Battery Box Sealing and Leak Testing: IP Ratings and Test Methods

Sealing is where battery enclosures fail most often, and it is the specification most often written badly. "Watertight" appears on drawings every week. It is not a criterion — nobody can test against it, and a supplier who accepts it has not agreed to anything enforceable.

Two separate things have to be specified. The first is the rating: what level of ingress protection the enclosure is designed to survive. The second is the test: the method and the pass/fail number that proves a given part meets it. A box can hold IP67 in design and still ship leaking units if the second one is missing.

This page covers both and links to the detailed articles on each.

Step 1: the rating

IP codes come from IEC 60529. The two digits are independent — the first is solid ingress, the second is liquid.

The distinction that matters for a moving vehicle: IP65 resists a jet, IP67 resists being under water. They test different failure modes and neither implies the other, which surprises people who assume the scale is linear.

One more point worth knowing — IP ratings are type tests, not production tests. A rating certifies a design passed once under laboratory conditions. It says nothing about whether your thousandth unit is sealed. That gap is what step 2 exists to close.

The full breakdown of what each code tests and which one to specify is in IP54 vs IP65 vs IP67 for battery boxes.

Step 2: the test

Four methods cover production leak testing. They differ by two orders of magnitude in sensitivity and roughly two orders in equipment cost, so the choice is a real decision rather than a detail.

Sensitivity and cost move together, and the right choice follows from your acceptance criterion. If the spec says 1 × 10⁻⁶ Pa·m³/s, pressure decay physically cannot see it — the test will pass parts that leak, and the number on the report will be meaningless.

The reference for tracer-gas methods is ISO 20485:2017. How each station is set up, what the numbers mean, and how to read a supplier's report are covered in how battery boxes are leak tested.

Where leaks actually come from

When a production tray fails leak test, the cause is usually not the weld.

Connector interfaces are the most common leak point — high-voltage connectors, communication ports, vent membranes. These are sealed with gaskets or form-in-place seals, and they fail more often than the metal joints do.

Fastener holes and threaded features come next, particularly where a tapped hole breaks into a sealed cavity.

Welds are third, and among welds the failure is almost always at the root rather than the visible bead.

That ordering matters when you write the test specification: a test that only pressurises the cavity will find all three, but a test chosen for the wrong sensitivity will miss the small ones regardless of where they are.

How to write it into a purchase order

Four lines, and all four are needed:

1. The rating — "IP67 per IEC 60529" or whichever applies.
2. The method — "helium vacuum chamber per ISO 20485" or "differential pressure decay", named explicitly.
3. The number — "≤ 1 × 10⁻⁶ Pa·m³/s" or "≤ 50 Pa drop over 3 minutes at 10 kPa". A number without a method is untestable; a method without a number is unenforceable.
4. The coverage — "100% of production units" rather than sampling. This is the clause that actually protects you, because a batch certificate covers the batch, not the part in your hands.

Ask for records traceable to each serial number. That single requirement makes the difference between a supplier who tests and one who reports testing.

Where this gets difficult

Two honest limitations.

A sealed box is not a maintenance-free box. Seals age, gaskets take compression set, and vent membranes clog. A design that passes IP67 on day one may not on year three. If the pack is serviceable, specify the seal as a replaceable item rather than assuming the rating is permanent.

Tighter is not free. Moving from 10⁻⁴ to 10⁻⁶ Pa·m³/s sensitivity changes the equipment, the cycle time and the scrap rate. Parts that were acceptable at one level start failing at the next. Specify the tightest criterion your application actually needs — over-specifying costs money without adding safety.

Frequently asked questions

Is IP67 enough for an e-rickshaw or e-scooter pack? For most road use, yes — it covers immersion to 1 m for 30 minutes, which handles flooding and deep water crossings. IP65 is often sufficient where the pack is mounted above the wading line and only sees rain and washing.

Can I skip the production leak test if the design is IP67 certified? No. The rating is a type test on one sample. Production variation — tool wear, seal seating, casting porosity — is exactly what production testing catches.

Which test should I specify for a mid-volume program? Differential pressure decay is the usual answer: better sensitivity than simple pressure decay, cycle times compatible with a line, and equipment cost that is justified from a few thousand units up. Move to helium when the acceptance criterion or the safety case demands it.

What does Pa·m³/s actually mean? It is a leak rate — pressure times volume per second. The smaller the number, the smaller the hole it can detect. 1 × 10⁻⁶ Pa·m³/s is roughly a thousand times tighter than what pressure decay can resolve.

Should the test be done before or after anodizing? After finishing and after hardware fitment, on the complete assembly. Testing the bare shell tells you the shell holds pressure; it says nothing about the connector seals installed afterwards.

Related products

• E-rickshaw battery boxes — three-wheeler enclosures
• E-scooter battery cases — two-wheeler enclosures
• Aluminium battery trays and covers — pack-level trays
• Custom OEM and ODM enclosures

Related guides

• Aluminum battery box materials — alloys, tempers and finishes
• How aluminum battery boxes are made — processes and volume thresholds

Specifying a sealing requirement? Send the rating, the environment and your annual volume, and we will quote against a criterion you can actually test. WhatsApp +86 133 0570 9557 or info@aymetals.com. Our integrated production network covers extrusion, CNC, welding and leak testing, every stage audit-welcome.