RobustPCBWhere Robust Design Starts
CalculatorsالعربيةUpgradeSign inSign up

What's new

Full changelogEngine 1.18.3

Copper Area & Thermal Vias

Thermal

Estimated RθJA from copper area, copper weight and the thermal vias under the pad.

Inputs

Board profile

Fills Copper weight, Board thickness and Ambient temperature from your board profile.

Still air 15–25 (with radiation), gentle forced air 40–80, strong airflow 100+
W/(m²·K)
QFN/DPAK with exposed pad ≈ 2–6, SOIC-8 with pad ≈ 3–8, SOT-223 tab ≈ 15–20
°C/W
The datasheet RθJA on a minimum pad, if given
°C/W
W
°C

Results

2 to check
Estimated RθJA

Not judged: a layout estimate with no target to check against. Check the junction temperature it gives in Junction Temperature & Safe Limits or Thermal Resistance Network.

RθJC 5.0 + [ copper 46.7 ∥ package 120 ]
38.6°C/W
Junction temperature
102.9°C
Allowable dissipation

for TJ ≤ 125 °C
2.071W
All results (5)
Copper spreader alone

top 91 ∥ (vias 4.7 + bottom 91)
46.7°C/W
Copper spreading length

√(k·t/h), useful pour radius
25.9mm
Per-via thermal resistance

25525 µm² of barrel copper
163°C/W
Via field total
10.2°C/W
FR-4 slab path

k ≈ 0.3 W/(m·K) through plane
8.9°C/W
16 vias give 10.2 °C/W through the board, against 8.9 °C/W through the FR-4. Add more vias. 163 °C/W each, in parallel with the FR-4 path. The vias carry most of the heat once their total is below the FR-4's.
A first-order estimate, ±30 % at best. Confirm with a thermocouple or a ΨJT measurement before release.

Your pour is 13.8 mm in equivalent radius, 0.53 × the 25.9 mm spreading length. Area beyond one spreading length helps much less.

RθJA vs copper area

Parameter sweep

Vary one input over a range and see the answer and verdict at each step, as a table and a curve.Pro

Worst-case corners

Put a tolerance on each input and get the worst-case band around the answer.Pro

Give at least one input a tolerance above zero.

Heat path

The principle

A surface-mount package sheds most of its heat into the board. The board then acts as a fin: copper spreads the heat sideways, and both faces lose it to the air by convection and radiation. The fin model shows when more copper stops helping.

Spreading length in a thin plate
Ls=kCu tCuhL_s = \sqrt{\frac{k_{Cu}\,t_{Cu}}{h}}
About 26 mm for 1 oz copper in still air (h ≈ 20 W/(m²·K)). Copper beyond 2–3 Ls from the source adds little.
Fin efficiency of a circular pour
η=tanh⁡(mr)mr,m=1Ls,  r=Aπ\eta = \frac{\tanh(m r)}{m r}, \qquad m = \frac{1}{L_s},\; r = \sqrt{\frac{A}{\pi}}
Convection resistance of one plane
Rθ=1h A ηR_{θ} = \frac{1}{h\,A\,\eta}
Thermal via (plated barrel)
Rvia=tboardkCu π (d+tp) tp,Rfield=RviaNR_{via} = \frac{t_{board}}{k_{Cu}\,\pi\,(d + t_{p})\,t_{p}}, \qquad R_{field} = \frac{R_{via}}{N}
Only the plated barrel conducts. A copper-filled via conducts several times better.
Through the bare laminate
RFR4=tboardkFR4 A,kFR4≈0.3  W/m⋅KR_{FR4} = \frac{t_{board}}{k_{FR4}\,A}, \qquad k_{FR4}\approx 0.3\;\t{W/m·K}
Copper spreader
Rspread=Rtop∥(Rvia∥RFR4+Rbottom)R_{spread} = R_{top} \parallel \left(R_{via}\parallel R_{FR4} + R_{bottom}\right)
Combined network
RθJA=RθJC+(Rspread∥Rθpkg)R_{θJA} = R_{θJC} + \left(R_{spread} \parallel R_{θpkg}\right)
Rθpkg is the path with no pour: package body, leads and nearby laminate. Without it the model would give absurd values on a minimum pad.
Why more copper stops helping
η→1mr  ⇒  Rspread∝1A(r≫Ls)\eta \to \frac{1}{m r} \;\Rightarrow\; R_{spread} \propto \frac{1}{\sqrt{A}} \quad (r \gg L_s)
Within one spreading length R ∝ 1/A; beyond it R ∝ 1/√A. RθJC is a floor that no copper removes.
  • kCuk_{Cu}385 W/(m·K), thermal conductivity of copper
  • tCut_{Cu}copper thickness: 1 oz/ft² = 34.8 µm = 1.37 mil
  • hhcombined convection + radiation coefficient, W/(m²·K)
  • AAcopper area on that face
  • NNnumber of thermal vias
  • tpt_pvia barrel plating thickness, typically 20–30 µm
More detail

What improves RθJA

  • Vias to the other side. A dense via field under the pad often beats doubling the top pour, because it brings the bottom copper into play.
  • A solid plane joined by vias spreads heat better than a top pour broken up by cut-outs.
  • Heavier copper beats more area once the pour is past about 2 Ls. Going from 1 oz to 2 oz lengthens Ls by √2 and halves the lateral resistance.
  • Airflow. Tripling h, from still air to modest forced air, cuts each copper face's resistance by √3 to 3×.
  • No thermal relief on a thermal pad. Relief spokes aid soldering and block heat.

Via practice

A typical field is a 0.3 mm drill on a 1.0–1.2 mm grid, tented or plugged on the bottom so solder does not wick away during reflow. Epoxy-filled, capped vias fix joint voiding; copper-filled via-in-pad is the premium option.

Calibrate it

Published still-air data for a small SMD power package on 1 oz copper run roughly 100–140 °C/W on a minimum pad, 60–90 °C/W on 100 mm² and 40–55 °C/W on 645 mm² (1 in²). The defaults land near these. If the datasheet gives RθJA on a minimum pad, enter it as Rθpkg to match your package.

Not modelled: heat into nearby parts and connectors, uneven board temperature, enclosure walls, board orientation and mutual heating. Use it to rank layouts, and confirm the absolute number with a thermocouple or ΨJT.
See also: JEDEC JESD51-5 and JESD51-7, which define the standard thermal test boards. Allegro MicroSystems AN295014 gives the still air around a package as up to 2000 °C/W per unit thickness. That is why the board, not the package, is the real heat path in surface-mount designs.

Engine version ⁨1.18.3⁩