A stitched seam cannot be waterproof, because the needle punches 5 to 7 holes per centimetre straight through the shell and every one of them is a capillary path. RF welding replaces thread with a molecular bond: two polar films are pressed between electrodes at 27.12 MHz, the molecules heat dielectrically, and the interface fuses into a single continuous layer in 1.2 to 3.5 seconds with no hole anywhere along it. The result is a seam 8 to 15 millimetres wide that reaches at least 90 percent of the base fabric tensile strength in peel, typically 150 to 300 newtons per 50 millimetre strip, and holds IPX6 to IPX7 without tape or sealant. Welded bodies start at 500 pcs per design and colourway, sample in 6 to 10 working days and ship in 35 to 50 days from FOB Xiamen.
Why a Stitched Seam Cannot Be Waterproof
The problem is the needle, not the thread. A lockstitch at 5 to 7 stitches per centimetre punches two holes per stitch through the shell, so a 400 millimetre seam carries roughly 400 to 560 holes, each around 0.3 to 0.5 millimetres across. Polyester and nylon thread then wicks: water enters at one hole, travels along the thread by capillary action, and appears at the next hole, which is why a stitched seam wets the inside of a pack even when the fabric itself is impermeable.
Tape over the stitch line is the standard workaround, and it works for a while. A polyurethane tape heat-pressed over the needle holes bridges them, but the bond is adhesive and it ages: hydrolysis, plasticiser migration and flexing lift the tape at the ends and at corners within two to four seasons. Welding removes the problem at the source by removing the holes, which is why every genuinely submersible pack body in the category is welded rather than sewn. The stitch-versus-seal argument is set out in full in why stitched seams leak.
| Seam method | Water path | Realistic level | Ages by | Cost index |
|---|---|---|---|---|
| Plain lockstitch | 400-560 needle holes per 400 mm, plus wicking | IPX2-IPX3 | Thread rot, hole elongation | 1.00 |
| Stitched and taped | Bridged by adhesive tape | IPX4-IPX5 | Tape hydrolysis and corner lift | 1.15 |
| Stitched, taped and sealed | Bridged, with sealant at the ends | IPX5 | Sealant cracking at flex points | 1.25 |
| RF welded | None, the interface is fused | IPX6-IPX7 | Film-to-fabric bond, slowly | 1.20 |
| RF welded plus airtight zip | None, one sealed opening | IPX7-IPX8 | Zip mechanism, not the seam | 1.55 |
How High-Frequency Welding Actually Works
Radio frequency welding, also called high-frequency or dielectric welding, heats the material from the inside rather than from the outside. Two layers of film are laid between a shaped upper electrode and a flat lower bed, the press closes at 2 to 4 bar, and a generator drives an alternating field at 27.12 megahertz through the stack. Polar molecules in the polymer attempt to align with that field and reverse with it millions of times per second, and the internal friction of that reversal is what generates heat precisely at the interface.
That mechanism is why the process is so well suited to coated fabrics. Heat is generated only in the polar material, so the electrode stays relatively cool, the outer faces of the film do not scorch, and the weld forms through the full thickness in seconds rather than being conducted inward from a hot surface. Heating stops the instant the field stops, which gives a repeatable cycle and a narrow, well-defined weld zone.
| Parameter | Typical range | Effect of too little | Effect of too much |
|---|---|---|---|
| Frequency | 27.12 MHz, ISM band | Not adjustable | Not adjustable |
| Generator power | 1.5-4 kW per station | Cold weld, peel below 90 percent | Burn-through, film thinning |
| Dwell time | 1.2-3.5 s | Incomplete fusion at the interface | Squeeze-out, weld under 8 mm |
| Electrode pressure | 2-4 bar | Void lines along the seam | Film squeezed to under 70 percent thickness |
| Cooling under pressure | 0.5-1.5 s | Springback, visible seam memory | Longer cycle, no quality gain |
| Weld width | 8-15 mm | Peel strength falls off sharply | Stiff seam, harder to fold |
Two parameters deserve more attention than they get in most tech packs. Dwell time and pressure interact, so a weld that is under-powered can sometimes be rescued by a longer dwell and vice versa, and the only way to find the window is to run a coupon matrix at sampling and keep it on file. Cooling under pressure is the other one: releasing the press while the interface is still above its softening point lets the seam spring back and leaves residual stress that shows up as a leak six months later.
What a Welded Seam Looks Like and How Strong It Is
A welded seam is visually obvious once you know what to look for: a flat, glossy band 8 to 15 millimetres wide with no thread on either face, slightly thinner than the surrounding film because the interface has been squeezed. There is often a faint witness line from the electrode edge. On the inside, the same band, with the two films no longer distinguishable from one another where they meet.
Strength is specified in peel, not in tensile, because peel is how a seam actually fails. The acceptance criterion used across the category is that the welded joint should reach at least 90 percent of the base fabric tensile strength, which on a 0.4 to 0.6 millimetre laminate means roughly 150 to 300 newtons per 50 millimetre strip. Below about 80 percent, corners begin to lift within a season; above 95 percent, the failure moves into the film itself, which means the weld is no longer the weak point and the design is balanced.
RF Welding Against the Other Joining Methods
RF is not the only way to join a coated fabric, and it is not always the right one. Hot air and hot wedge weld non-polar films that RF cannot, ultrasonic is fast on small parts, and adhesive works where the geometry defeats any thermal process. The choice is made from the polymer first and the part geometry second.
| Method | Heat source | Best for | Limits | Typical use |
|---|---|---|---|---|
| RF, 27.12 MHz | Dielectric loss in the polymer | PVC and TPU films, long seams | Only polar materials | Pack bodies, dry bags |
| Hot air | External hot air stream | PE, PP and non-polar films | Slower, wider heat-affected zone | Tarpaulins, liners |
| Hot wedge | Heated metal wedge | Thick geomembrane and PE | Contact marks, slower | Heavy covers |
| Ultrasonic | High-frequency vibration | Small parts, spot welds | Limited seam length | Tabs, valve housings |
| Adhesive and solvent | Chemical bond | Mixed materials, 3D corners | Cure time, solvent handling | Repairs, hybrid joints |
For pack bodies the comparison usually comes down to RF against adhesive, and RF wins on consistency and cycle time while adhesive wins on geometry. A three-dimensional corner where four panels meet is genuinely hard to weld, because the electrode has to reach the interface from both sides with even pressure; that is the one place a hybrid joint with a chemically bonded patch is still the practical answer. The trade-off is covered in welded against glued seams, and the alternative for lower-rated bodies is in taped against welded seams.
Which Materials Can and Cannot Be Welded
Polarity is the gate. A polymer has to have a dipole that can be driven by the field, which is why polyvinyl chloride and thermoplastic polyurethane weld beautifully and polyethylene and polypropylene do not. For the non-polar films the answer is hot air or hot wedge, which heats by conduction from the surface and therefore needs a longer cycle and produces a wider heat-affected zone.
| Material | RF weldable | Why | If not, use |
|---|---|---|---|
| PVC, plasticised | Yes, excellent | High dielectric loss, wide window | Not needed |
| TPU, ester or ether | Yes, good | Polar, narrower window than PVC | Not needed |
| PU-coated nylon or polyester | Yes, good | Coating is polar and fuses | Not needed |
| EVA foam and sheet | Yes, with tuned power | Polar, low melt strength | Reduce dwell and pressure |
| Polyethylene, polypropylene | No | Non-polar, no dipole to drive | Hot air or hot wedge |
| Silicone and rubber | No | Cross-linked, does not reflow | Adhesive or mechanical |
| Uncoated woven nylon | No | No thermoplastic film to fuse | Sew and tape, or laminate first |
Two material caveats matter in production. Plasticiser content shifts the dielectric behaviour of PVC, so two compounds with the same nominal hardness can need different dwell settings, and every new film lot should be re-qualified with a coupon. And moisture in the film turns to steam in the field, producing voids along the seam; film stored above 60 percent relative humidity should be conditioned or dried before welding. The comparison between the two main weldable films is in TPU against PVC shell.
How a Welded Seam Is Tested
Five checks cover the risk, and only one of them is the rating itself. A weld can pass an immersion test on day one and still be under-powered, so the mechanical and ageing checks are what separate a production-ready seam from a lucky one. The coated-fabric test methods used for peel, tensile and adhesion are collected by ASTM International in ASTM D751, and the ingress levels are defined by IEC 60529.
| Check | Method | Procedure | Pass criterion |
|---|---|---|---|
| Seam peel strength | ASTM D751 family | Pull a 50 mm welded coupon in peel | At or above 90 percent of fabric tensile |
| Weld cross-section | Microscope on a cut coupon | Measure remaining thickness at the interface | At least 70 percent of original film |
| Enclosure integrity | Inflation or IPX test | 3 kPa for 3 min, or the rated jet or immersion | Under 10 percent loss, dry indicator paper |
| Hydrostatic head | ISO 811 column | Rising water column on a fabric coupon | At or above 5,000 mm for the shell |
| Accelerated ageing | 168 h at 70 C and 95 percent RH | Age, then repeat the peel test | Peel retention at or above 85 percent |
| Cold fold | Fold at -20 C | Fold the seam, then inflate to 3 kPa | No cracking, no loss of pressure |
The cross-section check is the cheapest diagnostic in the list and the one most often skipped. Cutting a welded coupon and measuring the residual thickness at the interface immediately reveals whether the weld was over-pressed, which shows as film squeezed below 70 percent, or under-powered, which shows as a visible boundary line between the two layers. Neither is visible from the outside, and both pass a day-one immersion test.
Where Welds Fail, and Why
Weld failures cluster into four causes, and they are almost never the fault of the generator. Ranked by frequency in a real season: contamination on the film surface, drift in the tooling or the parameter window, material variation between lots, and geometry that the electrode cannot reach. Each has a specific signature, so a failed seam can usually be diagnosed in minutes rather than argued about for weeks.
- Contamination. Mould release, dust, print ink or finger oil on the mating surface produces a weld that looks right and peels at 40 to 60 percent. Wipe both faces with isopropyl alcohol and re-run the coupon.
- Parameter drift. A dwell shortened by 0.4 seconds to raise output leaves a cold weld that holds the first inflation test and lifts at the corner in month three. Lock the settings on the press.
- Tooling wear. An electrode edge worn by 0.2 millimetres reduces effective weld width and concentrates stress at the seam end. Measure electrodes every 50,000 cycles.
- Material variation. A new PVC lot with a different plasticiser loading changes dielectric loss and shifts the whole window. Re-qualify every film lot with a coupon.
- Geometry. A four-panel corner that the electrode cannot reach evenly welds thin on one side. Redesign the panel layout or specify a hybrid bonded patch there.
Saltwater deserves a separate note because it attacks the seam indirectly. Salt crystals forming at the weld edge and around hardware create local stress and wick moisture along the boundary, so a pack used in surf or on a boat should be rinsed in fresh water for two to three minutes per trip and dried open. The hardware side of that problem is covered in our saltwater corrosion guide, and the same process applied to fishing bodies is described in RF welded seams on fishing bags.
Design Rules That Make a Pack Weldable
Weldability is decided on the pattern table, not at the press. A panel layout that keeps weld paths in one plane, gives the electrode room to reach, and avoids crossings will weld consistently at high output; a layout that does not will fight the process for the life of the programme. The rules below are the ones that cost nothing at design time and save the most in production.
| Rule | Target | Reason | Cost of ignoring it |
|---|---|---|---|
| Weld flange width | At least 12 mm | Room for an 8-15 mm weld plus alignment tolerance | Seam edge creep, peel below 90 percent |
| Corner radius | At least 8-10 mm | Electrode pressure stays even through the arc | Thin corner welds that lift in use |
| Weld crossings | None where avoidable | A second pass re-melts the first | Local thinning and a visible weak point |
| Electrode access | Both faces reachable | The field needs even pressure on the stack | Hybrid joint needed, slower cycle |
| Panel count | Fewest that gives the shape | Each panel adds one seam to qualify | More tooling, more failure candidates |
| Film thickness tolerance | Plus or minus 5 percent | Thickness drives the energy requirement | Cold welds on thin lots, burn on thick lots |
One rule overrides the others: design for the fewest seams that still gives the shape. Every seam is a qualification exercise, a tooling item and a candidate failure point, and the packs with the best field record in the category are almost always the ones with the simplest panel layout. Buyers who want the rating and the reliability together should expect to give up an external pocket or two to get them.
Cost, Tooling and Ordering
Welding costs more per unit than sewing and less than buyers expect, because the cycle is measured in seconds and there is no thread, tape or sealant to buy. The real cost sits in tooling: each seam profile needs a shaped electrode, and a body with five distinct weld paths needs five of them. That is why the panel-count rule above is also the main cost lever.
| Item | Typical figure | Note | Programme impact |
|---|---|---|---|
| Electrode tooling per profile | 300-900 USD | One per distinct seam shape | Amortised from 500 pcs upward |
| Weld cycle per seam | 1.2-3.5 s plus 0.5-1.5 s cooling | Operator plus press time | About 2-4 min per body |
| Unit cost index, welded body | 1.15-1.30 | Against a stitched and taped body at 1.00 | Buys IPX6-IPX7 and no tape failure |
| MOQ per design and colourway | 500 pcs | Set by film, webbing and tooling minimums | Consolidate colours from 3,000 pcs |
| Sample timing | 6-10 working days | Includes a coupon matrix per film lot | Longer for five-plus panel bodies |
| Bulk timing | 35-50 days | FOB Xiamen or DDP into 100+ countries | Add 2-4 days for unit pressure checks |
- Every bulk order is inspected at AQL 2.5 before release, and welded bodies add a 3 kPa inflation check on a stated sample rate.
- Immersion-rated programmes at IPX7 or IPX8 add a per-unit pressure or vacuum check, which adds two to four days.
- Monthly output of 200,000 units is verified by SGS, and materials are screened against REACH and California Prop 65.
- A four-line brief starts a quotation: body shape, panel count, rating target and quantity.
The recommendation for a first programme is to weld the body and to be honest about the openings. A welded shell with a roll-top collar reaches IPX6 to IPX7 at a cost index near 1.20 and stays there for the life of the film; the same shell with three external pockets and coated zips lands at IPX5 regardless of how good the welds are. Buyers who want to see the process and the tooling in detail can look, and the category baseline is in our waterproof backpack specification guide; enquiries go to service@junyuanbags.com.
People Also Ask
Why are waterproof backpack seams welded instead of sewn?
Because a needle punches 5 to 7 holes per centimetre through the shell, so a 400 millimetre seam carries 400 to 560 holes that wick water inward. Welding fuses the two films into one layer at 27.12 MHz with no hole along the length, which is what allows a rating of IPX6 or IPX7.
How strong is a welded seam?
Strong enough to be the strongest part of the body when specified correctly. The acceptance criterion is peel strength at or above 90 percent of base fabric tensile, typically 150 to 300 newtons per 50 millimetre strip. Below about 80 percent the corners lift within a season.
Can all waterproof fabrics be RF welded?
No. The polymer must be polar, so PVC, TPU and polyurethane-coated nylon weld well, while polyethylene and polypropylene do not. Non-polar films are joined by hot air or hot wedge instead, which heats from the surface and needs a longer cycle.
What causes a welded seam to fail?
Four causes cover most cases: contamination on the mating surface, parameter drift on the press, material variation between film lots, and geometry the electrode cannot reach evenly. Over-pressed welds squeeze the film below 70 percent thickness and under-powered welds leave a visible boundary line.
Is a welded seam better than a taped seam?
Yes for rating and longevity. A taped seam bridges needle holes with adhesive that hydrolyses and lifts at corners within two to four seasons, while a welded seam has no holes at all and holds IPX6 to IPX7. Tape remains useful where geometry prevents welding.
How do you test if a welded seam is sound?
Peel a 50 millimetre coupon to confirm 90 percent of fabric tensile, cut a cross-section to confirm at least 70 percent residual film thickness, inflate the body to 3 kilopascals for 3 minutes, and repeat the peel after 168 hours at 70 degrees Celsius and 95 percent relative humidity.
Frequently Asked Questions
Q1. What is RF welding on a waterproof backpack?
A joining process that fuses two films into a single layer instead of stitching them. The panels are pressed between electrodes at 2 to 4 bar and driven at 27.12 megahertz, polar molecules heat by dielectric friction, and the interface melts and fuses in 1.2 to 3.5 seconds with no needle hole anywhere along the seam.
Q2. Why can a stitched seam never be fully waterproof?
Because the needle creates the leak path. At 5 to 7 stitches per centimetre a 400 millimetre seam has 400 to 560 holes, and thread wicks water from one hole to the next. Tape bridges those holes for two to four seasons and then lifts, so the holes are still there underneath.
Q3. How wide is a welded seam?
Typically 8 to 15 millimetres, on a flange of at least 12 millimetres so there is room for the weld plus alignment tolerance. Below 8 millimetres the peel strength falls off sharply; above 15 millimetres the seam becomes stiff and harder to fold on a roll-top collar.
Q4. What peel strength should a welded seam reach?
At least 90 percent of the base fabric tensile strength, which is roughly 150 to 300 newtons per 50 millimetre strip on 0.4 to 0.6 millimetre laminate. Above 95 percent the failure moves into the film itself, meaning the weld is no longer the weak point.
Q5. Which materials can be RF welded?
Polar ones: PVC, TPU and polyurethane-coated nylon or polyester all weld well. Polyethylene, polypropylene, silicone and uncoated woven fabric do not, because there is no dipole for the field to drive. Those need hot air, hot wedge or adhesive instead.
Q6. How long does a welded seam last?
Longer than the film around it in most cases. A correctly specified weld outlasts the shell and fails last, usually through slow delamination of the film-to-fabric bond after six to ten years. The failure modes that end a pack earlier are contamination, over-pressing and corner geometry.
Q7. Can a welded seam be repaired?
Yes, and it repairs better than most alternatives. A pinhole or a short split takes a peel-and-stick TPU patch in 60 to 90 seconds; a long seam separation needs a hot air or RF patch under a press, which is a workshop job of 20 to 40 minutes rather than a full rebuild.
Q8. Does RF welding make the pack heavier?
Marginally lighter overall. The weld itself adds no material, whereas a taped seam adds roughly 20 to 25 grams per metre of tape plus adhesive. The saving is small on a daypack and noticeable on a 60 litre body with several metres of seam.
Q9. Is welding more expensive than sewing and taping?
By about 15 to 30 percent on unit cost, against a stitched and taped body at 1.00. The premium buys IPX6 to IPX7, no tape to hydrolyse and no seam sealant to crack. The tooling, at 300 to 900 USD per seam profile, is amortised from 500 pcs upward.
Q10. What is the MOQ for a welded waterproof backpack?
500 pcs per design and colourway, the same as any other construction. Tooling is the practical constraint rather than quantity: each distinct seam profile needs its own electrode, so bodies with fewer panels cost less to set up and qualify.
Q11. How do I specify a welded seam in a tech pack?
State the weld width, the peel criterion at or above 90 percent of fabric tensile, the minimum residual film thickness at 70 percent, and the enclosure test at the target IPX level. Add a cross-section check and a post-ageing peel retest so the seam is qualified for year three, not only for day one.