Home / Blog / Waterproof Technology
Field vs Lab Testing for Dry Bags
Lab testing and field testing do not compete; they answer different questions. A lab immersion at one metre for 30 minutes, run to IEC 60529, tells you whether a specific closure geometry and seam set are sound today, with a measurable pass band and a cost of about two hours per batch. A twelve-week field trial with eight paddlers tells you whether the same bag still seals after creasing, salt, sun and sand have done their work, at roughly forty times the cost and with far more scatter in the data.
This guide sets out what each protocol actually controls, where the two methods disagree, how to size a sample so the result means anything, and how to correlate a line-side gate test with the returns you see a season later. The figures are working numbers: 100 litres per minute through a 12.5 mm nozzle, 20 to 50 mbar of inflation pressure, AQL 2.5 sample sizes of 200 pieces, and defect rates that show up in month four rather than week one. Use the tables to build one plan that uses both rather than choosing between them.
- Two testing worlds, two different questions
- What a laboratory test actually controls
- Immersion testing: the thirty-minute benchmark
- Spray and jet testing: IPX5 and IPX6 in practice
- Hydrostatic head: a fabric number, not a bag number
- Air inflation and pressure decay as a line-side gate
- What only the field finds
- Designing a field trial that produces usable data
- Sample sizes, confidence and the trap of three samples
- Correlating laboratory results with field returns
- Where saltwater changes the answer
- Building one plan that uses both
Two testing worlds, two different questions
A laboratory test asks a narrow question under fixed conditions: does water cross this barrier in this time, at this pressure, at this temperature? It is repeatable, it is cheap per unit of information, and it is the only way to compare two constructions fairly. Because the conditions are fixed, the answer is also narrow — it says nothing about how the bag behaves after three months of use.
A field trial asks the opposite question: does this bag keep working in the hands of people who over-fill it, sit on it, leave it in the sun and never rinse it? That answer is broad and messy, and it is the only kind of evidence that predicts warranty returns. The mistake programmes make is treating the lab result as a prediction of field life; it is not, it is a snapshot of a brand-new bag.
What a laboratory test actually controls
Every standard protocol fixes four variables and leaves everything else free: exposure type, intensity, duration and the sample condition at the start. IEC 60529 fixes immersion at one metre for 30 minutes in still water at ambient temperature with the sample dry on the outside. That is a strong, repeatable test and a weak model of a bag being dragged behind a kayak at three knots.
The value is comparability. Two labs running the same protocol get the same answer, which means a spec sheet written against the standard means something. The IEC publishes the grade definitions precisely so that a claim can be checked; a claim written as adjectives cannot be. Insist on the protocol, not on the prose, whenever you read a waterproof rating.
| Test | What it controls | Sample size | Time per batch | What it misses |
|---|---|---|---|---|
| IPX7 immersion, 1 m, 30 min | Closure and seam continuity | 3-5 pieces | 2 h | Wear, UV, over-filling |
| IPX6 jet, 100 L per min | Resistance to driven water | 3 pieces | 45 min | Crease fatigue |
| Hydrostatic head, fabric | Film and lamination resistance | 5 swatches | 1 h | Seam and closure behaviour |
| Air inflation, 20-50 mbar | Pinholes in welds | 100% of pieces | 40 s each | Where the leak will appear later |
| Seam peel, 50 mm strip | Weld energy consistency | 6 strips | 30 min | Geometry effects |
Immersion testing: the thirty-minute benchmark
Thirty minutes at one metre is the industry default for a reason: it is long enough for a capillary path to deliver a measurable amount of water and short enough to run in batches. The practical detail that changes results is air. A bag that goes into the tank with air trapped inside floats, and the part above the waterline is not tested at all; a weighted bag with the air pressed out tests the whole shell.
Water temperature matters more than people expect. Film stiffness rises as the water cools, so a closure that seals at 25 °C can weep at 5 °C simply because the fold does not conform as tightly. Run winter-market validation at the low end of the service range, and record the temperature on the test report; our immersion testing guide sets out the full setup.
Spray and jet testing: IPX5 and IPX6 in practice
Jet testing models a completely different exposure. IPX6 drives 100 litres per minute through a 12.5 mm nozzle from about three metres for at least three minutes, which is far more aggressive than the still-water immersion in one specific way: it applies local pressure at a point rather than uniform pressure everywhere. A closure that holds immersion can still leak under a jet if the fold is not tight enough.
This is why commuter and paddling bags are often specified to both grades. Rain at 40 km/h of coastal wind arrives as a jet event; a capsize arrives as an immersion event; a bag dragged alongside a boat sees both within one afternoon. Run the jet test with the bag in the orientation it will actually be used in, because a nozzle pointed at the closure and a nozzle pointed at the base produce very different results.
Nozzle distance is the variable most reports leave out, and it is the one that quietly decides the outcome. The standard fixes flow and nozzle diameter but the stand-off distance is set by the rig, and moving a bag from 3.0 m to 2.5 m raises the impact pressure at the surface enough to change a pass into a marginal fail. Photograph the rig, note the distance on every report, and use the same rig for every comparison, or the numbers are not comparable across seasons.
Hydrostatic head: a fabric number, not a bag number
Hydrostatic head is measured on a flat swatch under a rising column of water and reported in millimetres. A well-made 0.5 mm TPU laminate will run past 10,000 mm without penetration. That figure describes the film and the lamination, and it says nothing at all about the needle holes, the weld corners or the roll-top, which is where real bags leak.
Use hydrostatic head as an incoming material check and never as a product claim. It is genuinely useful for catching a bad lamination lot before it reaches the welding floor, and it is worthless for predicting whether a finished bag keeps a sleeping bag dry. The distinction is covered in more depth in our water column test notes.
Air inflation and pressure decay as a line-side gate
The fastest way to find a pinhole is to inflate the sealed bag to 20 to 50 mbar and watch a gauge for 30 seconds. A sound bag loses under 10% of its pressure in that window; a weld with a 0.3 mm channel loses it in under five seconds. Because the test takes roughly 40 seconds per piece, it can be run on 100% of production rather than on a sample.
Air also finds leaks that water hides, because gas moves through a smaller path than liquid does. The trade-off is that air testing does not tell you where a leak will matter in service: a pinhole in the base of a bag is fatal, while the same pinhole 5 cm below the fold line never sees water. Pair the air gate with a periodic water test on the same production to keep the two correlated.
What only the field finds
Four failure modes essentially never appear in a lab and dominate real returns. The first is crease fatigue: the fold line works, stiffens and finally cracks after hundreds of rolls. The second is grit in the fold, which creates a capillary path that no clean-bag test can reproduce. The third is buckle and webbing wear at the point where a strap rubs a deck. The fourth is UV stiffening of the whole shell.
None of these is a lab failure, and none of them means the lab test was wrong. They are time-dependent, use-dependent and contamination-dependent, which is exactly the class of problem a controlled protocol is designed to exclude. If your programme only ever runs lab tests, expect the first returns to arrive in month four, and expect them to be concentrated at the fold line.
Designing a field trial that produces usable data
A field trial is worthless unless it is instrumented. Give each tester a bag, a log card and a fixed inspection routine: a monthly air test, a monthly immersion check, and a photographic record of any crease or scuff. Without that structure you get anecdotes, and anecdotes cannot be turned into a design change or a specification tolerance.
Size the trial for the failure rate you expect. If a design change is meant to cut crease cracking from 8% to 2%, you need roughly 150 bags per arm observed over a season to detect that difference with any confidence. Eight enthusiastic testers will find the obvious problems and miss everything else; the number is the whole point.
| Variable | Laboratory setting | Field reality | Effect on the result |
|---|---|---|---|
| Water temperature | 20 °C, controlled | 5 to 28 °C | Cold film seals less tightly |
| Fill level | Around 60% | 70 to 100% | Over-filling shortens the fold |
| Closure rolls | Three, by a technician | Two to four, by the user | Two rolls leak under pressure |
| Contamination | Clean fold | Sand, salt, sunscreen | Grit creates capillary paths |
| Cycles | One | 200 to 400 rolls | Crease fatigue appears late |
Sample sizes, confidence and the trap of three samples
Three samples is a pilot, not a validation. It will catch a design that is badly wrong and it will miss everything subtle: a 5% defect rate has a good chance of not appearing in three pieces at all, and a 1% rate will never appear. For a production gate, the answer is statistical sampling against a defined acceptable quality level rather than a fixed number of pieces.
AQL 2.5 is the usual level for sewn and welded soft goods, and the sampling tables published by ISO give the sample size for each lot band. At general inspection level II, a lot of 3,200 to 10,000 pieces draws a sample of 200 with an accept number of 10 and a reject number of 11. That is a real gate; three bags pulled off the line is not.
Cost pushes in the other direction, and it has to be weighed honestly. A full immersion run on 200 pieces takes most of a working day and occupies a tank, which is why the air gate exists on the line and the water test exists in the audit room. The sensible split is cheap and fast at 100% coverage, slow and expensive at a few per cent, with the two linked by a monthly correlation check on the same production run.
| Lot size | Sample size | Accept | Reject | Typical defects seen |
|---|---|---|---|---|
| 281-500 pieces | 50 | 3 | 4 | Closure misalignment |
| 501-1,200 pieces | 80 | 5 | 6 | Weld skips at corners |
| 1,201-3,200 pieces | 125 | 7 | 8 | Buckle stiffness |
| 3,201-10,000 pieces | 200 | 10 | 11 | Print registration, pinholes |
Correlating laboratory results with field returns
The useful question is not whether the two methods agree, but how far apart they are and why. Build a simple matrix: for each failure mode, record whether the lab caught it, whether the field caught it, and at what point in service it first appeared. That matrix tells you where to spend money — on a better bench protocol, on a longer trial, or on a design change.
In practice the matrix is lopsided in a predictable way. Laboratory work catches material and geometry faults early and cheaply. Field work catches time- and contamination-driven faults late and expensively. A programme that spends 90% of its test budget on the bench and 10% on the field is normal and roughly right; one that spends nothing on the field is how a 4% return rate happens.
| Failure mode | Caught in the lab | Caught in the field | First appears |
|---|---|---|---|
| Weld pinhole | Yes, air test | Sometimes | Week 0 |
| Closure geometry error | Yes, immersion | Yes | Week 0 |
| Sand in the fold | No | Yes | Week 2 |
| Over-filled closure | No | Yes | Week 3 |
| Crease fatigue | Rarely | Yes | Month 4 |
| UV stiffening | Only accelerated | Yes | Month 6 |
Where saltwater changes the answer
Salt water at 35 grams per litre is only about 3% denser than fresh water and has very slightly higher surface tension, which means a bag that holds fresh water will generally hold sea water in a bench test. The difference is everything that happens after the test: crystals left in the fold, on the buckle and along the weld line, and accelerated corrosion of metal hardware.
Run coastal programmes with a salt-specific protocol: immerse in a 3.5% solution, dry without rinsing, then re-test. Bags that pass wet and fail after drying are telling you something a fresh-water test never will, because the crystal growth in a fold is a mechanical problem rather than a sealing one. Full protocol detail is in our saltwater versus freshwater testing guide.
Hardware is the second salt problem and the one that ends service life rather than just causing a damp load. A stainless spring in a cam buckle will pit in weeks if it is never rinsed, and a pitted spring loses the tension that keeps a strap from creeping under load. Specify 304 or 316 stainless for anything that contacts sea water, rinse it on the same schedule as the shell, and treat any stiffness in a cam lever as a hardware fault rather than a fabric one.
Building one plan that uses both
- Gate 100% of production with a 40-second air inflation test at 20 to 50 mbar.
- Run IPX7 immersion on three pieces per 500 and record water temperature on every report.
- Add IPX6 jet testing for any bag sold as a commuter, paddling or deck bag.
- Keep hydrostatic head as an incoming material check, never as a product claim.
- Sample finished lots at AQL 2.5, general level II, and log the defect categories.
- Run one instrumented field trial per platform per season with at least 150 bags.
- Review the lab-versus-field matrix quarterly and move budget to whichever side finds less.
- Compare the resulting specification against the full product line before launch.
Two protocols deserve a permanent place in that plan whatever the product: a destructive seam peel on six strips per shift, which measures weld energy directly, and a quarterly accelerated UV run, which is the only cheap proxy for the stiffening you will otherwise discover in month six. Both are covered in more detail in quality control for waterproof bags and warranty testing standards.
The economics are straightforward once the plan exists. Bench work costs a few hours per batch and catches the faults that would otherwise ship as a bad lot; field work costs a season and catches the faults that would otherwise become warranty claims. Our production partner has run both since 2014 in an SGS-verified 4,950 m2 facility with seven lines and a 200,000-piece monthly capacity, sampling in 6 to 10 working days and bulk in 35 to 50 days.
If your immediate question is narrower than a full programme, two shorter guides will help. For a single bag on a kitchen table, start with testing a dry bag at home, which uses a bath and a 30-second squeeze. For the fastest line-side check, read the air inflation leak test, then set your sampling level with AQL inspection for dry bags.
Frequently asked questions
Q1. What is the difference between lab testing and field testing for dry bags?
Lab testing asks whether a new bag seals under fixed conditions such as one metre for 30 minutes; field testing asks whether it still seals after creasing, salt and sun. The lab is cheap and repeatable, the field predicts warranty returns, and a programme needs both.
Q2. Is IPX7 immersion enough to validate a dry bag?
No. It validates the closure and seam continuity of a brand-new bag at ambient temperature. It will not find crease fatigue, grit in the fold, over-filling or UV stiffening, which are the four failure modes that dominate real returns.
Q3. How many samples should I test?
Three is a pilot, not a validation. For a production gate use AQL 2.5 at general inspection level II: a lot of 3,201 to 10,000 pieces draws 200 samples with an accept number of 10 and a reject number of 11.
Q4. Why does a bag pass in the lab and fail on the water?
Usually because of fill level, closure rolls or contamination. A technician closes with three clean rolls at 60% fill; a user closes with two rolls at 95% fill with sand in the fold. Cold water stiffening the film is the other common cause.
Q5. How long should a field trial run?
A full season, and at least twelve weeks. Crease fatigue typically appears around month four and UV stiffening around month six, so anything shorter will only surface the faults the bench could have caught anyway.
Q6. Does water temperature change immersion results?
Yes. Film stiffens as water cools, so a closure that seals at 25 °C can weep at 5 °C. Always record water temperature on the test report and validate winter markets at the low end of the service range.
Q7. Is hydrostatic head a useful number for a finished bag?
Only as an incoming material check. It measures a flat swatch, so it says nothing about welds, corners or the roll-top, which is where finished bags actually leak. Treat a 10,000 mm figure as film quality, not as product performance.
Q8. Why use an air inflation test instead of water?
Air finds smaller paths and takes about 40 seconds per piece, so it can be run on 100% of production rather than on a sample. The trade-off is that it does not tell you whether the leak location matters in service, so pair it with periodic water testing.
Q9. Does saltwater make a dry bag leak?
Not directly in a bench test, because sea water is only about 3% denser than fresh. The damage happens after: salt crystals grow in the fold and along the weld line, and hardware corrodes, both of which cause failures a fresh-water test never shows.
Q10. What is AQL 2.5 and why is it used?
It is an acceptable quality level of 2.5% used with statistical sampling tables for sewn and welded soft goods. It converts a lot size into a sample size and an accept number, which is a real gate, unlike a fixed handful of pieces.
Q11. How much should a test programme cost?
Roughly 90% of the budget on bench work and 10% on field trials is normal and about right. Bench testing catches material and geometry faults early and cheaply; field testing catches time-driven faults late but is what prevents warranty claims.
Q12. Can accelerated UV testing replace outdoor exposure?
It is the only cheap proxy and it correlates well for colour shift and stiffening, but it cannot reproduce wet-dry cycling or abrasion. Use it to rank materials, then confirm the winner outdoors on at least one season of real use.
Q13. Should every bag be immersion tested in production?
No. Immersion destructive-to-time ratio makes it a sample test: three pieces per 500 is typical. Inflate and pressure-decay every piece instead, and use immersion to audit whether the air gate is still correlated with real water ingress.
Q14. What is the fastest way to find a pinhole in a weld?
Inflate to 20 to 50 mbar and watch the gauge for 30 seconds. A sound bag loses under 10% of pressure in that window, while a weld with a 0.3 mm channel loses it in under five seconds.
Q15. Do laboratory tests predict how long a bag will last?
Not on their own. They measure the condition of a new bag, not its durability. Service life comes from combining bench results with accelerated ageing and a season of field data, which is why the two are run together.
Q16. How do I compare two bag designs fairly?
Fix every variable except the design: same water temperature, same fill level, same number of rolls, same operator, same orientation in the jet or tank. Any one of those left free will dominate the result and make the comparison meaningless.
People also ask
How do you test a dry bag in a laboratory?
Fill it to about 60%, close with three rolls, submerge at one metre for 30 minutes with the air pressed out, then open and check for internal moisture. Record water temperature, run three to five pieces, and follow with an IPX6 jet test for bags sold for paddling or commuting.
Why do dry bags fail in real use but pass tests?
Because the test conditions are cleaner than reality. Users over-fill, roll twice instead of three times, and get sand and salt into the fold; the film also stiffens in cold water and after UV exposure, none of which a bench protocol reproduces.
What does IPX7 mean for a dry bag?
It means the bag survived 30 minutes of immersion at one metre in still water as defined in IEC 60529. It is a snapshot of a new, correctly closed bag and does not describe performance under jets, abrasion, cold or age.
How many dry bags should a field trial include?
At least 150 per design arm if you want to detect a drop in failure rate from 8% to 2%. Eight or ten testers will find obvious faults and miss everything subtle, and the sample size is what makes the result usable.
Is saltwater testing necessary for all dry bags?
Only for bags sold into coastal or marine use. Run a 3.5% solution immersion, dry without rinsing, then re-test, because the failure mode is crystal growth in the fold and corrosion of hardware rather than the water itself getting in.
Can air pressure testing damage a dry bag?
Not at 20 to 50 mbar, which is well below any stress the film will see in service. Do avoid inflating and then leaving a sealed bag in direct sun, where internal pressure can rise 15 to 20% from heating alone.
Need this in production volume?
QUANZHOU JUNYUAN BAGS has produced waterproof bags since 2014 in an SGS-verified 4,950 m2 facility in Fujian, China. Minimum order 500 pieces, sampling in 6–10 working days, bulk in 35–50 days, FOB Xiamen — write to service@junyuanbags.com with your spec and target market.