Optical Communication: A Precision Heat Sink Held to ±0.05 mm Fin Pitch
A Tier-1 communications equipment maker needed a heat sink for a pluggable optical module. The thermal model assumed a fin array at a specific pitch and a base that sat flat on the module — both of which are manufacturing results, not design intentions.
- Industry
- Optical communication
- Process
- Extrusion + CNC base machining
- Alloy
- 6063-T5
- Key spec
- Fin pitch ±0.05 mm
What the customer needed
The customer's thermal simulation produced a fin geometry with a defined pitch and height, and a base that had to sit flat against the module to conduct heat out of the package. The heat sink itself was small, which is exactly what makes it difficult: at this scale, a 0.05 mm error in fin position is a meaningful fraction of the gap between fins.
The quantified constraint was ±0.05 mm on fin pitch, measured across the array rather than at a single point. A secondary constraint was base flatness, agreed per part because it depends on how the module is clamped.
The part also had to be produced without milling between fins, since a cutter cannot enter a gap of this size without leaving a burr that would change the airflow path.
Constraints that decided the part
- Fin pitch ±0.05 mm across the array
- Base flatness agreed per part for module contact
- No milling between fins — burrs would disturb airflow
- Alloy 6063-T5 for thermal conductivity
- Mounting holes positioned from the same datum as the fin array
How the part was made
The same route we would propose today, in the order it actually runs.
-
Section and die design
The fin ratio decided the feeder design. A finer array needs a die that fills each slot before the streams rejoin, or the tips come out short and the thermal area is smaller than modelled.
-
Extrusion
6063 run and cooled under control, then aged to T5. The fin array is formed in the die and is not machined afterwards.
-
First article on pitch
Pitch measured at the root and at the tip across the full array, not at one point. The report ships with the sample so the thermal assumption can be checked against a real number.
-
Base machining
The mounting face machined flat, and the module fixings drilled and tapped from the same datum as the fin array, so the sink seats where the model expects it.
-
Inspection and packing
Pitch, flatness and hole position recorded on the lot; parts packed to protect fin tips in transit.
The numbers the part had to hold
These are the tolerances and process values that governed the job. They come from our capability, not from a customer specification we cannot publish.
| Parameter | Value | Why it mattered |
|---|---|---|
| Fin pitch | ±0.05 mm | Sets the airflow gap the thermal model assumes |
| Profile cross-section | ±0.05 mm | Keeps fin height and base thickness consistent |
| Base flatness | Agreed per part | Controls contact resistance to the module |
| Hole position | ±0.01 mm | Fixings must align with the module footprint |
| Alloy | 6063-T5 | Thermal conductivity plus clean anodizing |
| Fin formation | Extruded, not milled | No burrs in a gap a cutter cannot clear |
The part of this job that nearly went wrong
This is the section worth reading. Every program has one problem that decides whether the parts come out right, and it is usually not the one on the drawing.
Fin pitch drifting across a long die run
A fine array is sensitive to die wear and to pulling tension. Over a production run, pitch can drift enough that the parts still look correct while the airflow gap has moved outside the model.
The risk
A heat sink that passes a visual check and underperforms thermally, discovered in system test rather than at goods-in.
What we changed
Pitch is measured at the root and the tip across the full array on the first article, and the die is corrected before the production run. The measurement is reported, so the customer can compare it with the thermal model instead of trusting a drawing.
Deburring between fins without touching them
If the array is machined after extrusion, a cutter has to enter the gap between fins. At this pitch there is no room for a deburring tool, and a burr at the fin root blocks the airflow path the design depends on.
The risk
Rework that cannot be inspected reliably, on a part where the defect is inside a 0.05 mm gap.
What we changed
Machining was restricted to the base, the mounting holes and the outer profile. The fins exist only as extruded surfaces, so there is no burr to remove between them.
Fin geometry that matches the model, measured and reported
The heat sink ships with the fin array formed entirely in the die, the base machined flat, and the mounting holes positioned from the same datum. Fin pitch is held to ±0.05 mm and measured across the array on the first article and on each lot.
What the customer gained was not a better number on a drawing — it was the ability to trust the thermal model, because the part in the box matched the geometry the model was built on.
We do not publish a thermal resistance for this part. That number depends on airflow, interface material and module power, and it belongs on the customer's test bench.
Common questions about this program
The questions this case generates from engineers and sourcing teams.
How is fin pitch measured on a part like this?
At the root and at the tip, across the full array rather than at a single point, on the first article and on production lots. Reporting one measurement from the middle of the array would hide exactly the drift that matters.
Can you supply a thermal resistance figure?
No. Thermal resistance depends on airflow, interface material and module power, and we do not publish a figure we cannot test. We hold the fin geometry and the base flatness and report them; the thermal performance is verified on your bench.
Why not machine the fins for a finer pitch?
Machining a gap this narrow leaves a burr that cannot be removed without damaging the fin, and the burr changes the airflow path. Extruding the array leaves clean surfaces and removes the operation entirely.
Have a Thermal Part Where Geometry Is the Constraint?
Send the fin geometry, the base flatness that matters and the module power range. We will confirm what the die can fill and what we will measure.
What to send
- 2D drawing or 3D model (STEP, IGES, DWG, DXF, PDF)
- Alloy and temper
- Surface finish and color
- Estimated annual quantity
Client names are withheld under confidentiality. Figures reflect our verified process capability.
Two more programs like this one
Integrated LED Heat-Sink Housing
Fins extruded with the housing body instead of a separate sink.
Profile ±0.05 mm
Read the case study0.3 mm Thin-Wall Housing
A different constraint, the same discipline about holding a dimension in production.
Wall 0.3 mm
Read the case study