, or low-temperature flexibility. Those five decide most field failures on a harness, and each is qualified by a different test. A part can hold a valid IP67 rating and still fail after two winters because the seal took a compression set, or fail at sixty thousand kilometres because the jacket hardened and cracked at the cold end of its range. Specify those separately, and ask for the qualification evidence for each.
What should I specify for a harness that has to work in a cold climate?
State the low end of the operating temperature explicitly and treat it as a materials decision rather than a note. That means checking the insulation material against the cold extreme rather than the ambient, re-checking the minimum bend radius at that temperature — a radius that is fine at 20 ℃ can crack a jacket at −30 ℃ — and asking for a thermal cycling result on the seals rather than an ingress test alone. If the assembly is also flexed, the flex qualification has to be run at the cold extreme too, because a cycle count established at room temperature does not transfer.