Contractor Job Estimator
$29.00
BuyMetal Roof Thermal Expansion Calculator
How far the panel moves, how much of it lands on one clip, and whether the travel you have is enough
A metal roof panel is a long thin bar of metal bolted to a building that does not move. Heat it and it grows; cool it and it shrinks. Do that a few hundred thousand times over a roof's life and the only question that matters is where the movement went — into a clip that was designed to slide, or into a fastener, a seam, a trim lap, or a bow in the panel face.
This works out the movement from the panel material, the run length, and the temperature range the panel itself sees, then splits it the way the roof actually splits it: expansion on one side of the installation temperature, contraction on the other. It gives you the travel a single clip has to have, the longest run that travel supports, and how many expansion joints a longer roof needs.
Any trade
Checkout opens here — you stay on this page.
The details that let the movement happen
Every number above is only worth something if the roof is built so the panel can actually move. That means the fixed point is genuinely fixed, the sliding end is genuinely free, laps are sealed with something that stretches rather than something that cracks, and the layout was snapped before the first panel went down rather than chased afterwards.
Metal Roofing Screws, Self Tapping Hex Washer Head
- Exposed-fastener panels have no clip, so the movement lands on these
- Bonded washers have to seal while the hole slots slightly over the years
- The force figure on this page is why screw-down runs get length limits
Malco S2R 3" Hand Seamer
- Hand seaming at the eave cleat, closures and trim returns
- The fixed end only works if it is genuinely locked, not just screwed
- Also the tool for dressing a panel end that got worked at the free end
Cofair Quick Roof Butyl Seam Tape
- Butyl stays flexible instead of going brittle and cracking at a lap
- End laps and expansion details are exactly where movement concentrates
- Non-curing, so it keeps sealing through tens of thousands of cycles
Malco M2006 Offset Aviation Snips
- Offset handles keep your knuckles off the panel on long cuts
- Expansion details mean more cut trim, not less
- Left-cut pairs with a right-cut for working both directions of a run
TAJIMA Chalk‑Rite CR301JF Jam Free Chalk Line
- Snap the fixed-point line before the first panel, not after the tenth
- A scribe mark at a clip is how you prove the roof is moving as designed
- Bold 1 mm line holds up on a bright metal deck
As an Amazon Associate, TestTalkHQ earns from qualifying purchases. Prices and availability can change.
Why a metal roof needs a plan for moving
Nothing on a roof is stationary. Shingles move too, but asphalt is a mat of discrete pieces with gaps between them, and each tab only has to absorb its own small share. A metal panel is one continuous piece from eave to ridge. Whatever the whole run grows by has to end up somewhere.
The Metal Construction Association is blunt about the size of the swing, and it is not the weather forecast:
“These changes in panel temperatures may range more than 200°F in a season, 150°F in a single day, and 100°F within minutes as cloud cover inhibits direct sunlight and a cool rain begins. To varying lesser severity, thermal cycling occurs tens of thousands of times in a single year, and hundreds of thousands over the life of the roof.” — MCA, Standing Seam Best Practices: Clips
That last sentence is the one that actually matters. The movement itself is small — a fraction of an inch on most roofs. It is the number of repetitions that destroys things. A fastener that moves four thousandths of an inch every afternoon is a fatigue test running for thirty years.
The standing seam roof exists because of this. Panels are held by concealed clips instead of face screws precisely so that the movement has a designed place to go, and the whole business of fixed clips, floating clips, pinned ends and expansion joints is the vocabulary for deciding where that place is.
The arithmetic, and where it comes from
There is only one equation underneath all of this, and it is the standard linear thermal expansion relation:
ΔL = α × ΔT × L
α = coefficient of linear thermal expansion, in/in per °F
ΔT = temperature change, °F · L = original length, inches
α = coefficient of linear thermal expansion, in/in per °F
ΔT = temperature change, °F · L = original length, inches
Building Research Systems states it in the form metal roofers actually use, with the run in feet and the answer in inches. Their worked Example 1 is an 80 ft steel rod going from 20 °F to 110 °F:
ΔL = (0.00065 / 100 °F) × (110 − 20) × (80 × 12)
ΔL = 0.5616 in
ΔL = 0.5616 in
That 0.00065 per 100 °F is just 6.5 × 10−6 per °F written the way a shop chart writes it. Multiply it by 12 and you get the number BRS carries into their roof examples — 0.000078 inches of movement per foot of run, per degree F, for steel. Every figure this calculator produces is that rate, your temperature difference, and your run length.
Coefficients this page uses
| Panel material | α (×10−6 in/in/°F) | Movement per 100 ft per 100 °F | E (psi) | Source |
|---|---|---|---|---|
| Steel (galvanized / Galvalume) | 6.5 | 0.78 in | 29,000,000 | BRS / MBMA practice; AISC 360 for E |
| Aluminum | 12.3 | 1.48 in | 10,000,000 | CDA Table 14 |
| Copper C11000 / C12200 | 9.4 | 1.13 in | 17,000,000 | CDA Table 14 and Table 1 |
| Stainless, Type 304 | 9.6 | 1.15 in | 28,000,000 | CDA Table 14 |
CDA figures are the Copper Development Association's expansion/contraction table for architectural metals, reproduced as Table 14 in the NOMMA/NEF Architectural Applications reference. The third column is CDA's own check figure scaled up — their table gives copper as 0.113 in per 10 ft per 100 °F, which is the same number. Aluminum sheet is usually taken at 12.8 × 10−6 in some references and 12.3 in others; the difference is about 4% on the answer, and the custom field is there if your panel supplier publishes a specific value.
One honest disagreement, and why steel is listed at 6.5. CDA's table gives carbon steel as 7.3 × 10−6, not 6.5. The metal building and metal roofing industry — BRS, MBMA, and the worked examples every panel manufacturer circulates — uses 6.5, which is also the structural-steel value in AISC and ASCE. This page defaults to 6.5 so its answers line up with the numbers your panel supplier will quote, and gives you a custom field if you want to run 7.3 and see the 12% difference for yourself. Neither number is wrong; they come from different sources and slightly different temperature ranges.
Why the installation temperature is a real input
Most published examples stop at a total movement figure. That figure is the full cold-to-hot range, and it is the right number for sizing an expansion joint or a trim gap. It is not the number a clip has to accommodate in one direction.
A clip is installed at one specific panel temperature and then has to work in both directions from there. If you set panels on a 120 °F August deck, almost the entire remaining movement is contraction. Set them on a 30 °F February morning and almost all of it is expansion. Same roof, same total range, completely different demand on the sliding element.
expansion still to come = α × (Thot − Tinstall) × L
contraction still to come = α × (Tinstall − Tcold) × L
clip travel needed = the larger of the two
contraction still to come = α × (Tinstall − Tcold) × L
clip travel needed = the larger of the two
MCA makes the same point from the hardware side:
“Clips must have enough travel in either direction… Some two-piece panel clips have an integral centering devise that prevents an installation where the traveling hook element is ‘bottomed out’ at the time of installation. This feature prevents ‘travel’ limitations and in-service jamming.” — MCA, Standing Seam Best Practices: Clips
A bottomed-out clip is not a clip. It is a fixed connection that nobody decided to make, in a place nobody chose, and the roof will find that out on the first cold night.
Worked example24 ga steel standing seam, 40 ft eave-to-ridge run, pinned at the eave. Panel temperature range −10 °F to 160 °F, installed at 70 °F. Total movement is 0.000078 × 170 × 40 = 0.530 in. But from the 70 °F install point the panel still has 0.000078 × 90 × 40 = 0.281 in of expansion and 0.000078 × 80 × 40 = 0.250 in of contraction ahead of it. A clip with 0.75 in of travel each way is comfortable; one with a quarter inch is already out of room on the hot side.
Fixity: where you decide the movement goes
A floating standing seam roof has to be pinned somewhere. If it is not, the whole roof can walk down the slope — what MCA calls the panels “clutching” off the roof. The choice of where to pin is a design decision, and it changes the answer above by a factor of two.
- Pinned at the eave. The traditional arrangement. Everything grows toward the ridge, and the ridge detail absorbs the full movement of the run. Gravity and the panel movement work in the same direction at the eave, which is convenient for the anchorage.
- Pinned at the ridge. Movement arrives at the eave, gutter and eave trim. Sometimes chosen when the ridge is the complicated tie-in and you would rather not have movement there.
- Pinned mid-run. Halves the arm length in both directions, so each end sees about half the movement. That is the setting to reach for when a long run is over a clip's travel and you do not want an expansion joint — but it also means both ends need to accept movement, not just one.
BRS puts the general case for choosing deliberately rather than by default:
“Regardless of which clip is utilized, you are not going to stop the expansion and contraction. You can, however, have some control of the direction…” — MBCI, Built to Move
What gets missed
The panel is usually the easy part, because somebody engineered the clip. The failures tend to be in everything the panel touches:
- Longitudinal trim. Eave trim, gutters, fascia and rake trim run the other way and are often screwed solid at both ends. They need expansion laps of their own even where the panels do not.
- Penetrations. A pipe boot has to be sized so the hole in the pan can move around the pipe. Cutting a tight hole in the panel and sealing it to the stack turns a vent pipe into a second fixed point.
- Tie-ins to adjacent structure. If a roof runs into a masonry wall, a canopy or an older building, the detail there has to let the roof move. Anything rigid becomes fixity you did not design.
- End laps. Where a run is spliced, the lap has to pass the expansion force from one panel to the next without buckling or slotting — which is why the detail wants a backup plate, tape sealant and enough screws.
The force number, and what it actually predicts
If the panel is fully restrained, the strain it wanted to make turns into stress instead:
σ = E × α × ΔT psi
F = σ × A lb · A = panel width × base metal thickness
BRS gives the same relation as F = A × E × ε × δT
F = σ × A lb · A = panel width × base metal thickness
BRS gives the same relation as F = A × E × ε × δT
Run the default case and it comes out at just under 17,000 psi in the steel and around 4,900 lb per foot of panel width. Nothing about a roof holds that. The number is useful precisely because it is absurd: it tells you that something is going to give, and the design question is only which thing.
Read the force figure as tension, not compression. A 24 ga sheet in tension really can carry that stress — it is well below yield. In compression it cannot, because a wide thin panel buckles first. That buckle is oil canning, and it is why a restrained panel goes wavy in the sun rather than snapping a fastener. On the contraction side there is no buckling escape, which is where torn fastener holes and popped seams tend to show up.
On a through-fastened roof there is no sliding clip at all, and BRS describes what happens instead: the panel pushes against each screw, the purlins roll in the direction of the force until the panel-to-fastener interface gives up, and the panel tears and piles up around the screw shank. That is the mechanism behind the slotted, elongated screw holes you find on old screw-down roofs, and it is the reason exposed-fastener systems come with a published maximum run length.
Picking temperatures you can defend
This is the input that carries the uncertainty, and it is worth being deliberate about. Two approaches are in common use.
Panel temperature directly
What the calculator asks for. The panel, not the air. Dark panels in full sun run substantially above ambient, and clear-night radiant cooling takes a panel below ambient. MCA's seasonal figure of more than 200 °F of range is the sanity check: if your two numbers are 60 °F apart you are almost certainly using air temperatures by mistake.
Air temperature with a published extreme
BRS's worked examples take the mean of the annual extreme dry-bulb temperatures from ASHRAE's climatic design data — −2.4 °F to 97.2 °F for Springfield, Missouri, and 31.5 °F to 95.0 °F for Tampa, Florida — and then apply a 0.80 factor for the difference between theoretical and observed movement. That is a defensible method with real numbers behind it, and if you use it, set the factor selector to 0.80 so you are running their method rather than a hybrid.
Two things this page deliberately does not do. BRS's examples also subtract the interior temperature swing of a conditioned building — 5 °F in the Springfield case, 10 °F in Tampa. And their 0.80 factor is, in their words, a reduction of “as much as 20%” observed in field studies. Both reduce the answer. Reducing the answer is the wrong direction to round when the thing you are sizing is clip travel, so this calculator defaults to the factor at 1.00 and offers no interior-swing credit. If you want BRS's exact figures for a conditioned metal building, switch the factor to 0.80 and subtract the interior swing from your two temperatures by hand.
Frequently asked questions
How much does a metal roof expand and contract?
Steel panels move about 0.000078 inches per foot of run for every degree Fahrenheit of panel temperature change, which works out to roughly 0.78 inches per 100 feet of run per 100 °F. Aluminum moves nearly twice as much at about 1.48 inches, copper about 1.13 inches and Type 304 stainless about 1.15 inches over the same run and temperature change.
What temperature range should I use for metal roof expansion?
Use panel surface temperatures, not air temperatures. The Metal Construction Association reports that panel temperatures can range more than 200 °F over a season, 150 °F in a single day and 100 °F within minutes when a cool rain hits a sunlit roof. A dark panel in direct sun sits well above ambient, and a clear night takes a panel below ambient by radiant cooling.
Why does the installation temperature matter?
Because a clip has to work in both directions from wherever it was set. A panel installed on a hot deck has almost all of its remaining movement on the contraction side; the same panel installed on a cold morning has almost all of it on the expansion side. Clip travel has to cover the larger of those two one-way movements, not half the total range.
How long can a standing seam panel run be?
It depends on the clip travel and the temperature range, not on a universal number. Divide the clip's one-direction travel by the movement rate for your material and the larger of your two one-way temperature differences. With 0.75 inches of travel, steel, and a 90 °F swing from installation to peak, that is about 107 feet from the fixed point. Always confirm against the panel manufacturer's published maximum run.
What is the difference between a fixed clip and a floating clip?
A fixed or one-piece clip is a single part seamed into the panel and screwed to the structure, so any movement has to come from flexure of the clip, the panel and the purlins. A floating or two-piece clip has a base that stays put and a hook that slides, so the movement happens inside the clip itself and is independent of the substrate. Floating clips allow far longer runs.
Does thermal movement cause oil canning?
It can. When a panel is restrained and heated, the growth it cannot make turns into compressive stress, and a wide thin panel buckles into visible waviness long before it reaches any stress that would break something. MCA notes that a clip interface with excessive frictional resistance can exacerbate oil canning for exactly this reason.
Why do screws on an exposed-fastener metal roof end up in slotted holes?
Because a through-fastened panel has no sliding clip, so it pushes directly against every screw. The purlins roll in the direction of the force until the panel-to-fastener interface gives way, and the panel tears and piles up around the screw shank. That is the published failure mechanism for screw-down roofs, and it is why they carry maximum run lengths.
Do I need an expansion joint in a metal roof?
Only when the run is longer than the clip travel supports, or where the roof crosses a structural expansion joint in the building beneath it. The calculator gives you the longest run your clip travel allows and the number of joints a longer run needs. Where the structure has a joint, the panels need one in the same place regardless of what the movement arithmetic says.
Does a metal roof over an unheated building move more?
Generally yes, and so does an uninsulated one. An insulated conditioned building holds one side of the panel closer to a steady temperature, which trims the swing the panel sees. BRS's worked examples subtract the interior temperature swing for that reason. This calculator does not apply that credit, because it reduces the answer and clip travel is the wrong place to be optimistic.
Does panel colour change how much it moves?
Not the coefficient, but very much the temperature. Colour changes solar absorptance, so a dark panel reaches a higher surface temperature in the same sun than a light or reflective one. That shows up in this calculation as a larger hot-side number, which is where the extra movement comes from.