Twelve-hour-old copper joints leaking after quenching
A plumber is making up twelve 22 mm copper joints for a geyser installation supplied at 2 bar (municipal main plus a 0.5 HP pump). He cuts the tube with a hacksaw, wipes the burr away with emery cloth, dips a brush into a tin of zinc-chloride flux, heats the fitting with a butane torch until the copper glows a dull red, feeds 50/50 tin-lead solder at the cup edge, and quenches each finished joint in a bucket of water so he can handle it and move on. Two days later three joints are leaking from pinholes on the underside, the tube near those joints is blackened, and two other joints show green staining but no leak. A pressure test holds for about twenty minutes before the gauge begins to fall.
1. Identify the process faults in the account and link each one to the blackening, the pinholes or the green staining.
Read the worked answer
At least five faults are visible:
- Hacksaw cut with the burr merely wiped. A hacksaw leaves a square shoulder and swarf. The burr must be reamed out and the tube end cleaned to bright metal for 20–25 mm. Solder cannot fill a capillary gap contaminated with swarf, giving voids that open up as pinholes.
- Overheating to dull red. Copper should never be brought above a dull cherry at most, and a normal soft-soldered joint should never glow. At red heat the flux burns off, black copper oxide forms (the blackening), the capillary gap is destroyed by oxidation, and the tube is annealed and softened.
- Quenching. Dropping a hot joint into water chills the thin solder film unevenly; the solder contracts against the tube and can crack or leave micro-voids, and residual stress remains. Pinholes on the underside - the side that cools fastest and fills least - are the classic signature. Joints must cool in air.
- Contaminated flux or a dirty brush. A brush used across several jobs carries grit, old flux and moisture into the tin, so the flux fails to dissolve oxide at the moment it is needed; solder then sits on the surface rather than being drawn in.
- Flux left on the pipe. Zinc-chloride flux is aggressive. Residue left un-wiped absorbs condensation and attacks the copper, producing the green copper-chloride staining. Those two joints are not yet leaking but will perforate.
Water left in the line would add a sixth fault - steam blowing holes through the solder.
2. Describe the correct procedure from cutting to final test, including a defensible test pressure and duration.
Read the worked answer
Prepare. Cut square with a wheel-type tube cutter (rotating and tightening gradually so the tube is not ovalised) or a fine hacksaw, then ream out the burr and chamfer lightly.
Clean. Polish the tube end and the inside of the fitting to bright metal for 20–25 mm with clean emery cloth or stainless wool; do not touch the cleaned surfaces with fingers.
Flux. Apply a thin, even film of fresh approved flux to the tube only (not the fitting socket), using a dedicated clean brush. Avoid excess, which is what later causes corrosion.
Assemble and heat. Push the tube fully home and rotate a quarter turn to spread the flux. Heat evenly around the fitting with a soft, brushing flame, keeping the flame moving and off the tube itself; do not overheat.
Solder. Touch solder to the joint at the point opposite the flame. When the joint is at temperature the solder melts on contact and is drawn in by capillary action. Feed until a fine, bright ring shows at the cup edge - no more. Use lead-free solder for potable water.
Finish. Remove the flame, do not disturb the joint until the solder has set, wipe the flux residue off while the joint is still warm with a damp cloth, and let it cool naturally. Never quench. Flush the line before commissioning.
Test. With air expelled and the highest point vented, hydrostatically test at 1.5 times the maximum working pressure - about 3 bar for this 2 bar system - and hold for at least one hour with no drop, then inspect every joint dry and bright. The twenty-minute hold that failed here was too short and at too low a pressure to be meaningful.
3. The owner wants only the three leaking joints re-sweated. Give your decision with reasons, and say what you would do about the two stained joints.
Read the worked answer
Decision: cut out and remake, do not re-sweat in place. Reasons:
- A joint that is already leaking has a void filled with flux residue, oxide and moisture. Reheating drives that contamination into the gap and traps it; the repair may pass a short test and fail within weeks.
- Lying a new flame on an old, overheated, annealed joint adds more oxide and more stress; you cannot see whether the second fill is complete.
- The leak is a symptom of a method fault applied to all twelve joints, not to three. Every joint in this run was overheated and quenched, so every joint has the same latent defect.
What to do:
- Drain and dry the line. Cut out each leaking joint with about 100 mm of tube either side, and remake with new fittings, correct flux and no quenching. Re-sweating over a leak is acceptable only on a large, hard-to-reach joint as a temporary measure, and it must be recorded as such.
- Cut out the two green-stained joints as well. Green staining means active flux-driven corrosion on the inside as well as the outside; these are future pinholes and leaving them is knowingly handing the owner a second failure.
- Because all twelve were quenched, treat the whole run as suspect: remake any joint that shows blackening or a poor solder ring, then flush thoroughly and re-test at 1.5 times working pressure for an hour with the geyser isolated.
- Tell the owner plainly that the first test that failed after twenty minutes was under-specified, and that the correct hydrostatic test, not a quick look, is what will prove the repair.