How to Troubleshoot Hot Runner Leaks: Resin and Melt Failure Points

A hot runner that leaks 2 g of glass-filled melt per shot sounds minor until you do the math. At a 20 second cycle that is roughly 360 g per hour of resin at 230-300 °C cooking against steel, heaters and thermocouple wires. Within a shift the leaked resin carbonizes, welds nozzle tips into bushings and shorts heater bands. The mold comes apart for 4 to 16 hours of teardown, and connector OEMs including TE Connectivity and Luxshare treat that downtime as a line-stop event, not a maintenance ticket. This guide walks the four leak failure points, the diagnosis order that finds the real root cause fast, and the numeric criteria mold shops use to fix each one.

The Snapshot

• Tool steel expands about 12 µm per meter per °C. A 600 mm manifold heated from 20 °C to 230 °C grows roughly 1.5 mm, and the seal stack must absorb all of it (typical figure).
• Manifold and nozzle seal faces should hold flatness within 0.02 mm and a surface finish of Ra 0.8 or better for leak-free seating (typical industry figure).
• Thermocouple readings drifting more than ±10 °C from setpoint signal heater decay or poor contact, a leading pre-cursor to freeze-off leaks.
• Manifold bolts get torqued in three stages, 30 percent, 60 percent, then 100 percent of spec, in a cross pattern, with hot re-torque per OEM instruction.
• A full hot half teardown and seal replacement typically takes 4 to 16 hours, versus 30 to 60 minutes for an in-press nozzle inspection.

How Hot Runner Leaks Start

Every hot runner leak is a pressure story. Injection pressure at the nozzle tip routinely peaks between 50 and 150 MPa, and pack pressure holds 20 to 60 MPa against the melt channel walls for seconds at a time. The melt only escapes where a sealing interface loses contact force or where a gap opens wider than the seal can bridge.

A hot half contains a small number of these interfaces. Melt flows from the machine barrel into the manifold through an inlet seal, splits through the manifold channels to each nozzle, and passes the nozzle tip to sprue bushing interface into the cavity. Each interface is a steel-to-steel joint held closed by bolt preload, thermal growth or a compression seal ring.

The joint only stays sealed while contact pressure exceeds melt pressure by a safety margin. Most hot runner OEMs design for a seating stress that is several times peak injection pressure. When that margin collapses, melt finds the path of least resistance and leaks.

Three mechanisms collapse the margin. First, thermal growth mismatch: the manifold expands more or less than the mold plates around it, and the designed crush on a seal ring disappears or overloads. Second, mechanical damage: worn seal faces, galled threads or a bent nozzle from a crashed mold reduce contact area. Third, temperature control drift: a cold spot freezes melt into a solid plug, and injection pressure then spikes behind the plug and blows out the weakest joint.

Understanding which mechanism is active matters because the fixes conflict. Adding torque to a joint that is already overloaded by thermal growth crushes the seal flat and makes the leak worse. The diagnosis sequence later in this guide exists to keep you from applying the wrong fix.

The 4 Leak Failure Points

Leaks cluster at four locations, and each leaves a distinct signature on the mold. Learning the signatures cuts diagnosis time from hours to minutes.

1. Nozzle tip to sprue bushing interface. This is the most common leak point in open-gate systems. Melt escapes around the nozzle tip and pools between the hot half and the cavity plate. You find resin drool at the parting line, brown streaks on the cavity plate face, or a nozzle tip frozen into the bushing at teardown.

2. Manifold seal interfaces. Melt leaks between the manifold and the mold plates, at the inlet seal, or around channel plugs. Signatures include resin in the ejector box, smoke from the mold vents, and carbonized deposits on manifold edges. Plugged manifold systems leak here more often than solid-machined ones because every plug is another potential path.

3. Heater and thermocouple failures. A burned-out heater band creates a cold zone. Melt freezes there, pressure spikes behind the plug, and the resulting blowout leaks at whichever joint is weakest. The signature is a leak that appears only after the system has run a while, plus a temperature deviation over ±10 °C on the controller.

4. Thermal expansion mismatch. The manifold and the mold plates grow at different rates or in different directions. Seal contact force rises or falls through the heat-up cycle. The signature is a leak tied to temperature: absent when cold, appearing at steady state, or worse after a soak period.

Real molds often show combinations. A heater failure can trigger an expansion-driven overload that permanently crushes a seal ring, so the leak continues after the heater is replaced. That is why the fix sequence matters as much as the part replacement.

← swipe to scroll →

Table note: Signatures and thresholds are typical industry figures. Always defer to the hot runner OEM manual for your specific system.

Leak Diagnosis Sequence: 5 Checks in Order

Work this order every time. Skipping ahead is how shops crush good seals and miss the real cause.

Check 1: Read the temperature profile first. Before the mold even comes down, log every zone. A healthy system holds each zone within ±5 °C of setpoint at steady state, and no thermocouple reads open or erratic. Any deviation beyond ±10 °C points at a heater or sensor problem, which can be the actual cause of a leak that looks mechanical. This check costs 10 minutes and saves an 8-hour teardown.

Check 2: Locate the leak path physically. With the mold open and the hot half warm, look for fresh resin at the parting line, at nozzle tips and around manifold edges. Resin that is amber and soft leaked recently. Resin that is black and brittle leaked earlier and carbonized, so it marks the history, not necessarily the active path. Clean the region and run a few shots to watch where melt actually appears.

Check 3: Inspect seal faces. Pull the hot half and examine every seal interface. Sealing surfaces should be flat within 0.02 mm with a finish of Ra 0.8 or better (typical figures). Look for three defects: a compression seal ring squeezed to less than 70 percent of its original thickness, dings or scratch tracks crossing the seal circle, and galling marks that signal metal-to-metal contact where a gap should exist.

Check 4: Verify torque and bolt condition. Manifold bolts lose preload through thermal cycling. Check every bolt against the OEM torque spec with a calibrated wrench. M12 grade 12.9 manifold screws commonly spec between 80 and 105 N·m, but your system's manual governs. Any bolt that turns without resistance has yielded and must be replaced, not re-tightened.

Check 5: Measure alignment and thermal gap. Nozzle tip to sprue bushing centerline alignment should be within 0.05 mm (typical). Cold gap between manifold and plate at assembly must match the OEM expansion table for your operating temperature, often in the 0.1 to 0.4 mm range per interface. Measure with feeler gauges at room temperature, cold, before the mold ever heats up.

Nozzle to Sprue Bushing Leaks

The nozzle tip to sprue bushing joint is where the hot half meets the cold half, and it leaks for four measurable reasons.

Reason 1: wrong nozzle installed height. Every nozzle has a specified installed height that sets the tip crush against the bushing. Too high and the tip never seats, leaving a gap that injection pressure fills instantly. Too low and the tip over-crushes, deforming the sealing land or bending the tip. Measure the height with a depth micrometer at assembly and hold the OEM tolerance, typically ±0.02 mm on the seated dimension.

Reason 2: worn or damaged bushing sealing land. The bushing face that contacts the nozzle tip wears with every thermal cycle, especially with abrasive resins. A 30 percent glass-filled nylon erodes sealing lands far faster than unfilled polypropylene. Inspect the land under magnification. A wear ring wider than the OEM contact spec, often 0.5 to 1.0 mm, means the bushing needs replacement or re-machining.

Reason 3: misalignment. If the nozzle axis is off from the bushing axis by more than 0.05 mm, contact pressure concentrates on one side and the opposite side opens a leak path. Misalignment comes from mold plate shift, worn leader pins or a hot half installed without checking the register. Measure with a dial indicator before blaming the nozzle.

Reason 4: cold nozzle tip. An open-gate nozzle tip runs intentionally cooler than the manifold to prevent drool, typically 20 to 40 °C below melt temperature. If tip temperature drops too low, melt freezes at the gate, pressure spikes behind it, and the joint blows. If your controller cannot hold tip temperature within ±5 °C, fix the control loop before touching the steel.

Quick containment for a nozzle-to-bushing leak in production: drop pack pressure 10 to 20 percent if the part still fills, and schedule the mold pull. Running hot with an active leak carbonizes resin into the joint and converts a $50 seal or bushing replacement into a nozzle and bushing rebuild. Precision shops like DieStrike machine replacement sprue bushings and nozzle seats to ±0.005 mm, and standard components such as sprue bushings ship in 3 to 7 days when the mold cannot wait.

Manifold Seal Leaks

Manifold leaks come in two families: seal ring failures at the manifold-to-plate interfaces, and channel plug leaks in plugged manifold designs. Both are assembly-quality problems, and both are preventable.

Seal ring compression. Most hot runner systems seal with steel or nickel alloy rings that compress between the manifold face and the mold plate. The design gives a controlled crush, often 0.05 to 0.15 mm of compression at operating temperature (typical figures). Two failure modes show up. Under-compression leaves a gap: the ring never seats and melt weeps from hour one. Over-compression flattens the ring into a pancake: it looks sealed when cold but springs back unevenly as it heats, and the leak appears after a soak period.

Measure ring thickness before installation. A ring compressed below 70 percent of its free height has lost its spring-back and must be replaced. Never reuse a crushed ring during reassembly, even if it looks fine.

Face flatness and finish. The manifold sealing face and the mold plate pocket must both hold flatness within 0.02 mm across the seal area, with a finish of Ra 0.8 or better (typical industry figures). Warped manifold faces come from uneven heating, over-torque, or a manifold that was dropped or clamped during handling. Check flatness on a surface plate with a dial indicator. A face out beyond 0.05 mm usually needs re-machining, and beyond that the manifold may need straightening by the OEM or replacement.

Channel plug leaks. Plugged manifolds have drilled channels closed with pressed or threaded plugs. A plug that walks out or a plug bore that was not cleaned of EDM recast will weep melt. Leaked resin near a plug leaves a distinct carbon trail radiating from the plug edge. Repair means drilling out the plug, cleaning the bore, and pressing a new oversized plug per OEM spec. Do not weld over a leaking plug as a shortcut, the heat distorts the channel and the leak returns.

Torque discipline. Manifold bolt preload sets seal compression, so torque error is directly leak error. Follow a three-stage sequence: 30 percent, then 60 percent, then 100 percent of spec, always in a cross pattern from the center bolts outward. Many OEMs also require a hot re-torque after the first heat cycle, when seals have taken their initial set. Use a calibrated wrench. Guessing torque by feel is how new seals get crushed on day one.

Heater and Thermocouple Failure Leaks

Heater failures cause leaks indirectly but reliably. The chain runs: heater decays, zone runs cold, melt viscosity climbs or melt freezes, injection pressure spikes behind the restriction, and the weakest joint in the system blows. Shops that only look for mechanical causes at the leak site miss this chain about as often as they find it.

Test heaters by resistance before pulling anything. With the system off and cool, measure each heater band and probe. Compare readings to the OEM table, usually printed on the wiring diagram. A band reading 30 percent below nominal resistance has shorted turns and will overheat locally. A band reading open is dead. Bands within 10 percent of each other in the same size class are healthy (typical figure).

Test thermocouples the same way. A J-type or K-type TC should read roughly its ambient value on the controller, usually 20 to 30 °C in a warm shop. A reading of zero, full scale or jumping by tens of degrees means a broken junction or a loose connection. A TC mounted with poor contact reads 15 to 30 °C low, so the controller overdrives the heater. The result is a zone that runs hotter than the resin can take, degrades material into gas, and gas pressure opens leak paths at seals.

Watch the deviation trend. A healthy zone overshoots on startup then settles within ±5 °C. A zone that cycles ±15 °C or more at steady state has a failing heater, a loose TC, or a controller tuning problem. Log deviations over a production run. A slow upward drift in deviation over weeks is heater decay, and replacing the band before it fails completely prevents the freeze-off blowout that leaks.

Check for resin insulation. Carbonized leaked resin coats heater bands and acts as insulation. A band covered in black char runs hotter internally while the melt channel stays cold, which accelerates burn-out and worsens the cold zone. Clean bands with a brass brush at every teardown. Any band with char baked into the windings gets replaced.

Heater and TC replacement is cheap insurance. Bands and probes cost tens of dollars each, while a freeze-off blowout can warp a manifold that costs thousands. Keeping a spare set of each heater size and TC type in the crib is standard practice at disciplined mold shops.

Thermal Expansion Mismatch

Thermal expansion is the most misunderstood leak cause because it is invisible when the mold is cold and only misbehaves at temperature. The physics is simple. Tool steels expand about 12 µm per meter per °C (typical figure). Heat a 600 mm long manifold from 20 °C to 230 °C and it grows roughly 1.5 mm. The mold plates around it, made of P20 or 4140 class steel and cooled to 40 to 90 °C, grow far less.

Hot runner designers plan for this. The manifold either floats on support pads and grows into designed gaps, or it is bolted rigidly with seal rings that absorb the differential. Either way, the cold assembly gap must match the OEM expansion table. If a mold was repaired and reassembled with the wrong shim, a missing support pad, or seal rings of the wrong thickness, the design crush becomes a design gap at operating temperature. The system leaks at steady state and seals again when it cools, which confuses operators and invites the wrong fix.

How to verify the expansion stack. Measure every interface gap cold, before heat-up, and compare against the OEM drawing. Common checkpoints: the gap between manifold edge and pocket wall, the compression allowance at each seal ring, and the nozzle installed height, since nozzle bodies also grow and change tip crush by 0.05 to 0.15 mm between cold and hot (typical figures). Write the cold measurements down. They are the baseline for every future reassembly.

Symptoms that point at expansion. A leak that appears 30 to 60 minutes after heat-up, once the manifold reaches steady state, is an expansion signature. A leak that seals itself after a cool-down and reopens at temperature is also expansion. A leak that worsens when you raise melt temperature by 10 to 20 °C is almost certainly expansion-driven, because the extra growth unloads the seal.

The fix is dimensional, not torque. Do not chase an expansion leak by tightening bolts. Extra preload at cold assembly becomes overload at temperature and crushes seal rings flat, which converts a recoverable leak into permanent damage. Correct the stack instead: restore the OEM shim thickness, replace compressed seals with the specified part, and re-measure cold gaps. If the manifold face warped from past overload, flatness repair comes first. A mold shop holding ±0.002 mm geometry tolerance on mold components can re-machine seal pockets and support pads back to print, which is the permanent fix for a repeat-expansion leaker.

Downtime Cost and Fast Containment

Leak economics decide how fast you must act. The resin loss itself is trivial. A leak of 2 g per shot at a 20 second cycle wastes about 8 to 10 kg per day, worth maybe $30 to $80 depending on the grade. The real cost is downtime and collateral damage.

Downtime tiers. An in-press nozzle inspection and seal swap takes 30 to 60 minutes. Pulling the hot half for manifold seal replacement runs 2 to 4 hours. A full teardown with manifold face repair, heater replacement and re-qualification takes 4 to 16 hours. If the mold must leave the press for off-site machining, add shipping and queue time, and a 1-day repair becomes a 1 to 2 week outage. Connector and automotive programs running two or three shifts feel every hour, since a stopped mold idles a press worth $30 to $80 per hour in burden.

Collateral damage escalates fast. Leaked melt carbonizes within minutes at 230 to 300 °C. Carbonized resin welds nozzle tips to bushings, so a $40 bushing replacement becomes a seized nozzle you must cut out. Char infiltrates thermocouple junctions and heater windings, turning an electrical system into scrap. Worst case, leaked melt reaches mold cooling lines or ejector pins and damages the cold half too. Every hour you run with an active leak multiplies the repair bill, typically by 2 to 5 times over a single shift.

Containment steps that protect the asset. First, purge the leaked region while the mold is still warm. Cold resin is mechanical to remove, but warm resin wipes clean and leaves no carbon seed. Second, coat nozzle tips and seal faces with a thin anti-seize or mold release rated for hot runner temperatures before reassembly. Third, protect heater bands and TCs with Kapton tape or ceramic sleeving where the leak path crossed them. Fourth, run a short trial after repair, 20 to 50 shots, and inspect every seal face before releasing the mold to production.

When to pull the mold out of the press. Any leak with visible smoke, any leak that recurs within an hour of a seal swap, and any leak with a temperature deviation over ±10 °C all justify pulling the hot half immediately. Continuing to run those cases converts a maintenance event into a capital event. Shops that track this honestly find that the molds pulled early cost a fraction of the molds run one more shift.

6 Design Rules That Prevent Leaks

Most hot runner leaks are born in the mold design and assembly phase, not in production. These six rules come from building and repairing hot halves for programs at connector OEMs including TE Connectivity and Amphenol, and they cover the failure points above.

Rule 1: Specify the thermal stack on the drawing. The mold drawing should call out cold assembly gaps for every hot runner interface, tied to the OEM expansion table and the actual operating temperature. A gap that is correct at 190 °C for polypropylene may be wrong at 300 °C for glass-filled PPS. Note the operating range next to the gap dimension so future repairs reassemble to the right number.

Rule 2: Hold sealing face flatness to 0.02 mm and Ra 0.8. Both the manifold face and the mold plate pocket get flatness callouts. Precision mold makers machine these faces on grinders or hard-mill them, then verify on a surface plate. A ±0.005 mm machining capability on mold dimensions makes the 0.02 mm seal flatness a controlled process, not a hope.

Rule 3: Write the torque spec into the mold manual. Every manifold bolt size, its torque value, the three-stage sequence and the hot re-torque step belong in the mold documentation that ships with the tool. M12 grade 12.9 hardware typically specs 80 to 105 N·m, but the OEM value governs. Technicians who cannot find the spec invent one, and invented specs crush seals.

Rule 4: Give nozzle tips a controlled land, not a knife edge. The tip sealing land should match the OEM spec, commonly 0.5 to 1.0 mm of contact width. A land too narrow cuts into the bushing under thermal load. A land too wide cannot maintain contact pressure across its full width when the face runs slightly out of flat. Specify tip and bushing as a matched pair, and see a dedicated nozzle guide like how to select hot runner nozzles before committing a gate design.

Rule 5: Build in alignment verification. Nozzle-to-bushing alignment within 0.05 mm should be a documented check at T1 and at every hot half reassembly. Leader pins and bushings that hold the plate stack in register are wear items. A mold that has run 500,000 shots may have shifted enough to load one side of every nozzle seal.

Rule 6: Schedule the preventive teardown. Heater resistance checks every production run, seal ring measurement at every hot half pull, and full manifold inspection at the OEM interval, commonly every 1 to 2 million shots (typical figure). A pre-run checklist like 7 hot runner maintenance checks turns these into a one-page habit. Repairs found early are seal swaps. Repairs found late are manifold replacements.

← swipe to scroll →

Table note: Criteria are typical industry figures. Confirm every value against the hot runner OEM manual for your system.

When a hot half needs machining work rather than parts, mold repair and maintenance services that hold tight tolerances on seal pockets, support pads and bushing seats are the difference between a repaired mold and a repeat leaker. For new tools, a full hot runner system integration at design time beats retrofitting seals after the first leak.

FAQ

Q1. How do I tell if a hot runner leak is mechanical or thermal?

Check when the leak appears. A leak from the first shot after heat-up is usually mechanical: worn seal face, wrong installed height or a crushed ring. A leak that starts 30 to 60 minutes into the run, once the manifold reaches steady state, points at thermal expansion mismatch. A leak that only shows up with a temperature deviation over ±10 °C points at heater or thermocouple failure causing freeze-off pressure spikes.

Q2. Can I keep running production with a small hot runner leak?

Only for minutes, not shifts. Leaked melt carbonizes within minutes at 230 to 300 °C, and carbon welds nozzle tips to bushings and fouls heaters. Every hour of running with an active leak typically multiplies the final repair bill by 2 to 5 times. If the part still fills, drop pack pressure 10 to 20 percent as temporary containment and pull the hot half the same shift.

Q3. What torque should manifold bolts get, and in what order?

The OEM spec governs, but M12 grade 12.9 manifold screws commonly run 80 to 105 N·m (typical figure). Always torque in three stages, 30 percent, then 60 percent, then 100 percent, in a cross pattern from the center outward. Many systems also need a hot re-torque after the first heat cycle. Replace any bolt that turns without resistance instead of re-tightening it.

Q4. How often should I replace hot runner seal rings?

Replace every seal ring at each hot half teardown, never reuse one. A ring compressed below 70 percent of its free height has lost spring-back and will leak again. For scheduled service, a full manifold inspection every 1 to 2 million shots is a common industry interval, with heater resistance and thermocouple checks every run.

Q5. What flatness do hot runner sealing faces need?

Seal faces should hold flatness within 0.02 mm with a surface finish of Ra 0.8 or better (typical industry figures). Check with a dial indicator on a surface plate. A face out by more than 0.05 mm generally needs re-machining before new seals will hold, otherwise the replacement rings will leak again within days.

The Bottom Line

Hot runner leaks are seal-contact failures with four repeatable causes. Diagnose in order: temperature profile, leak path, seal faces, torque, then expansion gaps. Fix dimensions, not torque, and never reuse a crushed ring. Running a leak past one shift multiplies repair cost 2 to 5 times. If your hot half needs seal pockets or bushing seats machined back to print, talk to our mold team.