Loading Pyodide...

Heatsink Designer & Analysis

Estimate what a straight plate-fin heatsink can actually do under natural convection, forced airflow, or a fan-limited operating point. The default view stays simple: define the thermal budget, define the sink, and see whether the design closes.

Tool Purpose & Scope

This rewrite replaces the old generic thermal calculator with a dedicated heatsink workflow. The current implementation focuses on steady-state, air-cooled straight plate-fin heatsinks and exposes the key design tradeoffs: thermal resistance, source-temperature margin, fin efficiency, airflow impedance, and convection-versus-radiation contribution.

  • Required sink thermal resistance is computed from the full thermal budget.
  • Achieved sink performance is solved from the heatsink geometry and cooling mode.
  • Natural convection, forced flow, and fan-curve operation are supported.
  • Derivations show substituted values for the core thermal equations.

The tool currently supports one geometry family (plate-fin) with airflow bypass modeling and single-source 2D spreading analysis. Additional heatsink geometries (pin fin, radial fin) may be added in the future.

Inputs

Expert Mode Show manual property overrides and ambient-pressure input

Read the temperatures as a chain. Source or junction means the hottest internal node of the mounted part. Case is the outer mounting surface, base is the heatsink metal under that mounting area, and ambient is the surrounding air away from the sink.

Use this as the hottest allowable internal source temperature. In electronics this is often the junction limit; in other hardware it may be the core, winding, or internal hotspot limit.
Thermal resistance from the hottest internal source region to the outer mounting surface. In electronics this is often called junction-to-case resistance.
Total thermal resistance from the case to the heatsink base. Use when you already know the absolute value for your specific contact area.
Area-normalized thermal impedance from the TIM datasheet. Measured end-to-end and includes contact resistance at both interfaces. The tool divides by the source contact area (or full base if no source is set) to get absolute K/W. Typical ranges: high-performance thermal paste 0.01–0.1, thermal grease 0.1–0.5, phase-change pad 0.3–1.0, gap pad 1–5, adhesive transfer tape 3–10.
Computes R = BLT / (k × A), where BLT is bondline thickness, k is thermal conductivity, and A is the source contact area. This estimate excludes contact resistance at the interfaces and will be optimistic compared to the measured thermal impedance. Use the datasheet impedance (K·cm²/W) when available. Typical conductivities: thermal paste 1–5, adhesive tape 0.5–1, gap pad 1–8, solder 25–60 W/m·K.
Which TIM mode should I use?
Heatsink on an IC/component package: Use Thermal impedance with the value from your TIM datasheet (e.g., thermal paste ~0.05, pad ~2.0 K·cm²/W). The source footprint is the package contact area.
Heatsink bolted to a motor housing, enclosure wall, or chassis: The "interface" here is the metal-to-metal contact plus any thermal compound applied. Use Thermal impedance if you have a measured value, or Conductivity + BLT with the thermal compound specs and an estimated bondline (0.05–0.15 mm for paste under clamping). If the mating surface is rough or has an air gap, the contact resistance can dominate — typical bare machined aluminum contact is 1–5 K·cm²/W depending on surface finish and pressure.
Heatsink epoxied or soldered directly: Use Conductivity + BLT with the adhesive or solder properties (epoxy ~1–3 W/m·K, solder ~50 W/m·K) and the bond thickness.
No interface material / direct mounting: Select None to set the resistance to zero. This is a lower bound — real contact resistance depends on surface quality and clamping force, but is often small relative to other thermal resistances.

Start with the heatsink envelope and fin layout. The geometry previews on the right update immediately from these values.

Spacing: --

Source Footprint

If the heat source is smaller than the heatsink base, spreading resistance adds temperature rise. Leave at 0 to assume the source covers the full base.
Expert mode lets you override the preset conductivity and emissivity directly.

Source Position

Position the source center on the base. Leave at 0 to center the source automatically.

Cooling mode picks the governing relationship. Natural convection solves buoyancy-driven channel flow. Forced flow uses known air delivery. Fan curve mode balances fan capability against heatsink pressure drop to find the operating point.

Orientation affects the natural convection correlation. Forced and fan modes ignore this setting.
Natural convection uses only the geometry and temperature rise; no explicit fan inputs are needed.
Used when the sink sits in a known stream but you do not know the exact channel flow rate.
If this is greater than zero, it overrides the approach velocity estimate.
If the heatsink is in open air without a duct or shroud, some airflow will bypass the fin channels.
Use expert mode to adjust air properties for altitude or sealed-enclosure pressure assumptions.
Inputs:
Sample Test Cases
Top Plan
Plan view shows the footprint, fin orientation, and flow direction when airflow is imposed.
Front Section
Front view is the only cross-section view. Use it to read fin count, spacing, thickness, and total height.

Enter the thermal budget and geometry, then press Calculate.

Status

    Note: The Spread View peak junction temperature exceeds the target, even though the headline result (centered-source model) shows an acceptable margin. Check the Spread View tab for localized hotspot details.
    Set a source footprint (length and width) in the Geometry section to see localized base spreading.

    Temperature ladder from ambient to source, plus the convection/radiation heat-rejection split.

    Run a calculation to see charts.

    Base Temp
    --
    Case Temp
    --
    Source Margin
    --
    Convection h
    --
    Radiation h
    --
    Re
    --
    Nu
    --
    Flow Rate
    --
    Channel Velocity
    --
    Pressure Drop
    --
    Rsink
    --
    Rrequired
    --
    Fin Efficiency
    --
    Surface Efficiency
    --
    Spreading Rθ
    --

    Thermal Path View

    Run a calculation to populate the source, case, and base temperatures.

    Source / Junction

    Hottest internal temperature node of the mounted part.

    Pending calc
    Above case: --

    Case

    Outer mounting surface where heat enters the interface material.

    Pending calc
    Above base: --

    Base

    Heatsink metal near the mounting footprint.

    Pending calc
    Above ambient: --

    Ambient

    Bulk air temperature away from the sink.

    --
    Reference node

    1D Sweep: Vary one parameter at a time to see how it affects the selected output metric. The baseline point is marked in red.

    Configure and run a sweep to see results.

    2D Trade Space: Sweep two parameters simultaneously to see feasibility regions and interaction effects.

    Configure and run a contour to see results.

    Design intent: this tool is not trying to be CFD in disguise. It is an explicit early-design model for straight plate-fin heatsinks, built to show what the dominant terms are and where the design margin comes from. Every correlation, assumption, and validity limit is documented below.

    1. Thermal Path Concept

    This diagram is the conceptual version of the temperature stack. It shows where each temperature lives physically and how the available temperature rise is divided between the internal source node, the outer mounting surface, the interface, and the heatsink itself.

    Set the thermal budget to see how much of the temperature rise must be handled by the heatsink itself.

    2. Thermal Nodes

    The temperatures in this tool describe different physical locations along the same heat-flow path. They are not interchangeable. Source or junction temperature is the hottest internal temperature of the mounted part. In electronics this is commonly called junction temperature, but in other hardware it may represent a core, winding, or internal hotspot limit. Case temperature is the outer mounting surface where heat leaves the part. Base temperature is the heatsink metal near the mounting footprint. Ambient temperature is the surrounding air away from the immediate hot boundary layer.

    The heatsink only controls the base-to-ambient part directly. The internal source-to-case path and the interface material spend the rest of the temperature rise before the heat ever reaches the fin array.

    3. Thermal Budget

    The sink does not own the full temperature rise from the source node to ambient. The mounted part usually spends part of the temperature budget internally and another part across the case-to-sink interface. The sink only gets whatever temperature rise is left after those upstream resistances are subtracted.

    $$R_{\theta,\mathrm{req,sink}} = \frac{T_{j,\max} - T_{\infty}}{Q} - R_{\theta,jc} - R_{\theta,cs}$$
    \(T_{j,\max}\): allowable source or junction temperature, \(T_{\infty}\): ambient air temperature, \(Q\): heat load, \(R_{\theta,jc}\): source-to-case resistance, \(R_{\theta,cs}\): case-to-sink resistance.

    If Rθ,req,sink is negative, the upstream resistances alone consume the entire budget — no heatsink can fix this. Reduce the heat load, raise the temperature limit, or improve the upstream path.

    4. Plate-Fin Surface Model

    The solver treats the sink as an isothermal base feeding straight fins. Base temperature is found by bisection: the solver iterates Tb until the total heat rejected equals the heat load. The total exposed area is the exposed base between fins plus the broad fin faces and tips. Fin efficiency discounts the fin area because real fins are not perfectly isothermal.

    $$Q = \eta_o\,h_{\mathrm{eff}}\,A_t\,(T_b - T_{\infty}), \qquad \eta_o = 1 - \frac{A_f}{A_t}(1-\eta_f)$$
    \(A_t\): total exposed area (base + fins), \(A_f\): fin area only, \(\eta_f\): single-fin efficiency (§10), \(\eta_o\): overall surface efficiency, \(h_{\mathrm{eff}} = h_{\mathrm{conv}} + h_{\mathrm{rad}}\): combined heat-transfer coefficient.

    Assumption: The base is isothermal. This is reasonable for aluminum bases with moderate heat loads but breaks down when a small source sits on a large, thin base — see §12 for the spreading correction.

    5. Air Properties

    Air density, viscosity, thermal conductivity, and specific heat are evaluated at the film temperature Tfilm = (Tb + T∞)/2.

    $$\mu = 1.716 \times 10^{-5} \left(\frac{T}{273.15}\right)^{3/2} \frac{273.15 + 110.4}{T + 110.4} \quad \text{(Sutherland law)}$$ $$k_{\mathrm{air}} = 0.02414 \left(\frac{T}{273.15}\right)^{3/2} \frac{273.15 + 194.0}{T + 194.0} \quad \text{(Sutherland-type fit)}$$ $$\rho = \frac{p}{R_{\mathrm{air}} \cdot T}, \qquad c_p \approx 1006 + 0.1(T - 300) \;\mathrm{J/(kg{\cdot}K)}$$
    All temperatures in Kelvin. \(R_{\mathrm{air}}\) = 287.058 J/(kg·K).

    Validity: These fits match NIST data within 1.5% from 200 K to 700 K. The cp linear fit is adequate for the 250–500 K range typical of heatsink design. Outside this range a full property package should be used. Ref: [1], [8].

    6. Natural Convection — Vertical Parallel Plates

    For vertical natural convection, the fin spacing matters at least as much as raw area. Tight spacing creates more area but can choke buoyancy-driven flow. This implementation uses the Bar-Cohen and Rohsenow composite isothermal parallel-plate relation [2].

    $$\mathrm{Nu}_s = \left(\frac{576}{Ra_m^2} + \frac{2.873}{\sqrt{Ra_m}}\right)^{-1/2}$$ $$Ra_m = \frac{g \beta (T_b-T_{\infty}) s^4}{\nu \alpha H}$$
    \(s\): fin spacing (clear gap), \(H\): fin height (channel depth), \(\nu\): kinematic viscosity, \(\alpha\): thermal diffusivity, \(\beta = 1/T_{\mathrm{film}}\): volumetric expansion coefficient. Convection coefficient: \(h_{\mathrm{conv}} = \mathrm{Nu}_s \cdot k_{\mathrm{air}} / s\).

    Validity: Isothermal vertical plates, Ram > 0. The composite form blends the fully-developed limit (Nu → const) with the isolated-plate limit (Nu ∼ Ra1/4). Channel velocity is estimated from a chimney-flow energy balance — treat as approximate.

    Ref: Bar-Cohen, A., Rohsenow, W. M. (1984) [2].

    7. Natural Convection — Horizontal Plates

    For horizontal orientations, the McAdams correlation [3] is used with the characteristic length Lc = A/P (area over perimeter of the base footprint).

    $$\text{Heated surface up: } \mathrm{Nu} = \begin{cases} 0.54\,Ra_L^{1/4} & 10^4 < Ra_L < 10^7 \\ 0.15\,Ra_L^{1/3} & 10^7 < Ra_L < 10^{11} \end{cases}$$ $$\text{Heated surface down: } \mathrm{Nu} = 0.27\,Ra_L^{1/4}, \quad 10^5 < Ra_L < 10^{10}$$
    \(Ra_L = g\beta\Delta T\,L_c^3 / (\nu\alpha)\), \(L_c = A_{\mathrm{base}}/P_{\mathrm{base}}\).

    Assumption: Uses the flat-plate correlation rather than the channel correlation (§6). This is a simplification — horizontal fin arrays have more complex flow patterns than vertical ones.

    Ref: McAdams, W. H. (1954) [3]; Incropera et al. [8], Table 9.1.

    8. Forced Convection

    Under forced flow, the fin channels are treated as rectangular ducts. The flow regime is determined by the channel Reynolds number.

    $$Re = \frac{\rho V_{ch} D_h}{\mu}, \qquad D_h = \frac{2 s H}{s + H}$$
    \(V_{ch}\): channel velocity = \(\dot{V}/A_{\mathrm{open}}\), \(D_h\): hydraulic diameter, \(s\): fin spacing, \(H\): fin height.

    Laminar (Re < 2300): Combined-entry developing-flow correlation from Muzychka & Yovanovich [4]:

    $$\mathrm{Nu} = \left(7.54^3 + (1.841\,Gz^{1/3})^3\right)^{1/3}, \qquad Gz = Re \cdot Pr \cdot D_h / L$$
    \(Gz\): Graetz number. At large Gz (short channels, high flow), entry effects dominate and Nu rises above the fully-developed value of 7.54.

    Turbulent (Re ≥ 2300): Gnielinski correlation [5] with Petukhov friction factor [6]:

    $$\mathrm{Nu} = \frac{(f/8)(Re - 1000)\,Pr}{1 + 12.7\sqrt{f/8}\,(Pr^{2/3}-1)}, \qquad f = (0.790\ln Re - 1.64)^{-2}$$
    Valid for 2300 < Re < 5×106, 0.5 < Pr < 2000. The same friction factor \(f\) is used for both heat transfer and pressure drop to maintain self-consistency.

    9. Pressure Drop

    Pressure drop includes Darcy friction losses through the channel plus entrance contraction and exit expansion losses.

    $$\Delta P = \left(f\frac{L}{D_h} + K_c + K_e\right)\frac{\rho V_{ch}^2}{2}$$
    \(K_c = 0.42(1-\sigma^2)\): contraction loss, \(K_e = (1-\sigma)^2\): expansion loss, \(\sigma = A_{\mathrm{open}}/A_{\mathrm{frontal}}\): open-area ratio.

    Laminar friction factor: Shah & London [7] rectangular-duct Poiseuille-number fit:

    $$f = \frac{24(1 - 1.3553\beta + 1.9467\beta^2 - 1.7012\beta^3 + 0.9564\beta^4 - 0.2537\beta^5)}{Re}$$
    \(\beta = \min(s/H, 1)\): channel aspect ratio. This is the exact Poiseuille solution for a rectangular cross-section, not the circular-pipe approximation.

    Turbulent friction factor: Petukhov correlation [6] as shown in §8.

    Ref: Shah, London (1978) [7]; Kays, London, Compact Heat Exchangers; Simons (2003) [9].

    10. Fin Efficiency

    Tall or thin fins underperform because the fin tip runs cooler than the root. The standard straight-rectangular-fin formula with a corrected tip length is used.

    $$\eta_f = \frac{\tanh(mL_c)}{mL_c}, \qquad m = \sqrt{\frac{h_{\mathrm{eff}} \cdot P}{k \cdot A_c}}$$ $$L_c = H + \frac{t}{2}, \qquad A_c = t \cdot L, \qquad P = 2(L + t)$$
    \(L_c\): corrected fin length (adds half the thickness to account for tip area), \(A_c\): fin cross-section area, \(P\): fin perimeter, \(k\): material thermal conductivity, \(h_{\mathrm{eff}} = h_{\mathrm{conv}} + h_{\mathrm{rad}}\).

    Overall surface efficiency combines fin and exposed-base contributions: \(\eta_o = 1 - (A_f/A_t)(1 - \eta_f)\).

    Assumptions: Uniform heat-transfer coefficient over the entire fin surface. 1D conduction along the fin height. Tip heat transfer modeled via the corrected-length approximation rather than a separate tip boundary condition.

    Ref: Incropera et al. [8], Table 3.5.

    11. Radiation

    Radiation is handled through a linearized coefficient so it can be added to convection as an effective surface coefficient. A view-factor correction accounts for fin-to-fin radiation blockage.

    $$h_{\mathrm{rad}} = \varepsilon \sigma (T_s^2 + T_{\infty}^2)(T_s + T_{\infty}) \cdot F_{\mathrm{eff}}$$ $$F_{\mathrm{eff}} = \frac{s}{s + H}$$
    \(\varepsilon\): surface emissivity, \(\sigma\) = 5.670×10−8 W/m2K4 (Stefan-Boltzmann constant), \(T_s, T_{\infty}\): surface and ambient in Kelvin, \(F_{\mathrm{eff}}\): effective view factor from the parallel-plate model. \(s\): fin spacing, \(H\): fin height.

    \(F_{\mathrm{eff}}\) models the fact that in deep, narrow channels, most radiation from one fin strikes the adjacent fin (at nearly the same temperature), reducing net radiation to the environment. For wide spacing, F → 1. For deep channels, F → 0.

    Assumptions: Gray-diffuse surfaces. Linearization is valid when Ts − T∞ is moderate (<100 K for most heatsink applications).

    Ref: Sparrow, Cess (1978) [10]; Siegel, Howell (2002) [11].

    12. Spreading Resistance & Base-Plane Solver

    When the heat source is smaller than the heatsink base, the heat must spread laterally before reaching the fin array. This adds an additional thermal resistance.

    Scalar Model (always applied)

    $$R_{\mathrm{sp}} = \frac{\psi}{k \sqrt{\pi A_s}}, \qquad \psi = \frac{(1-\epsilon)^{3/2}}{\epsilon\tau + (1-\epsilon)^{3/2}}$$
    \(\epsilon = \sqrt{A_s / A_p}\): relative source size, \(\tau = t / \sqrt{A_p}\): relative base thickness, \(A_s\): source area, \(A_p\): plate area, \(t\): base thickness, \(k\): conductivity.

    Ref: Yovanovich, Muzychka, Culham (1999) [12]; Lee, Song, Au, Moran (1995) [13].

    2D Field Solver (Spread View tab)

    When a source footprint is defined, the Spread View tab additionally solves a 2D base-plane conduction problem using Gauss-Seidel iteration with successive over-relaxation (SOR). The fin array is represented as a distributed sink term derived from the global solver’s Rθ,sink.

    $$\sum_{\mathrm{neighbors}} G_n(T_{ij} - T_n) + G_{\mathrm{sink},ij}(T_{ij} - T_{\infty}) = Q_{ij}$$
    \(G_x = k t \Delta y / \Delta x\): in-plane conductance (x-direction), \(G_y = k t \Delta x / \Delta y\) (y-direction). \(G_{\mathrm{sink},ij} = G_{\mathrm{sink,total}} \cdot A_{\mathrm{cell}} / A_{\mathrm{total}}\): distributed sink. \(Q_{ij}\): source heat input per cell (area-weighted overlap).

    Assumptions: 2D in-plane conduction only (fins not individually meshed). Heat rejection is a distributed sink term, not localized at fin roots. Adiabatic outer edges. One-way coupling: the global solver runs first, then the field solver uses Rθ,sink as input.

    13. Bypass Model

    In an unducted configuration, some approach airflow bypasses the fin array instead of entering the channels. The bypass fraction depends on the fin aspect ratio.

    $$f_{\mathrm{bypass}} \approx \frac{1}{1 + K\sqrt{H/s}}, \qquad K = 0.5$$
    \(H\): fin height, \(s\): fin spacing, \(K\): empirical constant. Deep, closely-spaced fins capture more flow; shallow, wide-spaced fins lose more to bypass.

    This is a first-order model. Real bypass depends on the exact shroud geometry, approach flow profile, and fin-tip shape. In a ducted configuration (shroud forces all air through fins), bypass is zero.

    Ref: Simons, R. E. (2003) [9].

    14. Fan Curve Intersection

    In fan-curve mode, the operating point is found by intersecting the fan pressure-flow curve with the heatsink system curve. The system curve is ΔP ∝ Q² from the pressure-drop model in §9.

    $$P_{\mathrm{fan}}(Q) = P_{\max}\left(1 - \left(\frac{Q}{Q_{\max}}\right)^2\right) \quad \text{(parabolic model)}$$
    \(P_{\max}\): fan shutoff pressure, \(Q_{\max}\): fan free-delivery flow rate. The operating point is found by binary search (60 iterations) where \(P_{\mathrm{fan}}(Q) = \Delta P_{\mathrm{sink}}(Q)\).

    Piecewise-linear fan curves can also be used if specific P-Q data points are available.

    15. Mixed Convection

    When forced-flow velocities are low enough that buoyancy forces are comparable to inertial forces, the flow is in the mixed-convection regime. The Richardson number quantifies this.

    $$Ri = \frac{Gr}{Re^2} = \frac{g \beta \Delta T\,D_h^3 / \nu^2}{Re^2}$$
    Ri > 0.1: buoyancy effects may be significant. The tool warns when this occurs but does not apply a mixed-convection correlation — this is a known limitation.

    16. Model Limitations & Validity Ranges

    Model Parameter Valid Range
    Air properties Tfilm 200–700 K (−73 to 427 °C)
    Bar-Cohen/Rohsenow Ram > 0 (composite form)
    McAdams horizontal RaL 104–1011
    Muzychka/Yovanovich Re < 2300 (laminar)
    Gnielinski/Petukhov Re, Pr 2300 < Re < 5×106, 0.5 < Pr < 2000
    Linearized radiation ΔT Best for ΔT < 100 K
    Yovanovich spreading Source Centered, rectangular, uniform flux
    2D field solver Grid 41×25 default; convergence < 10−4 K

    What this tool does not model: 3D conduction through fins, turbulence-driven flow instabilities, airflow maldistribution across the array, transient thermal response, contact resistance at fin-to-base joints, or manufacturing tolerances. Use CFD or test data for these effects.

    17. References

    1. [1] Lemmon, E. W., Jacobsen, R. T., Penoncello, S. G., Friend, D. G. (2000). Thermodynamic Properties of Air and Mixtures of Nitrogen, Argon, and Oxygen from 60 to 2000 K. J. Phys. Chem. Ref. Data 29(3). NIST baseline for the Sutherland fits.
    2. [2] Bar-Cohen, A., Rohsenow, W. M. (1984). Thermally Optimum Spacing of Vertical, Natural Convection Cooled, Parallel Plates. J. Heat Transfer 106(1), 116–123. doi:10.1115/1.3246622. Used in §6.
    3. [3] McAdams, W. H. (1954). Heat Transmission, 3rd ed. McGraw-Hill. Horizontal plate correlations used in §7.
    4. [4] Muzychka, Y. S., Yovanovich, M. M. (2004). Laminar Forced Convection Heat Transfer in the Combined Entry Region of Non-Circular Ducts. J. Heat Transfer 126(1), 54–61. doi:10.1115/1.1643752. Used in §8 (laminar branch).
    5. [5] Gnielinski, V. (1976). New Equations for Heat and Mass Transfer in Turbulent Pipe and Channel Flow. Int. Chem. Eng. 16, 359–368. Used in §8 (turbulent branch).
    6. [6] Petukhov, B. S. (1970). Heat Transfer and Friction in Turbulent Pipe Flow with Variable Physical Properties. Adv. Heat Transfer 6, 503–564. Friction factor used in §8 and §9.
    7. [7] Shah, R. K., London, A. L. (1978). Laminar Flow Forced Convection in Ducts. Academic Press. Poiseuille-number polynomial for rectangular ducts (§9).
    8. [8] Incropera, F. P., DeWitt, D. P., Bergman, T. L., Lavine, A. S. Fundamentals of Heat and Mass Transfer, 7th ed. Wiley. General reference for correlations, fin efficiency (Table 3.5), horizontal convection (Table 9.1).
    9. [9] Simons, R. E. (2003). Estimating Parallel Plate-Fin Heat Sink Pressure Drop. Electronics Cooling 9(1). Pressure-drop methodology and bypass model (§9, §13).
    10. [10] Sparrow, E. M., Cess, R. D. (1978). Radiation Heat Transfer. Hemisphere. View factor model (§11).
    11. [11] Siegel, R., Howell, J. R. (2002). Thermal Radiation Heat Transfer, 4th ed. Taylor & Francis. View factor reference (§11).
    12. [12] Yovanovich, M. M., Muzychka, Y. S., Culham, J. R. (1999). Spreading Resistance of Isoflux Rectangles and Strips on Compound Flux Channels. J. Thermophys. Heat Transfer 13(4). Scalar spreading model (§12).
    13. [13] Lee, S., Song, S., Au, V., Moran, K. P. (1995). Constriction/Spreading Resistance Model for Electronics Packaging. ASME/JSME Thermal Eng. Conf., vol. 4. Additional spreading-resistance formulation (§12).

    18. Worked Examples

    Example 1: Natural Convection

    Given: 15 W heat load, 25 °C ambient, target junction limit 100 °C. Aluminum 6063-T5 heatsink (k = 201 W/m·K), base 100 × 80 × 5 mm, 10 fins at 30 mm tall × 1 mm thick, black anodized (ε = 0.85). Rθ,jc = 0.50 K/W, Rθ,cs = 0.20 K/W. Vertical orientation, natural convection.

    Step 1 — Thermal Budget:

    Available rise = 100 − 25 = 75 °C. Total budget = 75 / 15 = 5.00 K/W.
    Upstream consumption = 0.50 + 0.20 = 0.70 K/W.
    Required sink Rθ = 5.00 − 0.70 = 4.30 K/W.

    Step 2 — Geometry:

    Fin spacing: s = (80 − 10 × 1) / (10 − 1) = 7.78 mm.
    Hydraulic diameter: Dh = 2 × 7.78 × 30 / (7.78 + 30) = 12.35 mm.
    Total wetted area: ≈ 680 cm².

    Step 3 — Results:

    The Bar-Cohen/Rohsenow correlation (§6) gives hconv ≈ 7.9 W/m²K. Linearized radiation (§11) adds hrad ≈ 1.19 W/m²K (corrected for fin-channel view factor Feff = 0.206).
    Fin efficiency (§10): 97.3%. Sink Rθ = 1.66 K/W — well within the 4.30 K/W budget.
    Base temperature: 49.9 °C. Junction temperature: 60.4 °C. Margin: 39.6 °C.
    Heat split: 13.0 W convection, 2.0 W radiation.

    Enter these values in the tool to verify. Results should match within rounding precision.

    Example 2: Forced Convection (Same Heatsink)

    Given: Same heatsink and thermal path as Example 1, but with 2 m/s approach velocity (forced convection mode).

    Results:

    At 2 m/s approach, the channel velocity is 2.29 m/s (due to contraction into fin channels). Re ≈ 1770 (laminar, §8).
    hconv jumps to 24.0 W/m²K — roughly 3× the natural-convection value. Radiation drops to 0.66 W because the base runs much cooler.
    Fin efficiency: 92.9%. Sink Rθ = 0.63 K/W. Pressure drop (§9): 0.6 Pa.
    Base temperature: 34.4 °C. Junction temperature: 44.9 °C. Margin: 55.1 °C.
    The forced airflow cuts the thermal resistance by nearly 3× compared to natural convection.

    Switching from natural to forced convection is often the single largest lever in a thermal design.