Last updated: 2026-09-05 · 5 min read

A weld line forms wherever two or more melt fronts meet inside the cavity and have to fuse — and whether it ends up as a harmless witness or a rejected part depends almost entirely on how hot the fronts still are when they meet. Weld lines cannot always be eliminated, but they can always be predicted, relocated and weakened: the levers are gate and runner design first, process second, resin last. This article covers the mechanism, the difference between a weld line and a meld line, the places they will appear in your part, the process and tooling changes that actually move them, and the acceptance evidence to demand at T1.

Key facts

<>

  • A weld line is the seam where two flow fronts converge head-on; a meld line is where fronts converge at an angle and keep flowing. Meld lines are usually stronger and less visible.
  • Severity tracks the melt-front temperature at the meeting point: hot fronts interdiffuse and recover most of the resin strength; cold fronts leave a weak interface with a visible notch.
  • On filled resins the strength loss is larger, because fibers cannot re-orient across the seam — check the weld-line strength factor on the resin datasheet instead of assuming.
  • Flow simulation predicts weld-line location and meeting temperature before steel is cut; the DFM stage is where relocation is nearly free.
  • Acceptance needs two tracks: cosmetic (limit samples, agreed viewing distance) and functional (tensile specimens cut with the seam inside, tested to a standard such as ASTM D638).

    • What is a weld line — and how is it different from a meld line?

    Weld line (knit line). Two melt fronts approach each other head-on — most commonly on opposite sides of a hole, a boss or an insert — and collide. The front surfaces have already started to cool into a skin, so the seam is a place where two skins have to weld through each other. The result is a visible line and, on structural parts, a local strength reduction.
    Meld line. Two fronts converge at a shallow angle and continue to flow together after meeting. Because the fronts shear past one another instead of stopping dead, the molecular interdiffusion across the seam is better and the residual notch is shallower. Functionally a meld line is the same physics executed more gently.

    The distinction is not vocabulary for its own sake: it tells you what to change. A head-on weld line responds strongly to front temperature and to the geometry that created the head-on collision; a meld line is often acceptable as molded. When a supplier says “that is just a weld line, every part has one,” the useful question is which of the two it is, at what temperature the fronts met, and whether the drawing called that surface cosmetic.

    Why does a weld line weaken the part?

    Mechanism at the seam What it does to the part What you can observe
    Cold-skin fusion. Each front carries a thin, cooled outer layer; the two layers must remelt into each other. When the fronts arrive cold, the interface stays as a boundary — the part can crack along it under load. Tensile fractures that follow the seam; a faint parting in the cross-section.
    Trapped air and contamination. The convergence point is the last place air, moisture volatiles and debris can escape. Micro-voids and residues block molecular bonding at exactly the seam. Matte spots, fine burning marks, or a dull halo near the line.
    Surface V-notch. The two fronts meet below the nominal surface, leaving a shallow groove. The notch is a stress riser that concentrates load along the seam. A line you can feel with a fingernail on high-gloss surfaces.
    Fiber misalignment (filled resins). Fibers orient with flow and cannot turn across the seam. Strength loss is markedly larger than in unfilled resins of the same family. Weld-line strength factors printed on filled-resin datasheets, well below 1.0.

    Two consequences follow. First, the visible line and the weak seam are related but not identical: a line can be cosmetically invisible after texturing while the weak interface remains, and a faint line on an unfilled resin can carry near-nominal strength. Second, the strength conversation belongs to the resin datasheet, not to folklore — suppliers publish weld-line strength factors for filled grades precisely so that a molder can size ribs and bosses around the seam rather than argue about it.

    Where will weld lines appear in your part?

    Most locations are predictable from geometry alone, which is why the DFM review flags them before the first chip is cut:

  • Downstream of holes, slots and windows — the flow splits around the obstruction and re-joins behind it; the longer the path around it, the colder the re-union.
  • Between multiple gates — every additional gate adds at least one convergence; symmetric multi-gating puts the seam at the part’s mid-plane, where it shows on both faces.
  • Around bosses and ribs — the flow wraps local features and closes behind them, often at a thickness transition.
  • At wall-thickness changes — a thin section slows and cools the front, so the seam downstream of any thickness change arrives colder than the nominal wall predicts.
  • Beside inserts and overmolded features — metal inserts chill the front locally, so seams closing over them are the coldest in the cavity.
  • Where runners re-join in family tools — unbalanced runners deliver fronts at different pressures and temperatures to the same convergence.

    • Flow simulation turns that list into a map: modern molding simulation plots weld-line locations and, more usefully, the temperature of each front at the moment of meeting — a seam predicted at high front temperature often needs no action at all, while the same geometry predicted cold is where the design work goes. If a supplier cannot show you that map during DFM, the weld-line conversation is being deferred to trial-and-error on your steel.

    Which process changes actually improve weld line quality?

    Process lever What it changes at the seam Watch the limit
    Melt temperature Hotter fronts arrive hotter, remelt each other’s skin and interdiffuse — the single most effective lever on strength. Resin degradation window from the datasheet; discoloration is a sign you traded a seam for a different defect.
    Mold temperature A warmer cavity delays skin formation, so the fronts stay molten longer at the meeting point. Cycle time rises; on semi-crystalline resins the window matters for shrinkage and warpage too.
    Injection speed Faster filling shortens the time fronts spend in transit and adds shear heating at the front. Too fast and the convergence traps air — you exchange a weld line for burns and gas marks.
    Packing pressure Consolidates the seam and reduces voids at the notch; helps cosmetics and mildly helps strength. Overpacking builds residual stress and sinks the area around the gate instead.
    Venting at the convergence Removes the trapped air that blocks fusion and stains the line. Vent lands sized to the resin, not opened indefinitely — deep vents feed flash.

    The honest summary: process can recover most of what geometry gave away, but it cannot repeal the geometry. If the fronts must cross a long, thin, cold section to meet, no reasonable setting makes that seam disappear — it can only be made thinner, hotter or less visible. That is why the next section belongs before the process window tuning, not after it.

    How does mold design reduce or relocate weld lines?

    1. Decide where the seam should land — then aim the gates there. Gate position is the strongest lever in the whole exercise: moving or re-aiming a gate drags the convergence out of the cosmetic face and into a rib root, an internal wall or an edge that assembly hides. The gate type and location decision should name the expected weld lines explicitly, not discover them at T1.
    2. Trade a weld line for a meld line. Aiming two fronts to converge at a shallow angle instead of head-on converts the worst seam into a milder one; sometimes one extra degree of approach angle is worth more than any process change.
    3. Sequence the gates. On parts where a family of gates is unavoidable, valve-gated hot runners open in a programmed order so the fronts meet hot and late, or sequentially fill rather than race — the seam still exists but arrives molten.
    4. Feed the seam locally. A slight local thickening or a flow leader near the expected convergence keeps the fronts hot where they meet, at the cost of a small volume change that the part design usually tolerates better than the seam.
    5. Vent where you know the fronts will meet. Predicted convergence points get vent land priority — steel expected to trap air is vented on purpose, not discovered on the trial bench and patched with a maintenance note.
    6. Plan the surface around the line. Where a seam must remain on a visible face, texturing that region deliberately makes the line read as geometry instead of a defect — a decision for the drawing, made once, instead of a polishing argument at every trial.

    Can a weld line be made invisible — and should it be?

    Three honest answers coexist. On a textured or matte surface, a weld line often disappears to the eye at normal viewing distance, which is why industrial housings live with seams that would scream on a piano-gloss face; the higher the SPI polish level, the more honestly the seam shows. On a painted, plated or coated surface, the line can vanish cosmetically — but paint hides the geometry, not the interface, so the structural question remains open underneath. And on a structural part, “invisible” is the wrong goal: the goal is a seam whose strength the design tolerates, verified with specimens, not polished until someone stops noticing it. The acceptance question is therefore never “is there a line?” — there usually is one somewhere — but “is the line where the drawing agreed it would be, does it read within the agreed cosmetic standard, and does the part carry load across it?”

    That framing is also what keeps the acceptance conversation calm. General acceptance conditions for moulded parts — which lines count as process witnesses, which count as defects, and how dimensional and appearance requirements are read together — are standardized in ISO 20457, and it is worth asking a prospective supplier which edition they quote from before the RFQ turns into an argument at trial. The standard will not tell you whether your seam is acceptable; your drawing, your cosmetic classification and your load path will. It will stop the two of you from arguing about vocabulary while the tool waits on the bench.

    When does the resin choice matter for weld lines?

    More than most buyers expect. Unfilled, amorphous resins forgive: their seams recover a large share of nominal strength when the fronts meet hot, and the datasheet usually does not even list a factor. Filled and reinforced grades are a different conversation — glass beads, glass fibers and mineral fillers raise stiffness but cannot re-orient across the seam, so the weld-line strength factor drops and the part must be ribbed and bossed with that number in mind, not with the unreinforced number. Color matters too: dark, high-gloss parts show seams that light textures hide, and some pigments accent the convergence differently at different front temperatures. The practical discipline is to read the weld-line data for the exact grade — a searchable property database such as MatWeb carries datasheets from most major suppliers — and to treat any molder who quotes “the weld line strength of the resin family” instead of the grade as a molder you have not yet caught in a rounding error. Resin substitution is a legitimate fix for a stubborn seam, but it arrives with its own requalification; it belongs at the end of the decision list, not the beginning.

    The weld-line arguments I never have are the ones where the DFM map showed the customer exactly where the seams would land, and they picked which ones they could live with before we cut steel. Then the trial is a confirmation, not a negotiation. The arguments I always have start with someone discovering a seam on a gloss face at T1 and asking why nobody mentioned it — and the answer is always the same: nobody asked where the fronts would meet. Give me the cosmetic faces and the load paths, and I will draw you the seams on the part before we order steel. That conversation costs one meeting; discovering it after hardening costs a re-trial and a repaint.

    — Arthur, Tooling Engineer at aCatMould

    Frequently asked questions

    Are weld lines always a defect?

    No. A weld line is a consequence of how melt fills a cavity, and most parts have several. It becomes a defect when it lands where the drawing says it must not appear, or when the part must carry load across a seam that was never sized for the local strength loss. Everything else is a witness of the process.

    How much strength does a weld line lose?

    It depends on the resin, the fillers and how hot the fronts met. Unfilled grades meeting hot can approach their nominal strength; filled grades lose more, which is why datasheets publish a weld-line strength factor below 1.0 for reinforced grades. Design against the datasheet number for your exact grade, measured the way standards such as ASTM D638 measure it — not against a family average.

    Can painting or plating hide a weld line?

    Cosmetically, usually yes — a coating covers the surface notch and the seam stops reading. Structurally, no: the interface underneath is unchanged, so any load the part carries across the seam still needs the seam to be strong. Paint solves the appearance question only.

    Do more gates mean more weld lines?

    As a rule, each additional gate adds at least one convergence somewhere, so yes — but the trade is often worth it. A second gate can cut the flow length so dramatically that both fronts arrive hot, turning one cold, weak seam into one hot, mild one. Gate count is an engineering decision, not a moral one.

    Will a higher melt temperature alone fix a bad weld line?

    Rarely by itself. Temperature is the most effective single lever, but a seam created by a long cold flow path needs geometry help — gate position, local thickening, venting — and pushing temperature alone into a degraded or discolored molding is the usual failure mode. Raise temperature, yes; expect it to do the work of mold design, no.

    What should I ask for at T1 to evaluate weld lines?

    Three things: the predicted-versus-actual seam map from the simulation so you know whether the tool did what the DFM promised; cosmetic judgment against agreed limit samples at the agreed viewing distance; and, for structural seams, specimens cut with the seam inside and tensile-tested to ASTM D638 so the conversation carries numbers. A T1 report without the seam map is a report that deferred your hardest question.

    Key takeaways

    1. Weld lines form wherever fronts converge head-on; meld lines are the milder, angled version — and the difference is a design variable, not vocabulary.
    2. Severity tracks front temperature at the meeting point: hot fronts fuse and recover strength, cold fronts leave a notch and an interface.
    3. Location is predictable — behind holes, between gates, at thickness changes and over inserts — and simulation maps both position and meeting temperature during DFM.
    4. Fix order is fixed: gate and runner design relocate the seam, process settings heat it, resin choice is the last resort with its own requalification.
    5. Accept in two tracks: cosmetic against limit samples, functional against tensile specimens carrying the seam — verified at T1, not assumed in the shipment.

    If your part carries holes, multiple gates or a gloss face — and every part does — send us the 3D model and the cosmetic drawing before the deposit. You will get the predicted seam map, a proposal for where each weld line should land, and the acceptance evidence list for T1, from a named tooling engineer, before any steel is ordered.

    Sources
    ASTM D638 — Standard Test Method for Tensile Properties of Plastics, ASTM International
    ISO 20457:2018 — Plastics moulded parts: Tolerances and acceptance conditions, International Organization for Standardization
    Autodesk Moldflow — injection molding simulation for weld-line prediction, Autodesk
    Plastics Industry Association — SPI surface finish sets and moldmaking resources
    MatWeb — material property database for resin datasheets and weld-line strength factors

    Sources: Plastics Technology · Plastics Technology Online · Society of Plastics Engineers (SPE) · MoldMaking Technology