Copper Pour Area: Metric Layout, Imperial Thermal Spec
The layout tool reports the top-layer pour under your regulator as 19.4 cm². The regulator datasheet plots junction-to-ambient resistance against copper area in square inches, with tick marks at 0.5, 1 and 2. Nothing on either page meets in the middle, and the thermal review is tomorrow.
Silicon vendors have kept the square inch in their thermal curves for decades, while every ECAD area report, every mechanical drawing and every stack-up table you own is in millimetres and centimetres. This page bridges that one gap so a pour figure lifted from the layout can be dropped straight onto the datasheet curve.
Why the Two Units Collide on a Thermal Review
Derating curves are drawn in square inches
Layout reports come back metric
Copper weight is a thickness, not an area
Test boards are defined in inch dimensions
From Layout Report to Datasheet Curve
Keep the polygon report open on one screen and this page on the other. Three passes usually settle a thermal question: convert the pour you have, look up what the curve promises, then work backwards to the pour you actually need.
Enter the pour area you measured
Type the polygon figure into the left box — 19.4 for a 19.4 cm² pour. If the tool reported millimetres, divide by 100 first, or pick mm² from the searchable dropdown and skip the mental step. Both boxes update while you type; there is no button to press.
Land the result on the curve's axis
The inch figure appears with up to eight decimals, thousands separated by a space. Round it to the resolution the plot can actually show — 3.007 is a 3 in² tick — and read the resistance off the trace above it.
Work the target backwards
When the curve tells you 2 in² buys the resistance you want, type 2 into the right-hand box and read 12.9 cm² on the left, or press the swap arrows to flip the pair for a whole session of reverse lookups.
Move the figure into the review notes
The copy button next to each field lifts the bare number — no unit, no thousands spacing — so it lands in a spreadsheet cell or a review comment as a value. Selecting inside the field and pressing Ctrl + C behaves the same way.
Copper Area Against Junction Temperature Rise
The numbers below trace a representative single-layer 1 oz pour under a tab-style package such as a DPAK or SOT-223. Exact values differ part to part, but the shape of the curve does not: the first square centimetre of metal buys far more cooling than the tenth.
| Pour on the layout | Datasheet axis | Typical θJA | Rise at 1 W |
|---|---|---|---|
| 0.65 cm² (landing pad only) | 0.10 in² | ~125 °C/W | 125 °C |
| 1.6 cm² | 0.25 in² | ~95 °C/W | 95 °C |
| 3.2 cm² | 0.50 in² | ~78 °C/W | 78 °C |
| 6.45 cm² (the classic figure) | 1.00 in² | ~60 °C/W | 60 °C |
| 12.9 cm² | 2.00 in² | ~50 °C/W | 50 °C |
| 19.4 cm² | 3.00 in² | ~45 °C/W | 45 °C |
| 32.3 cm² | 5.00 in² | ~41 °C/W | 41 °C |
| 87.1 cm² (76.2 × 114.3 mm coupon) | 13.5 in² | ~38 °C/W | 38 °C |
Follow the last two rows and the flattening is obvious: going from 32.3 to 87.1 cm² nearly triples the metal and claws back three degrees. That is the moment to stop widening the pour and start looking at via arrays, a plane on the far side, airflow, or a package with a lower internal resistance from junction to tab.
What Helps During a Thermal Sign-Off
Both directions, one page
Type on either side. The design pass goes metric to imperial; the target pass goes the other way, from an inch figure on the plot back to the pour you have to draw.
Thirteen area units in the list
The searchable dropdowns cover mm² and dm² alongside cm² and in², which saves a division whenever the polygon report hands you square millimetres.
Numbers that paste cleanly
Copy delivers digits only, so a converted area drops into a thermal spreadsheet as a number rather than as text you have to strip afterwards.
Calculated in the page
The arithmetic happens in your browser, so dimensions from an unannounced board never travel anywhere, which matters while a project is still under NDA.
Copper Area and Thermal Questions
How much copper do I need to reach a 60 °C/W junction-to-ambient figure?
For a tab package on a single-sided 1 oz board, that target usually lands near the datasheet's one-square-inch mark, which is 6.45 cm² of connected pour — about 25 × 26 mm if you draw it square. Two caveats decide whether you actually get it. The metal has to be electrically and thermally continuous with the tab, so a plane sliced apart by signal traces counts for far less than its raw polygon figure. And the curve assumes still air with the board lying flat; forced airflow across the same pour can pull the effective resistance down by a third or more, while standing the board vertically inside a sealed enclosure pushes it the other way.
What does "1 oz copper" mean, and does thickness matter as much as area?
It is a weight per unit area — one ounce of copper distributed over a square foot — which comes out at roughly 35 µm of finished thickness. Half-ounce foil is about 17 µm, two-ounce about 70 µm. Thickness governs how far heat travels sideways before it has to leave through the board surface, so doubling the plating widens the useful radius of a pour rather than adding a fixed number of degrees. The practical consequence: 6.45 cm² of 2 oz copper beats the same 6.45 cm² in half-ounce foil, and a vendor curve is only quotable if the copper weight printed in its caption matches your stack-up.
What size are the JEDEC thermal test boards in square centimeters?
The standard high-conductivity coupon measures 76.2 × 114.3 mm, which is 87.1 cm² of outline, or 13.5 in² in the units the standard was written in. Smaller packages are characterised on a coupon roughly half that length. Those boards carry buried planes as well — the familiar 1s0p and 2s2p labels count signal and plane layers — which is why a two-plane result can be around half the single-layer number for the very same part. Treat a headline resistance as a way of comparing packages against each other, not as a promise about your product, unless your board genuinely carries that much continuous metal near the device.
If I double the pour from 6.45 cm² to 12.9 cm², does the temperature rise halve?
No, and expecting it to is the most expensive assumption in this whole exercise. In the table above that step moves the resistance from about 60 to about 50 °C/W — a saving near 17 %, nowhere near half. Heat spreads outward from the tab and the temperature of the metal falls with distance, so copper added at the rim of a large pour already sits close to ambient and sheds almost nothing extra. Beyond roughly 6 cm² on a thin single layer the curve is visibly bending over, and past 30 cm² it is nearly horizontal. At that point the cheaper wins are elsewhere: more via stitching, a plane on the opposite side, or splitting the dissipation across two devices.
Do thermal vias buy more than simply enlarging the pour?
Once the top-side pour is past a few square centimetres, usually yes. A grid of plated vias under an exposed pad ties the upper copper to whatever is on the far side, so the bottom layer starts working as a second spreader instead of sitting idle. A common arrangement is 0.3 mm drills on a 1.0 to 1.2 mm pitch, which puts nine to sixteen of them beneath a typical exposed pad; the gain per via tails off quickly after that. Report the pair of numbers together in a review — "12.9 cm² (2 in²) top pour, twelve vias to a bottom plane" — because an area quoted on its own tells a reviewer almost nothing about the path the heat is taking.
No comments yet. Be the first to comment!