A PP-R pipe pressure test goes wrong in a particular, repeatable way. The riser is fused, the line is filled, the pump takes it up to pressure, and half an hour later the gauge has slipped. Somebody says leak. The screed gets held, a fitter walks the run with a torch, and on a bad day a sound joint gets cut open to prove a point. Nothing was leaking. The pipe was doing what polypropylene does under pressure.
PP-R is visco-elastic. Put it under test pressure and it creeps — the bore dilates slightly, the water inside occupies more room, and with no more water going in, the gauge falls. The UK Water Industry’s own guidance is blunt about the scale of this on polyolefins: creep stress relaxation effects “dominate the pressure decay until some 24 hr. – 36 hr. after reaching STP” on PE mains, which is why a test built around a flat gauge reading tells you about the pipe wall rather than the joints. The procedures below are the published ones that account for it, with the figures attributed to the documents that print them.
Key Takeaways
- EN 806 sets the test pressure at 1.1 × maximum design pressure — where MDP is itself 1.5 × maximum operating pressure, so 1.65 × operating. A 5 bar system tests at 8.25 bar, not 30.
- The widely republished “1.5 × working pressure, 30 bar on PN20” figure is not the EN 806-4 test-pressure basis. Check what your specification actually cites before pumping.
- A falling gauge on a plastic line is not by itself a leak. Creep makes pressure decline without any water escaping.
- EN 806-4 Procedure A drops the pressure to one third deliberately, then holds 90 minutes: the pass is that it falls no further.
- EN 806-4 Procedure B allows a 0.6 bar drop over 30 minutes, then a further 0.2 bar over 120 minutes.
- On the visco-elastic three-phase method a rising curve after a deliberate 10–15% pressure cut is the pass signal.
- Gauge resolution of 0.1 bar, connected at the lowest point, and temperature equalisation whenever the difference exceeds 10 K.
Why a PP-R Line Fails a Test It Should Have Passed
Steel and copper behave the way a pressure test assumes. Pressurise them, close the valve, and the gauge holds — the pipe is stiff enough that its dimensional response to internal pressure is negligible over a test window. Any measurable decay means water is going somewhere it should not. That assumption is baked into how most site crews were trained, and it is the wrong assumption for polypropylene.
What the pipe is doing while you watch the gauge
The Plastics Pipe Institute describes the mechanism for polyethylene in terms that apply to the whole polyolefin family: the material “dilates in diameter (creep-strains) when subjected to higher stress during hydrotest,” so “for a fixed volume of clean fill water, the hydrostatic pressure will decline slightly during the test time, as the polyethylene molecular chains stretch and align under high stress.” Their conclusion is the sentence every site engineer testing PP-R should have in mind — “This pressure decline does not mean the polyethylene is leaking. It is a visco-elastic material parameter that requires adjustments to the hydrostatic test procedure as compared to rigid elastic metallic pipes.”
That source is written about PE. PP-R is a different polymer with a different modulus and it is not interchangeable with PE in any design calculation — but it shares the visco-elastic mechanism, which is precisely why the standards that deal with plastic pipework treat PE and PP together. UK Water Industry guidance IGN 4-01-03 notes that BS EN 805 singles out “PE and PP pipes as being special cases of materials with viscoelastic properties and thereby in need of different assessment,” because the pressure “will still decay in a main without any leakage because of stress relaxation.”
How long the effect lasts
On buried PE and PVC mains, the same guidance puts creep in charge of the reading for a long time: the effects “dominate the pressure decay until some 24 hr. – 36 hr. after reaching STP,” and a result acceptable with confidence arrives only “after at least 10 hr. of pressure decay for PVC and some 36 hr. for PE pipes.”
Those hours describe large-diameter buried mains, not a PP-R riser in an apartment block, and nobody should transplant them to a building test. The transferable point is the shape of the problem: for the first part of any hold, a plastic line’s gauge is reporting creep and leakage mixed together, and no single decay number separates them.
Every method below exists to separate creep from leakage. That is the only problem they are solving.
This is also why joints get blamed unfairly. Genuine fusion defects have their own signatures — cold joints, collapsed bores, over-insertion — and they are worth knowing on sight, which is the subject of our guide to PP-R fusion joint failures and how to diagnose them. But a gauge that drifts down during a hold is not evidence of any of them.
What Test Pressure PP-R Actually Requires
Search this topic and the same number comes back from several pipe-factory blogs: test at 1.5 × the maximum working pressure, which they render as 30 bar for PN20 pipe and 15 bar for PN10. It is worth being precise about what that figure is, because it is not the EN 806 test-pressure basis and the gap is not small.
The basis EN 806 actually sets
Northumbrian Water’s Guidance Note 3 on pressure testing new water service pipework, revision 4 dated 02/10/2019, states the requirement directly: to satisfy the recommendations in BS EN 806, “fittings and pipework must be pressure tested at 1.1 x (1.5 x maximum design pressure) or 1.65 x maximum operating pressure.” The multiplier is applied to the pressure the system is designed and operated at — not to the pipe’s PN class.
Their worked example is the clearest way to hold it. A system with a maximum working pressure of 5 bar carries a statutory minimum test of 1.5 × 5 = 7.5 bar under the Water Supply (Water Fittings) Regulations 1999 and the Scottish Water Byelaws, Schedule 2 Paragraph 5. Under the BS EN 806 recommendation it is 1.1 × [1.5 × 5 bar] = 8.25 bar. Water Regs UK reaches the same 8.25 bar from the same basis independently.
| Basis | Applied to | 5 bar system | Source |
|---|---|---|---|
| 1.1 × MDP (= 1.65 × operating) | System design pressure | 8.25 bar | BS EN 806, per Northumbrian Water GN3 rev4 |
| 1.5 × max working pressure | System working pressure | 7.5 bar | UK statutory minimum, Water Fittings Regs 1999 Sch.2 Para.5 |
| 1.5 × PN class of the pipe | Pipe pressure class, not the system | 30 bar on PN20 pipe | Widely republished on supplier blogs; not the EN 806 basis |
The third row is where jobs go wrong. Testing a domestic riser at 30 bar because the pipe is stamped PN20 loads every fitting, valve and appliance connection on the line to a pressure the system was never designed for, and the components that fail first are usually not the PP-R.
Where your design pressure comes from
You need the system’s design pressure before any of this resolves to a number, and the pipe’s class is an input to that, not a substitute for it. Our published PP-R pipe and fitting range lists the self-cleaning ceramic-core and fiber-glass reinforced PPR-FB-PPR constructions at 10 bar (class 1) and 8 bar (class 2), and the PPR-AL-PPR aluminium composite at 6 bar (class 1) and 4 bar (class 2), across OD 20–110 mm.
Those classes are tied to service conditions and temperature, which is the subject of EN ISO 15874 service classes — and because PP-R derates with temperature, the design pressure of a hot line is not its cold rating. If you are working that back from first principles, our explainer on how plastic pipe pressure ratings are derived covers the SDR and safety-factor arithmetic.
One honest caveat on our side: Hitze does not publish an approved field test pressure, hold time or acceptance threshold for its PP-R systems, and this article does not invent one. Every procedural figure here belongs to the standard or guidance document credited beside it. Where a project specification names a different figure, the specification governs.
The Three Published Procedures, Side by Side
There is no single world procedure for plastic pipework, and pretending otherwise is how a contractor ends up running a European test on a job that is specified to a North American one. Three method families are in general use, and each solves the creep problem differently.
| Method | Sequence | Pass criterion |
|---|---|---|
| EN 806-4 Procedure A (contracted / expansion) | Pump to test pressure, maintain by pumping 30 min; reduce carefully to one third of test pressure; hold 90 min | Pressure does not drop further over the 90 min, no visible leakage |
| EN 806-4 Procedure B (two-window decay) | Pump to test pressure, maintain by pumping 30 min; then no further pumping | Falls no more than 0.6 bar over the next 30 min, then no more than a further 0.2 bar over the following 120 min |
| Three-phase visco-elastic method | Relax 60 min; pressurise in under 10 min; hold 30 min; allow 60 min creep; cut pressure 10–15%; observe 30 min | Pressure curve rises after the cut; drop over the creep window must not exceed 30% |
| ASTM F2164 (rebound, North America) | Up to 1.5 × working pressure at the lowest point; 4 h hold with make-up water; reduce 10 psi; monitor 1 h | Pressure stays within 5% of the reduced pressure |
Procedures A and B are quoted here as Northumbrian Water reproduces them for plastic systems, captioned in their guidance as “In accordance with Test Procedure A. BS EN806-4:2010” and “In accordance with Test Procedure B. BS EN806-4:2010” respectively. The three-phase method is published by Essex & Suffolk Water under the heading “Test method for pipework with visco-elastic behaviour (plastic pipework)”. The ASTM F2164 summary is the Plastics Pipe Institute’s, which notes it is the prevailing US and Canadian field method and is referenced in AWWA M55 (2020).
Choosing between them
- Follow the specification first. If the contract names a procedure, that is the one you owe, and a “better” test run instead of the specified one is still a non-conformance.
- Procedure B is the easiest to witness — it needs only a gauge and a clock, and its two drop allowances are unambiguous numbers a client’s engineer can check.
- Procedure A and the three-phase method suit long or large-bore runs, where creep is large enough that a decay-limit test throws false fails.
- The three-phase method is the most diagnostic: a rising curve is positive evidence of a sound line, not merely the absence of bad evidence.
A note on figures you will meet elsewhere. The “minimum 11 bar for 30 minutes” strength test often quoted against DIN EN 806-4 comes from Fränkische’s manual for its alpex multilayer system with crimp and press metal fittings, where the same document gives 4 to max. 6 bar over 60 minutes for heating systems to DIN 18380. Those are that manufacturer’s values for that system. They are not a PP-R requirement, and a system-specific figure has to come from the pipe maker or the specification, not from a neighbouring product’s manual.
Running the Contracted Test Step by Step
This is the three-phase visco-elastic sequence as Essex & Suffolk Water publish it, with the decision rule stated at each step. It is worth running in this order even where a simpler procedure is specified, because the intermediate readings tell you where a problem is rather than only that one exists.
Phase 1 — prepare, pressurise, and let it creep
- Test while the pipework is still exposed. EN 806-4 Section 6 calls for testing after installation is complete but before it is covered — once screed or plaster is on, a failure costs demolition instead of a re-fuse.
- Fill with filtered water, flush and vent to atmosphere. Trapped air is compressible and will make every subsequent reading meaningless.
- Let the pipe relax for 60 minutes before you pressurise, and make sure no air re-enters during that window.
- Pressurise to the system test pressure in under 10 minutes. A slow ramp lets creep begin before the clock starts and blurs the phases together.
- Maintain the test pressure for 30 minutes, pumping as required to hold it.
- Then stop pumping and leave it 60 minutes to allow the pipe to creep.
- Decision point: measure. If the pressure has dropped by more than 30%, phase 1 must be repeated — that is the published trigger, and it is a repeat rather than an immediate fail.
Phase 2 — the deliberate cut
Rapidly reduce the pressure by 10–15% of the system test pressure by drawing water off, and measure the volume you removed. It should not exceed the allowable water loss, which on this method is a figure the designer specifies for the line — it depends on pipe volume, so there is no universal number to quote and anyone printing one has stopped being careful. This step is counter-intuitive to a crew trained on copper: you are deliberately taking pressure out of a line you are trying to prove.
Phase 3 — read the curve, not the number
The rapid reduction causes the pipe to contract. Observe the pressure for 30 minutes. The curve should be pointing upwards — the stretched pipe wall recovering is pushing pressure back up, and only a line with no leak path can do that. In case of doubt, continue observing for up to 90 minutes. An upward curve is a pass of the entire procedure.
A worked example makes the shape concrete. Take a small commercial hot and cold installation operating at a maximum 5 bar, so the EN 806 test pressure is 8.25 bar. Fill, vent, and leave it an hour. Pump to 8.25 bar within ten minutes, hold it there with the pump for 30 minutes, then stop and wait an hour — suppose it settles to 7.4 bar, a drop of about 10%, comfortably inside the 30% repeat trigger, and entirely creep.
Now draw off water to cut roughly 1 bar (about 12%), checking the drawn volume against the designer’s allowance, and watch. If over the next half hour the gauge climbs off that reduced figure, the line is tight and you can sign it off. If it sits flat or keeps sinking, you have a leak path to find — and now you know it is a leak, because a sound PP-R line recovers.
Where the curve stays flat, the fault is more often at a transition than in the fused work — a brass-insert threaded joint, a valve, or a connection to another material. Fusion defects are a real category with real signatures, but they show up differently, and the sequence for making those joints correctly in the first place is covered in our guide to installing PP-R pipe and fittings.
Temperature, Gauges and the Errors That Fail Good Pipework
Creep is the effect people miss, but it is not the only thing that moves a gauge. Four error sources account for most disputed results, and all four are controllable before the pump comes out.
Temperature
A closed line full of water responds to temperature, and on a site where cold mains water meets a warm plant room the difference is easily enough to swamp a pass criterion. EN 806-4 practice requires temperature equalisation whenever the difference between the installation and the test medium exceeds 10 K, with a typical waiting period of 30 minutes after filling before the test begins.
Supplier technical content commonly puts the sensitivity at roughly 0.5 to 1.0 bar per 10 °C — treat that as an unverified estimate rather than a design figure, but note what it implies. At that order of magnitude, an unequalised line can move further on temperature alone than Procedure B’s entire 0.2 bar second-window allowance.
The gauge and where it sits
- Resolution: the gauge must indicate a pressure difference of 0.1 bar. A gauge that reads in whole bar cannot resolve a 0.2 bar acceptance limit at all.
- Position: connect at the lowest point of the system under test, so static head is included rather than discovered later.
- Recording: where the result has to be witnessed or approved remotely, a data logger removes the argument about what the needle did while nobody was watching.
Air, and what else is on the line
Air is compressible and water is not; a pocket of it turns a hydrostatic test into a spring and makes both the drop figures and the drawn-off volume unreadable. Flush and vent completely, and re-vent after any repressurisation. Equally, appliances, pressure vessels and water heaters must be disconnected, and every outlet capped with metal plugs, caps, blanking plates or blind flanges rather than left to a service valve of unknown provenance — a weeping test cap has failed more systems on paper than the pipework it was capping.
Two points apply whatever the method. A hydrostatic test stores energy, so restrain the test section against movement. And VDI 6023 practice is to commission drinking water systems immediately after testing and flushing — a line left full and static for weeks is a hygiene problem, which is when a compressed-air test becomes the better choice.
Factory Type Testing Is Not Your Field Test
A recurring conversation on site: the pipe is certified to EN ISO 15874, so why is a pressure test needed at all? The two things measure different objects, and a specifier who can articulate the difference wins the argument quickly.
What ISO 15874-5 covers
ISO 15874-5:2013 — “Plastics piping systems for hot and cold water installations — Polypropylene (PP) — Part 5: Fitness for purpose of the system”, second edition dated 2013-02-15 with a corrected version of 2014-01-15 — specifies system-level tests on assembled representative samples: an internal pressure test, a bending test, a pull-out test, thermal cycling, pressure cycling, and leak tightness under vacuum. It sits alongside Part 1 (General), Part 2 (Pipes), Part 3 (Fittings) and Part 7 (conformity assessment guidance).
Those are type tests. They demonstrate that a pipe-and-fitting system, made and joined correctly, is fit for its intended service classes — thermal cycling and pressure cycling over a laboratory programme, not an afternoon on a gauge. What they cannot speak to is the several hundred joints a particular crew made on a particular building, with a particular welding machine, in whatever the ambient conditions were that week.
What to ask a supplier for
Material credentials and installation quality are separate lines of evidence, and a buyer should collect both. On the material side, Hitze holds SKZ testing on PP-R pressure pipe and fittings and a DVGW type examination for PP-R drinking-water pipe, and runs in-house hydrostatic testing against DIN and EN ISO thresholds — our quality control process sets out that sequence. Certificate references are quoted per enquiry against the covered models, so ask for the scope covering the sizes you are buying and verify it with the issuing body.
On the installation side, nothing a supplier can send you substitutes for your own witnessed test. That is the honest division: the certificate covers what left the factory, the record you write on handover covers what was built.
The Handover Record: What to Write Down
A test nobody recorded properly is a test you will run again. The value of the record is not administrative — when a leak appears in year three, the question is whether the system was proven sound at handover, and a record that names its acceptance criterion answers it while a signed sheet saying “pressure test passed” does not.
| Field | Why it has to be on the sheet |
|---|---|
| Procedure named and dated edition | “Passed” is meaningless without the criterion it passed against |
| System design and operating pressure | Lets anyone re-derive the test pressure and check it |
| Test pressure applied | Proves the multiplier used, and that the line was not over-pressured |
| Water and ambient temperature, and the difference | Shows whether the 10 K equalisation rule was satisfied |
| Readings at each phase boundary, with times | The curve, not the endpoint, is the evidence on a plastic line |
| Gauge resolution and calibration reference | A 0.2 bar limit is unprovable on a coarse gauge |
| Section tested: whole system or part, with a sketch | Sectional testing is normal; an untested section must be visible as such |
| What was isolated, capped or disconnected | Records what the test did not cover |
| Both signatures, dated | Contractor and client representative — a countersigned sheet is the defensible one |
Two practical habits are worth adopting alongside the sheet. Photograph the gauge at each phase boundary with the time visible, because a photograph settles a dispute that a transcribed number reopens. And where the system is tested in sections as floors are completed, keep the sheets together with the sketch, so the handover pack shows the whole installation was covered rather than leaving a reader to work out which riser was tested when.
If you are procuring for a project rather than running one, this is the point to confirm what the specification demands, because acceptance criteria and witnessing arrangements are set per project and per market — not by the pipe supplier. On our side there is no minimum order quantity on standard PP-R ordering and samples are available before you commit; lead time, and any print minimum on a private-label run, are set against the size mix and destination and are confirmed in the quotation rather than published as one figure.
Conclusion
The two mistakes that cost the most on a PP-R pressure test are related: taking the test pressure from the pipe’s PN class instead of the system’s design pressure, and reading a creeping gauge as a leaking one. Get the pressure basis right and the number usually comes down; use a procedure written for visco-elastic pipe and the reading starts telling you something true. Both are decisions made before the pump is connected.
If you are specifying or buying PP-R for a project, work back from your system’s design pressure and temperature to the class and construction you need, then confirm the acceptance criteria your specification demands before commissioning is booked.
Frequently Asked Questions
What pressure should I test a PP-R system at?
BS EN 806 sets it at 1.1 × maximum design pressure, which works out at 1.65 × maximum operating pressure — 8.25 bar for a system operating at 5 bar. Apply it to the system’s design pressure, never to the pipe’s PN class.
How long does a PP-R pressure test take?
EN 806-4 Procedure B runs 30 minutes pumped plus a 30-minute and a 120-minute observation window. The three-phase visco-elastic method takes longer once its 60-minute relaxation and 60-minute creep allowances are included.
Is a pressure drop during the test always a leak?
No. PP-R creeps under pressure, so the bore dilates and the gauge falls with no water escaping. This is why the plastic-pipe procedures allow specific drops or look for a rising curve after a deliberate pressure cut.
Can I pressure test PP-R with compressed air?
Air testing is used where freezing or long standing times make water a hygiene or damage risk, but gas is compressible and stores far more energy, so it follows different safety rules and much lower pressures. Follow the procedure your specification names.
When should the test be done — before or after screeding?
Before. EN 806-4 Section 6 calls for testing after installation is complete but while the pipework is still exposed, so a failure costs a re-fuse rather than breaking out finished floors or walls.
Does an EN ISO 15874 certificate mean I can skip the field test?
No. ISO 15874-5 covers system type tests on representative samples in a laboratory. It says nothing about the joints your crew made on your building, which only a witnessed field test can demonstrate.
