Injection Molding Defects: Causes, Fixes and How to Diagnose Them

A defect that appears on the first shot and a defect that appears on the four hundredth are different problems, and treating them the same is what makes defect chasing expensive. In commercial molding, roughly 60 percent of defects start in the part design or the mold design, about 30 percent start in the process window, and about 10 percent start in the machine or the material batch. Start from that split and the diagnosis stays short: if the defect was present on trial one, stop adjusting settings and go straight to the tool.

Why most defects are fixable, and why most are misdiagnosed

A defect that appears on the first shot and a defect that appears on the four hundredth are different problems, and treating them the same is what makes defect chasing expensive. In commercial molding, roughly 60 percent of defects start in the part design or the mold design, about 30 percent start in the process window, and about 10 percent start in the machine or the material batch. Start from that split and the diagnosis stays short: if the defect was present on trial one, stop adjusting settings and go straight to the tool.

The defect map: which symptoms point where

Before anything else, it helps to sort symptoms by the part of the system they implicate. The table below is the working reference we use when a buyer sends us a part and asks whether the mold or the settings are at fault. The symptom column is what shows up on the part; the likely origin column is where the cause usually is; the first check column is the cheapest test that can confirm or eliminate it.

Two things in that table are worth calling out. First, none of the symptoms is unique to one cause. Sink marks can come from insufficient packing, from unbalanced cooling, or from a gate sitting where it should not. Silver streaks can come from moisture in the resin or from gas trapped at the end of fill. That ambiguity is precisely why the fixed order in the seventh section exists.

Second, notice which symptoms have the widest blast radius. Warpage and short shot are not one-part problems; they change how the whole assembly fits, and they are the two symptoms that most often end a project between the buyer and the mold maker. Anything else is a conversation about a specific feature on a specific part.

Short shot and the trapped-air family

A short shot is an incomplete fill. The part is missing material where the flow front stopped, usually at the end of the flow path or behind a weld line. It is the most misdiagnosed common defect, because the instinct is to raise the pressure, and raising the pressure on a venting problem makes the flash worse while leaving the short shot in place.

There are two distinct causes and they look similar on the part. Insufficient pressure produces a short shot with a frozen, matte surface where the flow stopped, and the unfilled region follows the geometry of the flow path. Trapped air produces a short shot with a burn mark at the end of the flow, often with a visible burn ring or scorch, because the air was compressed rather than vented.

The distinguishing test is cheap: reduce the injection speed by half and re-run. If the short shot either disappears or moves to a different, further location, the part was under-packed or the flow was freezing early, and the fix is in the process window. If the short shot stays exactly where it was and still shows a burn mark, air is trapped in the cavity and the fix is in the tool: deep, correctly positioned vent slots at the end of fill, and a check for a nozzle that is buried too deep or pointing straight at a blind spot.

There is a third cause worth separating out: a flow front that froze before it could fill. It looks like a short shot but the cause is melt temperature, mold temperature or the thickness of the flow path. The signature is that the short shot reappears at a different thickness transition each time you change something, and that the tool passes a cold-flow test on a slower, hotter setting. Raising melt temperature and slowing the fill usually resolves it; if it needs more than about 15 percent additional pressure, the flow path is too long or too thin for the material and the tool needs repackaging.

Sink marks, flow marks and weld lines

These three are the classic “the tool is not quite right” family, and they are also the three most often fixed by process changes.

A sink mark is a depression on the surface, usually visible from the other side of the wall as a corresponding bulge. It forms where the wall is locally thick and the core has pulled material out of the pack. The first response should be packing and holding: raise hold pressure toward the machine limit and extend hold time until the part weight stops rising. If the sink reduces and then comes back at a different thickness, the real cause is wall thickness variation, and the tool needs a core cooling change rather than a settings change.

A flow mark is the faint ripple that follows the direction of melt flow across the part. It is caused by the boundary layer between hot melt and the cooler mold wall being stretched as the front advances. Higher mold temperature and faster injection speed both reduce it; if it will not go away with those, the tool needs more or better-placed cooling to keep the surface temperature uniform.

A weld line is a visible seam where two flow fronts meet, usually with reduced strength and often a visible line on the cosmetic surface. The weld line is not really a defect; it is a consequence of the gate position relative to the geometry. Moving the gate so the flow fronts meet where the part is not visible, or adding a differential cooling treatment across the weld line so it does not become a visible blemish, is the normal remedy. On a cosmetic part, the weld line should be discussed at design review, because no amount of process work makes a visible weld line invisible on a polished surface.

Warpage: the most expensive defect to misdiagnose

Warpage is a dimensional distortion caused by uneven shrinkage or uneven cooling. The part may be perfectly within every tolerance when it comes off the machine and still not fit, because the distortion is the difference between the nominal shape and the shape the customer needs. It is also the defect most likely to be misdiagnosed, because warpage looks like a tool problem and is often a material or design problem.

The four mechanisms in that table need to be separated because the remedies do not overlap. Differential shrinkage is intrinsic to the material and scales with wall thickness, so it is a design constraint: uniform wall thickness is the cure. Cooling imbalance is a tool problem: one region of the cavity reaches temperature faster than another, so it shrinks more in the same cycle time. Ejection stress is a handling problem: the part sticks, the ejector pushes it off, and it relaxes into a different shape after it is out. Oriented shrinkage comes from fiber-reinforced resin: the flow direction shrinks differently from the transverse direction, so the tool has to place the flow deliberately rather than accidentally.

One test distinguishes most of them quickly. Measure the part while it is still in the cavity, or measure it before it has cooled to room temperature. If the part is straight in the cavity and warped on the bench, the cause is cooling or ejection, not material. If it is already warped in the cavity, the cause is differential shrinkage or orientation, and it is a design and tool question. That single measurement saves more time than any other diagnostic step, and it is the one most often skipped.

Jetting, burn marks and silver streaks

These three are all about how the melt enters the cavity, and they are frequently confused with each other.

Jetting is a visible tail or rib where the melt arrives through the gate as a poorly conditioned stream rather than a filled channel. It is caused by a cold slug or by the melt being pushed through a long cold runner. A cold slug shows up as a solid lump of unmelted material; a flow disturbance shows as a rippled tail. Moving the gate away from the visible surface, re-locating the cold slug well, and raising melt temperature all help. Jetting on a cosmetic surface cannot be fully corrected in process; it needs a gate change.

Burn marks are brown or scorched streaks at the end of fill, and they are almost always trapped air being compressed rather than vented. The fix is venting, and specifically the vent slots being deep enough and positioned at the last point to fill. A vent that is too shallow will produce a burn mark that grows as the cavity fills, because the air has nowhere to go until it is too late.

Silver streaks are thin bright streaks along the flow direction, caused by moisture in the resin volatilizing at the melt front. The test is simple and takes five minutes: run the same tool with freshly dried, virgin resin and no regrind. If the streaks disappear, it was material. If they persist, look at back pressure and at the screw’s decompression settings, which control how much trapped air is carried forward with the melt. Silver streaks are frequently blamed on the tool and are frequently a material-handling problem.

Flash, voids and the parting-line family

Flash is a thin fin of material along the parting line. It is a mold problem in almost every case: a parting line that does not close evenly because of wear, debris, or a clamp force that is too low for the projected area. Debris in the parting line is the most common cause and is also the cheapest to fix. Beyond that, flash means the injection pressure is exceeding what the parting line can hold, which means the cavity is overpacked, which often means the packing is fighting a sink problem. Raising clamping force to compensate is a short-term answer that shows up later as flash on the same tool at a different machine.

Voids and gas pockets are bubbles trapped inside the wall. They come from two places: the gas already in the melt, and the gas generated by the reaction of regrind or contaminated material. Voids clustered near the gate point to the second cause. Voids distributed through the thick section point to the first. Neither is a tooling problem, and both respond to material handling and screw settings.

Weld lines were covered above, and flow marks too. One more member of this family is worth naming: burn and flame. A flame that appears only on the first shots and then disappears is trapped air in the cavity, which means the tool needs better venting. A flame that appears after a few thousand shots is usually a vent that has been partially blocked by regrind dust, and it will keep getting worse until the tool is cleaned.

The fixed troubleshooting order that saves the most time

The reason defect investigation drags on is that people change things in whatever order the symptom suggests. The symptom order is wrong. Material first, then mold temperature, then speed and pressure, then packing and hold, then cooling time, and only then the tool. Each step has a one-line reason why it comes before the next.

Two rules make this work in practice. Change one variable at a time and write down what changed and what happened, because the dominant reason a defect survives on the floor is that three settings were adjusted on the same trial and the trial record cannot say which one mattered. And stop after two adjustments: if a single variable change does not move the symptom at all, twice in a row, the cause is not in the process window. Go to the tool and check venting, cooling, gate position and the parting line.

That stopping rule is the one teams skip, and it is where the money leaks. A defect that should have been diagnosed as a venting problem gets three rounds of parameter tuning, and each round costs a day of production and builds a record that the tool is “sensitive”. The tool was never the variable being tested.

Tool-side versus process-side: how to tell them apart

The single most valuable diagnostic is the one that separates tool problems from process problems, because the two have completely different cost implications. A process problem costs a shift. A tool problem costs days and money, and the moment you start reworking a tool without a diagnosis, you have bought a repair you may not need.

There are four tests, in order of cost.

The trial record test. If the tool was delivered with a parameter set, and the defect appears on the first trial under that parameter set, the cause is in the tool. If the defect appears after material, settings or regrind percentage was changed, the cause is in the process. Most disputes with suppliers are settled by this one piece of paper, which is why the trial record should be handed over with the tool rather than kept in the factory.

The same-part, different-tool test. Run the same part on a second tool of similar geometry. A defect that follows the part is caused by the material or the design; a defect that follows the tool is caused by the tool.

The material test. Run the same tool with dried virgin resin and zero regrind. Defects that vanish were never tool problems. This takes one trial and it eliminates the single most commonly misdiagnosed category.

The cool-down test. Measure the part warm, inside or immediately out of the cavity, and again after it has reached room temperature. A part that is straight warm and warped cold is a cooling or ejection problem. A part that is already warped warm is a shrinkage, orientation or design problem. This test takes a caliper and five minutes, and it splits the largest single category of defects into two completely different work streams.

What to collect before you call the mold maker

The quality of the answer you get depends entirely on what you send with the question. A photograph and a description produces a guess and a bill. A complete package produces a cause and a quote, and it is the difference between a five-minute email and a two-week argument.

The package has six parts. The defective part, plus a good part from the same tool for comparison if you have one. A photograph of the defect under even light, with a scale reference in frame, and a second photograph showing where the defect sits on the part. The printout from the machine: actual part weight against the theoretical weight, injection profile, peak pressure, hold pressure, hold time, melt temperature, mold temperature, cycle time and cushion position. The material: grade, batch, regrind percentage, and the drying record. The tooling record: the trial parameters the tool was released with, and the maintenance history including shots since last service. The history: when the defect started, and what changed immediately before it started.

That last item is worth emphasising. “It started after we changed the resin batch” and “it started on the tool’s four hundredth shot” are answers. “The parts have been inconsistent since we received them” is a starting point. The most useful question in the whole exchange is simply what changed immediately before this defect appeared, because the answer is usually in the trial record and it points straight at one cause.

How to prevent the same defect from coming back

Prevention is mostly a documentation problem, and it is worth doing because the same defect reappearing is what turns a one-day issue into a recurring cost. Three things carry most of the weight.

First, review the part design before the tool is cut. Uniform wall thickness, ribs at roughly half the thickness of the wall they sit on, no abrupt section changes, fillets on every corner that would otherwise be sharp, and draft everywhere the tool needs to release. Most defects that appear on trial one are traceable to one of those five things, and all five are free to fix before the steel is cut.

Second, balance the cooling so the whole cavity reaches working temperature at the same rate. A cavity that runs hot in one region and cold in another produces warpage, sink and flow marks simultaneously, and the fix is in the tool rather than in the settings. Cooling layout is the one tool-side decision that pays back across every defect category.

Third, and most neglected, write the process window down and hand it to production. The majority of repeat defects we see are not repeat tool faults. They are the same settings problem reappearing when an operator changes the settings, or when a shift runs with a different set of numbers because the original set was never written down. A tool that depends on one operator’s memory is a tool that will produce defects.

None of this is exotic. It is the same discipline the plant uses to keep a stable process running once a tool is working: hold the material steady, hold the window steady, hold the record, and change one thing at a time. Applied deliberately before a tool is built rather than after it is failing, it removes most of the defect problems that come to us from buyers after the tool has already been paid for.

Three defect investigations and what they turned out to be

Defect disputes usually stall because both parties are arguing about the wrong variable. These three are real investigations from our own toolroom, and in each one the cause was named before anything was changed to the tool, because a rework without a diagnosis is a purchase with no guarantee attached to it.

Case 1: The warpage that was actually the ejectors

A European appliance customer reported that housings from a tool we had built were twisting in assembly, and that the twist was inconsistent enough to stop the line. Their incoming inspection measured parts at different angles and rejected them on flatness. The mould had been running for eleven months with no complaints, and nothing about the tool or the material had changed.

We asked for the parts warm rather than cold, and the picture changed immediately. Warm, the parts were flat and within every dimension. Cold, they were twisted. A part that is straight warm and distorted cold has not been distorted by shrinkage; it has been distorted by the stress of being taken out of the tool. The cause was not warpage at all. The draft on the two long side walls had been specified at half a degree, which is inside tolerance for a tool that has not yet been through a full thermal cycle and outside it once the cavity has polished itself in. The ejectors were pushing a part that was gripping harder than the design expected.

The fix was a draft change on two walls and a polish of the cavity face, not a new tool and not a cooling rework. The general lesson is that the cheapest diagnostic in injection molding is a caliper and a warm part, and that a defect which appears only after a tool has run for months is usually a wear or draft problem rather than a material one. Our mould maintenance reference covers the wear patterns that produce exactly this class of late-life drift.

Case 2: The silver streaks that were never a tool problem

A medical device customer rejected an entire batch with silver streaks running the length of the flow direction and asked us to repair the tool. The photographs showed the classic appearance of volatiles from a wet resin, and the parts had been moulded with a regrind ratio that had risen from 15 percent to 40 percent in the two months before the complaint, because the customer’s own scrap rate had gone up and the regrind was being reprocessed to recover material.

We asked them to run one trial with dried virgin resin and zero regrind. The streaks disappeared. The tool had never been involved. The general lesson is that silver streaks are the single most commonly misdiagnosed defect in injection molding, because they look like a tool surface problem and are usually a material handling problem; the reversing screw barrel that was left in for the regrind was the real culprit. It took one trial to prove it, and it took three weeks of misdirected tool repair requests to arrive at that trial, which is why we now ask for the drying record before anything else. Our materials reference sets out how resin grade and regrind ratio change the behaviour of the same tool.

Case 3: The short shot that turned out to be vent slots

A consumer electronics customer came back with a short shot on a new tool, in the last third of the flow path, with a burn mark at the end of fill. Their plant had responded the way most plants do: they had raised the injection pressure by 15 percent and the hold time by 20 percent, and the short shot had stayed in exactly the same place while flash appeared around the parting line.

That response is the clue. Flash and short shot appearing together means the cavity is being over-packed without the air ahead of the flow front being able to escape, which means the vents are not doing their job. The vents on this tool were the shallowest we had cut on a tool of that size, because the cavity volume was low relative to the projected area and the vent depth had been calculated against the fill volume rather than the compression ratio of the trapped air. We opened the vent slots, and the short shot resolved at the original pressure and hold settings, with the flash going with it.

The general lesson is that a short shot with a burn mark is an air problem, and that raising pressure against an air problem converts it into a flash problem rather than fixing it, which means the diagnosis moves further away instead of closer. Our defect reference lists the symptoms that separate the two failure families.

Injection molding defect terms quick reference