What Was Happening
A customer in Hammersmith W6 got in touch about a desktop gaming PC — a Chillblast build with a Palit GeForce RTX card — that had just come back from a repair elsewhere for an unrelated CPU problem. That original fault appeared to have been sorted. What arrived back, though, was a machine that wouldn’t put anything on screen.
The symptoms were oddly encouraging. Press the power button and the PC came to life: fans spinning, case and component lighting on, no unusual noises, no burning smell, nothing to suggest something had died. But the monitor sat there reporting no DisplayPort signal, and on the motherboard the VGA debug LED stayed stubbornly lit.
They had already done some sensible investigating before calling us. Looking inside, they had noticed the graphics card was sitting loose in its slot — and when they gently pressed it back down, the card’s own lighting came on for the first time. Understandably, that read like progress. But the screen stayed black, and at that point they stopped rather than start forcing things, which was exactly the right instinct.
They asked whether a technician could come out and diagnose it, and were open to us taking the machine back to the workshop if it turned out to need more than a home visit could cover. We had a slot free that same afternoon and booked them in at £90 per hour.
Our Diagnosis
The two facts that mattered most were already in the customer’s description, and they pointed the same way.
The debug LED told us where to look. Most modern motherboards run a short power-on self-test and light a small LED for whichever stage fails — CPU, DRAM, VGA or BOOT. A VGA light that comes on and stays on means the board got past the processor and memory and then couldn’t establish communication with a graphics device. That narrows things enormously: the CPU and RAM are almost certainly fine, and the problem lives at the graphics card, its slot, or the link between them.
The lighting coming on was a red herring — a useful one. A graphics card draws power from two separate places: the PCIe slot itself, and the supplementary power leads that run directly from the power supply. The card’s fans and RGB run happily off those supplementary leads alone. So a card can look alive — lit up, fans turning — while its edge connector isn’t properly engaged in the slot. That is precisely what “I pressed it back in and the lighting came on” describes.
Taken together, the two symptoms told a consistent story before we touched anything: a card that was receiving power but not communicating. That’s a seating problem far more often than it’s a dead card.
We checked the cheap and reversible things first. The monitor and its DisplayPort cable were verified as working, and we confirmed the display was on the right input — worth ruling out, because it costs nothing and occasionally saves an hour. Then we opened the case and looked at how the card was actually sitting.
It wasn’t home. The card had been pressed down at the front, but the retention catch at the end of the slot had never engaged, and the card was sitting very slightly proud along its length — enough that the contacts weren’t consistently making. That final click is the tell, and it hadn’t happened.
The likely cause was straightforward. The machine had just travelled by courier for its CPU repair. Graphics cards are heavy, cantilevered off a single slot, and vibration over a long journey is exactly the thing that walks them loose.
How We Fixed It
The fix was mechanical and quick, but worth doing properly rather than pressing harder.
We disconnected power and removed the supplementary PCIe power leads from the card, released the retention catch at the end of the slot, and took the card out completely. That matters: reseating properly means out and back in, not pushing down on something already jammed at an angle.
With the card out, we checked the slot for debris and looked over the gold edge contacts for damage or corrosion. Both were clean — no bent pins in the slot, no scoring on the contacts, nothing to suggest the card had been damaged by being run half-seated.
We then refitted the card squarely, pressing evenly along its length until the retention catch clicked home, and secured the bracket to the case so the card’s weight was properly supported rather than hanging on the slot alone. The supplementary power leads went back on, fully clipped.
On the first boot after refitting, the VGA debug LED went out during the self-test and the monitor picked up a signal immediately.
We verified rather than assumed: several restarts from cold, checking the display came up every time, and testing both monitor inputs so the customer wasn’t left with a setup that only worked one particular way. We also confirmed the earlier CPU fault hadn’t returned once the machine could actually be seen on screen.
The Result
The customer had a working gaming PC back the same afternoon they contacted us — booked and finished inside a few hours, with about an hour spent on site. No parts were needed and nothing had to go back to the workshop, so the whole thing came to the standard onsite hour at £90.
Worth stating plainly, because it’s the honest outcome here: there was nothing wrong with the graphics card. It didn’t need replacing, and had they gone straight to buying a new one, they’d have spent several hundred pounds fixing a problem that was mechanical.
Why This Happens
Graphics cards are the most awkward component in a modern desktop. They’re often the heaviest single part in the case, and they hang off one connector with only a small bracket at the back taking any of the load. Modern cards are heavy enough that sag is a normal, expected thing rather than a defect.
That geometry makes them uniquely vulnerable to movement. A tower that stays put on a desk will generally keep its card seated for years. A tower that gets carried down stairs, put in a van, driven across London and carried up again is a different proposition — and one that’s just been couriered for an unrelated repair is close to a worst case. We see loose cards after house moves, after machines come back from someone else’s bench, and after any occasion where a desktop travelled on its side.
The reason this fault is so confusing to diagnose from the outside is the split power path described above. Almost every other component in a PC is either connected or it isn’t. A graphics card has two independent supplies, so it has a genuine halfway state: powered and visibly alive, but electrically invisible to the motherboard. Every intuition built on “is it getting power?” leads you astray here.
This is also why the debug LEDs on a motherboard are worth knowing about. They cost nothing to read and they tell you which stage of start-up failed, which is often more information than a black screen and a lot of guessing. If yours lights VGA and stays there, the graphics path is where to look — and reseating the card is the first thing to try, well before assuming it’s dead.
One caveat on doing it yourself: reseating a card is not difficult, but it does need the machine unplugged, the retention catch released rather than fought, and the card pulled fully out rather than levered. If a card resists coming out, something is still latched — forcing it is how slots get broken, and a broken PCIe slot is a motherboard replacement rather than a five-minute job.
Local Help in Hammersmith W6
If a desktop is powering on but showing nothing on screen — especially after it’s been moved, transported or worked on by someone else — it’s very often recoverable without new parts. We’ll tell you honestly if that’s what we’re looking at, and equally honestly if the card really has failed.
Hammersmith is well inside our callout area, and for a desktop that isn’t displaying, coming to you is usually quicker than moving the machine again. Book an onsite visit in Hammersmith or call 020 7610 0500 — mention the make of PC and whether any debug lights are showing on the motherboard, and we can often tell you what we expect to find before we arrive.
