How to Word an IP Requirement So Every Supplier Bids the Same Test

An IP requirement turns bid-ready when it reads as a test definition instead of an adjective: name the standard, state the full two-digit code, spell out the conditions the second digit stands for, and describe the exact configuration the pack is in when the water reaches it. Add the sample count, the build stage the tested units come from, the evidence deliverable by name, and the retest trigger. Do that, and four quotes that all say 'IP67' finally describe four comparable bids. The rating is the easy part. Comparability is the work.

There's a specific silence that lands in a conference room when four suppliers have all quoted IP67 and nobody can say which one is telling the truth. I've sat in it. Bid tab on the screen, four rows, an IP rating column with the same four characters in every cell, and four attachments that might as well have come from four different industries.

Row one tested a bare pack: no harness, connector bay taped over, gland still in a bag on the bench. Row two tested a fully dressed unit, connectors mated, breather vent installed. Row three sent a two-page report with no lab name and no photographs. Row four ran the immersion for twenty-four hours because that was their internal standard, which is either more rigorous than anyone asked for or a completely different test, depending on how you read it.

Procurement wanted to sort by price. Engineering couldn't sign the 'technically equal' box for any of the four. We spent nine days re-quoting a line item that should have taken an afternoon.

Since then I write the IP requirement like a work instruction.

The second digit is the whole argument

Everyone writes IP67. Almost nobody writes what X7 means inside their own document. X7 is immersion: one meter below the surface to the top of the enclosure, thirty minutes. X6 is a 12.5 mm nozzle at roughly 12.5 liters per minute and about 100 kPa, held three meters away for at least three minutes.

Those are not interchangeable, and the direction of difficulty surprises people. A pack on a mower deck or a scooter floor sees directed jets, not standing water. A pack that passes immersion can fail a jet, because the jet carries pressure and the gasket was never designed to hold a pressure differential across the lid. If your service environment is a hose or a pressure washer, ask for the code that tests a hose.

So write both digits and write their conditions. Then know where the standard goes quiet.

IPX8 has no fixed conditions, so you have to fix them

IPX8 is continuous immersion, and the depth and duration are agreed between the manufacturer and the user. That phrase in the standard is your problem. Leave it blank and the supplier fills it in. One bidder prices three meters for sixty minutes, another prices one meter for thirty and calls it X8, and both are defensible. Write the depth and the time yourself or you have bought a word, not a test.

For road vehicles, ISO 20653 is the reference rather than IEC 60529, and the K-suffixed codes mean something different: IP6K9K is a hot, high-pressure jet at close range, the kind of thing a steam cleaner does. If your customer washes the machine, that's the code in the spec, not X7.

Dust has a similar hole

IP6X wants a talc chamber with a vacuum drawn on the enclosure so internal pressure sits below ambient. IP5X is the same chamber without the vacuum. Suppliers occasionally run the 6X chamber with no vacuum because the fixture is simpler, and the report won't always tell you which one they ran. Put that line in the report format and you'll find out.

The configuration line is where bids diverge the most

This is the highest-value sentence in the whole requirement, and it's the one most specs leave out. State exactly what is installed on the pack at the moment of test. Connectors mated or capped and with which caps. Cable gland installed and torqued to the drawing callout. Breather vent in place or blanked. Enclosure fasteners at specified torque. Potting or sealant cured per the drawing. Temporary shipping caps named, and named by location.

Mated connectors seal the interface. Capped connectors test the cap. A gland installed finger-tight is a hole with hardware around it. None of that is a trick question — it's just never written down, so each bidder answers it their own way and each answer is honest.

While you're there, state the orientation. A lid gasket that holds with the pack upright and drains freely can behave differently when the pack is inverted, and the standard doesn't care which way you hang it.

Price the same scope

Sample quantity, build stage, destructive disassembly after the test, and the post-test electrical checks all belong in the RFQ, not in a later conversation. Insulation resistance, dielectric withstand, capacity, and a functional check after immersion are real hours in a lab and someone has to quote them.

One sample is a coin flip, not a qualification. I'd rather see three pilot-build units than one golden sample built by an applications engineer with a hand torque driver. The golden sample tells you the design can pass. The pilot units tell you the line can, and those are different claims that get conflated constantly.

Witness travel and unit return are line items too. Ask whether the price includes a customer witness and whether the tested units come back or get scrapped. Both change the number, and neither shows up unless you ask.

Name the evidence deliverable in the RFQ

Don't leave the report to the quality agreement. Put it in the RFQ as a named document with its contents listed: standard and edition, full code, lab name and accreditation scope, test date, the drawing revision and unit serial numbers tested, a conditions table, a photo set showing the sealed interfaces before and after rather than a wide shot of a tank, pass or fail, and the post-test functional and disassembly findings.

The drawing revision is the line people forget. A report that doesn't name the revision it tested validates a pack that may no longer exist by the time the report lands. When the revision moves, the report goes stale and nobody notices until an audit.

Accredited third-party versus the supplier's own lab is a genuine tradeoff, not a rule. Third-party reports travel further with your customers and add lead time and cost. An in-house lab is workable if you witness the test and agree the report format before the water goes on. Agreeing it afterward is how you end up with row three.

Ask what happens when a unit fails

The answer sorts suppliers faster than any capability presentation.

The weak answer is: we would test another sample. The strong answer sounds like a process. They describe a pre-test leak check — pressurize the enclosure, watch the decay, find the leak before the lab time is spent — and they can tell you which interface they'd suspect first, the gasket compression they designed for, and where the freeze point sits on the design. That answer means they built to the requirement instead of testing for it at the end.

Then ask the follow-up that matters more: is that leak check on qualification samples only, or on every unit coming off the line? The second answer tells you whether the rating is a property of the product or a one-time event in a lab.

Write the retest trigger into the spec

Define it: any change touching a sealed interface triggers a retest to the same procedure with the same evidence package. Gasket material or durometer, gland, vent, connector family, weld, potting, adhesive, fastener torque, enclosure tooling. Add production deviations that touch a sealed interface — a seam gap, a substituted screw, a different gasket supplier. Without that sentence the rating drifts with the first unannounced change and you find out from a field return.

Put it on the scorecard

If IP is a checkmark in the evaluation, the cheapest bid scores as technically equal to an accredited report with a full photo set, and price decides. Weight it as a scored line with anchors you write in advance. No report scores at the bottom. A self-declaration with no test behind it sits just above. An in-house report with no revision traceability sits in the middle. An accredited report in the correct configuration, with the photo set, the serials, and a production leak check, scores at the top. Keep it as a gate as well as a score, so a bid without evidence can't win on arithmetic.

Strike these words

Waterproof. Water-resistant. Sealed. Weatherproof. Splash-proof. 'IP67 compliant.' 'IP67 or equivalent.' 'IP67 or better.' Each one hands interpretation back to the bidder, and the bidder will hand it back to you at the worst moment. 'Or better' is the worst offender, because IP68 is not better than IP67 unless you defined the depth and duration yourself.

I don't write these lines to make the pack better. I write them so the bids come back describing the same test and I can compare price to price, line to line, without a week of follow-up emails. A tight IP paragraph won't rescue a bad gasket design. It will keep you from paying for the argument later, and that's a different kind of money.