Electrofusion Short Circuit and Smoke: Why Fittings Burn

Nov 09, 2026

Leave a message

An electrofusion fitting that smokes is not a fitting that was welded badly. It is a fitting in which the heating coil stopped being a resistor and started being a short, and the difference matters because the two have almost nothing in common in the way they are prevented. A badly welded joint is a heat balance problem. A smoking fitting is an electrical fault that happens to occur inside a joint.

The failure is sudden and it is usually loud: the machine alarms, smoke appears at the fitting body or at the terminals, and the cycle stops. What follows is the part that decides whether the same thing happens on the next joint, and it starts with reading the fault code rather than clearing it.

What the coil is, and what a short circuit actually means

An electrofusion coupler is a moulded PE body with a resistance wire wound into its inner surface, close to the bore, with the two ends brought out to terminals on the outside. The welder applies a defined voltage across those terminals for a defined time, both taken from the fitting's own data. The wire heats, the PE around it melts, the melt expands, and the pressure that builds up is what fuses the pipe to the fitting.

That whole design assumes the wire stays a single continuous resistor of a known value. A short circuit is any path that lets current bypass part of it. The turns are close together by design, separated by a millimetre or two of polymer; the moment that polymer stops insulating, the current has a shortcut, the resistance drops, the current rises, and the local temperature runs away faster than the machine can react.

Section through an electrofusion coupler showing the fitting body, the heating coil at the bore, two terminals, the normal current path and two adjacent turns bridged by a short
The turns are separated by a millimetre of polymer. That separation is the insulation, and it is the first thing to fail.

Once the polymer carbonises, the failure is self-sustaining. Carbon is conductive, so the burnt path becomes a better conductor than the wire it bypassed, which concentrates even more current in the same place. That is why a fitting that starts to smoke rarely stops at a small defect: the mechanism accelerates. The visible result is a black track, a hole in the body, or a coil that has been pushed out through the side. There is no equivalent mechanism in butt fusion, where the heat comes from outside the joint, which is one of the practical differences between the two processes compared in electrofusion versus butt fusion.

Four ways the current finds a path it should not

Water is the first and the most common. Pure water is a poor conductor, but the water on a site is not pure: groundwater, mud, road salt and drilling fluid all carry ions. A film of it across the terminals, or along the wire where the bore has been contaminated, bridges turns that were never meant to touch. It also does something else: it flashes to steam at melt temperature, and the expansion displaces the coil. In a fitting that survives the cycle, the same steam leaves a ring of voids at the interface, which is covered under voids and porosity in HDPE welds.

The second is mechanical. A pipe pushed past the insertion stop can scrape or displace the coil, and a fitting that has been dropped, stood on, or stored in a warm container can have a deformed bore in which the wire is no longer at a uniform depth. Neither is visible from outside, and both change the local heat balance.

The third is at the terminals. Corroded pins, bent pins, or a cable connector that is pushed on but not seated give a contact resistance that is higher than the coil's own. The heat then appears at the pin rather than in the joint, which is why a fitting can burn at the terminal while the coil itself looks intact.

The fourth is the parameter set. If the barcode cannot be read and the operator types a fitting code from memory, the voltage and time that reach the coil may belong to a different size or a different manufacturer. Too much energy produces smoke without any short circuit at all, and the damage looks similar.

The gap: less heat sink, not just more volume

A gap between pipe and fitting is usually discussed as a volume problem, as if the only consequence were that the melt has further to travel. That is not the mechanism that burns fittings. The wire is cooled by the material in contact with it, and where the pipe is not touching, there is nothing on one side to take the heat away.

So a wire with a gap runs hotter for the same energy, and it runs hotter on one side only. That is why the resulting damage is asymmetric: a burn on the side where the pipe sat away from the bore, and a perfectly normal joint on the other. The most common sources are an oval pipe, a scraper that took too much off one side, and a fitting that is simply at the loose end of its tolerance.

Chart of wire temperature against time for three cases: no contact giving no heat sink, over-energy from wrong parameters, and correct parameters, with a degradation band at the top and a fusion range below it
Same energy in, three different wire temperatures. The resistance check passes in all three.

This is also the reason why measuring the fit is worth the ten seconds it takes. A pipe that will not enter the fitting without force, and a pipe that drops in with visible clearance, are both wrong, and neither will be caught by the resistance check because neither is an electrical fault at the start of the cycle.

Temperature: the fitting that was left in the sun

The parameters on a fitting assume a starting temperature. A coupler that has been on the dashboard of a vehicle, in a black box in direct sun, or on the ground next to a running generator does not start at ambient. With the same voltage and the same time it receives the same energy, but it needs less to reach fusion, so the excess goes into overheating the wire.

The cold case is the mirror image. A fitting at minus five degrees takes more energy to reach the same state, and whether the machine compensates depends on the machine and on whether the operator told it the ambient temperature. Where compensation is available and not used, the joint is under-fused rather than burnt, which is the more dangerous of the two outcomes because it fails later and silently.

The practical rule is to keep fittings at the temperature the procedure assumes: out of the sun, off frozen ground, and out of the bag only when they are about to be used. The environmental limits themselves are the same ones that apply to the pipe, and they are set out under weather conditions for PE pipe welding.

Where the burn is tells you which one it was

The damage pattern is the most reliable diagnostic available, because the fitting is scrap either way and the only thing left to learn from it is the cause. Cut the failed fitting out, keep it, and look at where the damage sits before it goes in the bin.

Where the damage isWhat it usually meansWhat to check before the next joint
Burnt or melted at the terminal pinsContact resistance at the connection, not in the coilPin condition, cable connector, corrosion, moisture in the terminal well
One localised black spot in the coilTwo turns bridged, or a carbon track started at one pointBore cleanliness, water, coil displacement from over-insertion
Smoke along the whole coil, no black spotOver-energy: the parameters were not this fitting'sBarcode read, fitting code, voltage and time against the nameplate
Burn on one side only, joint otherwise correctThe wire had no heat sink there: gap or ovalityPipe roundness, scraper depth, fitting clearance
Alarm before any heating, coil coldResistance out of window: open circuit, water bridge, or wrong fittingMeter the coil, compare with the nameplate, check for water
Four coupler sections showing damage at the terminal, at one turn, along the whole coil, and on one side only, each labelled with its cause
Where the damage sits points at the cause. The fitting is scrap either way; the pattern is what you keep.

Why the fitting cannot be re-welded

The rule that any fitting which smoked or shorted must be discarded is not a manufacturer's caution, it is a consequence of what has happened inside it. Three things have changed and none of them can be reversed.

The coil geometry has moved. The wire that was wound at a uniform pitch is now displaced where the melt pushed it, so the next cycle would heat unevenly even if the electrical fault were somehow cleared. The polymer around the damaged zone has been pyrolysed, which means its fusion behaviour is no longer the behaviour the parameters were written for. And the resistance of the coil is no longer the value the machine expects, so the next cycle would either be refused or run at the wrong energy.

There is also a practical argument that matters more on site: a second cycle on a damaged fitting takes the same time as cutting it out, and produces a joint that nobody can inspect and nobody will trust. The only correct action is to cut the joint out completely, replace the fitting, and re-make the connection with a fresh length of pipe.

What to do in the first five minutes

The sequence matters because some of the evidence disappears as soon as anything is touched.

CheckWhat a bad reading looks likeWhat it points at
Fault code, before clearingResistance check failed, current fault mid-cycle, or supply interruptionWhen the fault happened, which separates electrical from thermal causes
Coil resistance, at the terminalsFar below nominal means bridged turns; far above or open means a broken wire or a corroded pinWhether the coil itself is damaged or the connection is
Terminal conditionCorrosion, water in the well, a bent or recessed pinContact resistance, which burns at the pin rather than the coil
Insertion markNot visible at the fitting faceOver-insertion, which can scrape or displace the coil
Fit and roundnessVisible clearance, or a pipe that had to be forced inMissing heat sink, which burns one side and leaves the other sound
Parameter sourceTyped rather than scanned, or a code that does not match the nameplateOver-energy, which smokes without any short circuit

Stop the cycle and disconnect the cable, but do not clear the fault code. Read it, write it down, and photograph the display. The code identifies whether the machine detected an out-of-window resistance, a current fault mid-cycle, or an interruption, and those point at different causes.

Then measure the coil across the terminals with a meter and compare it with the value on the nameplate. A reading far below nominal suggests the turns are bridged. A reading far above, or open, suggests a broken wire or a corroded pin. A reading close to nominal on a fitting that has already been heated tells you the fault was in the connection rather than the coil.

Check the terminals for corrosion, water and bent pins. Check the pipe insertion mark: if it is not visible at the fitting face, the pipe went in too far. Check the pipe roundness and the fit, because a loose fit is the one cause that leaves no electrical trace at all.

Finally, check the parameter source. If the barcode was scanned, the question is whether the label was the right one for that fitting. If it was typed, the question is where the numbers came from. The nameplate on the fitting is the reference, and how to read it is worth ten seconds of anyone's time before the next cycle. The parameters themselves, and the way a joint is meant to progress from check to cool, are set out in the PE pipe fusion welding guide.

Five step diagnostic sequence after a failed electrofusion cycle: read the fault code, meter the coil, inspect the terminals, check the insertion mark, check the fit and roundness
Five minutes, in this order. Some of the evidence disappears as soon as anything is touched.

Prevention that survives a busy day

Most of the rules are the ones already written in every fitting installation guide. They fail in practice for a different reason: they are all small, and each one is easy to skip once.

  • Keep the fitting in its bag until the moment of assembly. The bore is a prepared surface and it collects everything the pipe collects, including condensation. The handling rules for fittings are set out in daily care of electrofusion fittings.
  • Mark the insertion depth on the pipe, not on the fitting, and check the mark at the fitting face after insertion.
  • Keep the terminals dry and capped. A terminal well full of groundwater is a short waiting for a voltage.
  • Scan the barcode rather than typing it, and check that the fitting code on the screen matches the nameplate before pressing start.
  • Do not re-use a fitting that has been dropped or that has a deformed bore, whatever it looks like.
  • Shelter the joint in rain, fog and heavy dew. Water is the leading cause and it does not need to be visible to be present.
  • Clamp and align tees and saddles properly. A saddle that is not seated square loads the coil on one side before the cycle even starts, and the correct seating procedure is in electrofusion saddle construction.

The other half of prevention is knowing what the machine is telling you before it fails. The machine-side faults, including the ones that look like fitting faults but are control board or output problems, are covered under electrofusion machine troubleshooting. Distinguishing the two is the difference between replacing a fitting and replacing a machine.

What this looks like in a quality system

A shorted fitting is a recordable event, not a nuisance. The fault code, the coil resistance, the fitting batch, the ambient conditions and the operator belong in the weld log, because the pattern across a week of joints is what identifies a batch problem rather than an operator problem. A single smoked fitting is an incident. Three in a day from the same box is a batch.

That record is also what makes the cut-out defensible. A joint that has been cut out and replaced without a note is indistinguishable from a joint that was never made, and the next inspector has no way to know the run was disturbed. The wider defect catalogue, including the non-electrical failures that look similar from outside, is in HDPE pipe fitting defects and their causes. Where a contract names a joining standard, the reporting and acceptance provisions differ between the main documents, and those differences are compared in DVS 2207 and ISO 21307.

All of it, in the end, is an operator question. A machine reads resistance correctly every time; the reason a burnt fitting still reaches a pressure line is that someone skipped a step they had done correctly a hundred times before. On a butt fusion machine the same assumption is made about the pressure gauge, which is why the instrument side of that process is checked separately under butt fusion machine calibration. That is what welder qualification and operator training is meant to hold in place once the crew is under time pressure.

Frequently asked questions

Can I restart the welding cycle on the same fitting?

No. Once the coil has been heated, its geometry and its resistance have changed, and a second cycle runs at the wrong energy through a displaced wire. The joint is cut out and the fitting is replaced. This applies even when the first cycle was interrupted by a power failure rather than by a fault.

Why did the fitting smoke when the parameters were correct?

Because correct parameters only fix the energy going in. If the wire had no heat sink on one side because of a gap or an oval pipe, or if a turn was bridged by water, the same energy produces a much higher local temperature. The parameters were right and the joint was still wrong.

Does water really cause a short circuit?

Not pure water, but site water carries ions. A film across the terminals or along the bore gives the current a parallel path, and the machine sees a resistance below the window or a current rise mid-cycle. The water also turns to steam and displaces the coil, which is a mechanical fault caused by a chemical one.

What does the fault code tell me that the damage does not?

It tells you when the fault was detected. A resistance check failure happens before any heating, which points at moisture, a broken wire or a wrong fitting. A current fault during the cycle points at a short developing under heat. An interruption points at the supply or the cable. The damage pattern tells you where; the code tells you when.

Is smoke always a failure?

Not always. Some fittings vent melt or steam from an indicator hole by design, and a small extrusion there is normal. Smoke from the fitting body, from the terminals, or from the joint line is not, and neither is any smoke accompanied by an alarm. The distinction is in where it comes from, not in how much there is.

Can a burnt fitting be repaired with a new coil?

There is no such repair in the field. The coil is moulded into the fitting body and cannot be replaced without destroying it. Any procedure that involves cutting open a fitting, re-wiring it or filling the burnt area is not a repair, it is an untested joint with no rating.

Send Inquiry