Put one flat metal cover over a PCB and look at what sits underneath. The board is a tiny skyline. VRAM packages, MOSFETs, controllers, coils, and capacitors all rise to different levels. Some parts run hot. Some are simply tall. And the heat sink above them does not bend down politely to meet each one.
That is where thermal putty for electronics gap filling can make sense. It forms around local height differences and replaces the air between components and a heat sink, cold plate, spreader, or housing.
Still, putty is not magic clay. Too little leaves a gap. Too much can hold the whole cooler up. The right approach is to map the real assembly, control the material volume, and check the contact after the screws go in.
Why Different Component Heights Become a Cooling Problem
A cooler can look flat while the parts below it are anything but flat. Package tolerances, solder thickness, PCB bow, housing flatness, fastener position, and assembly load all add variation.
This means component height and interface gap are not the same thing. A 1.2 mm-tall device does not automatically need 1.2 mm of material. What matters is the remaining space between the top of that device and the cooler after the product is assembled.
The complete path is simple:
heat source → interface material → cooling surface
If a tall inductor touches the enclosure first, a nearby power IC may still sit under an air pocket. A firm material may also bow the board before the shorter part makes contact.
A good thermal interface material must therefore do two jobs. It must move heat, of course. It must also accommodate the geometry without loading delicate components like a badly fitted doorstop.
When Thermal Putty Fits a Mixed-Height Assembly
Thermal putty is a strong candidate when several devices share one cooling surface and the gap changes from one location to the next. Its soft body deforms under relatively low pressure. That helps it fill local steps without requiring a separate pad thickness for every component.
It is especially useful when:
- component-to-cooler gaps vary across the board;
- the original pad thickness is unknown;
- the PCB or housing has meaningful flatness tolerance;
- low assembly stress is important;
- the layout contains awkward shapes or tight spaces;
- prototypes change too often to justify new die-cut pads;
- the interface may need to be opened and reworked later.
There are limits. A thin CPU or GPU die interface normally needs grease or phase-change material. A stable gap may be easier to control with a pre-cut pad, while a permanent joint may suit a curing filler. Very soft putty can also slump in a vertical joint.
| Material | Best fit in a mixed-height design | Main strength | Main watch-out |
| Thermal putty | Variable gaps and stepped components | Conforms with low pressure | Volume, slump, and squeeze-out must be controlled |
| Thermal pad | Known, repeatable gaps | Clean placement and fixed thickness | One thickness may not contact every height |
| Thermal grease | Very thin, tightly clamped interfaces | Low bond line and good wetting | Cannot safely bridge a large gap |
| One-part gel | Variable gaps with controlled dispensing | Process-friendly and conformable | Flow and long-term stability need checking |
| Two-part curing filler | Permanent, complex gaps | Stable shape after cure | Mixing, cure, stress, and rework become part of the process |
For a deeper material-by-material decision, see Haktak’s guide to thermal putty versus thermal pads for uneven gaps. The choice is not about which material sounds more advanced. It is about which one fits the actual joint.
Build a Height Map Before Choosing the Putty
“The gap is about 1 mm” is rarely enough. About where? At room temperature or operating temperature? Before or after the screws are tightened?
Record the minimum, nominal, and maximum assembled gap at each important location. Include component and solder tolerances, PCB flatness, housing tolerance, and the fastener stack. For existing hardware, a height gauge, witness material, pressure film, or section measurement may help.
Then add the thermal information. The tallest component is not always the hottest. A large choke may touch the lid easily while a lower power stage needs the better heat path.
| Location | Assembled gap | Thermal concern | Mechanical concern | Electrical need |
| VRAM package | Min/nominal/max | Memory temperature | Cooler seating near the GPU die | Usually insulating |
| MOSFET or power stage | Min/nominal/max | High local heat flux | Package and solder-joint load | Dielectric margin may be critical |
| Inductor or choke | Min/nominal/max | Moderate or variable loss | Often one of the taller parts | Check exposed conductors |
| Controller IC | Min/nominal/max | Lower area for heat flow | Small package, concentrated load | Confirm insulation requirement |
| Two-part curing filler | Permanent, complex gaps | Stable shape after cure | Mixing, cure, stress, and rework become part of the process |
This worksheet catches mistakes early and gives a supplier something useful. A photo and a request for “the best 12 W/m·K putty” is not much of a specification, honestly.
Select Thermal Putty by the Real Gap Window
Thermal conductivity is useful, but it is only one piece of the interface. The rough bulk relationship is:
$$R_{material}=\frac{BLT}{kA}$$
Here, BLT is bond line thickness, k is thermal conductivity, and A is the effective heat-transfer area. Thicker material raises resistance. More contact area lowers it. Real joints also have contact resistance at both surfaces, which the simple equation does not capture.
So a fully contacting 6 W/m·K material can beat a poorly fitted 12 W/m·K material. Compare the assembled result at the required thickness and pressure. Haktak’s guide to conductivity versus thermal impedance explains why.
Look at the mechanical behavior as carefully as the thermal number:
- Softness or modulus: Can the putty form around the lowest component without overloading the tallest one?
- Viscosity and yield behavior: Will it dispense cleanly and stay where it is placed?
- Slump resistance: Can it hold position when the board is vertical or hot?
- Squeeze-out: Does the material move into keep-out zones when the housing closes?
- Pump-out and bleed: Will thermal cycling push binder or material away from the contact area?
- Rework: Can the assembly be opened, cleaned, and filled again without damaging nearby parts?
Chemistry matters too. Silicone systems stay flexible across a useful temperature range. Silicone-free grades may suit optics, relays, sensitive contacts, or later bonding areas where siloxanes or bleed are a concern. Common conductive fillers include alumina, boron nitride, and zinc oxide. The formulation—not the word “putty”—determines dielectric behavior.
Industry guidance from Laird on gap fillers also highlights low pressure, varying topography, silicone and non-silicone options, and the need to evaluate bleed and outgassing in sensitive equipment.
How to Apply Putty Across Components of Different Heights
Do not spread one arbitrary blanket over the whole PCB. Divide the layout into zones. Each zone should receive enough material to bridge its worst-case gap after the cooler reaches its intended final position.
For manual work, use small preforms, rolls, dots, or beads. Repeatable volume and coverage matter more than the shape. In production, a dispenser can control shot size, location, speed, and pattern. Keep material away from connectors, test points, screw bosses, optical paths, and other no-go areas.
The cooler is the final forming tool. Install it with the planned hardware and torque sequence. Do not press one corner down by hand and call the contact proven. Uneven fastening can create a false result.
Reopen early prototypes and inspect the witness pattern. Look for:
- a clear imprint on both mating surfaces;
- coverage over every intended heat source;
- no bare patch or trapped void;
- controlled side flow rather than a large spill;
- no board bend, cracked package, or shifted part;
- proper seating at the primary CPU, GPU, or power-module interface.
Watch the Shared-Cooler Trap
Here is a common GPU example. Putty is placed on the VRAM and VRM parts around the main die. The technician adds extra material “just to be safe.” Once the screws are tightened, that thick ring acts like a soft spacer. The cold plate no longer lands correctly on the GPU die. Memory contact looks great, but core and hotspot temperatures rise.
The same can happen in an ECU or telecom housing. Too much fill above a tall component holds the lid away from shorter, hotter devices. It changes the mechanical stack instead of improving cooling.
When bond line control is central to the problem, review how bond line thickness affects thermal performance before fixing a dispense volume.
Common Failure Modes and Practical Fixes
The first thermal run is only a starting point. If something looks wrong, use the symptom to work backward through geometry, material, and process.
| Symptom | Likely cause | How to check | Practical correction |
| One lower component stays hot | Too little material, a void, or a larger local gap | Open a prototype and inspect the imprint | Increase controlled volume in that zone or revise the pattern |
| CPU/GPU core runs hotter after assembly | Putty or pad is holding the shared cooler up | Check die paste imprint and cooler seating | Reduce excess material and repeat the specified torque sequence |
| Putty leaks or creeps | Overfill, unsuitable rheology, heat, gravity, or weak containment | Run hot storage in the real orientation | Choose a more stable grade, reduce volume, or add suitable containment |
| PCB bows | Putty is too firm or too thick; screw load is uneven | Measure board deflection during assembly | Use a softer material, correct the gap, or revise the fastener design |
| Performance drifts after cycling | Pump-out, bleed, dry-out, or loss of contact | Compare thermal results and interface images before and after aging | Change formulation or mechanical control and repeat reliability tests |
| Results vary by operator | Portion size or placement is not controlled | Weigh shots or inspect dispense dimensions | Add fixtures, preforms, work instructions, or automated dispensing |
Also watch for stacked mistakes. A slightly warped housing, excess material, and uneven screw torque may look like one “bad putty” problem. Often it is the combination.
Applications Across Electronics Industries
Mixed-height cooling appears wherever one surface collects heat from several parts.
- GPUs, laptops, consoles, and SSDs: Putty can cover memory, controllers, and power devices without several pad thicknesses. Direct-die contact still needs checking.
- AI servers and accelerator boards: Repeatable dispensing, low assembly force, and service access may matter as much as conductivity.
- Telecom equipment: Vibration, temperature swings, and vertical orientation make shape stability important.
- Automotive and power electronics: ECUs, ADAS units, BMS boards, inverters, chargers, and drives bring stricter dielectric and lifetime demands.
- LED and compact consumer devices: Clean assembly, bleed, and material compatibility become important near lenses and sensitive contacts.
If putty cannot meet the process or reliability need, the wider family of thermal conductive gap fillers includes other dispensable and preformed routes.
Standards and Validation Before Production Release
Coupon data is helpful. The finished assembly is the real test.
Measure temperature or interface impedance at the actual gap extremes. Test at realistic power, pressure, and orientation. ASTM D5470 is widely referenced for TIM measurements, but thickness, pressure, area, temperature, and method must be disclosed. Haktak’s overview of common TIM testing standards adds context.
Then run the reliability checks that match the product:
- thermal cycling and high-temperature aging;
- humidity exposure where relevant;
- vibration and mechanical shock;
- vertical slump and squeeze-out;
- pump-out, oil bleed, and dry-out inspection;
- dielectric testing before and after aging;
- removal, cleaning, and refill trials if the unit is serviceable.
If flame performance is required, ask for evidence for the exact formulation and tested thickness. UL 94 classifies material behavior under specified laboratory conditions. It does not certify that the finished device is safe in every end-use condition.
For products sold into the EU, request current substance documentation against the applicable RoHS Directive and REACH obligations. Do not assume that “halogen-free,” “RoHS compliant,” and “UL 94 V-0” mean the same thing. They answer different questions.
What to Send a Thermal Putty Supplier
A useful sample request starts with the assembly, not a shopping list of conductivity numbers. Send:
- a drawing or section view of the thermal stack;
- component locations and minimum, nominal, and maximum gaps;
- power loss, contact area, and target component temperature;
- cooler material, surface finish, and mounting method;
- the allowable force on components, solder joints, and PCB;
- electrical insulation and flammability requirements;
- silicone-free, low-bleed, or low-outgassing needs;
- manual or automated application method and expected cycle time;
- operating temperature, orientation, vibration, and aging profile;
- packaging preference, production volume, and rework needs.
This helps the supplier narrow the chemistry, softness, viscosity, and format before sampling. Haktak supports material selection and testing for electronics around real gap and reliability conditions.
Conclusion
Thermal putty works well with uneven component heights because it can form around local steps with relatively low pressure. But its flexibility does not remove the need for engineering control.
Map the assembled gaps. Separate the hot parts from the merely tall ones. Choose the material by thermal, mechanical, electrical, and process behavior. Then verify contact after the cooler is installed and repeat the check after aging.
Put another way, let the geometry choose the material. Not the biggest W/m·K number on the page.
Frequently Asked Questions
Can thermal putty handle components with different heights?
Yes. A suitable putty can deform around local height differences and fill variable spaces between several components and one cooler.
How much height difference can thermal putty fill?
There is no universal limit. It depends on formulation, unsupported span, orientation, pressure, and reliability conditions. Follow the qualified range and test both extremes.
Do I still need to measure the gap when using thermal putty?
Yes. Putty reduces the need for an exact sheet thickness, but volume, compression, and final bond line still depend on the minimum and maximum assembled gaps.
Can thermal putty replace several thermal pad thicknesses?
Sometimes. It can simplify a layout that needs several pads, but contact, pressure, electrical behavior, process, and aging still need validation.
Can too much thermal putty stop a heat sink from seating correctly?
It can. Excess material may hold the cooler above a CPU, GPU die, or shorter component. Check the final seating and witness pattern after a trial assembly.
Is thermal putty safe around VRAM, VRM, MOSFETs, and exposed contacts?
Only if the grade has suitable electrical properties. The product name alone does not prove dielectric strength or volume resistivity.
Should thermal putty be used directly on a CPU or GPU die?
Usually not. Direct-die joints are thin and tightly clamped, so grease or an approved phase-change material is normally more appropriate. Use putty only when the manufacturer approves that interface.
How can I tell whether every component touches the cooler?
Open a prototype and inspect both surfaces. Look for full coverage, consistent deformation, no bare zones, and correct seating at critical interfaces.
Does higher W/m·K always cool mixed-height components better?
No. A higher-conductivity material can perform worse if it is too thick, leaves voids, contacts poorly, or changes the cooler position. Judge the assembled thermal resistance.
What causes thermal putty to leak, slump, or pump out?
Causes include overfill, unsuitable rheology, heat, vertical orientation, repeated expansion, poor containment, and an uncontrolled gap.
Is silicone-free thermal putty better for sensitive electronics?
It may be better near optics, relays, contacts, or later bonding areas where silicone bleed is a concern. It is not automatically better in every other respect.
Can thermal putty be dispensed automatically?
Yes, if the grade and packaging suit the equipment. Validate shot size, bead shape, speed, placement tolerance, separation, and restart behavior.
Does thermal putty need to cure?
Many putties remain soft and do not cure. Some putty-like products crosslink after dispensing. Check the technical datasheet instead of relying on inconsistent naming.
Should putty be replaced after removing the heat sink?
For reliable work, replace material that is contaminated, displaced, hardened, or no longer forms a controlled interface. Reuse should follow the supplier’s guidance and the project’s risk level.
What tests should be run for automotive or industrial use?
Test thermal performance, cycling, heat aging, humidity, vibration, shock, slump, bleed, pump-out, dielectric behavior, material compatibility, and rework under the actual assembly conditions.
