Graphics Card Not Working: How to Fix Display Issues

You press the power button, the fans spin, and the monitor stays black. Perhaps Windows starts in the background, or perhaps the machine restarts when a game loads. The obvious conclusion is that the graphics card has failed, but a black screen alone doesn't prove that. A loose cable, incorrect display output, missing PCIe power, damaged driver installation, or BIOS setting can produce almost exactly the same symptom.

A professional repair bench works from the outside in, then separates software faults from physical faults before anyone recommends spending money on a replacement GPU. That approach matters in the UK, where available retailer evidence suggests that genuine graphics card failures are a much smaller part of the problem than online headlines can make them appear.

Table of Contents

• Recognising that your graphics card is not working - Why failure headlines need context

• Checking cables, power, and external displays first - Follow the display chain

• Ruling out software and driver problems - What to check in Windows - Use a clean driver reinstall

• Testing seating, power, and basic hardware diagnostics - Compare simple tests with proper diagnosis

• Deciding when to seek board-level GPU repair or replacement - What a Sheffield repair assessment should establish

Recognising that your graphics card is not working

The first sign is often dramatic. You switch on the PC and get no signal, coloured blocks appear across the screen, Windows reports that the display driver has stopped responding, or a game suddenly becomes a slideshow. Some faults are intermittent, with the computer working normally on the desktop before losing display output under load. Others appear as persistent artefacts, flickering, crashes, or a card that disappears from Windows and BIOS.

Those symptoms justify a diagnosis, not an immediate replacement. Overclockers UK told Digital Trends that its average return rate across popular graphics cards was about 3.5%, with roughly half of those returns classed as unwanted rather than genuine faults. For that retailer's sample, the implied fault-driven return rate was closer to 1.75%. The same discussion reported zero failures among RTX 2080 Ti systems from one established UK system builder, while another UK retailer said it wasn't seeing unusually high failure rates for AIC RTX 2080 Ti cards. The UK retailer discussion and reported return figures are useful context, although they aren't a universal failure rate for every card or computer.

Why failure headlines need context

Graphics card failure stories can reflect a particular launch period, supplier batch, or model rather than normal ownership. During the 2018 RTX 2080 Ti discussion, a forum survey of around 70 owners reported a failure rate above 55%, while a separate subreddit survey of more than 100 RTX 2080 owners reported 2.5% failures. Those figures weren't UK retail statistics, and the contrast shows why survey populations and return definitions matter. Historical coverage of GPU failure rates and warranty claims helps explain why a headline number shouldn't decide the diagnosis.

In practical terms, a card that produces a black screen but remains visible in BIOS or Device Manager may still have working core hardware. A card that displays corruption before Windows loads, vanishes from BIOS, or fails in a known-good test system is more concerning. The pattern matters more than the fear attached to the phrase graphics card not working.

Workshop rule: Treat “no display” as a symptom. Prove whether the monitor chain, configuration, driver stack, power delivery, or card itself is responsible.

Checking cables, power, and external displays first

Start without opening the case. Confirm that the monitor is switched on, set to the correct input, and connected to a port on the graphics card rather than a motherboard video output. If your PC has integrated graphics, the motherboard ports may be active only under certain BIOS settings, so moving a cable between outputs can change the result without fixing the underlying fault.

Remove and reconnect the HDMI or DisplayPort lead at both ends. Try another cable if one is available, then test a different port on the GPU. Adapters and docking hardware add another possible failure point, so bypass them temporarily. A television with a suitable input can work as a quick alternate display, while a second monitor can show whether the problem follows the PC or stays with the original screen.

Follow the display chain

Use this order:

1. Monitor input: Select the input manually instead of relying on automatic detection.

2. Cable path: Remove switches, adapters, extension leads, and capture devices.

3. GPU output: Test another physical port on the graphics card.

4. Alternate screen: Connect a known-working monitor or television.

5. Power connection: Check that the computer starts reliably and that every required PCIe power connector is attached to the GPU.

A card can light up or spin its fans and still lack the auxiliary power needed for proper operation. Fan behaviour isn't a reliable pass or fail test, because many modern cards control their fans according to temperature and may keep them stopped at idle.

If the PC has recently been moved, transported, cleaned, or upgraded, suspect disturbed connections before electronic failure. Check the mains lead, monitor power cable, and power supply connections without forcing plugs into place. If the system shuts down, clicks, smells hot, or shows signs of unstable power, stop repeatedly restarting it and consider whether the power supply needs assessment. Guidance on the wider issue is available in this power supply replacement guide.

A useful result at this stage is not necessarily a working screen. If a different cable and monitor produce the same symptom, you've narrowed the fault to the computer or its configuration. If the alternate display works, focus on the original screen, input selection, cable, or adapter rather than replacing the GPU.

Ruling out software and driver problems

Once the physical display path is confirmed, move to Windows and driver diagnosis. A corrupted graphics driver can leave the system apparently dead, especially after a Windows update, driver change, interrupted installation, or failed graphics application. A simple “update driver” over the existing installation often doesn't remove damaged files or conflicting components, so it may not change anything.

If Windows loads but the normal desktop is black, try restarting into Safe Mode. Safe Mode uses a minimal driver environment, which makes the result valuable. If the screen works there, the panel, cable, and basic display output are less likely to be the cause, and the normal driver stack becomes the main suspect.

What to check in Windows

Open Device Manager and expand Display adapters. Look for the graphics card name, a warning symbol, or a generic Microsoft display entry. The result isn't conclusive on its own, but it tells you whether Windows can communicate with the device and whether the driver reports an error.

Also check BIOS or UEFI. Look for the primary display setting and confirm that the system is configured to initialise the PCIe graphics card when that is the intended output. If the card is present in firmware but fails only after Windows starts, software remains a strong possibility. If firmware doesn't detect it after physical checks, the problem is more likely hardware or PCIe communication.

Use a clean driver reinstall

With the card physically fitted correctly, a clean reinstall using Display Driver Uninstaller, commonly called DDU, is usually the highest-value software step. Download the correct driver from the GPU manufacturer before removing the existing one, disconnect the PC from the internet if Windows might automatically install a driver, and run DDU from Safe Mode according to its instructions. Then install the appropriate driver for the exact model and operating system.

Don't install several driver packages over one another while the fault is still unclear. Record the original symptoms and test after each meaningful change. If the display returns after the clean installation, the GPU wasn't necessarily defective. The driver stack was preventing normal operation.

This software and driver troubleshooting video can help visualise the process, but take care with BIOS changes and driver removal. If you aren't confident recovering from a failed boot, stop before making the system harder to diagnose.

Independent UK repair guidance recommends starting with external causes, then checking PCIe seating, power connectors, BIOS display settings, and finally using Safe Mode with DDU for a clean driver reinstall. It also notes that failure to detect the card in BIOS or UEFI after these checks points more towards board-level diagnosis. The complete UK graphics card troubleshooting sequence mirrors the order used in a practical repair assessment.

Testing seating, power, and basic hardware diagnostics

If the driver route changes nothing, shut the computer down fully, switch off the power supply, and disconnect the mains lead. Press the power button briefly to discharge residual power, then remove the side panel and inspect the card. A heavy GPU can sag or shift in its PCIe slot, particularly if the case has been moved.

Release the PCIe retention clip before lifting the card. Don't pull at an angle or use force against the motherboard. Inspect the slot for dust, damage, or foreign material, then reinstall the card evenly until the bracket sits correctly and the retention clip engages. Reconnect every required PCIe power lead firmly. Modular power supplies should use the correct original cable for that power supply, not a cable borrowed from another unit.

Compare simple tests with proper diagnosis

A home bench test can answer useful questions, but it has limits.

Test	What it can indicate	What it can't prove
Reseat the card	Poor slot contact or movement	That the board is electrically healthy
Try another PCIe power lead	Cable or connection trouble	That the power supply is stable under load
Test another slot	Slot or motherboard communication issue	That the GPU's components are good
Test in another compatible PC	Whether the fault follows the card	The exact failed component
Run a controlled load test	Instability under demand	The location of a board-level fault

If the card works in another compatible system, investigate the original motherboard, power supply, BIOS configuration, and Windows installation. If a known-good GPU works in your computer, that comparison also shifts suspicion towards the original card, although power requirements and driver differences still need consideration.

Don't repeatedly stress a card that's producing severe artefacts, shutting the machine down, or becoming unusually hot. A controlled test is useful after basic stability has been established, not as a way to force a visibly failing board to continue operating. This GPU stress testing guidance is relevant once the system can display normally and the hardware connections have been checked.

Professional diagnostics go further than swapping parts. A repair centre can check whether the GPU is detected at the electrical and firmware level, inspect power rails, test memory and voltage regulation, and identify shorted or failed components. Microsoldering may resolve a damaged power circuit, memory fault, connector issue, or board-level break, but it requires proper equipment and board knowledge. A home reseat is low risk when done carefully. Component diagnosis is not a sensible extension of that same test.

Deciding when to seek board-level GPU repair or replacement

A graphics card becomes a strong hardware-fault candidate when it fails in a known-good computer, remains absent from BIOS or UEFI after correct seating and power checks, or produces repeatable artefacts before the operating system loads. At that point, another driver reinstall is unlikely to help. Continuing to swap cables can create confusion rather than new evidence.

Repair and replacement involve different trade-offs. Replacement is straightforward when the card is older, inexpensive, widely available, or still covered by a useful warranty. A specialist repair can make more sense when the card is valuable, difficult to source, matched to a custom system, or showing a fault that can be isolated to replaceable circuitry. A technician may test VRAM, power delivery, the PCIe interface, and individual board components before deciding whether repair is technically and financially sensible.

A general repair provider may decline a non-working GPU because it doesn't perform component-level work. For comparison, this graphics card diagnostic service from Nerds 2 You Edmonton shows the kind of specialist assessment customers should look for, namely diagnosis before a repair recommendation rather than an automatic replacement quote.

What a Sheffield repair assessment should establish

Ask for clear answers to these points:

• Detection: Does the card appear in BIOS, Windows, or a test system?

• Reproduction: Can the technician recreate the black screen, crash, or artefact?

• Fault area: Is the issue linked to power delivery, memory, the PCIe interface, or another board section?

• Repair route: Are genuine, compatible components available, and is the work economical?

• Outcome: Will the card be tested under normal and controlled load before collection?

A local specialist such as Steel City IT can assess graphics cards, perform component-level diagnosis, and handle GPU or VRAM repair where the fault is suitable. Its GPU repair service information is relevant if you're in Sheffield and the card has already failed the basic checks. Turnaround depends on diagnosis, parts, and board complexity, so a few days is a reasonable practical expectation rather than a guarantee.

Don't authorise a replacement until you know whether the card is defective and whether repair is viable. The best next step is the one supported by evidence: restore the driver if software is responsible, correct the installation if the card was not seated or powered correctly, or obtain board-level testing when the fault follows the GPU.

---

Steel City IT provides PC and laptop diagnostics, graphics card and VRAM repair, upgrades, and custom gaming PC support for customers in Sheffield. If your graphics card still shows a black screen after the checks above, visit Steel City IT to arrange a practical assessment before committing to replacement hardware.