BRENTWORTHBRENTWORTHMail-in electronics repair
Search the store
AccountCart
Overview

AMD Radeon RX 6000 Series GPU Repair

The AMD Radeon RX 6000 series — built on RDNA 2 architecture and released from 2020 through 2022 — spans from the RX 6950 XT and 6900 XT at the flagship end through the 6800 XT, 6800, 6750 XT, 6700 XT, 6650 XT, 6600 XT, and 6600. These cards introduced AMD Smart Access Memory support, significantly improved ray tracing performance over RDNA 1, and are now three to five years into service — old enough for thermal interface materials and fan bearings to show meaningful degradation.

The most common repair presentation for high-end RX 6000-series cards is load-dependent artifacting caused by degraded thermal pads on the GDDR6 memory chips. These cards use a large memory array with thermal pads contacting both the heatsink on the front and the backplate on the rear. As those pads age and compress, individual chips lose adequate heat transfer and run above safe junction temperatures under load — producing artefacts, crashes, and instability that closely resembles hardware VRAM failure but is actually a thermal management problem. This is one of the most cost-effective GPU repairs Brentworth handles: a full pad replacement at the correct thickness typically resolves the fault entirely.

The dual-BIOS switch present on most RX 6000 AIB designs also makes BIOS corruption events more recoverable on this generation than on platforms where no backup BIOS exists.

Ready to start?

Use the intake form to describe the device, the fault, and the result you want. The more specific you are, the easier it is to give you a useful answer.

Start Repair
Common Issues

What we fix

Memory overheating and thermal pad degradation on RX 6900 XT and 6800 XT

RX 6000-series flagship cards use GDDR6 memory chips distributed across the front of the PCB, with thermal pads transferring heat to both the front heatsink and the rear backplate. These pads, when new, keep junction temperatures within an acceptable range during sustained GPU workloads. After several years of thermal cycling, the pads compress, dry out, and lose their thermal conductivity — causing individual GDDR6 chips to spike in temperature under load. The result is artefacts, screen corruption, driver crashes, or hard system resets that occur specifically during demanding workloads and disappear at idle. A full pad replacement — including the backplate-side pads that many previous repair attempts overlooked — resolves this in the majority of cases.

BIOS corruption and dual-BIOS switch recovery

RX 6000-series AIB cards nearly universally include a dual-BIOS switch that selects between a primary performance BIOS and a quieter secondary BIOS. When the primary BIOS is corrupted by a failed driver update, a botched flash attempt, or a power event during initialisation, the card may produce no display output or fail to be recognised by the system. Switching to the secondary BIOS position frequently restores operation, after which the primary chip can be safely reflashed. Cards where both BIOS chips are affected require hardware programmer recovery.

No display output and POST failure

RX 6000-series cards that produce no display output may have a BIOS corruption issue, a display output circuit fault, VRAM damage from prolonged overheating, or a connector-related power fault. The dual-BIOS switch is always worth trying first on cards with no display, as it eliminates the most common recoverable cause before physical inspection is needed. Cards where the switch does not recover display output warrant bench testing to determine the actual root cause.

Fan failure and cooler wear on AIB designs

RX 6000-series AIB cards from Sapphire, ASUS, PowerColor, XFX, and MSI used a wide range of dual and triple-fan cooler designs. Fan bearings on cards three to five years old are showing wear — grinding noise, intermittent stalls, or fans that stop under load are the typical signs. A stalled fan on a card already affected by degraded thermal pads is particularly problematic, as both failures compound to produce extreme temperatures. Fan replacement and pad service are often addressed together on RX 6000-series cards showing both symptoms.

PCIe power connector damage

High-end RX 6000-series cards — the 6900 XT, 6800 XT, 6800, and 6950 XT — use dual 8-pin PCIe power connectors. Damage from forced insertions, cable stress, or years of connection cycles can produce intermittent power delivery that causes crashes under load. The RX 6950 XT at 335 watts is particularly sensitive to marginal power delivery quality. Connector condition is assessed as part of any diagnostic process on these models.

Common Questions

Frequently asked questions

My RX 6900 XT or 6800 XT artefacts under load. Is it a VRAM fault?

Almost certainly not, at least not initially. Load-dependent artifacting on RX 6000-series flagship cards is one of the most common repair presentations Brentworth sees for this generation, and the cause is overwhelmingly thermal pad degradation on the GDDR6 memory chips rather than failed chips. The pads on both the heatsink-facing and backplate-facing sides of the memory array need to be assessed and replaced — many prior repair attempts that only addressed the GPU die paste while leaving the memory pads untouched failed to resolve the issue. A correctly executed full pad replacement resolves the majority of these cases. If artefacts persist after pad service, VRAM diagnosis is the next step.

What thickness thermal pads do RX 6000-series cards need?

Pad thickness requirements vary by card model and the specific component position — memory chips, VRMs, and inductors may all require different pad thicknesses on the same card. Using the wrong thickness is one of the most common reasons a pad replacement attempt does not resolve the issue: pads that are too thin leave a gap and provide no thermal transfer; pads that are too thick can prevent the cooler or backplate from seating correctly and apply uneven pressure. Brentworth uses measured reference data for each supported card model to ensure correct pad selection.

My RX 6700 XT or 6600 XT is also artifacting. Does it have the same pad issue as the 6900 XT?

RX 6700 XT and 6600 XT cards have different cooler and memory array configurations compared to the 6900 XT and 6800 XT, and the thermal pad degradation pattern is less severe on these mid-range models. That said, any RDNA 2 card now three or more years old can develop thermal instability from degraded TIM materials. Artifacting on a 6700 XT is more often caused by GPU die paste degradation than memory pad failure, though both are assessed during diagnosis.

My RX 6900 XT shows no display. The dual-BIOS switch did not fix it. What next?

If switching to the secondary BIOS did not restore display output, and the card shows no signs of life on either BIOS position, the fault is likely hardware rather than firmware. The most common hardware causes of complete no-display on an RX 6900 XT are a VRAM fault from prolonged overheating, a display output circuit failure, or a board-level component failure. Bench testing in a known-good system is the next step to determine whether the card produces any output at all or whether it is completely non-functional. Describe the full history of the card — including whether it had prior artifacting — in the intake form.

Is the RX 6950 XT worth repairing?

The RX 6950 XT is the fastest RDNA 2 card and carries 16GB of GDDR6, which means it retains meaningful value for gaming and professional workloads. Thermal service, fan work, and connector repair are generally justified on a 6950 XT. Board-level microsoldering is evaluated on a case-by-case basis depending on the fault and extent of damage. Brentworth will assess viability before any chargeable work begins.

Some RX 6000-series cards have a quiet BIOS and a performance BIOS. Does the BIOS position affect repair?

The dual-BIOS switch on most RX 6000 AIB cards selects between a performance mode BIOS and a quieter, lower-power BIOS. For repair purposes, either BIOS position can be used for bench testing — the diagnostic results are valid regardless of which profile is active. If you have been running the card on the secondary BIOS due to a corrupted primary, note this in the intake form so the technician understands the card's current state.

How long does RX 6000 series thermal pad replacement take?

Full thermal pad replacement on a 6900 XT, 6800 XT, or 6950 XT involves complete disassembly, cleaning of all pad and paste contact surfaces, application of new pads at the correct thickness to each position on the memory array and VRM components, reassembly, and load testing to confirm the repair. Most straightforward thermal service cases complete within a few business days of the card arriving. Cases where the thermal assessment reveals additional faults — damaged chips, connector issues, or prior repair complications — take longer.

Is there a diagnostic fee for RX 6000 series GPU repair?

Yes. Brentworth charges a non-refundable diagnostic fee that is credited toward the repair cost if you proceed. For high-end RX 6000-series cards, thermal pad assessment is an early step in the diagnostic process, since this resolves the majority of artifacting cases before more invasive diagnosis is needed.

Ready to start?

Use the intake form to describe the device, the fault, and the result you want. The more specific you are, the easier it is to give you a useful answer.

Start Repair

Related services

Specialist Work

GPU's

Diagnosis-first GPU repair for no display, artifacting, overheating, fan failure, and damaged connectors.

View all services