Injection Molding Defects: Spotting and Preventing the Six That Reach a Customer

The single rule behind every defect

Injection molding defects are almost always traceable to three variables: material shrinkage, process setting, or tooling geometry. Plastic shrinks as it cools, and anything that makes that shrinkage uneven - a thick section, a cold mold zone, trapped gas, or a distant gate - shows up as a visible fault on the finished part. Reading a defect is therefore a matter of asking where the shrinkage or the flow went wrong, not of guessing at random.

A useful mental model is to treat every defect as a question about where the plastic went, or where it failed to go. A depression asks whether the inside shrank more than the outside; a burn asks whether air had somewhere to escape; a short shot asks whether the flow front was ever given enough pressure and venting to reach the end. Asking in that order is faster than scrolling a troubleshooting menu.

That framing matters to a buyer because most defects are not random bad luck. They are repeatable, diagnosable, and either designed out before the steel is cut or managed at the machine during the run. A molder who reads the symptom and isolates the single variable responsible can usually fix it before a drift becomes a rejected batch. The six defects below cover the overwhelming majority of what reaches a quality dock, and each one points back to a specific cause you can act on rather than a mystery you have to live with.

This is also why a defect report should never stop at the symptom. The part number, the cavity, the shot count and the last tool service date belong on the same line as the photo, because the cause is almost always recoverable from that context. Buyers who ask suppliers for that context get fewer repeat defects than buyers who simply return the part and wait.

1. Sink marks

Localized depressions, usually opposite thick ribs or bosses, are caused by excessive local shrinkage as the core of the part cools and pulls the solidified skin inward. They appear most on thick sections, bosses, gear hubs and the area opposite a gate, because those are exactly the places where the interior stays molten longest and keeps contracting after the surface has set.

The reason sinks track thick features is simple physics: thick sections hold heat, stay liquid longer, and so they continue to contract after the thin skin around them has already frozen and locked its size. The skin wins the size argument locally and gets pulled inward. That is why a sink almost never appears on a uniform thin wall - it appears where the wall is not uniform.

The standard fixes are process and geometry together: raise holding pressure and hold time so the cavity stays packed while the gate freezes; add fillets or thin the offending rib so the section simply shrinks less; lower melt temperature slightly; and improve gate size and location so packing pressure actually reaches the thick area. Push too far and you trade a sink for flash, so the balance between pack pressure and clamp force is the real skill, not a single number on the screen.

In practice the first move is almost always to add hold time before adding pressure, because pressure without time still lets the gate freeze before the cavity is topped up. Only after hold time is near the gate-freeze limit does more pressure help, and that is the point at which flash becomes the risk to watch.

2. Warpage

Distortion comes from non-uniform shrinkage - differential orientation and cooling across the part. A flat panel warps because one side cools faster than the other and freezes a different level of internal stress; the part then relaxes out of shape as it ages, sometimes days after it left the mold. Long, thin or asymmetric parts are the usual victims.

It helps to think of warpage as frozen stress rather than a dimensional error. The part leaves the mold apparently flat, then moves as the locked-in orientation relaxes - which is why a warped part can pass a fixture check on the press and fail it on the inspection bench an hour later. Cooling both faces at the same rate is what prevents the stress from being locked in at all.

The fixes are mostly design: keep wall thickness uniform (the single biggest lever), balance mold-temperature zones so both faces cool alike, lengthen cooling time, and design symmetrical geometry. Warpage is largely designed out at CAD; once the tool exists you are managing it, not eliminating it, and a symmetric gate layout is worth more than any process tweak.

For a buyer, the practical takeaway is to ask about the mold-temperature strategy, not just the cycle time. A supplier who can hold both cavity halves at a matched temperature is controlling warpage at the source; one who only quotes a fast cycle is probably trading that control for speed you did not ask for.

3. Flash

A thin fin of material at the parting line comes from insufficient clamp force or a worn, dirty or dented parting line that lets melt escape under pressure. It is often the first sign of tool wear rather than a process problem, which is why it is worth treating as a maintenance event instead of a setting to chase.

Flash also scales with viscosity in the obvious direction: a resin that is easier to push is also easier to push where it should not go, so a free-flowing grade that seems like a processing convenience can be the thing that starts a parting line weeping. That is another reason resin choice is a defect decision, not just a cost decision.

Fixes: increase clamp tonnage, service and clean the parting line, lower injection pressure, and check for degraded resin that flows too easily and finds every gap. Flash is usually a maintenance signal - a parting line that needs attention - more than a setting, so check the tool before chasing process numbers.

The maintenance discipline that prevents flash is unglamorous and constant: keep the parting line clean, keep the vent depth correct, and treat any new fin as a tool event to log. Shops that log flash as a tool event instead of a pressure tweak keep their tools longer and their reject rates lower.

4. Short shot

An incomplete fill means the cavity never fully packed. The fixes are to raise melt and mold temperature, increase injection speed and pressure, enlarge the gate or runner, and add vents - because trapped air compresses and physically blocks the advancing melt front long before the plastic runs out of places to go.

It is worth separating the two kinds of short shot, because they need opposite responses. A short shot near a gate is usually a pressure or temperature problem and responds to process. A short shot at the far end of a long wall is a flow-length problem and will not respond to process at all - it needs geometry or material. Treating the second kind as the first is how hours get wasted.

A short shot at the far end of a long thin wall usually means the wall is simply too thin for the resin's flow length; no amount of pressure fixes a bad flow-length-to-thickness ratio, and the real fix is a design change or a hotter, faster-filling grade. This is why flow length belongs in the design review, not the troubleshooting meeting.

The design lever that prevents most flow-length short shots is unexciting: specify a minimum wall thickness for the chosen resin and keep long flows away from thin sections. A designer who knows the resin's practical flow length at the quoting stage removes the problem before tool steel is committed.

5. Weld (knit) lines

A visible line where two flow fronts meet - around holes, inserts or multiple gates - is a knit line, and mechanically it is a weak seam because the two fronts never fully weld where they meet. Fixes: raise melt temperature, move the gate so fronts merge earlier and hotter, add vents, and increase injection speed so the two fronts weld before they chill.

The weakness at a knit line is not cosmetic only. Because the two fronts meet with a skin already formed on each, they never fully intermix, so the seam is a genuine mechanical discontinuity - which is why a knit line on a loaded snap or a structural rib is a different problem from one on a cosmetic face. The location decides how much you care.

Weld-line strength matters most on load-bearing features. Where a knit line lands on a structural wall, ask the molder how it was managed - a hot-runner and a relocated gate often turn a visible seam into an invisible, strong join. DAYIN's field guide to the six molding defects walks through this with production examples, and it is the reference most useful when a knit line sits somewhere it should not.

Improving a knit line is mostly about making the two fronts hot and pressurised when they meet: higher melt temperature, a gate that brings them together sooner, and vents that stop air from sitting in the way. A molder who shows you a relocated gate to move a knit line off a load path is managing the defect at the level that matters.

6. Burn marks

Brown or black scorching is compressed, trapped gas igniting as the cavity fills. The fixes are to add or clear vents, reduce melt temperature, slow the injection rate so air escapes ahead of the front, and reduce decompression so the screw does not draw air into the melt on the return stroke.

The ignition itself is a small, local version of what happens in a diesel engine: enough pressure and heat on trapped gas and it combusts against the plastic. The cure is never more pressure - it is always a path for the gas. That is why the fix is mechanical, a vent, rather than a parameter, a lower number.

Burn marks cluster at the last place to fill - exactly where venting is missing. They are the clearest possible signal that a vent is blocked or absent, and they disappear the moment the air has somewhere to go. Because they are pure venting problems, a burn mark is usually the cheapest defect on this list to actually fix: clear the vent and the part is clean.

Because the burn appears at the last place to fill, it is also a useful map of the mold's venting coverage. A new burn mark in a new spot is the mold telling you a vent has clogged or a feature has changed the flow path - read it as a maintenance cue, not just a scrap reason.

Defect, cause and first fix at a glance

The table below compresses the six into a single diagnostic reference. Use it as the first screen when a part comes back from the floor - confirm the tool is clean and the resin is dry, then pick the first fix and change only that one variable before re-checking. Keep a printed copy at the press; the fifteen seconds it saves on the first suspect check pays for itself the first time it stops someone from over-trimming a process and creating a second defect.

Diagnose one variable at a time

Chase defects one variable at a time: confirm the tool is clean and the resin is dry, verify shot size and cushion, then change a single setting and re-check. Changing three things at once tells you nothing about which one mattered, and it can hide a real fix behind a coincidence you will never reproduce.

The discipline is boring but it is the only one that scales: a logbook that records one change and one result builds a private model of the tool that no supplier memo ever will. After a few runs that notebook is worth more than the original process sheet, because it captures how this specific cavity actually behaves rather than how the resin datasheet says it should.

Modern cells log cavity pressure and automatically reject parts outside the window, so a drift toward sink or flash is caught at part one rather than at final audit. A monitored, documented process is what keeps these problems off the floor in the first place, and it is the difference between a supplier who prevents defects and one who hopes they do not happen.

Cavity-pressure monitoring turns that manual discipline into an automatic one. When the cell rejects a part the moment pressure leaves the window, the drift is caught before it becomes a batch, and the data it records is exactly what you need to argue a tool or a resin change with the supplier.

Prevent them from shipping at all

The cheapest defect is the one that never ships. That means specifying wall thickness and gate location in the CAD review, qualifying the tool on a first article, and agreeing on a cavity-pressure window with the molder before volume runs. A partner who can show you the process data, not just the parts, is the one who will not be on the phone about a rejected batch six weeks into production.

None of this is expensive to set up; it is a matter of agreeing the acceptance method before the tool is cut, not after the first bad lot. The cost of a first article and a documented process window is trivial next to the cost of a recall, a line stoppage at a customer, or a relationship that ends over avoidable scrap. Prevention is a paragraph in the purchase order, not a capital project.

For production-volume plastic parts, a supplier with in-house tooling, a documented molding process and a quality lab is worth more than a low unit price. DAYIN runs tooling, presses and inspection under one roof, and a 35-year molding and OEM track record is the kind of history that shows up as fewer defects rather than as a line on a quote.

And the supplier choice is the lever that makes the rest real. A partner with the tool room, the presses and the lab in one building can act on a defect in hours rather than in a chain of subcontractors. That continuity is what turns 'we found a defect' into 'we prevented it', which is the only version of the story you want on a production program.

Frequently asked