Rain Test vs Submersion Test: Which One You Need

What Each Test Simulates

The two tests model different real events, and the confusion between them starts when a buyer treats both as generic evidence of waterproofness. A rain test models a product standing or being carried in falling water for a limited period. A submersion test models a product wholly under water, under pressure, for a sustained period, and then being retrieved.

Rain is directional. Water arrives from above and to one side, runs down the faces of the product, and collects at low points, at the base and along any horizontal seam or closure. It arrives with kinetic energy from falling droplets, but with essentially no static pressure differential, because there is no standing column of water above the product. That combination, low pressure and directional flow, is what makes a rain test good at finding closure faults.

Submersion is omnidirectional. Water presses against every surface at once, with a pressure differential that increases with depth. There is no runoff and no dry side, and there is no place for water to drain away from. The product is also handled while submerged and lifted out, which applies forces that do not occur in a rain test. That combination is what makes submersion good at finding pinholes, weld voids and capillary paths.

The real-world events behind each method

• Rain test models: a shower at a bus stop, a downpour on a commute, spray on a boat deck, a bag left on wet ground
• Submersion models: a bag dropped in a river, a dry bag strapped to a kayak, a bag submerged in a flooded car boot
• Neither models: a bag closed damp with a wet towel inside, which is a condensation and wicking problem
• Neither models: sustained contact pressure, for example kneeling on a wet bag, which is a hydrostatic problem

Note the pressure figure in that table, because it is the crux of the difference. One meter of water is roughly 0.1 bar, or about 10 kilopascals, which is small in absolute terms but applied to every square centimeter of the product simultaneously and continuously. A rain droplet arrives with far more local instantaneous pressure at the impact point but for a tiny fraction of a second over a tiny area. Sustained low pressure and momentary high pressure find completely different defects.

The practical consequence for a buyer is that you should never accept a rain test result as evidence for an immersion claim, and you should not pay for submersion testing on a product that only claims shower resistance. Match the method to the promise, and put both the claim and the method on the same line of the specification. Our ingress and seam test protocol is organized on exactly that principle.

Energy Levels Compared

Energy is the cleanest way to think about why the two tests find different things. There are two relevant forms: kinetic energy from moving water, and static pressure from a standing column. Rain testing has mostly the first and almost none of the second. Submersion has almost none of the first and a sustained amount of the second.

This matters because a closure behaves differently under the two. A zipper with a storm flap is designed to shed running water, and it does that well: water arrives, runs down the flap, and continues past. The same closure under submersion has water on both sides of the flap, and there is no downhill direction for it to shed to. Pressure then drives water through the coil by whatever path exists, and the flap contributes nothing.

Rough energy comparison

• Light rain: low kinetic energy, negligible static pressure, intermittent exposure
• Heavy rain or a shower cabinet: moderate kinetic energy, negligible static pressure, continuous exposure
• Partial immersion to 100 mm: negligible kinetic energy, about 0.01 bar static, continuous
• Immersion to 1 meter: negligible kinetic energy, about 0.1 bar static, continuous for the full duration
• Immersion to 3 meters: about 0.3 bar static, which is where most soft closures are decisively beaten

A hydrostatic head test sits outside this comparison but helps calibrate it. A fabric rated at 10,000 mm withstands a column equivalent to about 1 bar of static pressure on a flat specimen. That is an order of magnitude above what one meter of submersion applies. So if a bag at one meter leaks, it is almost never because the fabric failed; it is because the seam, the closure or a needle hole provided a path at far lower pressure than the fabric would have needed.

That single observation explains most buyer and supplier arguments in this category. The supplier quotes a hydrostatic result that is genuinely excellent and genuinely irrelevant to the failure. The buyer reports water inside the bag that is genuinely there. Both are correct, because the ingress path was never tested. Ask which path, then choose the method that exercises it.

There is also a duration dimension to energy. Capillary wicking is slow: water moving along a zipper tape or through a needle hole may take tens of minutes to produce a visible quantity inside a bag. A five-minute rain test will not find it. A thirty-minute submersion test will. If your product has any textile path from the outside to the inside, duration is as important as pressure, and it is the parameter most often shortened to save time.

One caution when comparing published figures: energy calculations depend on droplet size and velocity distribution, which vary between apparatus. Do not treat the numbers above as calibration data. Use them to reason about which method exercises which mechanism, and rely on the named standard with its stated parameters for anything contractual.

Failure Modes Each Test Exposes

Each method has a signature set of defects it finds reliably, and a set it reliably misses. Knowing both lists is more useful than knowing which test is harder, because it tells you what a pass actually means.

Read the middle of that table carefully. Three of the most common real-world failures, capillary wicking, needle holes and weld voids, are essentially invisible to rain testing. That is not a criticism of rain testing; it is a statement about scope. If your product has a welded wet compartment, an embroidered logo or a sewn label, you have potential paths that rain testing will not find, and only submersion will.

What rain testing finds better

Rain testing is superior for anything geometry-dependent. Because water runs downhill and collects, it finds a storm flap that is too short, a pocket that faces upward and fills, a grommet positioned where water pools, and a seam line that happens to be horizontal on the carried orientation. These are design faults that submersion will often miss entirely, because under water there is no downhill and no collection point.

It is also better for anything about how the product is used. A bag is carried in an orientation, opened and closed, and set down. A rain test can reproduce that; a submersion test cannot. If your concern is whether a commuter gets wet papers during a downpour, the rain test is the more faithful model even though it is the gentler test.

The third advantage is statistical. Because a rain test is usually non-destructive and fast, you can run many units. Twenty units through a rain cabinet gives you a distribution; three units through a submersion tank gives you an anecdote. Where the failure rate is low and the consequence is moderate, a larger rain sample is often better evidence than a small submersion sample.

The honest conclusion is that neither test is a superset of the other. They are complementary, and a programme that runs only one has a blind spot that is predictable in advance. Name the blind spot in your quality file so that when a failure appears in the field, you know immediately which test should have caught it.

Why a Rain Pass Does Not Imply Submersion

This is the point that causes the most commercial damage, so it is worth stating plainly: a product that passes a rain test can fail a submersion test by a wide margin, and this is normal rather than exceptional. There are four structural reasons, and none of them indicate a bad supplier.

Four structural reasons

1. No runoff: a storm flap works by shedding water downhill. Under water there is no downhill, so the flap contributes almost nothing.
2. Sustained pressure: a droplet impact lasts milliseconds. Thirty minutes at one meter applies continuous pressure to every path, however small.
3. Time for wicking: capillary movement along a tape or through a needle hole is slow. A short exposure ends before the water arrives inside.
4. Handling: retrieval applies forces that do not occur in a spray cabinet, and a bag lifted by one handle stresses its closure differently.

The second reason is the one to internalize, because it explains why the margin can be large rather than marginal. Consider a zipper coil with a gap of a fraction of a millimeter. Under rain, a droplet may cross that gap and stop, because there is no pressure to push it further and surface tension holds it. Under a meter of water, the same gap has continuous pressure behind it and the water keeps moving until it reaches the inside. The defect is identical; the outcome is completely different.

There is a marketing consequence that buyers control directly. If a hangtag says waterproof and the file only contains a rain test, the claim is unsupported, and in several jurisdictions that is a consumer protection exposure rather than a quality question. Either upgrade the test or soften the claim. Both are legitimate; the unsupported claim is not.

Note also what happens with enclosure codes. IEC 60529 defines the IP code and its test sequence for enclosures, and the immersion grades in that standard are often quoted on soft goods. The relevant point is scope: the standard was written for rigid enclosures, and applying it to a soft bag involves interpretation about how the sample is supported and closed. Many in the industry use the depth and duration as a reference. Say so in the documentation rather than printing a bare rating you cannot defend in a dispute.

A useful internal rule: never let a marketing claim be drafted before the test protocol is agreed. The protocol is a constraint on the claim, not a consequence of it. When the claim is written first and the testing is commissioned to support it afterwards, the protocol gets shaped to produce a pass, and that is how unsupported claims reach the market.

Finally, watch for the reverse error too. A product that passes submersion is not automatically good in rain, because submersion does not test runoff behavior or orientation. A dry bag that is genuinely submersible can still let water into an upward-facing pocket during a downpour. Test the claim you make, and if you make two claims, run two tests.

Sequencing Both in One Protocol

When a programme needs both, sequence matters. The general rule is to run the gentler, non-destructive, more informative test first, then the destructive one, because the destructive test consumes the unit and you want the cheaper information extracted first.

A workable sequence

1. Condition all units at 20 plus or minus 2 degrees C and 65 plus or minus 4 percent relative humidity for 24 hours
2. Run a visual and dimensional inspection on every unit and record it
3. Run the rain test on the full sample set, in the carried orientation, for the stated duration
4. Dry the exterior, inspect the interior, and record the result for every unit
5. Draw a sub-sample for destructive submersion testing, taken from units that passed the rain test
6. Submerge at the stated depth for the stated duration, retrieve, dry the exterior, and inspect
7. Photograph every unit interior before and after, and log the results against unit numbers

There is a deliberate choice in step five. Drawing the destructive sub-sample from units that passed rain testing tells you whether rain is a valid predictor for your product, which is genuinely useful information. Drawing it at random from the whole lot tells you the marginal failure rate, which is what you want for a lot-release decision. Decide which question you are asking before you sample, because the two give different answers.

For a lot-release decision, sample at random from the packed cartons, after packing, and record the carton numbers. For a design-validation decision, sample from units you know passed the earlier stage. Mixing the two is a common protocol error and it makes the result uninterpretable, because a failure could mean either a design problem or a lot problem.

Build in a re-test clause. State in advance whether a failure triggers a retest, a doubled sample, or a lot rejection, and who pays for the extra testing. Without that clause, every failure becomes a negotiation and the party with the most schedule pressure usually wins, which is not the party who should be making a quality decision.

One operational detail: allow drying time between stages. A unit that goes into submersion while still damp from the rain test carries free water on its exterior, and that water can enter during handling and be recorded as ingress. Dry the exterior thoroughly between stages and log the drying time, or your second-stage result will be contaminated by the first.

Keep the whole sequence within the production schedule. Sampling runs 6 to 10 working days for a standard construction and 12 to 15 working days when the build is complex, and bulk production runs 35 to 50 days after sample approval. Build the test window into that plan rather than discovering it at the end, and issue the report before the balance payment is due under T/T 30/70 terms.

Interpreting Borderline Results

Borderline results are the hardest part of ingress testing and the part most protocols ignore. A unit with three drops of water inside has failed a zero-tolerance criterion and passed a five-milliliter one, and the difference between those two statements is the difference between rejecting and shipping a container.

The first step is to define the criterion numerically before testing, not after. Choose a threshold, state it in the protocol, and record the measured quantity against it. A criterion expressed as visible moisture is not a criterion; it is an argument waiting to happen, because two observers will disagree about what is visible under poor lighting on a dark liner.

Common borderline cases and how to handle them

• Condensation: water droplets formed inside a humid bag are not ingress. Pre-dry the interior and compare against a control unit.
• Water at the closure only: indicates a closure path rather than a fabric or seam path. Record the location, not just the quantity.
• A damp patch on a liner: may be wicking through a needle hole from branding. Photograph and mark the position.
• One unit out of five failing: usually a process variation rather than a design fault. Double the sample before deciding.
• A result just over threshold on every unit: usually a marginal design. Treat it as a fail and fix the design.

That last case is the one buyers get wrong most often. If every unit is just over the threshold, the instinct is to relax the threshold, because the product is clearly close. That is the wrong call. A marginal design with a normal distribution of variation will ship a predictable percentage of units that are clearly over, and those are the returns. Fix the design or lower the claim.

Use a control unit in every run. Include one sealed unit with no path to the interior and one unit with a deliberate known defect. If the control with no path shows water, your handling or your drying step is contaminating the result. If the deliberate-defect control passes, your method is not sensitive enough to find the failure you are looking for. Both outcomes invalidate the run and both are cheap to detect.

Record location, not just quantity. A milliliter of water at the base seam is a different problem from a milliliter at the closure, and the fix is different. Photograph the interior with a marker showing where the water was found, and log the position against the unit number. Over a season, the pattern of locations tells you more than the pass rate does.

Finally, resist the temptation to average. An ingress result is not a property with a mean; it is a distribution with a tail, and the tail is what reaches the customer. Report the worst unit as prominently as the average, and set your acceptance rule on the tail rather than the middle.

Cost and Duration Differences

The two tests differ materially in cost, and the difference is mostly labor and consumed product rather than equipment. Understanding where the cost sits helps you decide how much of each to run.

The consumed-product line is the one that surprises buyers. At a realistic sampling rate over a season, submersion testing consumes a meaningful number of units, and those units have a fully absorbed manufacturing cost. Budget for them explicitly rather than absorbing them as a surprise, and remember that a unit consumed in testing is a unit you cannot sell.

Labor is the second surprise. A rain cabinet can run unattended during exposure, but mounting, inspecting, photographing and logging are all manual. A submersion run needs the same plus retrieval and drying. In practice the labor per unit is similar once the exposure is excluded, which means the cost difference between the two methods is smaller than the equipment comparison suggests.

There is a way to reduce both: use a fast screening method to select which units go into the expensive test. An air decay check is quick and non-destructive, and units that show a decay anomaly are the ones most worth submerging. This targeted approach finds more failures per unit consumed, and it is the usual reason to add an air decay detector to a bench.

Decide the mix by consequence rather than by cost. If a leak ruins a customer's electronics, the cost of the test is irrelevant and you should run submersion on a meaningful sample. If a leak mildly annoys a commuter, a large rain sample is better value than a small submersion sample. Match the rigor to the consequence of the failure, not to the price of the bag.

Choosing per Product Promise

The decision rule is short: choose the test that matches the promise, and run both only when the promise includes both. Below is the mapping we use, based on the claim language that appears on hangtags and listings.

Apply the same rule to your own documentation. If a listing says waterproof, the file needs a submersion result with a stated depth, duration and threshold. If a listing says water resistant, a rain result plus a hydrostatic number is sufficient and honest. The mismatch between claim and evidence is the single most common compliance failure in this category, and it is entirely within the buyer's control.

Then write it into the RFQ in one block: claim language, method name, parameter set, sample size, acceptance threshold, and who pays for re-testing. Six lines, and they convert an ambiguous quality discussion into a contract. Ask your supplier to quote against that block rather than against a generic request for waterproof testing.

The commercial terms close the loop. Bulk production runs 35 to 50 days after sample approval, quoted FOB Xiamen with T/T 30/70 payment terms, with the inspection report issued before the balance payment is due. That linkage is the most effective quality lever a buyer has, because it aligns evidence with money rather than with goodwill.

If you want a single recommendation: for any bag sold above a mid retail band with a waterproof claim, run both. Use rain testing on a large sample to catch geometry and closure faults, and submersion on a smaller destructive sample to catch the paths rain cannot see. The combined cost is modest, and the blind spot you remove is the one that generates returns. Reference the named methods, from AATCC 22 and AATCC 35 through IEC 60529 and ISO 811, and start from the waterproof bag product range or the custom programme path to scope your own protocol. Both methods run inside the production and inspection workflow, alongside our OEM waterproof bag programmes and custom dry bag manufacturing.

Frequently Asked Questions

Q1: Does passing a rain test mean a bag is waterproof?

A: No. A rain test applies directional water with almost no static pressure for a short period, so it finds closure and runoff faults. It does not find weld voids, needle holes or capillary wicking, which need sustained pressure from all directions. An immersion claim requires a submersion test with a stated depth, duration and threshold.

Q2: What is the difference between a rain test and a submersion test?

A: Rain testing delivers low-pressure water to one face of a product for minutes and is usually non-destructive. Submersion surrounds the product with water at depth for a sustained period, applying about 0.1 bar per meter continuously, and usually consumes the unit. They find different failure modes and neither replaces the other.

Q3: How deep and how long should a submersion test be?

A: One meter for thirty minutes is the common reference point for an immersion-grade claim on soft goods. Specify depth, duration, water temperature, whether the bag is loaded, and the failure threshold together. Shorter exposures miss slow capillary paths, so a five-minute dip is screening rather than qualification.

Q4: Can a bag pass submersion but fail a rain test?

A: Yes. Submersion does not reproduce runoff, orientation or water collecting at low points. A genuinely submersible dry bag can still fill an upward-facing pocket during a downpour. If the product carries both claims, run both tests, because neither result implies the other in either direction.

Q5: Which test is more expensive?

A: Submersion, mostly because it consumes the unit and needs retrieval and drying labor. Equipment capital is comparable. The labor per unit is similar once exposure time is excluded, so the real difference is the cost of the units destroyed, which should be budgeted explicitly rather than absorbed as a surprise.

Q6: Should I run both tests on every programme?

A: Run both when the claim is waterproof or submersible. Run rain only for water resistant or shower proof claims. The combined protocol is modest in cost and removes the blind spots that generate returns: rain catches geometry faults, and submersion catches pinholes, wicking and needle holes.

Q7: How should I define a failure threshold?

A: State a numeric quantity before testing, such as less than five milliliters of water inside, rather than a subjective phrase like visible moisture. Record the measured quantity and the location against each unit number. A criterion set after seeing the results is not a criterion, it is a negotiation.

Q8: What causes condensation inside a tested bag?

A: A humid interior forms droplets when the bag is opened and cooled, and those droplets are not ingress. Pre-dry the interior before testing and always run a sealed control unit alongside the sample. If the control shows water, your handling or drying step is contaminating the result and the run is invalid.

Q9: How many units should I submersion test?

A: Use 3 to 5 for design validation, 5 to 8 for process validation, and a destructive sub-sample drawn at random from sealed cartons for lot release alongside an AQL 2.5 plan under ISO 2859-1. Never draw destructive units from the shippable sample, because the two plans interfere with each other.

Q10: What is the right order for running both tests?

A: Run the non-destructive rain test on the full sample first, dry the exterior, inspect, then draw a destructive sub-sample for submersion. Dry thoroughly between stages, or water carried on the exterior will enter during handling and be recorded as ingress. Log the drying time for every run.

Q11: Why does a good fabric still leak at one meter?

A: Because the water almost never comes through the face fabric. A 10,000 mm hydrostatic rating is roughly an order of magnitude above the pressure at one meter, so the path is nearly always a seam, a closure, a needle hole or a wicking tape. Test the path, not just the material.

Q12: Is an IPX7 claim valid on a soft bag?

A: IEC 60529 defines the IP code and its test sequence for rigid enclosures, so applying it to a soft bag involves interpretation about support and closure. Many in the industry use the depth and duration as a reference. Document the interpretation you used rather than printing a bare rating you cannot defend.

Q13: What should I include in an RFQ for ingress testing?

A: Six lines: the claim language, the method name, the parameter set, the sample size, the acceptance threshold, and who pays for re-testing. That block converts an ambiguous quality discussion into a contract, and it lets a supplier quote against defined work rather than a generic request for waterproof testing.

Q14: How do test results fit the production schedule?

A: Sampling runs 6 to 10 working days for a standard construction and 12 to 15 working days when complex, and bulk production runs 35 to 50 days after sample approval. Testing happens at first article and pre-shipment, and the report should be issued before the balance payment is due under T/T 30/70 terms.

People Also Ask