The service call comes in August, and the voice on the phone is apologetic, as if the homeowner did something wrong. From the driveway the meter base on the side of the house looks fine. Up close, the gray PVC riser that used to disappear into the bottom of the enclosure now stops an inch and a half short of it. The conductors bridge the gap in open air, insulation faded by sun, and the end of the conduit sits there like a sleeve pulled back from a wrist. Nobody hit it. Nothing was loose. It was installed in April by someone careful, and it has been slowly tearing itself apart every afternoon since.

Here is the uncomfortable part: the installer did almost everything right. Straight run, proper straps, glued joints that will outlive the house. The one thing they got wrong was the belief that the job was finished when they drove away. The job was a snapshot. The conduit was going to keep living in a world that gets hot and cold, and nobody gave it anywhere to go.

Plastic moves, and it moves a lot

Everything expands when it warms. The question is how much, and for PVC the honest answer is more than your intuition allows. The NEC publishes the number in Table 352.44: the coefficient of thermal expansion for rigid PVC conduit is 0.0000338 inch per inch per degree Fahrenheit. That looks like nothing. Multiply it through a real run and it stops being nothing.

A 100-foot straight run is 1,200 inches of pipe. Warm it by 100°F and it grows 4.06 inches. The table lists the same thing in friendlier steps: about two inches per 100 feet for a 50-degree swing, about one inch for a 25-degree swing. Steel conduit, by comparison, has a coefficient roughly one-fifth as large, which is why electricians who learned on EMT and rigid carry an intuition that quietly betrays them on PVC.

A hundred-degree swing sounds extreme until you think about where conduit actually lives. The run along the south wall of a warehouse sees a January morning and a July afternoon. Gray pipe in direct sun runs well above the air temperature around it. A rooftop run across a black membrane sees worse. The relevant number is not today's forecast. It is the difference between the coldest the conduit will ever be and the hottest.

Where the inch and a half went

Glued PVC joints are rigid. Once the cement cures, a run between two boxes is one continuous piece of plastic, and plastic that cannot stretch in length has to put that growth somewhere. It has two choices.

If the straps let it slide, the run pushes against whatever it is terminated in. On a warm day it shoves on the box. On a cold night it pulls. The pull is the one that wins. A male adapter threaded into a meter base, or a cemented hub, holds against compression all day, but repeated tension on the glue line, thousands of cycles over a few seasons, is how a riser walks out of its enclosure. That is the August service call.

If the straps are tight, the run cannot slide, so it bows. You have seen this: a long horizontal PVC run that looks like a sine wave between straps, or the one that popped a strap off the wall and now hangs away from it. Buckling is the plastic telling you it needed to grow and you gave it no permission.

Both failures look like damage. Neither one is. They are the predictable outcome of a well-built rigid system meeting a temperature cycle it was never allowed to move with.

What NEC 352.44 actually requires

The code does not ask you to guess. Section 352.44 requires expansion fittings for PVC conduit to compensate for thermal expansion and contraction wherever the length change, figured from Table 352.44, is expected to be 1/4 inch or greater in a straight run between securely mounted items such as boxes, cabinets, elbows, or other conduit terminations.

A quarter inch is the threshold. Work backward from it and the result is sobering. At the table's rate, each foot of PVC grows about 0.02 inch across a 50-degree swing. A quarter inch arrives at roughly 12 feet. For a 100-degree swing, it arrives at about 6 feet. The fitting is not an accessory for long industrial runs. It is a requirement on a great deal of ordinary exterior work that gets built without one every day.

Notice the phrase about securely mounted items. An elbow counts. A run that turns a corner and heads off in a new direction is two straight runs for this purpose, and each one gets its own arithmetic. A long run with a 90 in the middle will not relieve itself through the bend. It will try to push the elbow off the wall.

There is a companion rule most people skip. Section 352.30 says PVC conduit shall be fastened so that movement from thermal expansion or contraction is permitted. The strap is supposed to hold the pipe to the wall, not clamp it in place. Two-hole straps cranked down until the pipe is dented are doing the opposite of what the code intends. The expansion fitting is where the length change is absorbed. The straps are how the pipe is allowed to reach it.

Installing one so it actually works

An expansion coupling is a piston inside a sleeve, with enough travel to absorb the computed change and a seal to keep the raceway closed. The thing that separates a working install from a decorative one is where you set the piston on the day you glue it.

If you install on a cold morning, the pipe is near its shortest. It is only going to grow, so the piston needs to be pushed nearly all the way in, leaving the travel available for expansion. Install on a hot afternoon and the opposite holds: the pipe is near its longest and will mostly contract, so the piston should sit near full extension. Manufacturers supply a chart that maps installation temperature to piston position. Use it. A fitting set wrong bottoms out halfway through the first winter and then behaves exactly like a glued coupling.

Mount the fitting so the conduit on at least one side can slide, and strap the run so the movement flows toward the fitting rather than toward a box. On a riser, that usually means the expansion fitting sits just below the enclosure, with the pipe beneath it free to move within its straps. Underground is a different problem: buried PVC lives at a nearly constant temperature, and 300.5(J) is concerned with settlement and frost rather than heat. The place thermal expansion bites a buried run is the riser, where it leaves the stable ground and climbs into the sun.

The snapshot problem

There is a well-documented quirk in how people think about time that explains why this failure is so common. Psychologists Jordi Quoidbach, Daniel Gilbert, and Timothy Wilson named it the end-of-history illusion: people readily acknowledge how much they have changed in the past, yet predict they will change very little going forward. We treat the present as the finished version.

An install day is the same trap. The run is straight, the straps are square, the joints are cured. It looks done, and done feels permanent. But the conduit is not at rest. It is at one point in a cycle it will repeat every day for decades. The electrician who builds for the moment is not careless. They are human. The electrician who builds for the cycle has simply learned to picture the same wall in January and July before gluing a single joint. It is a useful habit well beyond conduit: anything rigid that is given no room to move will eventually find room on its own, usually at the weakest connection.

Your next moves

  • Walk your current job and find the longest straight PVC run between terminations. Measure it in feet and multiply by 0.0406 to get the growth in inches for a 100°F swing, then scale to the realistic swing for that location. If the answer is 1/4 inch or more, it needs a fitting.
  • On any exterior or rooftop PVC run, assume a swing of at least 100°F unless you have a real reason to use less. Sun on gray pipe adds heat the forecast never shows.
  • Check the straps on an existing PVC run with your hand. If the pipe cannot shift a hair under pressure, back the straps off until it can, and cite 352.30 if anyone asks why.
  • Before gluing an expansion coupling, read the manufacturer's temperature chart and set the piston for today's conditions. Write the install temperature on the pipe in marker next to the fitting so the next person understands what they are looking at.
  • Treat every elbow and box as the end of a run. Do the arithmetic on each straight segment separately rather than on the whole route.

The number in your pocket

None of this requires an app. It requires a coefficient, a tape measure, and the habit of imagining the wall in a different season. But the coefficient lives in a table in Chapter 3, and the math has to happen on a ladder at two in the afternoon with a glue can open. Voltly keeps Table 352.44, the support spacing in 352.30, and the rest of the field calculations in your pocket with no signal required, so the question of whether this run needs a fitting takes ten seconds instead of a trip to the truck. If you spend your days building things that have to survive weather, it is worth a look at https://voltly.lumenlabs.works.