Answers › Flow lines and jetting

What causes flow lines (flow marks) and jetting on molded parts, and how are they fixed?

Reviewed by the JCH tooling team · Updated

Flow lines are ripples or bands across the flow path where the melt front slowed, chilled and pressed unevenly onto the steel. Jetting is a snake-shaped strand starting at the gate, left when melt shoots into open cavity without touching a wall. Flow lines respond mostly to heat and speed; jetting usually needs a gate change or a slow start.

Flow lines and jetting both start as marks near the gate on early trial parts, and both are a record of how the melt front moved through the cavity. They need to be told apart before anyone adjusts the press, because their cures pull against each other. Fictiv's jetting article states the conflict plainly: more injection speed might fix flow lines but will almost certainly make jetting worse. Treat every wavy mark as one problem and a trial can turn into a back-and-forth between the two.

What the flow front is supposed to do

Melt doesn't push through a cavity like a piston. It advances by fountain flow: hot material from the middle of the stream rolls out at the front and lays itself onto both mold walls, where it freezes into a skin, and the melt behind travels inside that skin. So every point on the surface is a print of the front at the moment it passed. A front that moved steadily, stayed hot and stayed pressed against the steel copies the mold finish evenly. A front that slowed, stalled, pulsed or never touched the wall leaves that history in the surface.

That gives the first sorting rule. Flow lines lie across the flow direction, like ripples or tree rings, because they record moments when the front itself faltered. Jetting runs along the flow as a winding strand that begins at the gate, because it is a stream of melt that got out ahead of the front.

Flow lines: four patterns from one root

Different shops use different names. Where the marks sit and how they look matter more than the label.

The common thread is a front that lost speed, heat or pressure at the wall. Process cures for flow lines all put one of those back: a warmer mold, a hotter melt, a faster or steadier fill, and enough pressure in reserve that the set speed is actually reached.

Two top views of the same flat cavity. Left: a small gate fires into open space and the melt leaves as a folded strand that runs ahead of the flow front. Right: the gate stream strikes a core pin within a few millimetres, a normal flow front forms, and ripples called flow lines can still appear across the flow if the front slows and cools. Jetting runs along the flow; flow lines run across it small gate free jet folds on itself and chills before the main front arrives nothing ahead of the gate to hit Gate into open space Strand along the flow, starts at the gate Fix: aim the gate at steel, fan or tab it, slow the first stage core pin flow front lines: ripples form here if it stalls Gate stream hits a pin A normal front forms near the gate Flow lines, if any, lie across the flow Fix: heat, speed, pressure at the front

Jetting: melt that never met a wall

Jetting is a different event. When melt leaves a small gate at high speed and the cavity opens up in front of it, nothing slows it down. BASF describes the melt squirting into the free cavity as a jet that crumples where it happens to touch the wall. The strand's surface has already cooled, so it doesn't bond properly to the melt that follows. What you see is a snake-shaped, often rough or matt trail starting at the gate. Fictiv adds that in a clear part it can look like a tangled cord inside the wall, and that the poorly fused strand leaves weak interfaces much like a weld line. That makes jetting a possible strength problem, not only a cosmetic one.

Jim Fattori, writing on tunnel gates in Plastics Technology, makes two points worth remembering. Most gates jet to some degree until the stream hits an obstruction, and the most common obstruction is the wall opposite the gate. And gate size and angle shift the odds. Tunnel gates are typically cut 30° to 50° from the parting line, and the steeper the angle, the more the stream shoots up into open space instead of striking the core. On size, he notes that into a 0.100 in (2.54 mm) cavity, a 0.030 in (0.76 mm) gate jets more than a 0.060 in (1.52 mm) one. Size alone isn't the whole story, though: he also describes a 3/4 in by 1/4 in (19 by 6.4 mm) edge gate that still jetted because it fired into a wide-open cavity. Pelley adds that gating into a large, thick area can cause jetting, depending on the gate type.

Worked example: a small gate into an open face

Take a hypothetical ABS lid with a 2.5 mm wall and a flat face 80 mm wide. An edge gate 1.0 mm deep and 3.0 mm wide feeds the face from one side, aimed straight into open cavity. Each shot puts 24 cm³ through the gate in a 1.2 s fill.

The melt has to slow by a factor of about 67 within a few millimetres of the gate. Fountain flow manages that when the melt touches both walls. A 1 mm ribbon entering a 2.5 mm gap with nothing in front of it doesn't, which is the same thin-gate, thicker-cavity geometry as Fattori's comparison. Expect a snake. Three ways out, with the arithmetic:

  1. Slow start. Profile the first part of the stroke at one-fifth of the speed until a small puddle of melt forms past the gate, then ramp up. Gate speed drops to 6,667 / 5 = 1,333 mm/s. That often softens jetting, but the stream still points at open space, and Fictiv's view is that severe jetting usually needs a gate or mold change.
  2. Fan gate. Widen the gate to 25 mm and deepen it to 1.5 mm: 25 × 1.5 = 37.5 mm², so gate speed falls to 20,000 / 37.5 = 533 mm/s, one-twelfth of the original. The apparent shear rate at the gate wall, using the 6Q / (W × h²) estimate in the gate size calculator, goes from 6 × 20,000 / (3.0 × 1.0²) = 40,000 1/s to 6 × 20,000 / (25 × 1.5²) ≈ 2,130 1/s. The melt now enters as a wide sheet that reaches both walls almost at once.
  3. Tab gate, or a gate aimed at steel. Keep the small gate but fire it into a sacrificial tab, or move it so the stream strikes a wall or core pin right away. A jet still forms, but it crumples against steel and becomes a proper front before it reaches the show face. Fictiv calls aiming the gate at a wall the most effective fix and describes tab gates as the industry standard for demanding cosmetic parts.

A deeper gate also stays open longer during pack, so check the freeze ratio in the calculator before the change is cut. The numbers above are an illustration; run your own part volume and fill time.

BASF's brochure shows the same logic applied to tiger lines on a molded part. In its before-and-after drawing the round gate went from 2 mm diameter with a 2 mm land to 3.3 mm with a 0.4 mm land, the sprue from 5 to 6.5 mm, and the runner section was opened up. Gate area rose by (3.3 / 2)² ≈ 2.7 times, so average speed through it fell to about 37% of the original, and the land became one-fifth as long. BASF credits the even surface of the second part to the gating system's lower pressure loss.

A clear molded plastic part being lifted off a polished mold core by hand, with the mold open in the press
A clear part shows the full depth of its wall, so a jetted strand or flow ripples that would hide on an opaque housing are plain to see. On transparent parts, gate position and fill profile get decided early.

Fixes, sorted by where they live

SymptomProcess firstTool changePart design
Record grooves, often late in fillFaster fill, hotter melt and mold, more pack pressure and rateLarger gates where the gating loses a lot of pressureNo thick sections at the end of flow; no sudden wall steps
Tiger stripesHotter melt; mold at the top of the recommended rangeBigger gate, runners, sprue and nozzle boreThicker wall if allowed; ask about a more elastic grade
Halo or blush at the gateSlow-then-fast profile; hotter meltRounded gate with enough section, or gate through a runner outside the show faceGradual transitions after the gate
JettingSlow first stage, hotter meltAim at a wall or pin, enlarge, radius the gate entry, fan or tab gateRemove abrupt thick-to-thin changes and sharp corners on the path

The rows come from BASF's recommendations, with Fictiv's jetting fixes and Iannuzzi's grade finding added. On record grooves, Bozzelli's checks are worth copying before any setting moves. Peak first-stage hydraulic pressure should sit 200 to 400 psi below the set limit; if it touches the limit, the press can't hold the speed you set and the front slows on its own. Transfer to pack should come with the part 95% to 99% full, and he isolates fill by dropping pack to about 300 psi or setting second-stage time to zero. He checks the real melt temperature with a hot probe or an infrared sensor, and if the mold is the cause he raises its temperature by 20 to 30 °F (about 11 to 17 °C) or shortens the cycle. The 200 to 400 psi margin is a hydraulic reading and the 300 psi pack is plastic pressure; the method carries over, but the numbers vary by press.

Part design does a lot of the work. BASF lists sudden wall thickness changes as a cause of jetting, record grooves and dull spots alike, and Fictiv adds that sharp corners make the flow break away from the wall and that thick-to-thin steps accelerate the melt. Generous radii and gradual transitions, covered in designing wall thickness, remove many of these marks before the tool exists. Where the gate goes, and what kind it is, is the other half; the trade-offs between edge, fan, tab, tunnel and pin gates are laid out in choosing a gate type.

Is it flow lines, jetting, a weld line or splay?

A short-shot series settles most arguments. Stop the fill at several points, keep the parts in order, and look at the first few. A jet shows as a thin strand reaching across the cavity ahead of any puddle. Hesitation shows as successive shorts that barely advance in one area while the rest fills.

Machined steel mold half with the convex core of a housing-type part, a guide pillar and guide-bushing holes
The steel decides where the melt can stall and whether a gate stream has anything to hit. Steps, ribs and pins on a core like this are where hesitation marks and jetting usually start, so they are worth reviewing before the tool is cut.

Checklist before anyone cuts steel

  1. Draw the gate and the flow direction on a rejected part. Are the marks across the flow or along it?
  2. Does the mark start at the gate? Fictiv notes jetting almost always does, though its tail can run far into the part.
  3. Run the short-shot series and photograph it.
  4. Confirm first stage isn't pressure-limited and that transfer happens before the part is full.
  5. Measure real melt and mold surface temperatures, not set points.
  6. Try a slow-through-the-gate, then fast, profile.
  7. If jetting survives a slow start, plan a gate change: aim it at steel, enlarge it, or fan or tab it. A fill study (see what mold flow analysis tells you) can check the new location first.
  8. For tiger stripes in filled or blended grades, ask the resin supplier whether a different grade is less prone to them.

When the answer changes

Color, gloss and clarity. Pelley notes that surface finish can amplify or reduce how visible these defects are, and that an off-white color hides them better than gray, with black worse still. A glossy black or a clear part will show flow lines that a textured light-gray part never would, so set the cosmetic limit by the real color and finish. See clear molded parts and mold texture before deciding a mark is a reject.

Hot runners with valve gates. The speed at which the valve pin opens affects hesitation at the front, and Pelley notes it can be simulated. A halo that appears only on valve-gated drops may be a sequencing setting rather than the gate size.

When the gate can't move. If the only workable gate fires into a thick section, Pelley lists gas assist, coining and two-shot molding as ways to remove that source of blemish. Those change the tool and process substantially, so weigh them early.

When it carries load. Because the jetted strand is poorly bonded, jetting on a clip, hinge or boss is a test question, not a looks question. For the wider question of whether a given defect belongs to the tool or the process, see molding defects: tool or process?, and the defect troubleshooter ranks likely causes for other symptoms.

Common questions

Do flow lines or jetting make a part weaker?

Record grooves and tiger stripes are differences in surface gloss and texture, so they are mainly a cosmetic issue. Jetting is different. BASF and Fictiv both note that the cooled strand bonds poorly to the melt around it, which can lower strength. If jetting sits on a clip, boss or other load path, test the parts instead of judging them by eye.

Are tunnel (submarine) gates more likely to jet?

Not by nature. Jim Fattori argues in Plastics Technology that the blame is usually misplaced: most gates jet until the stream hits something. What matters is the angle and what the stream faces. Tunnel gates are typically cut 30 to 50 degrees from the parting line, and a steeper angle sends the melt into open space instead of onto the opposite wall.

Can mold flow analysis predict jetting and flow lines?

It can flag them. Brian Pelley of Autodesk writes that molding simulation can predict jetting effects, weld-line positions and air traps, and that hesitation shows up in the flow-front results. Fictiv adds that a fill study can estimate gate velocity and jetting risk before steel is cut. It is worth asking for while the gate location can still change.

Why are the flow lines worst at the far end of the part?

The front is coolest and slowest at the end of the flow path. BASF shows record grooves forming there when melt viscosity is too high, and lists thick sections at the end of flow and large pressure loss in the gating as causes. Raising mold and melt temperature, filling faster, or enlarging the gate keeps the front hotter when it arrives.

Will a textured surface or a different color hide flow lines?

Often, for mild cases. Pelley notes that surface finish can amplify or reduce how visible defects are, and that off-white hides them better than gray, while black is worse. Texture does not hide jetting on a load-bearing part, and it does not fix the cause, so agree the cosmetic limit on a sample in the real color and finish.

Sources

Send the part file and the quantity you expect to run, and we'll review it before you commit to tooling. NDA first if you want one.

Send a part for reviewBook a 30-minute call