Cold Fusion And False Fusion: Why HDPE Joints Leak After Backfill

Nov 03, 2026

Leave a message

Failure Analysis
Cold Fusion And False Fusion: Why HDPE Joints Leak After Backfill

A joint that looks perfect and fails three weeks later is not bad luck. It is a joint that was never fused, only pressed together. The two names for it are cold fusion, where the heat was insufficient, and false fusion, where heat was present but the surfaces never diffused into each other.

Both leave the same evidence: an acceptable-looking bead on the outside and a visible interface line when the joint is cut through.

9
Documented causes
5 s
Changeover limit, dn160
1 in 100
Typical destructive test rate
Causes

The Nine Causes, In The Order We Find Them

# Cause What you see
1 Timer started at first contact, not at bead Bead too large, joint brittle
2 Heat soak shorter than required Small bead, unbonded line on cut face
3 Plate temperature low (worn element, display gap) Every joint from that machine is weak
4 Changeover too slow Bead fine, bend test snaps at interface
5 End face touched or contaminated after facing Localised voids, porosity
6 Pipe ends wet or damp Steam voids, honeycomb section
7 Drag pressure not measured, pressure under-applied Undersized bead, no collar roll-over
8 Misalignment or excessive offset Uneven bead, thin wall on one side
9 Mixed grades or manufacturers in one joint Passes test, creeps later

Number 4 is the one that ends in disputes. The bead can look immaculate after a twenty-second changeover and the joint will still pass a thirty-minute hydrostatic test. It fails under pressure cycling six months later, when the trench is closed and the road is open.

Diagnosis

Cutting The Joint Tells You Which One It Is

Cold fusion: a clean sharp line

The two surfaces never mixed. Bead height is usually correct, because the heat was there; the time at temperature was not. Look at the machine, the timer and the plate thermometer.

False fusion: a rough granular interface

Sometimes a thin white or dull band. Heat and pressure were both present, but contamination or premature cooling separated the surfaces. Look at handling, cleaning and weather.

The distinction changes the fix. Cold fusion is a machine and procedure problem; false fusion is a handling problem. Cutting one sample per crew per week costs less than a single cut-out on a live main. When the joint is already made and the question is whether to cut it out, the tests that settle it are set out in cold weld detection and repair.

Timing

The Changeover Window Is Smaller Than Crews Believe

Diameter Maximum changeover
Up to dn160 5 s
dn200 to dn250 6 s
dn315 to dn400 8 s
Above dn400 10 s, two operators

On dn315 SDR11 the exposed melt surface begins to skin over within seconds of the plate being withdrawn. Two operators, one on the carriage and one on the plate, is standard practice above dn315 for this reason alone. Parameter detail for each diameter is in our butt fusion parameter reference.

Preparation

The Four Tasks That Get Skipped

  • Facing is not cleaning. Facing removes the oxide layer, not oil. Wipe with alcohol and a lint-free cloth, then do not touch the face.
  • Ends must be dry. Water flashes to steam at 225 °C and creates voids in the fusion plane. In winter, condensation forms on cold pipe within seconds of facing.
  • Check offset before heating. Above 10 % of wall thickness the pressure distribution across the face is uneven: one side over-pressured, the other starved.
  • One grade, one maker, one batch where possible. PE100 from two suppliers can differ in melt flow. The joint tests fine and creeps under load.

Fittings suffer the same contamination problem in a different place: the socket bore. Butt fusion fittings and spigot ends are shipped protected, and should stay wrapped until the moment of fit-up.

Prevention

What Actually Works On Site

  • A written welding procedure on site with the calculated pressure already filled in, not left to the operator.
  • A contact thermometer on the plate at the start of every shift.
  • A stopwatch for heat soak and cooling, not a phone.
  • Bead photographs logged against joint numbers.
  • Destructive testing at the specified frequency. One in a hundred is common on gas work and it is the only way to catch a systematic fault before it is buried.

Acceptance criteria belong to the project specification, usually written around ISO 21307 or ASTM F2620. The equipment side of the same problem is covered in electrofusion output and resistance faults.

FAQ

Frequently Asked Questions

Can a cold fusion joint pass a pressure test?

Yes, routinely. A thirty-minute hydrostatic test proves the joint holds at that moment, not that it is fused. Cold joints fail under bending, pressure cycling or settlement months later.

Does a bigger bead mean a stronger joint?

No. An oversized bead usually indicates over-pressure or over-heating, both of which thin the effective wall at the fusion plane.

Should I re-weld or cut out a suspect joint?

Cut out. Re-heating a made joint does not restore the interface, and grinding back to re-weld leaves a wall discontinuity.

How many joints should be destructively tested?

The rate is set by the project specification; one in a hundred is typical on gas distribution work and one in five hundred on water mains.

Equipment That Removes The Variables

Contact thermometers, calibrated pressure gauges, sharp facers and machines with logged joint data: the four things that turn a parameter document into joints that hold.

Zhejiang Qingfa Pipe Technology Co., Ltd. · www.qingfaindustry.com · +86 151 6813 1515

Send Inquiry