IP-Rated Battery Packs: Designing for Outdoor and Rugged Environments

Quick answer: an IP67 or IP68 number on a datasheet describes the enclosure, not the finished battery pack. Connector rating, cell format, altitude, vibration life, and thermal design under a sealed housing all sit outside that number and have to be engineered separately. Dan-Tech Energy, a custom Li-ion battery pack manufacturer with production in Germany and the US, treats all of them as part of the same design decision as the enclosure itself.

A pack that passes IP67 immersion testing on the bench still fails in the field. Usually not because the enclosure leaked. Because the connector wasn't rated to the same level, the pressure-equalizing vent was never specified, or the seal was never tested against three years of vibration instead of one afternoon in a test tank. The enclosure passed. The pack didn't.

Treated as a checkbox, an IP rating is where field failures come from. A pack ships with a sealed, tested enclosure. Six months later it's back on an RMA form: corrosion at the connector, a swollen cell that cracked the potting, a seal that held on day one and stopped holding after four hundred vibration cycles. None of that shows up in an IP67 immersion test. All of it shows up in the field, usually after the pack is already integrated into the customer's system and the fix is no longer cheap.

IP Ratings, Briefly: What the Numbers Actually Mean

The two-digit IEC 60529 format

IP ratings come from IEC 60529 and use two digits. The first covers solid particle (dust) ingress, 0 to 6. The second covers water ingress, 0 to 9 (9K for high-pressure, high-temperature washdown, tested under a related standard). "IP67" means level 6 dust protection (dust-tight, no ingress at all under vacuum) and level 7 water protection (temporary immersion, defined test conditions). Each digit is tested independently, and a rating only certifies what was actually tested: a housing rated IP6X for dust and never immersion-tested has no water rating at all, regardless of how sealed it looks.

IP65 vs IP67 vs IP68 vs IP69K: which one an application actually needs

The gap worth noticing is IP68. Unlike every level below it, IEC 60529 doesn't define a fixed depth or duration for level 8: the manufacturer sets the test condition and states it separately. Two enclosures can both carry "IP68" and mean different things by it. Before treating an IP68 spec as sufficient, ask the supplier what depth and duration it was actually tested to, not just the two digits on the sheet.

For most of Dan-Tech's customers, the real decision is narrower than the table suggests. Ground-based outdoor drones and AGVs operating in rain, mud, and standing water need IP67. Marine or fully submersible applications need a properly scoped IP68. IP69K almost never applies to airborne or mobile robotics use cases, it's a washdown-industrial spec, and asking for it by default adds enclosure cost with no matching failure mode to justify it.

The Connector and Cable-Gland Trap

A fully potted IP68 enclosure is worthless if the connector isn't rated to the same level

This is the gap most IP-rating guides skip entirely: the enclosure is one path into the pack. The connector, the cable gland, and any external vent are the others, and each one needs its own ingress rating for the pack's overall rating to mean anything. A housing that's IP68 on its own and paired with an unsealed connector is, in practice, rated to whatever the connector allows in. Ingress doesn't care which part of the assembly it entered through.

This is also where "IP-rated" claims get loose in marketing copy generally. A pack should never be described as simply "IP67 certified" as a blanket statement; certification and rated performance apply to a tested configuration, connector and gland included, not to the enclosure in isolation. The accurate claim is scoped: designed to meet IP67 with a specific connector and gland combination, verified for that build.

Rating context for Dan-Tech's actual connector options

None of the connectors commonly used in Li-ion pack integration carry the same ingress rating by default, and mixing them into an IP-rated design without checking that is the single most common version of this trap.

Look at the pattern: even the most rugged industrial connector option Dan-Tech uses, the Harting Han MPC series, ships at IP40 in its standard configuration, well below IP67. IP40 stops tools and wires above 1 mm; it does nothing against water or fine dust. If a project genuinely needs the pack's connector interface to hold IP67, that's a different housing/insert selection at the connector level, and it needs to be specified as its own requirement, not assumed because the enclosure around it is sealed.

The practical question to ask a supplier isn't "is the pack IP67?" It's "which connector, gland, and vent configuration was actually tested together, and to what rating did each element in that path hold?" A pack that answers that question with a real answer per component is a different level of engineering than one that answers with a single number on a spec sheet.

Sealing Technology vs. Cell Format: The Trade-off Competitors Skip

Pouch cells vent and swell; a fully sealed IP68 housing with no pressure relief traps that expansion

Pouch cells expand during normal cycling, not just in failure modes. A foil pouch cell like the Amprius SA08 (360 Wh/kg) or SA504 (383 Wh/kg) has no rigid can constraining that swelling; the pouch itself flexes. In an open or vented enclosure, that's a design parameter you allocate clearance for. In a fully potted or fully sealed IP68 housing with no pressure relief path, that same swelling has nowhere to go, and the pressure builds against the potting and the seal itself over the pack's cycle life.

This isn't a defect in the cell. It's a mismatch between a sealing decision and a cell-format decision made independently of each other, which is exactly how the two usually get decided in practice: mechanical engineering picks the enclosure strategy, and the cell gets selected on energy density and cost. Neither call is wrong on its own. Made without reference to the other, they can conflict.

A rigid cylindrical can behaves differently than a pouch stack inside the same sealed enclosure

A cylindrical cell's steel or aluminum can is a rigid mechanical boundary. Cells like the Molicel M65A (322 Wh/kg, 21700 format) or the Amprius SA112 (308 Wh/kg, 21700 format) still vent through their designed pressure-relief feature under a genuine fault condition, but they don't flex and swell against a sealed housing the way a pouch does during ordinary cycling. That's a meaningful part of why Dan-Tech defaults to 21700 cylindrical NMC cells for the majority of applications, reserving pouch formats for cases that specifically need the higher energy density a pouch format like SA08 or SA504 can reach, or a non-standard form factor a cylindrical cell can't fill.

For a sealed, IP-rated enclosure specifically, that default has a second justification beyond energy density and packaging: less internal mechanical margin has to be designed into the enclosure to accommodate cell behavior over the pack's life. It's not that pouch cells can't go into a sealed enclosure. It's that doing so correctly means designing venting and internal clearance for pouch swelling as a first-class requirement, not an afterthought discovered during qualification testing.

Application-Specific Guidance: Drones, Robotics, and High-Discharge Use

Drones and UAVs: altitude and pressure-differential venting

An IP67 rating is verified at or near sea level, in a static immersion tank. It says nothing about what happens to that same seal when a drone climbs from ground level to a few hundred meters of altitude and the pressure differential across the enclosure changes. Ground-based outdoor and marine applications, the use cases most generic IP-rating guides are written for, never encounter this. Flight does.

A sealed enclosure with no pressure equalization can stress the seal on every climb and descent cycle, and temperature drops at altitude can drive condensation inside a housing that was never designed to breathe. A Gore-type pressure-equalizing vent lets air pass while blocking liquid water and dust, keeping the enclosure at IP67/IP68 for ingress while removing the pressure differential that would otherwise work against the seal on every flight. For a drone/UAV pack, this vent isn't optional hardware bolted on after the fact; it belongs in the enclosure spec from the start, alongside the connector.

Robotics and AGV: vibration-driven seal degradation over field life

An IP67 test happens once, on a new unit, under controlled conditions. A robotics or AGV deployment runs for years under continuous mechanical vibration. Gaskets take a compression set under sustained load: the material's ability to spring back and maintain sealing pressure degrades over repeated compression cycles. A seal that held IP67 on day one of a multi-year deployment isn't guaranteed to hold it in year three, and nothing on the original test report tells you when it stopped.

This matters specifically for robotics and AGV integrations because the mechanical environment (motor vibration, wheel/track shock, structural resonance) is sustained and repetitive in a way a drone's flight loads aren't. A pack spec built for this application should ask about gasket material and expected compression-set life under the actual vibration profile, not just the initial IP rating, and should build a seal-inspection interval into the maintenance plan rather than assuming the day-one number holds indefinitely.

High-C-rate discharge under a sealed housing: the thermal trade-off

Sealing a housing for ingress protection and cooling a pack under high discharge current pull in opposite directions. A vented, open-frame enclosure moves heat by convection almost for free. A sealed IP67/IP68 housing blocks that airflow path along with the water and dust, and the heat generated by cells discharging hard, cells like the Reliance RS60 (50 A continuous / 80 A pulse) or the Amprius SA124 (50 A continuous / 180 A pulse, 2-second rating) in drone and robotics applications, has to go somewhere else.

There is no cost-free answer here, only a trade-off to make deliberately: active cooling (adds weight, power draw, and a new component to seal or vent correctly), larger pressure-equalizing vent area sized for thermal exchange as well as pressure relief, or a discharge derating curve that keeps the sealed pack inside its thermal limits at the cost of peak power. Which one is right depends on the duty cycle. A pack spec that names a target IP rating without naming a thermal strategy for high-current use hasn't actually finished the design.

What to Watch Out For: Integration Failures Nobody Mentions Upfront

Field-serviceability vs. IP rating for hot-swap and quick-release packs

A fully potted, permanently sealed pack is the easiest configuration to rate and hold at IP67/IP68, and the worst configuration for a robotics or AGV fleet that needs to pull batteries in the field and swap in a charged one in minutes. A quick-release connector interface has to be mated and unmated hundreds or thousands of times over the pack's service life, and every mating cycle is a chance to wear the connector seal, misalign the interface, or introduce debris at the exact point where ingress protection depends on a clean, repeatable seal.

This is a real design requirement to raise with a supplier before the enclosure is finalized, not after: how many mate/unmate cycles is the connector interface rated for while holding its ingress rating, and does that number actually match the fleet's real swap frequency over the deployment's planned life? A connector chosen for current rating and cost alone, without checking its rated mating-cycle life against IP performance, is a common source of a pack that ships sealed and stops sealing within months of field use.

Where IP testing sits in the certification program timeline

IP ingress testing is not part of the same certification family as UN38.3 (transport safety) or IEC 62133-2 (cell and battery safety), and it doesn't run on the same clock. UN38.3 and IEC 62133-2 are electrical and mechanical abuse tests performed on the cell or finished pack under defined fault and stress conditions. IP testing is an enclosure ingress test, and it depends on the mechanical design, connector, gland, and vent selection all being finalized first, since changing any one of them after the fact invalidates the previous ingress test.

Practically, that means IP testing sits later in the program than it might look on a generic certification checklist: it can't start meaningfully until mechanical integration is locked, while transport and cell safety testing can often proceed against the cell and BMS design in parallel, earlier. A program schedule that treats "certification" as one undifferentiated block risks discovering the IP-testing dependency on mechanical finalization only after the schedule has already been committed. Sequencing this correctly, and scoping which standards apply to a given project, is exactly the kind of conversation to have before a build starts, not during it.

Summary: Key Decisions

• An IP number describes the enclosure, not the pack. Connector, gland, and vent each need their own ingress rating for the whole assembly to hold the stated level.
• IP68 has no fixed test condition in IEC 60529. Ask what depth and duration a given IP68 claim was actually tested to before treating it as sufficient.
• Match connector rating to the enclosure rating deliberately. Bullet connectors (XT60, XT90, AS150U) carry no ingress protection of their own; even Harting's Han MPC30/MPC60 industrial connectors ship at IP40 standard, well below IP67.
• Pouch cells swell during normal cycling; sealed enclosures need to account for it. A cylindrical cell's rigid can removes one variable from that trade-off, which is part of why cylindrical is Dan-Tech's default recommendation.
• Drone and UAV packs need pressure-equalizing venting, not just an ingress-tested housing, to handle altitude-driven pressure and condensation.
• Robotics and AGV packs need a seal-life plan, not just a day-one IP test. Gasket compression set degrades sealing performance over years of vibration.
• High-C-rate discharge under a sealed housing forces a real thermal trade-off between active cooling, vent sizing, and derating. Name the strategy, not just the target rating.
• Field-serviceable, hot-swap packs need a rated mating-cycle life on the connector, checked against real fleet swap frequency, alongside its ingress rating.
• IP testing depends on final mechanical design and doesn't run on the same schedule as UN38.3 or IEC 62133-2. Sequence it into the program plan explicitly.
• Never describe a finished pack as blanket "IP67 certified." Scope the claim to the tested connector/gland/vent configuration, or use "designed to meet IP67."

FAQ

What's the practical difference between IP67 and IP68 for a battery pack?

IP67 means the enclosure survives temporary immersion to 1 meter for 30 minutes, a fixed, defined test. IP68 means continuous immersion beyond that, but IEC 60529 leaves the exact depth and duration to the manufacturer to specify, so two IP68 claims aren't automatically equivalent. Ask what depth and duration a specific IP68 rating was tested to.

Does an IP68-rated enclosure mean the whole battery pack is IP68?

Not by itself. The connector, cable gland, and any vent are separate ingress paths, and each needs its own matching rating. A sealed IP68 housing paired with an unrated connector performs at whatever the connector allows in, not at the housing's rating.

Can a pouch cell go into a sealed IP-rated enclosure?

Yes, but the enclosure has to account for pouch swelling during normal cycling as a design input, with clearance and, where needed, pressure relief. A cylindrical cell's rigid can removes that specific variable, which is one reason Dan-Tech recommends 21700 cylindrical cells by default and reserves pouch formats for applications that specifically need the added energy density.

Do drone batteries need pressure-equalizing vents even if they're already IP67?

For altitude-capable UAVs, yes. An IP67 rating is verified near sea level. Climbing changes the pressure differential across the enclosure on every flight, which stresses a fully sealed seal and can drive internal condensation. A Gore-type pressure-equalizing vent removes that stress while keeping the ingress rating intact.

Where does IP ingress testing fit relative to UN38.3 and IEC 62133-2?

It's a separate test family on a separate schedule. UN38.3 and IEC 62133-2 evaluate cell and pack safety under abuse and fault conditions and can often proceed in parallel with mechanical design. IP testing requires the mechanical design, connector, gland, and vents to be finalized first, since any change afterward invalidates the ingress test. Sequence it into the program plan rather than assuming it runs alongside the other standards on the same timeline.

Designing the enclosure, connector, venting, and cell format as one decision instead of four separate ones is exactly the kind of integration work a generic IP-rated enclosure supplier won't do for you. Dan-Tech's ToolBox is where that conversation starts. To see the cell formats and configurations available to build from, browse the Lithium-Ion Battery Packs catalog.