What a heat sink thermal test report must include: test setup, power and airflow, thermal resistance calculation, and measured vs simulated figures.
Heat Sink Thermal Test Report: What the Data Package Contains
Direct answer: A usable heat sink thermal test report states five things: the dissipated power, the ambient temperature, the airflow at the fin inlet, the thermal interface material and bond line, and the measured base temperature at the component footprint. From these, thermal resistance in C/W is calculated. A report that gives only a thermal resistance figure, with no test conditions, cannot be compared with anything and should not be accepted.
What a complete data package contains
| Section | Contents | Why it matters |
|---|---|---|
| Part identification | Drawing revision, material, finish, sample source (prototype or production lot) | Without the revision and lot, the result may not describe the part you are buying |
| Geometry record | Measured fin thickness, pitch, base thickness, overall dimensions | Explains variance between this report and the next one |
| Test setup | Heater type and size, mounting force, interface material and thickness, orientation | Orientation alone can change natural convection results by 15-25% |
| Conditions | Power in watts, ambient in C, airflow in m/s or CFM at fin inlet | The three variables that define the operating point |
| Measurements | Base temperature at footprint centre, ambient, and the thermocouple positions | Position determines whether you measure spreading or interface resistance |
| Calculated result | Thermal resistance in C/W, with the formula shown | Must be reproducible from the raw data |
| Simulation, if any | CFD model assumptions, mesh, boundary conditions, and the delta versus test | Simulated and measured results must be labelled separately |
Measured versus simulated: keep them apart
This is the most common way a data package misleads. A CFD result is a prediction based on assumptions; a bench test is a measurement with uncertainty. Both are useful and they are not interchangeable.
A competent report labels every number as measured or simulated, states the delta between them, and explains the delta. A 10-20% difference between CFD and bench test is normal and expected; it comes from interface resistance assumptions, contact pressure and airflow distribution. A report showing perfect agreement is either using the test to calibrate the model, which should be said, or is not showing you both.
When you receive a data package, the first question is not "is the thermal resistance good" but "what is measured and what is modelled". If the supplier cannot separate them, the entire package is a design study, not a test report.
How thermal resistance is calculated
The calculation is simple, which is why omitting the raw data is indefensible:
Rth = (T_base - T_ambient) / P
where Rth is in C/W, temperatures in C, and P is the dissipated power in W. A heat sink with a 62 C base at 25 C ambient dissipating 40 W gives (62-25)/40 = 0.925 C/W.
Three subtleties that a good report will address:
- Which base temperature. The value must be measured at the centre of the component footprint. Measured at the edge of a large base, the number includes spreading resistance and will look worse; measured on the component case, it includes the interface and will look worse still. State the location.
- Which ambient. Ambient measured 100 mm from the heat sink in still air is not the same as ambient inside a closed enclosure at 60 C. For enclosure applications, use the internal air temperature as ambient, because that is what the heat sink sees.
- Steady state. A reading taken five minutes after power-on is not steady state. Expect a stated stabilisation criterion, typically less than 0.5 C change over 10 minutes.
Test conditions worth requesting for a real procurement decision
| Condition | Why request it |
|---|---|
| Natural convection, heat sink vertical | The reference case for passive designs; orientation matters |
| Natural convection, heat sink horizontal, fins up | Common mounting, sometimes significantly different from vertical |
| Forced air at your actual airflow | A fan curve and an enclosure restriction are not the same as a datasheet CFM |
| Two power levels | Thermal resistance is not constant; a single point hides the trend |
| With your interface material | Interface resistance can be 20-40% of total; using a different material invalidates comparison |
| Aged part, if relevant | Anodising and interface materials change; a fresh sample is best case |
Reliability data that belongs alongside the thermal report
Thermal performance is one axis; durability is the other. For heat sinks in automotive, outdoor or industrial environments, the data package should also cover:
- Thermal cycling. Typically -40 to +125 C, 100-1,000 cycles, with thermal resistance measured before and after. The failure mode is interface degradation and, in bonded assemblies, delamination.
- Vibration. Relevant for transport and for fan-mounted or vehicle-mounted parts. Fin fatigue at the root is the concern on thin-fin extrusions.
- Salt spray. 96-500 hours neutral salt spray for coated parts, with the acceptance criterion stated, usually no base metal corrosion.
- Flatness retention after cycling. A base that was 0.05 mm flat and relaxes to 0.12 mm after thermal cycling has effectively failed the interface requirement, even though nothing cracked.
Numbers without acceptance criteria are not evidence. Insist that the report states what constitutes a pass.
Red flags in a supplier data package
- A single thermal resistance figure with no conditions. Unusable for comparison.
- Airflow stated in CFM only, with no measurement location. Fan datasheet CFM and inlet CFM differ substantially.
- CFD presented as test data. Common and damaging, especially when the model was never validated.
- No part revision. You cannot tell whether the report describes the part you are buying.
- Interface material unspecified or described as "thermal grease". The bond line thickness and conductivity dominate the result.
- Results reported to three decimal places. A bench test with a thermocouple and a wattmeter does not justify that precision; it usually indicates a simulation output dressed as a measurement.
Frequently asked questions
Can you provide a thermal test report for a custom heat sink?
Yes, as a charged service on a prototype or first-article part. It requires a heater, a controlled ambient, an airflow source and instrumentation, so it is not something that is done for free on a sample.
Is CFD enough, or do I need a bench test?
CFD is enough for design direction and for comparing geometries. It is not enough to commit to a thermal guarantee. If performance is contractual, test it.
What accuracy should I expect from a bench test?
Realistically +/-5-8% on thermal resistance, driven mainly by interface and airflow uncertainty. A supplier claiming +/-2% is either using a very controlled rig or overstating precision.
Do I need a report for every production lot?
No. A report on the first article plus periodic verification, typically annually or after a process change, is normal practice. Per-lot thermal testing is unusual outside safety-critical applications.
Tell us the application, power and airflow and we will scope the test conditions before quoting, so the report answers the question you actually have.
Heat sink sourcing guides
- Aluminum Heat Sink Price per kg: 2026 Cost Breakdown
- Custom Heat Sink MOQ and Lead Time by Process
- Heat Sink Extrusion Die Cost and Tooling Amortisation
- Heat Sink RFQ: What to Send for an Accurate Quote
- Heat Sink Tolerance and Flatness: What to Specify
- Heat Sink Surface Finish Options and Cost Impact
- Heat Sink RoHS, REACH and UL Compliance Explained
- Heat Sink Volume Pricing: How Breakpoints Really Work
- Heat Sink Prototype and Sample Service: What to Expect
- Heat Sink Packaging, Shipping and DDP Terms
- Heat Sink Supplier Audit and Second Sourcing
