The PVC Joint That Let Go in January

Why this matters

The diagnosis on this call was right, the fittings were right, and the plumber who made the joints has made thousands of them. The repair still came apart about eleven hours later and put water in a crawlspace overnight. That combination is what makes cold-weather solvent welding dangerous to a competent tech: nothing about the work looks or feels different in January, and the chart on the can that says otherwise is the one thing nobody reads on a job they could do blindfolded.

The first visit was a good repair

A homeowner on a private well called on a January morning after a cold snap: no water, and a hiss under the house. The plumber found a freeze-cracked PVC tee on the 1-inch line from the well entry to the pressure tank in the unheated crawlspace, and a cracked 2-inch PVC drain fitting a few feet away where the same cold had reached it.

Both diagnoses were correct, and both were confirmed by eye: a split running along the tee's body, and a hairline crack in the drain fitting that wept when the plumber ran water upstairs. The repair plan was correct too. Cut out both fittings, replace with new, primer and cement, restore the water.

He did it in about two and a half hours. He switched the pump off at its breaker, opened a faucet to drain the line, cut out the damage with a ratcheting cutter, deburred and beveled the pipe ends, dry fit the new fittings, primed and cemented. Then he waited about five minutes after the last supply joint, turned the pump breaker back on, watched the tank build to its 60 psi cutout, looked over every joint for drips, saw none, ran water to check the drain, and left at 3:00 p.m.

At about 2 a.m., one of the new supply joints separated. The pump ran continuously into the crawlspace until the homeowner woke at 5:30 to the sound of it and no pressure at the tap.

6:15 a.m.: water, wiring and a sealed space

The callback started with three things that had nothing to do with PVC.

• Power to the pump went off before anyone went under the house. The plumber had the homeowner switch off the pump breaker by phone, then confirmed at the panel on arrival that the breaker was off and the pump had stopped, because the pressure gauge at the tank was falling and the pump was silent. A crawlspace with standing water may hold a lit work lamp, heat tape on the pipes, a pressure switch and the pump wiring, and water joins all of them. Any circuit feeding the crawlspace that he could not identify was switched off at the panel too, and he tested the crawlspace light and the heat-tape outlet with a non-contact tester, proved on a live outlet upstairs before and after, before going in.
• The space was ventilated before it was worked in. He would be cutting and cementing again, in a low, enclosed space where solvent vapor is heavier than air and pools near the floor. He set a fan at the access hatch blowing air through the crawlspace and wore an organic vapor respirator for the cementing, and he kept the pump breaker off until the vapor had cleared, because a pressure switch makes a spark every time it closes.
• The water was dealt with as its own problem. Roughly three and a half hours of pump flow had soaked the vapor barrier. He told the homeowner on the spot that the crawlspace would need pumping and drying, and the office set that up that morning, since a wet crawlspace left to dry by itself in January stays wet for weeks.

What the joint looked like

The failed joint was the downstream side of the new tee. He cut it out with a few inches of pipe either side and split it lengthwise with a saw to see inside.

• The pipe had backed out of the socket by about three quarters of an inch. The witness mark of the primer on the pipe showed how deep the pipe had originally gone, and it fell short of the socket's bottom by a further eighth of an inch or so. The joint had never fully bottomed.
• The cement was still soft. More than fifteen hours after it was applied, a fingernail could dent it. It sat on the pipe in a thin, uneven layer with ropy streaks and bare patches, the way cold, thickened cement goes on.
• There was no fused zone. In a well-made solvent weld that has cured, a lengthwise cut shows pipe and fitting walls run together over much of the socket depth, with the boundary hard to find. Here the pipe slid away from the fitting cleanly; the two surfaces had softened and then stopped.

The other new joints, cut out and split as well since nobody was going to trust them, showed the same soft, partly bottomed cement in varying degrees. The drain joint had held only because a drain carries no pressure.

Why being right was not enough

Solvent cement for PVC, made to ASTM D2564, works by dissolving the surfaces of the pipe and the socket so the two can fuse as the solvents leave. Primer, made to ASTM F656, softens the surfaces first. Both are chemistry, and both slow down as temperature drops. Three things on this job pushed it outside what the can's instructions allow.

The cement had spent the night in the truck bed. The overnight low had been 18 F, and the can was still well below freezing when it was opened. Cold cement thickens. It does not spread into a full, even layer, and it resists the pipe being pushed into the socket, which is why the joints stopped short of the bottom. Cement that has turned jelly-like is meant to be discarded, not used.

The crawlspace was about 34 F. The plumber measured it on the callback with a thermometer at the joint: the space sits near the outdoor temperature because it is vented and unheated.

The cure time on the can was given by temperature band. The cure chart on that can, like most PVC cement charts, split the set and cure times by pipe size and by temperature, with its coldest band for work below 40 F. For 1-inch pipe at a well system's pressure, the cold-band cure was several times the five minutes he gave it, and the chart's notes said its times assume the cement and the pipe are within the temperature range on the label, with more time needed in cold and humid conditions. Five minutes is a habit formed on warm jobs in heated basements, where a short wait before a low-pressure check is usually forgiven, and it is the number a plumber carries in his head into a 34 F crawlspace.

Then the system loaded the joint in the way it was least able to take. The pump brought the line to 60 psi within minutes, and then cycled the line between its cut-in and cut-out pressures through the evening as the house used water. A joint that has set enough to hold a static test can still creep out of a tapered socket under that repeated push when its cement is soft. The 3:00 p.m. look for drips was honest and meaningless: an under-cured joint is watertight long before it is strong.

The thermal side made it slightly worse. PVC contracts as it cools, and the crawlspace fell a few more degrees overnight. On a short repair the movement is small, but it pulls in the same direction the pressure was already pushing.

The redo

The plumber rebuilt the supply repair the same morning, with fresh, sealed cans of cement and primer from the shop's heated storage, kept closed in the warm house away from any flame or heater for the hour it took to clear the water, replacing the ones that had sat in the bed all night.

• He confirmed the cement flowed freely off the dauber before every joint.
• He primed and cemented each joint, pushed the pipe fully to the bottom of the socket with a quarter turn, and held it there for about 30 seconds so the tapered socket could not push it back out.
• With the crawlspace still near 34 F, he did not trust the chart minimum. The shop's new rule, below, set the wait at 2 hours. The homeowner was told plainly why there would be no water until after lunch, and filled jugs from a neighbor in the meantime.
• After 2 hours, and after confirming with his nose and the fan that the vapor had cleared, he restored the pump breaker, let the pressure build slowly by cracking a faucet, held at cutout for 10 minutes watching every joint, and came back at the end of the day to look again.
• The crawlspace light and heat-tape circuits stayed off until the drying crew had pumped out the standing water; then he restored them, tested the heat-tape outlet's GFCI, and confirmed the tape was warming the line it protects.

The callback took about 4 hours on site, plus the drying contractor, against the 2.5 hours of the original repair.

The cold-weather joining procedure the shop adopted

The full method is taught in How to Solvent-Weld PVC in Cold Weather. What the shop put on the truck card after this job is short, and each line answers one thing that went wrong here:

• From November through March, solvent cement and primer spend the night in heated indoor storage, in the flammable cabinet, never in the truck, and ride to the job in a closed box (not in the cab, where any vapor is breathed by the driver). They are never warmed with a flame, a heat gun or a space heater, all of which can ignite the vapor.
• Measure the temperature where the joint is, with a thermometer, not the forecast. A crawlspace, a garage, an exterior wall cavity and a trench each have their own.
• Use the chart band you measured, and double it below 40 F, never less than 1 hour, unless a cement made and labeled for cold use states its own figure. Where the customer cannot be without water that long, the fork is a fitting the code as adopted locally allows for that pipe and use without solvent, not a shorter wait.
• Bottom every joint and hold it for 30 seconds. A joint that will not bottom is a sign of cold cement, a poor cut or a bad dry fit, not something to push past.
• Restore pressure slowly and come back. Watch at full pressure for 10 minutes, and either return before leaving the area or tell the customer what a failing joint sounds like and when to shut the pump off.

Most of the material-handling side is covered for every trade in The Cold-Weather Materials Handling SOP. The lesson that does not travel from a warm basement is the cure band: the joint does not care how many thousands the plumber has made.

References

• ASTM D2564, the standard specification for PVC solvent cements, and ASTM F656, for primers used with them; both bind through product labeling and the plumbing code as adopted by your jurisdiction.
• Cement and primer manufacturer's label directions and set-and-cure chart for the product in use.
• See related: PVC and Plastic Pipe When the Temperature Drops
• See related: How to Solvent-Weld PVC in Cold Weather
• See related: Materials Riding in the Truck Through Heat and Freeze