How To Choose Thermal Pad Thickness For VRAM And VRM Cooling

Thermal pads look simple, but selecting the wrong thickness can turn a graphics card repad into a hotter, noisier system. The pad must bridge the gap between a memory chip, voltage regulator component, and heatsink while compressing enough to create reliable contact. Thickness, softness, thermal conductivity, and surface fit all matter together.

VRAM and VRM areas are particularly sensitive because their components rarely sit at exactly the same height. A graphics card cooler may use several pad sizes across the PCB, while a motherboard heatsink can cover MOSFETs, chokes, and controller components with different clearances. A single thickness across every location is rarely the correct approach.

For Australian PC builders, replacement pads can also be an expensive exercise once GST, shipping, and limited local stock are factored in. A pad kit ordered from overseas may take weeks to arrive, while a premium silicone sheet from a local retailer can cost more than expected. Measuring first is cheaper than buying several unsuitable packs.

The safest method is to identify the original pad thickness, confirm the gap with measurements or contact marks, and choose a pad that compresses properly. This guide explains how to make that decision without relying on guesswork or assuming that a higher W/mK rating solves every cooling problem.

Why Thickness Matters More Than Conductivity

A thermal pad transfers heat only when both surfaces make broad, even contact. If the pad is too thin, it may fail to touch the heatsink across part of the component. Air pockets then act as insulation, raising VRAM or VRM temperatures even when the product claims excellent thermal conductivity.

An oversized pad creates a different problem. It may hold the cooler away from the GPU core or prevent the heatsink from sitting flat. On a graphics card, that can increase core temperatures, reduce mounting pressure, and cause uneven contact with the die. A pad advertised at 12 W/mK cannot compensate for a physical gap that is incorrectly filled.

Softness and compressibility are important because the printed circuit board, heatsink, and components have small height variations. Two pads with the same nominal thickness can behave differently if one compresses easily and the other remains stiff. Thermal putty can be useful for irregular gaps, but conventional pads are usually easier to control on broad, flat memory packages.

Identify The Components Under The Cooler

VRAM chips are the rectangular packages arranged around the GPU die. GDDR6 and GDDR6X memory can produce substantial heat during gaming, rendering, and memory overclocking, so their pads usually contact a dedicated section of the cooler. The required thickness may differ between the front and rear of the card.

VRM cooling usually covers the MOSFETs or power stages near the GPU power input, often in a long row beside the core. Some coolers also contact inductors, commonly called chokes, but others leave them uncovered or use a different interface material. Do not assume every component beneath a large pad needs the same treatment.

The card’s backplate may use pads for rear-mounted memory, PCB hotspots, or structural heat spreading. A backplate pad can be thinner or thicker than the front-side VRAM pad because its gap is determined by the board layout and backplate shape. Treat every contact zone as a separate measurement rather than applying a universal thickness.

Find The Original Gap Without Guesswork

Manufacturer service manuals, teardown photographs, and community repad reports can provide a useful starting point. Search for the exact graphics card model and cooler revision, because a factory-overclocked version or revised board may use different components. Hardware news and repair coverage can also reveal cooler changes, including details found through current hardware news.

Before removing anything, photograph the cooler and mark where each original pad sits. Measure the untouched pad with digital callipers, but remember that an old pad may have compressed permanently or expanded slightly after years of heat cycling. Its measured height is evidence, not absolute proof of the original specification.

A more reliable approach is to make a soft imprint. Place a small piece of modelling clay, thermal putty, or a carefully controlled non-conductive material over the contact area, reinstall the cooler with normal screw pressure, then measure the compressed result. The process must avoid contaminating the PCB and should be performed only when the cooler can be removed safely.

Understand Pad Compression And Tolerance

Thermal pads are commonly sold in thicknesses such as 0.5, 1, 1.5, 2, and 3 mm. The labelled figure describes the uncompressed sheet, while the working thickness depends on its hardness and the pressure available from the cooler. A 1 mm pad may compress significantly; a firm 1 mm pad may barely close a slightly larger gap.

For VRAM, the best thickness is generally the thinnest pad that fully contacts both the memory package and cooler after assembly. For VRM components, the correct size may need to accommodate height differences along a long power stage row. Excessive compression can put mechanical stress on small components, while insufficient compression leaves hotspots.

Thermal pad sheets can be cut with a sharp craft knife and a steel ruler. Keep protective films on the material until the pieces are ready to install, and avoid stacking pads to create a custom thickness. Layering introduces extra interfaces and can make the assembly unstable, especially when a cooler uses spring-loaded screws.

Choose Materials For The Heat And The Budget

A mid-range silicone pad with consistent thickness is often a better choice than an extremely high-conductivity product that is hard to compress or difficult to cut cleanly. For VRAM and VRM use, reputable pads in the approximate 6–12 W/mK range are commonly adequate when the contact is correct. The label should be treated as a comparison point, not a guarantee of real-world performance.

Very soft pads can spread around components and lose shape during installation. Very firm pads may fail to conform to minor surface irregularities. Some specialist products use phase-change behaviour or putty-like consistency, which can suit uneven VRM arrays but may be messier to remove during future maintenance.

Australian buyers should compare the final cost in Australian dollars, including delivery and GST. A product available from a Melbourne, Sydney, or Brisbane PC retailer may arrive faster and simplify warranty issues, even if its headline price is higher than an overseas listing. Avoid generic sheets with unclear specifications from marketplace sellers when the component is a valuable graphics card.

Match Memory And Cooling Goals

Thermal pads help move heat away from the package, but they do not determine memory performance by themselves. GDDR6 modules can tolerate substantial operating temperatures, yet lower temperatures may improve stability and overclocking headroom. Memory type also affects tuning behaviour; background reading on Hynix A-die and Samsung B-die illustrates why silicon characteristics matter alongside cooling.

GDDR6X cards deserve particular attention because their memory can run hotter than many GDDR6 designs. If a card has a history of high memory junction temperatures, correct pad contact should be verified before simply choosing a thicker or higher-rated product. Improving case airflow, cleaning dust from the heatsink, and setting a sensible fan curve may deliver gains that a premium pad cannot.

Room temperature also changes the result. A gaming PC in a warm Perth room or an unairconditioned Brisbane summer office has less thermal margin than the same system in a cool Melbourne winter. Compare temperatures at similar ambient conditions, and record GPU core, memory junction, and hotspot readings where the card exposes them.

A Practical Installation Checklist

Careful installation prevents many thickness-related failures. Remove residue with high-purity isopropyl alcohol and lint-free materials, but avoid scraping components or forcing a cooler away from the PCB. Cut each replacement piece to cover the intended contact area without overlapping nearby solder points.

Use this short checklist before tightening the cooler:

• Record the location and thickness of every original pad.
• Check whether the replacement material has protective films on both sides.
• Confirm that each piece covers the entire memory or VRM contact surface.
• Tighten screws gradually in a cross pattern.

After assembly, inspect temperatures at idle and under a repeatable workload such as a game benchmark or GPU stress test. A sudden rise in core temperature, unstable boost clocks, or a memory junction result worse than before can indicate that the cooler is being held away from the die.

Do not use thermal paste as a substitute for a large missing pad gap. Paste is designed for thin, tightly controlled interfaces and will pump out or fail when asked to bridge several millimetres. Thermal putty is the more suitable option for an irregular gap, provided it is electrically safe and compatible with the manufacturer’s materials.

Warning Signs Of The Wrong Thickness

The symptoms of poor pad selection can appear immediately or after several gaming sessions. High VRAM temperature with a normal GPU core temperature often points to incomplete memory contact, while a hotter core after repadding can indicate that the cooler is no longer sitting flat. A pad that has shifted during installation can produce inconsistent readings across workloads.

A useful diagnostic list includes:

• GPU core temperature rises sharply after the repair.
• Memory junction temperature remains unusually high.
• Fan speed increases while performance or boost clocks fall.
• The cooler rocks or leaves visible gaps during a contact check.

Disassemble only after the card has cooled fully, and inspect the pad impressions. A clear, even imprint across the component indicates contact. A shiny untouched patch or a narrow edge mark suggests that the pad thickness, placement, or compression is wrong.

The goal is a stable interface rather than the thickest possible material or the highest number printed on the packet. When the pad matches the real clearance, the cooler remains flat, VRAM and VRM heat can escape efficiently, and the graphics card has a much better chance of delivering consistent performance through long Australian gaming sessions.