PPR Pipe Failures: Diagnosing Fusion Joints That Leak

The call usually comes in the same shape. A riser that was signed off eight months ago is dripping into a ceiling void, the contractor says the pipe was faulty, the supplier says the joint was welded badly, and nobody can prove either because the evidence is buried in a wall. Meanwhile the building owner wants a number and a date.

PP-R fusion joints do not fail randomly. Six things go wrong, each leaves a different signature in the material, and each tends to surface at a different point in the system’s life. Once you can read those two facts together — what the cut section shows, and when the leak appeared — the argument about whose fault it is usually settles itself in about ten minutes.

The six ways a joint actually fails

Almost every article on this subject gives you the same four-item list: too hot, too cold, dirty, twisted. That list is not wrong, but it is not diagnostic — it does not tell you which one you are holding. Six distinct defect classes matter in practice, and they fail for different physical reasons.

• Cold joint. The interface never reached fusion temperature for long enough. Two pieces of plastic are touching, pressed together, and holding by friction and a partial bond.
• Overheated joint. Too much heat, or too long on the iron. Melt displaces where it should not, the bore deforms, and the material at the interface is degraded before the joint is even made.
• Insufficient insertion. The pipe never reached full socket depth, so the fused length is shorter than the fitting was designed around.
• Contaminated socket. Dust, oil, swarf or water sat on one of the two faces. The polymer could not interdiffuse across the contaminated patch.
• Wrong dwell at low ambient. The crew used summer timings in winter conditions. The result is a cold joint by a different route, which matters because the fix is different.
• Over-insertion with an intruding bead. The pipe was driven past the stop, and displaced melt has formed a ridge standing proud into the flow area.

What unites all six is worth stating plainly, because it changes how you think about the repair. A correctly made PP-R socket joint is homogeneous — the pipe wall and the fitting wall have melted into one continuous material, and there is no O-ring, no gasket and no seal anywhere in it. That is the structural advantage of heat fusion, and it is also why a failed fusion joint cannot be re-tightened, re-sealed or repaired. Either the two walls became one material, or they did not.

When each defect shows up: commissioning versus six months in

This is the axis nobody publishes, and it is the most useful evidence you have. Defects sort themselves into two populations by when they announce themselves, and the sorting is not arbitrary — it follows directly from how much load-bearing bond area the defect left behind.

The ones that fail immediately. A grossly under-inserted joint, a badly overheated joint with a collapsed bore, a joint that was rotated while the melt was still soft — these tend to leak during filling or during the commissioning test. There is simply not enough sound material to hold working pressure at all. These are the good failures: they cost a fitting and an hour, they happen while the pipework is still visible, and nobody argues about them.

The ones that pass the test and fail later. A cold joint or a contaminated socket often has enough partial bond and enough interference fit to hold static pressure for a couple of hours. It will pass a tightness test. What it cannot survive is time under thermal cycling — every hot draw-off expands the pipe, every cooldown contracts it, and the partial bond works loose at the weakest part of the interface. Six months of that, in a hot-water circuit, is a fundamentally harsher test than the one the joint passed at handover.

Read backwards, this gives you a genuinely useful rule. A leak in the first week points to gross execution error. A leak between three months and two years, in a joint that passed its test, points to a marginal fusion defect — cold joint or contamination — and it almost never appears alone, because whatever produced it was probably producing it all week.

What each defect looks like when you cut it open

If you have a failed joint in your hand, cut it longitudinally through the leak path. What you are looking for is simple: can you still see two materials, or one?

A cold joint shows an interface. There is a visible line where the pipe outside diameter met the socket bore, and the two surfaces often separate cleanly with hand pressure or a blade. The material either side is undamaged and correctly coloured — nothing was degraded, it was just never joined. This is the diagnostic opposite of an overheated joint, and confusing the two sends the corrective action in exactly the wrong direction.

An overheated joint shows too much melt in the wrong place. The bore is narrowed or ovalised where molten material was pushed inward, the bead is oversized and irregular rather than a neat uniform collar, and the interface may be discoloured. A joint that looks over-welded is not a strong joint — the polymer at the fusion face has been thermally damaged before it ever carried pressure.

Insufficient insertion is the one you can measure. Measure the fused length on the cut section and compare it against the socket depth the standard specifies for that diameter — the table in the next section gives the figure. If a 32 mm joint shows 10 mm of fusion where 16.5 mm was required, you are not making a judgement call, you are reading a measurement.

Contamination separates cleanly and leaves evidence. The failure surface is smooth rather than torn, and under decent light you can usually see what caused it — a film, a dust band, or the shadow of a fingerprint in oil. Torn, stringy, fibrous failure surfaces mean the opposite: the material was genuinely bonded and had to be ripped apart, which is what a sound joint looks like when you destroy it deliberately.

The parameters each failure comes from — and the one page one gets wrong

Every defect above traces back to a number somebody did not work to. DVS 2207-11 is the technical code governing heated-element welding of PP pipes and fittings, and it fixes insertion depth and timings per diameter. The table below is the general guideline as published in the Wefatherm PP-R specification manual, which reproduces the DVS 2207 Part 11 values.

Look at the spread and the defect mechanism becomes obvious. Heating time runs from 5 seconds at OD 20 to 60 seconds at OD 125 — a twelvefold range. A crew that has internalised “about ten seconds” is heating a 20 mm joint for double its time and a 90 mm joint for a quarter of it. That single habit produces overheated small-bore joints and cold large-bore joints on the same site, on the same day, which is exactly the confusing mixed pattern that makes people blame the material.

Three other parameters carry more weight than their obscurity suggests:

• Setpoint. The welding tools are heated to 260 ± 10 °C for PP socket fusion, per SIMONA’s heated-element welding guidance to DVS 2207-1, -11 and -15. Note that PP butt fusion runs at 210 ± 10 °C — a genuinely different number for the same material, and setting a socket iron to the butt figure is a reliable way to manufacture cold joints all day.
• Change-over time. The interval between pulling the parts off the heating element and pushing them together is a named phase of the process, and the guidance is that it must be kept as short as possible, because the plastified surfaces solidify. This parameter appears on no wall chart, is not in the table above, and is where a great many cold joints are actually created — the fitter who turns to find his pipe has lost the joint before he starts.
• Diameter limit for hand work. Manual socket welding is specified as suitable for d16–63 mm; from d75–125 mm the process is a machine operation. Hand-welding a 90 mm joint is not heroic, it is out of specification.

Two procedural details prevent the remaining defects. The pipe end should be chamfered on the outside at roughly 15°, to a width of 2 mm up to OD 50 and 3 mm above, so the leading edge does not plough contamination into the socket. And the parts must be pushed together without rotation — they may be aligned during the processing time, but never twisted relative to each other, because rotating a soft melt shears the bond that is trying to form.

On insertion, correct depth is reached when the depth mark is covered by the bead that forms. Pushing beyond that point is a named defect in the procedure itself, which warns against inserting the pipe too far into the fitting in order to prevent the internal diameter of the pipe being reduced.

Was it the weld, the handling or the pipe?

This is the question that actually needs answering, because it decides who pays. The evidence sorts into three buckets and the pattern of failure across the building tells you which bucket you are in far more reliably than any single joint does.

Points to fusion. Failures are at joints and only at joints, the pipe barrel is sound everywhere, and the failures cluster — one crew, one shift, one diameter, one floor, or one cold week in the programme. Clustering is the strongest single signal available, because a fusion defect is a habit, and habits repeat. If your failures map neatly onto the two weeks a particular subcontractor was on site, you have your answer.

Points to handling. Failures are away from joints, or at joints that show sound fusion on the cut section. Look for pipe that sat in the sun on site before installation, impact damage, pipework strained into position against its own stiffness, missing or over-tight supports, and joints loaded before the cooling time expired. A joint moved during its cooling window is a handling defect, not a fusion defect, even though it fails at the joint.

Points to material. Failures appear in the pipe barrel rather than at joints, they cross batches and crews, or the pipe fails a dimensional check. This is the rarest of the three and the easiest to test, because it is documentary.

Any competent PP-R manufacturer verifies resin before extrusion, checks outside diameter, wall thickness at multiple points and ovality against the S-series tables, and pressure-tests pipe, fittings and fused joint assemblies on hydrostatic benches against DIN and EN ISO thresholds — the practice at the Hitze plant, and the standard expectation you should hold any supplier to. If a batch genuinely failed, that evidence chain exists and can be produced. Ask for it by batch number. The relevant product standard family for PP-R hot and cold water systems is EN ISO 15874 and its service classes, which is also where a “the pipe was not rated for this” argument has to be settled.

Before anyone repairs anything, document it. Photograph the joint in situ before removal, cut it longitudinally rather than crushing it out, keep the failed joint labelled with location and date, and record the pipe’s printed markings — the PP-R pipe and fitting print line carries the size, class and standard reference that a supplier will ask for first. A failed joint in a bin is worth nothing in a dispute.

Corrective action: what to do with the joints you have

Start from the constraint: a fused PP-R joint cannot be repaired. There is no seal to replace and no thread to tighten, so a suspect joint is cut out and remade with a new fitting. That is not a counsel of perfection, it is the only available option, and it is why the scoping decision below matters more than the repair itself.

Scope the population, not the leak. The joint that leaked is a sample, not the problem. If the cut section shows a cold joint and the failure sits inside a cluster, the sensible assumption is that every joint made by that crew, at that diameter, in that period shares the defect and simply has not failed yet. Cut and inspect a small number of additional joints from the same population before deciding how far to go. Replacing one joint in a run of forty identical ones buys you a month.

Then re-test to a named procedure. A commissioning test only means something if the procedure, the pressure and the duration are stated. The three-stage tightness test published for PP-R drinking-water systems runs like this: an initial test at 1.5 times the highest operating pressure, restored twice at 10-minute intervals within 30 minutes, after which the pressure must not fall by more than 0.6 bar over the following 30 minutes; a main test beginning immediately and lasting two hours, during which the pressure must not fall by more than 0.2 bar relative to the end of the initial test; and a final test alternating 10 bar and 1 bar at intervals of at least five minutes with no leakage anywhere.

EN 806-4 takes a different route to the same question. A published installer guide to its Section 6 runs the strength test for drinking-water installations at a minimum of 11 bar held for 30 minutes, with no drop over that period, after an impermeability test and a visual check of every connection. Two details of the method matter more than the number: the gauge must resolve to 0.1 bar, and where ambient and water temperatures differ by more than 10 K the system needs a temperature-equalization period before the clock starts — skip it and the pressure decay you record is thermal contraction, not a leak.

Work to whichever procedure your project specification names, and record which one you used. Watch for one substitution in particular: the short ten-minute holds quoted in some guides belong to the compressed-air procedure at a few bar, not to the water test, and the two are not interchangeable as evidence.

Two details decide whether that test is meaningful. The test medium and the pipework must be at the same temperature before you start, because polypropylene expands under heat and pressure and a warming system will disguise a genuine pressure drop. And the gauge must resolve to 0.1 bar — a 0.2 bar acceptance criterion read off a gauge marked in bar is not a measurement. Detailed test sequencing belongs to the commissioning procedure rather than this page; the point here is only that a test run casually is what let the defect reach the wall in the first place.

Prevent the recurrence at the cause, not the symptom. Rank the fixes by what actually produced the defect. Mark insertion depth on every pipe before heating. Post the per-diameter timings at the workface rather than trusting memory across a size range. Check the iron’s actual face temperature rather than trusting its lamp. Move to machine fusion above 63 mm. Add the 50% heating-time increase to the method statement for any work below +5 °C. And clean both faces immediately before heating, not at the start of the shift.

If the diagnosis lands on the material side, or if repeated failures make you doubt the supply rather than the crew, that is a sourcing question rather than a site one — the specification, certification and batch-documentation side is covered on the distributor and importer supply pages, and the practical comparison of joint methods by diameter sits in the fusion welding methods guide. The procedure itself — the correct sequence rather than its failure modes — is covered in how to weld PP-R pipe, and the material’s broader behaviour, including why a fused joint is an advantage in the first place, is set out in PP-R advantages and disadvantages.

What to do next

If you are holding a failure right now, work in this order. Cut the failed joint longitudinally and decide the single question — one material or two. Check the fused length against the insertion depth for that diameter in the table above. Map every failure you know about onto crew, date and diameter, and look for the cluster. Cut two or three sound-looking joints from the same population to find out whether the defect is systemic. Then scope the remediation to the population you found, not to the leak you were called about, and re-test to a named procedure with a gauge that reads in tenths of a bar.

On specification and supply, what you can reasonably establish before ordering: PP-R is available in an OD 20–110 mm range for the Hitze line, standards scope for the range is documented against DIN 8077/8078 and EN ISO 15874 with SKZ testing and DVGW type examination for drinking-water pipe — ask any supplier for the certificate reference and the models each one covers rather than accepting a logo.

Hitze operates with no minimum order quantity and supports samples, which makes verifying a batch before committing to it straightforward; lead time is not published as a single figure because it is set per project against your size mix and order volume, so ask for it in writing with the quote; the same applies to sample timing and to what a quotation includes and excludes. If you are specifying rather than repairing, the specification and submittal checklist is the shorter route.

Frequently Asked Questions