Home / Blog / Waterproof Technology

How Water Still Gets Into a Dry Bag

Four physical routes, and only one of them is the fabric

Water crosses a waterproof barrier in four ways: through an opening that was never sealed, along a capillary channel, through a hole, or by being carried in. Only the third involves the fabric failing, and it is the least common of the four in service. Understanding which route you are dealing with is the whole diagnostic problem, because each one has a different signature and a different fix.

An unsealed opening is the roll-top that was closed over a strap, over a stuff sack corner, or with too little film left to fold. A capillary channel is the classic grain of sand: it holds the two film faces a fraction of a millimetre apart and lets surface tension do the rest. A hole is a puncture, a pinhole at a weld edge, or an abraded patch. Carried-in water is a wet item placed inside the bag, or condensation forming on the inner face.

The signatures differ. An unsealed opening delivers water fast and in volume, usually within minutes of the bag being submerged or hit by a wave. A capillary path delivers a slow trickle that only shows after an hour or more. A hole delivers a spray under pressure and almost nothing at rest. Carried-in water appears as a fine film of droplets on the inner surface, evenly distributed, with no point of entry anywhere on the bag.

Overfilling: the failure that outnumbers all the others

A roll-top closure needs a flat, clean band of film at least 55 mm deep to make three folds. Pack a bag to 95% of its nominal volume and that band no longer exists: the film is already taut around the contents, so the first fold becomes a crease rather than a wrap, and the second and third folds pull on the contents instead of on each other. The result is a closure that looks closed and behaves like an open one.

The geometry is unforgiving. On a 20L bag with a 240 mm mouth, filling to 19 litres leaves a band of roughly 40 mm, which allows two folds and no more. Two folds survive splash and light spray but not immersion, because there is only one continuous contact line between the water and the interior. The practical rule is to pack to 85% of nominal volume and never to let the contents touch the fold line.

A single grain of sand in the fold line

Capillary action does not need a hole. A sand grain around 0.5 mm across, trapped between the two film faces of a fold, holds them apart by its own diameter and creates a channel that water will travel along given time and a pressure difference. Once the channel is wetted, surface tension keeps it open, and the flow rate rises as the channel widens through flexing.

This is why a bag can be bone dry for two hours and then deliver 150 ml. The grit path fills slowly, then the whole load wicks. Hair, pine needles and dried salt crystals do the same thing, which is why paddlers who rinse the closure before rolling have dramatically fewer wet bags than those who do not. The fix is a fresh-water rinse of the fold line and a thumb run along it before every closure, which takes about four seconds.

Trapped air and the pump effect under pressure

Air inside a bag is not neutral. When a bag is submerged, hydrostatic pressure rises by roughly 0.1 bar per metre of depth, and a sealed bag with trapped air resists that pressure until the closure or the seam deflects enough to equalise. At that moment the pressure difference across the closure reverses for an instant and water is drawn in along the path of least resistance, which is usually the fold.

The second half of the pump effect is mechanical. A bag squeezed between a hatch rim and a gear load acts as a bellows: every wave flexes the body, and each flex cycle moves a small volume of air out past the closure and draws a small volume of water back in. Over an hour at 20 flexes per minute that is 1,200 cycles, and a leak path that passes 0.1 ml per cycle delivers 120 ml. This is precisely the failure mode an immersion test in still water does not reproduce.

Pressing the air out before rolling is therefore not just a packing-tidiness habit, it is a leak prevention measure. A bag rolled with the air expelled has no compressible volume to pump, and it sits flatter in a hatch, so it is squeezed less by the surrounding gear. If the bag has a purge valve, use it after packing and before the final fold, then close the valve.

Where a welded seam stops being continuous

A radio-frequency weld is continuous only if the electrode delivered even energy along its whole length. Where film thickness varies, where two layers became three at a corner, or where the electrode cooled between cycles, the bond is thinner and the seam is effectively narrower. Those thin spots are not leaks at rest; they are leaks under pressure, because the unbonded interface opens by a few microns when the film is stressed.

Corners are the classic location. A mitred corner has film folded over itself, and if the weld does not fully penetrate both layers the corner becomes a small pocket that opens outward under load. The same applies to the point where a D-ring patch meets the body seam: three layers, a step in thickness, and a stress concentration where the bag is lashed. Look for whitening at a weld edge after a load test; that is the bond starting to peel.

Age matters as much as geometry. A weld that was marginal on day one becomes a leak after two seasons of flexing, because each flex cycle advances the peel front a little. That is why seam failures cluster in years three to five rather than in the first season, and why a bag that has been stored tightly folded for months should be re-tested before a trip rather than trusted.

Hardware penetrations: valves, windows and zips

Every feature that passes through the film is a potential route. A purge valve has a body, a seal and a cap, each of which can fail independently; a clear window is a second material welded into the panel, with a different melt behaviour and a different expansion rate; a waterproof zipper is a mechanical closure with a long perimeter and a slider that must be fully seated to seal. None of these are bad ideas, but each adds a route that a plain welded body does not have.

Valves are the most commonly misused. A valve left open or half-closed is simply a hole, and a valve whose cap seal has picked up grit will weep slowly for an entire trip. Windows are the most commonly damaged, because a clear TPU panel is usually thinner than the body film and abrades faster, and because it is often placed exactly where a strap crosses the bag. Zips fail at the ends, where the tape terminates and the slider stops.

The decision rule is simple: a feature is worth its risk only if you use it every trip. A valve earns its place on a compressible load that must sit flat in a hatch. A window earns its place on a bag stored out of sight in a dark compartment. A zip earns its place where one-handed access matters more than immersion protection. Otherwise, specify the plain body.

Pinholes, abrasion and damage you cannot see

A puncture from a fishing hook, a tent pole end or a knife tip is obvious. What is not obvious is abrasion damage: a patch of film polished smooth against a boat rail, a rack bar or a concrete step loses its surface layer and with it most of its puncture resistance, long before it perforates. A 0.5 mm film abraded to 0.2 mm will still hold water at rest and will fail the moment it is flexed under load.

Pinholes at weld edges are the third invisible category. These are defects measured in tens of microns, produced when the weld energy was slightly too high and the film was thinned at the edge of the bond. They pass a visual inspection and a bath test, then open under pressure. The way to find them is a pressurised air test rather than a water test, because air at 0.1 bar escapes through an aperture that water will not enter against surface tension.

Creases, cold and the fatigue that arrives quietly

Film fatigues at creases. Every time a bag is folded along the same line, the polymer at that line flexes, and repeated flexing eventually produces a whitened band where the material has crazed. A bag stored folded flat for a winter can develop a permanent crease that no longer seals even when rolled, because the film has taken a set and will not conform to itself.

Cold makes this worse. TPU remains flexible well below freezing, but PVC-coated fabrics stiffen noticeably, and a stiff fold will not conform around grit or around a strap. A bag that seals perfectly at 20 °C can weep at -5 °C for exactly that reason. If you paddle or travel in winter, store the bag loosely rolled, warm the closure in your hands before folding, and add a fourth fold.

Water that never crossed the wall at all

A meaningful share of reported leaks are not leaks. Warm, moist air sealed inside a bag at 24 °C will deposit condensation on the inner film when the temperature drops to 8 °C overnight, and the droplets that form look exactly like ingress. The giveaway is distribution: condensation appears evenly across the inner surface and on the top of the contents, while a real leak is heaviest near the entry point and pools at the bottom.

The second source is carried-in water: a damp towel, a wet swimsuit, or a water bottle with a weeping cap placed inside a bag that was then sealed. A sealed body has no drain and no vapour path, so that moisture stays in the bag and spreads through the whole load. Twenty millilitres from a damp cloth is enough to make a fleece feel wet by the morning.

Depth, time and the numbers behind each route

Pressure is what converts a marginal seal into a leak, and it rises fast. At 0.3 m the differential across a closure is about 0.03 bar, which almost nothing leaks through. At 1 m it is 0.1 bar. At 3 m it is 0.3 bar, which is enough to open any capillary channel and any thin weld. This is why a bag can be reliable on a paddle board and unreliable when strapped to a boat that rolls in a swell.

The last row is the one that catches people out. No ingress protection rating tests cyclic flexing under load, because the IEC ingress code in IEC 60529 describes static jets and static immersion. A bag can legitimately carry an IPX7 mark and still deliver water when it is strapped into a hatch and flexed for an hour. That is a use case, not a defect, but it is worth knowing before you blame the bag.

Time matters as much as depth. A capillary path delivers a few millilitres per minute, so a 20-minute immersion shows almost nothing and a four-hour immersion shows 200 ml. When you test at home, match the duration to the trip, not to the standard. A 30-minute bath test predicts a wet exit; a four-hour weighted test predicts a day strapped to a deck.

A five-step field diagnosis in fifteen minutes

Run these in order at camp, with the bag still as you found it. Do not open it first, because the position of the water inside tells you where it came from. The whole sequence takes about fifteen minutes and needs no tools beyond a towel and a phone torch.

Step four deserves a note on method. Submerge the bag with the closure above water where possible, press gently, and look for a stream of bubbles, which locates the defect precisely. If the closure must be submerged, expect a few bubbles from the fold as trapped air escapes; a continuous stream is the signal, not the first few bubbles. Mark the spot with a pencil before the bag dries, because a pinhole is very hard to find once it is dry.

The first ten minutes with a wet bag

1. Keep the bag closed and upright; note where the water sits before opening it.
2. Tip the water out and weigh what came in by decanting into a measuring cup.
3. Separate wet items immediately and dry them; do not leave them in the bag.
4. Inspect the fold line for grit, then rinse it with fresh water and re-roll.
5. Run the squeeze test, then a 30-minute weighted bath before the next paddle.
6. If a hole is found, dry the area fully and apply a patch from the repair kit.

The measuring cup matters more than it sounds. Fifty millilitres points to condensation or a capillary path; 400 millilitres points to an open fold or a hole, and the two require completely different responses. Recording the volume also tells you whether the problem is getting worse across a trip, which is the clearest signal that a seam rather than a closure is at fault.

Drying matters as much as diagnosis. A bag that has taken water must be emptied, turned inside out and dried in shade before it is stored, because trapped moisture at a weld edge will start to lift the bond within a week. Full routines are in our roll-top cleaning guide, and repair options for a confirmed hole are covered in patch kits for TPU bags.

Design choices that close each route

Each route has a design countermeasure, and the good ones cost very little. Overfilling is addressed by sizing the body so that 85% of nominal volume leaves a full 60 mm fold band, which means the mouth should be narrower relative to the body height than many bags make it. Grit is addressed by a smooth, wide mouth with a single reinforced fold line rather than two stiff edges that hold particles.

On the inspection side, an air inflation test is the single highest-value check a production line can run, and it is cheap. A bag inflated to 0.1 bar and held for 30 seconds will reveal pinholes, thin welds and valve defects that a water test will not, because air escapes through apertures water cannot enter. The step-by-step method is set out in the air inflation leak test, and the sampling logic behind it in AQL inspection for dry bags.

Closure geometry is the other half, and it is the half users interact with every trip. Three folds is the standard because it gives three contact lines, and the third one is the one that survives grit in the first. The mechanics are explained in how roll-top closures seal, and the argument for a fourth fold in cold or rough conditions in how many rolls are actually needed.

If you are writing a specification, require the test standard to be named rather than the adjective, and require the test to match the use case. The ISO and IEC methods define what a rating means, and a bag that will be strapped to a deck needs a cyclic flex test that no rating currently covers.

Two further reading points close the loop. The pressure behaviour behind every route here is set out in pressure differential in dry bags, and the question of how long a seam stays sound across a service life is answered in seam integrity over service life. Compare any bag you own against the roll-top dry bag range to see which of these countermeasures are built in, and bring a field versus laboratory testing mindset to your own kit before the next trip.