Building thermal systems that behave the same across every manufactured unit

Focus: Mounting pressure control, TIM selection, screw torque, mechanical tolerance stack-up, and unit-to-unit thermal repeatability

Introduction

Thermal design failures rarely happen because a heatsink was completely ignored.

More often, they happen because the design worked perfectly on the first prototype.

One prototype had excellent thermal performance. The processor temperature stayed within limits. The thermal camera showed uniform heat spreading. The system passed validation.

Then production started.

A few units showed higher temperatures. Some devices throttled earlier. Some heatsinks felt warmer. Some products passed factory testing while others failed under the same workload.

The hardware design had not changed.

The thermal interface had.

This is one of the most underestimated challenges in embedded hardware design: thermal performance is not only a material problem. It is a mechanical consistency problem.

A heatsink, thermal pad, thermal paste, mounting bracket, and PCB are not independent parts. Together, they form a thermal system where small mechanical variations directly affect electrical reliability.

At Hoomanely, thermal design is approached from a repeatability perspective.

The goal is not:

"Can this heatsink cool the processor?"

The goal is:

"Will every manufactured unit transfer heat in the same way after assembly variations, component tolerances, and years of operation?"

A reliable thermal architecture is one where the first prototype, the thousandth production unit, and the field-return unit behave predictably.

Thermal Performance Is a Mechanical System

When engineers think about thermal performance, the first parameters usually considered are:

• Heat sink size
• Thermal conductivity
• Airflow
• Surface area

These are important.

But the final thermal path is determined by the complete stack:

Processor Die

↓

Package Surface

↓

Thermal Interface Material

↓

Heat Spreader / Heatsink

↓

Mechanical Mounting Structure

↓

Ambient Environment

Every interface introduces thermal resistance.

Even a high-performance thermal pad cannot compensate for poor mechanical contact.

A small air gap can create a much larger thermal barrier than expected because trapped air transfers heat poorly compared to solid materials.

This is why mounting pressure, surface flatness, and mechanical tolerance become part of thermal engineering.

Mounting Pressure: The Missing Thermal Specification

A common thermal design mistake is specifying the heatsink but not specifying the mounting force.

The thermal interface material works by filling microscopic surface imperfections between two surfaces.

However, the effectiveness depends heavily on compression.

Too little pressure:

• Air gaps remain
• Contact area decreases
• Thermal resistance increases

Too much pressure:

• PCB bending risk increases
• Package stress increases
• Components may experience mechanical fatigue
• TIM can be squeezed excessively

The objective is not maximum pressure.

The objective is controlled pressure.

A good thermal design defines:

• Required compression range
• Acceptable mounting force
• Mechanical limits of the PCB and components

For example:

A thermal pad may specify:

• Original thickness: 1 mm
• Recommended compression: 30–50%

The mounting system must consistently achieve that compression across production units.

Without this, every assembled device may have a different thermal path.

TIM Selection: Choosing for Real Assembly Conditions

Thermal Interface Material (TIM) selection is often based only on thermal conductivity numbers.

A datasheet may advertise:

"12 W/mK thermal conductivity"

But that number alone does not define real-world performance.

The actual thermal behaviour depends on:

• Surface roughness
• Compression ratio
• Contact pressure
• Operating temperature
• Long-term aging
• Assembly repeatability

Different products require different TIM strategies.

Thermal Pads

Thermal pads are commonly used in embedded products because they are:

• Easy to assemble
• Electrically insulating
• Mechanically forgiving

Advantages:

• Good tolerance handling
• Suitable for production assembly
• Less mess compared to thermal grease

Challenges:

• Require correct compression
• Thickness tolerance affects pressure
• Higher thermal resistance than direct compounds

Thermal Grease

Thermal grease provides excellent thermal performance when applied correctly.

However:

• Application thickness varies
• Manufacturing control is harder
• Pump-out can occur over thermal cycles

For high-volume products, consistency is often more important than peak thermal conductivity.

Phase Change Materials

These can provide a balance between:

• Production repeatability
• Thermal performance
• Long-term stability

The correct TIM is not always the material with the highest specification.

It is the material that maintains predictable performance after assembly.

Screw Torque: Turning Mechanical Assembly Into a Thermal Parameter

Screws are not just mechanical fasteners in a thermal design.

They define the pressure applied to the thermal interface.

The same heatsink installed with different screw torque values can have different thermal performance.

Low torque:

• Reduced compression
• Higher thermal resistance
• Possible hotspot formation

Excessive torque:

• PCB deformation
• Mechanical stress
• Component damage

This is why torque specification should be treated like an electrical parameter.

A production instruction should define:

• Screw type
• Screw size
• Tightening sequence
• Torque value
• Acceptable tolerance

The tightening sequence also matters.

For example, tightening one corner completely before others can create uneven pressure distribution.

A cross-pattern tightening sequence usually provides more uniform compression.

A good thermal assembly process considers:

• Torque variation between operators
• Tool calibration
• Production environment
• Screw and washer variation

The thermal design is only as repeatable as the assembly process.

Mechanical Tolerance Stack-Up: The Hidden Thermal Variable

Most thermal failures caused by manufacturing variation come from tolerance accumulation.

Consider a simple stack:

Heatsink height tolerance

+

TIM thickness tolerance

+

PCB thickness variation

+

Component package height variation

+

Mounting bracket tolerance

Each individual tolerance may look insignificant.

Together, they determine the final compression.

A design may target:

1 mm thermal pad compression

But actual production units may experience:

• 0.6 mm compression
• 1.3 mm compression

The same thermal design now has two different behaviours.

This is why thermal stack-up analysis should happen during mechanical design.

Important parameters include:

• PCB thickness tolerance
• Standoff height
• Heatsink flatness
• Pad thickness variation
• Screw hole position accuracy
• Component height variation

Thermal design is not complete until mechanical tolerances are understood.

Designing for Unit-to-Unit Thermal Repeatability

A production thermal solution should not be judged by the best-performing unit.

It should be judged by the worst expected unit.

During validation, thermal testing should include:

• Multiple production samples
• Different assembly operators
• Maximum workload conditions
• Temperature chamber testing
• Long-duration operation

The objective is finding variation.

Questions to ask:

• What is the temperature difference between units?
• Are hotspots appearing randomly?
• Does performance change after thermal cycling?
• Does removing and reinstalling the heatsink change behaviour?

A thermal design that only works on one carefully assembled prototype is not a production solution.

Mechanical Features That Improve Thermal Consistency

Good thermal architectures often include mechanical features specifically designed for repeatability.

Controlled Compression Stops

Adding mechanical stops prevents excessive TIM compression.

The heatsink reaches a defined position every time.

Spring-Loaded Mounting

Spring mechanisms can maintain more consistent pressure compared to rigid screws.

They compensate for:

• Material expansion
• Component height variation
• Assembly differences

Reference Surfaces

Machined or controlled mounting surfaces improve:

• Flatness
• Contact consistency
• Thermal transfer

Defined Assembly Sequence

The thermal process should specify:

1. TIM placement
2. Heatsink alignment
3. Screw sequence
4. Torque application
5. Verification

Thermal performance should not depend on individual technician preference.

Thermal Aging: The Mounting System Must Survive Time

Thermal performance changes over product life.

Factors include:

• TIM drying
• Material compression changes
• Screw relaxation
• Thermal cycling stress
• Mechanical fatigue

A design that works after initial assembly but degrades after years is incomplete.

Long-term thermal validation should include:

• Thermal cycling
• Vibration testing
• Extended operation
• Repeated heating and cooling cycles

The mounting system must maintain contact throughout the product lifetime.

Hoomanely Design Perspective

At Hoomanely, thermal management is not treated as adding a bigger heatsink when temperatures increase.

The thermal system begins with mechanical intent.

A heatsink is only effective when:

• The pressure is controlled
• The TIM is correctly selected
• The assembly process is repeatable
• The tolerance stack-up is understood
• The lifetime behaviour is predictable

A thermal design that works once is a prototype solution.

A thermal design that works thousands of times is a product architecture.

The difference is not the heatsink.

The difference is the engineering discipline behind the mounting system.

Final Thoughts

Reliable thermal performance is not created by selecting the highest conductivity material or the largest heatsink.

It comes from controlling every variable that influences heat transfer.

Mounting pressure defines contact quality.

TIM selection defines interface behaviour.

Torque defines assembly consistency.

Mechanical tolerances define production variation.

Lifetime analysis defines field reliability.

The best thermal designs are invisible to users because every unit behaves the same.

The product does not become hot because one unit had a better assembly process than another.

It simply works — consistently, predictably, and for years.