Graphics card errors and how to fix them

The graphics card is one of the hardest-working parts of any gaming or workstation PC — and one of the most common to fail. The good news: most GPU problems are repairable, and some you can sort yourself.

Artifacts and visual glitches

Strange colours, flickering textures, lines or blocky patterns are "artifacts". They usually mean overheating, ageing memory (VRAM), or a driver problem. Start by updating (or cleanly reinstalling) your graphics drivers, and check the card's temperature under load.

Crashes and black screens while gaming

• Overheating — dust-clogged fans or dried-out thermal paste
• An underpowered or failing power supply
• Driver conflicts after an update
• A loose card or power cable

Cleaning dust from the card and case fans fixes a surprising number of "random" crashes.

Loud, grinding or dead fans

GPU fans wear out. A grinding fan is on its way out; a dead fan means the card will overheat fast. Fans are replaceable — you don't need a whole new card.

No display at all

If you get nothing on screen, it isn't always the GPU — it can be the monitor cable, the motherboard, RAM or power supply. Try the motherboard's own display output (if it has one) to narrow it down.

"Buy a new one" is often the wrong advice. Fans, thermal pads and even VRAM can be repaired at board level for far less than a replacement card.

When to get it looked at

If cleaning and driver reinstalls don't help, or the card runs hot, artifacts or won't display, we repair graphics cards at board level across London — fans, thermal work and component-level fixes, usually the same day.

How to tell a GPU fault from a driver fault

Not every graphics problem is a dying card. A surprising number of "GPU faults" are actually driver conflicts — especially after a Windows update or a driver that half-installed. Before assuming the worst, it's worth doing a clean driver reinstall: use a tool like DDU (Display Driver Uninstaller) in safe mode to strip the old driver completely, then install the latest version fresh. If the artifacts or crashes vanish, it was software. If they survive a clean driver install, the fault is more likely in the hardware itself — and that's where we come in.

Overheating: the most common cause we see

By far the most frequent reason a graphics card starts crashing, throttling or artifacting is heat. Over a few years, dust builds up in the fans and heatsink, and the thermal paste between the GPU chip and its cooler dries out and loses its grip. The card then can't shed heat fast enough, so it either throttles down (killing your frame rates) or shuts the system down to protect itself. A proper clean and a fresh thermal repaste often brings a "failing" card completely back to life — a fraction of the cost of replacing it. If your GPU idles fine but crashes under load, or runs much hotter than it used to, heat is the first thing to check.

When a card is genuinely worth repairing

People are often told a faulty graphics card means buying a new one. That's frequently wrong. Fans are replaceable. Thermal pads and paste are renewable. Even some board-level faults — failing VRAM, power-delivery issues — can be repaired at component level by someone with the right equipment. A mid-range or high-end card that's only a couple of years old is almost always worth diagnosing properly before you write it off, and we'll give you an honest view on whether a repair or a replacement makes more sense for your specific card.