My monitor shows no signal

First, the message is better news than it sounds: your monitor is working. A monitor that can display the words “no signal” has power, a functioning panel, and a working input chain — it is simply reporting what it is not receiving: a video signal on the input it is currently watching. That means the break is somewhere between the monitor's input port and the PC's graphics hardware, and it is one of a short list of places: the cable, a port, the input selection, or the PC's graphics side — the GPU, its output, its driver, or a PC that powered on but never reached the point of producing a picture at all.

One distinction drives everything that follows: are the PC's fans spinning and its lights on? If yes, the PC has power and the graphics side is the suspect (driver, output, or a POST failure). If the PC itself is dead, that is a power problem wearing a display costume — and no monitor-side testing will ever fix it. “No signal” does not mean the monitor is broken; a genuinely failed monitor usually shows nothing, no message, not even the standby LED behaving normally.

Work the path in order — monitor, input, cable, port, source — because each step has a cheap test, and the tests are arranged so that every one of them either fixes the problem or removes a suspect from the list. The isolation articles near the end show how the results of those tests prove which component owns the fault, so you never buy a part on a guess.

Before anything else, confirm the monitor is actually on rather than in standby, because the two states matter: “no signal” with a lit LED means the monitor is powered and waiting for video; a dark screen with no LED at all means the monitor has no power, which is a different problem that will look identical once you get past it. Check the LED behaviour — on most monitors a solid light means on and a blinking light means standby or no signal, and the manual's front page usually lists the exact pattern, which is worth thirty seconds of reading.

If there is no power at all, the chain to check is short and mundane: the outlet (prove it is live with something else), the power cable, the power brick if the monitor uses an external adapter (a dead brick looks exactly like a dead monitor), the rear power switch — many monitors have a physical rocker on the back that people forget exists — and finally the power button itself, which is sometimes a small, unlit button rather than an obvious control. One of those five is the answer in essentially every no-power case.

If the monitor powers on cleanly, do not stop here — a powered monitor is only the first check on the path, and the message it is showing is the clue for the next one. Move on to the input selection, which is where the next most common cause lives, especially if anything was recently plugged in or moved.

The full power cycle is the single highest-success fix for “it worked yesterday,” and it takes about a minute. Shut the PC down properly (not sleep — sleep keeps all the stuck state you are trying to clear), turn the monitor off, unplug the PC's power cable, and wait thirty seconds so the capacitors on the motherboard and power supply drain fully. That drain is the point: a wedged graphics state, a failed link negotiation on a DP or HDMI connection (those connections re-train every power-up, and a bad training result can latch), a half-applied driver state, or a confused dock all reset when the hardware loses all power, not just when the operating system restarts.

Reconnect, give the PC a full minute (it must complete POST before any display exists), and test. Expect “no signal” to appear briefly while the PC boots and then clear when the display initializes — that is normal, not the problem recurring. The difference between a restart and a power cycle is worth internalizing: a restart re-runs the operating system on top of the same hardware state, while a power cycle resets the hardware state itself, which is why it clears things a restart cannot.

If the problem survives a clean power cycle, do not repeat it five times — repetition only tells you the stuck-state theories are wrong. Move to the input and the cable, which is where the next most likely causes are, and keep the power cycle in your pocket for the dock and the multi-monitor situations, where wedged middleman hardware is a live possibility again.

The most common cause of “no signal” in all of computing, and the most embarrassing: the monitor is watching the wrong input. Monitors with multiple inputs (HDMI 1, HDMI 2, DisplayPort, USB-C) frequently do not auto-switch when you plug something in — “auto-detect” works reliably until the day it stops, and it has a way of stopping precisely when you need it. If you recently changed cables, changed ports, or connected a second device, the monitor is very likely still watching the input you just vacated, showing a perfectly honest “no signal” for a cable that is no longer there.

Find the physical controls — usually a joystick or a row of small buttons on the bottom edge or side of the screen (the manual tells you which, and the buttons are often labelled only with icons) — and open the menu's Input, Source, or Signal option. Cycle through every input once; on the correct one, the message disappears. Some monitors remember the last active input per port, and a fresh power-up of the source sometimes re-triggers detection on its own, which is why a reboot “fixed” it for people who never found the input menu.

If cycling every input changes nothing, the monitor is doing its job and the break is in the cable, the port, or the source — continue down the path. One habit worth forming if you use two inputs regularly (a work PC on DisplayPort, a home PC on HDMI, for example): label the cable ends. A wrong input is a five-second error that routinely costs an hour of real diagnosis.

HDMI is the most common connection in existence and one of the most common failure points, for a mundane mechanical reason: the connector is a press-fit with no latch, so a cable can sit in both ends, look fully connected, and not actually be making contact on the pins. Pull each end out completely and push it back in until it seats with firm, even pressure — you should feel it bottom out. Then note the second HDMI-specific fact: cables have bandwidth generations. An HDMI 1.4-era cable tops out well below what a 4K high-refresh display wants; 2.0 covers the modern mainstream; 2.1 goes further. A “no signal” that appears only when you raise the resolution or refresh is very often a bandwidth ceiling, not a dead cable — and that has its own article below.

The working order for an HDMI path: re-seat both ends, try the monitor's other HDMI port (ports are independent and one can be dead while another works), test a different cable — ideally one proven on another device, because a “spare” that never worked is not a control — and check the adapter if you are using one. Adapter direction matters: an active adapter can drive an HDMI display from a DisplayPort source, but the reverse direction is not what most passive “HDMI to DisplayPort” cables do, and buying the wrong direction is a classic self-inflicted no-signal.

Useful corollary: if the same PC displays fine over DisplayPort but not over HDMI, suspect the HDMI port or that cable before you suspect the computer — HDMI is the statistically weaker link on most displays. And if “no signal” appears exactly when you bump the resolution or refresh rate, stop treating it as a connection problem and read the resolution and refresh article instead.

DisplayPort behaves differently from HDMI in two ways that matter for diagnosis. First, its connector has a latch — so if a DP cable has come loose, it has come fully loose, and the failure mode is less “barely touching” and more “definitely not connected.” Second, and more subtly: a DisplayPort connection requires the source and the monitor to train a link — a negotiation of lanes and bandwidth — every time the connection is established, and a failed or half-completed training produces no picture at all. The classic fix is almost embarrassing in its simplicity: unplug the cable at the monitor end for ten seconds and plug it back in, which forces a fresh training attempt. That single re-plug resolves a surprising share of DP no-signal cases.

The working order: seat the cable fully and make sure the latch clicks, force a re-train with the ten-second re-plug at the monitor end, try a different DP port on the monitor, and check the adapter situation if you are crossing connector types. The rule to memorize: a DisplayPort source can drive an HDMI display through a passive adapter, but an HDMI source cannot drive a DP display passively — that direction needs an active adapter, and a lot of “DP to HDMI” cables in circulation are one-way only, oriented the wrong way for the job they are being asked to do.

One DP-specific situation to flag now: some GPUs — older desktop cards and many laptop GPUs in particular — limit how many displays a single port can drive, or disable one output when another is in use. If your first monitor works and a second DP monitor says no signal, that is a multi-monitor configuration problem, not a cable problem, and the dedicated article below covers it. If DP fails on every port, every cable, and every source you have, the monitor's DP input is the suspect.

USB-C is where “no signal” hides behind an assumption, and the assumption is that all USB-C does video. It does not: some USB-C ports carry data and charging only (many on older laptops, and data-only ports exist on desktops too), and some USB-C cables carry only charge or only data, despite being physically identical to a video-capable one. If your “video” connection over USB-C shows no signal while the same cable happily charges a device, the leading suspect is not a fault at all — it is a cable or a port that never did video in the first place, and no amount of re-plugging will change that.

How to tell what is capable: video-capable ports are usually marked — a lightning-bolt icon (Thunderbolt) or a display symbol (a small rectangle, sometimes labelled with “DP” for DisplayPort alt mode) next to the port — but the manual is the definitive answer, and it takes less time to check than to guess. Then run the tests in this order: a known video-capable cable in a known video-capable port; a direct connection first, bypassing any dock (a dock is an entire extra failure point — its own article below); and a check of whether the operating system sees the display at all (Windows + P on Windows, the display menu on a Mac). If the OS lists the monitor, the video path is physically working and the problem is configuration; if it does not, the path is broken and the cable/port capability question is first in line.

Thunderbolt docks bundle power, data, and video into one chain, which means a wedged or failed dock presents itself as “monitor no signal” even when the dock's monitor port is fine and the direct connection would work. The direct-connection test is what separates “the dock is the problem” from “the monitor is the problem,” and it is the single most useful test in any docked setup. And if USB-C video never works with any cable or any device, the port may genuinely be data-only — in which case the fix is a different port or a different connection type, not a repair.

These two connections are still common on office PCs, older monitors, and projectors, and they fail in their own characteristic ways. VGA is analog — it carries a signal the monitor has to lock onto — which makes it the most resolution-sensitive connection in this list: a mode the monitor cannot display tends to show as “no signal” or “out of range” rather than a degraded picture, and a marginal VGA connection degrades in ways a digital one simply does not. DVI is digital in most of its common forms, but the type matters: DVI-D carries digital only, DVI-I carries both, and a 2560×1600 display over DVI needs a dual-link cable — a single-link cable in a dual-link port is a silent no-signal at high resolutions.

The mechanical checks are the same as everywhere else, plus the specifics: a VGA cable is held by two screws, and stripped, missing, or loose screws mean the connector is held by friction alone — a loose VGA connection is the single most common VGA failure, and the fix is tightening or replacing the cable. For adapters, direction is everything: getting an HDMI display from a VGA source is not something a passive cable does at all, and HDMI-to-VGA conversion requires an active converter that works one way only (digital source to analog display). If your setup involves a converter box, that box is a full extra component on the failure list, and bypassing it with a native connection is the cleanest test.

The good news with legacy connections is that the isolation logic is identical to the modern ones: another cable, another port, another source, in that order. If the monitor displays fine from the same PC over a modern connection, the legacy port, its cable, or its adapter is the confirmed weak point — and on older hardware that is usually the honest answer, worth a cable replacement rather than deeper investigation.

The another-cable test is the cheapest test in this entire tutorial and the one people skip the longest — most would rather open the computer than borrow a cable. It deserves to be done early because a bad cable is among the most common causes of “no signal” that survive every other check. Cables fail in quiet ways: an internal pair breaks (on a USB-C cable, the data pair can die while the charging pair works, so the cable “still works” for everything except video), a pin bends, a connector wears so it no longer fully seats, or a cable is simply below the bandwidth of the mode being requested. All of those present as no picture, and all of them are invisible until a known-good cable proves the point.

Do it properly: use a cable of the same connector type as the path you are testing (an HDMI replacement for an HDMI path — switching connection types tests two variables at once and muddies the result), and ideally one you have used successfully on something else, because a spare that never demonstrably worked is not a control, and “two cables both fail” does not mean the cable is cleared. Plug it in, give the connection ten seconds (DisplayPort will re-train), and test. New cable works: the old cable is the confirmed cause and you are done. New cable fails: the cable is cleared, and the suspect list shrinks to the ports, the input, and the source.

One nuance that makes cables the top suspect in “it worked yesterday” cases: a cable that works sometimes — after wiggling it, after a re-plug, only at certain bend angles — is a failing cable even if it passes a clean test. Failing connections are intermittent by nature, and intermittent is exactly the signature that points at a worn connector or a broken pair rather than a configuration, which would fail consistently. If your no-signal comes and goes, the cable test is not optional.

Monitors usually carry two or more inputs, and the one you are using might be the dead one. Monitor ports fail less often than cables do, but they do fail — a bent HDMI tab, a worn DisplayPort latch, a port whose detection circuitry has given out — and a dead input port produces exactly the same “no signal” as a dead cable, which is why the port test earns its place in the sequence. The test is quick: if the monitor has a second port of the same type (two HDMI ports, or an HDMI and a DP when you have cables for both), move the cable to it and select that input.

The result carries real diagnostic weight, because it distinguishes cases: the same “no signal” on every port of the monitor means the ports are not the cause, and the suspect moves to the cable, the input selection, or the source — one less thing to suspect is progress. A different port works means the original port is the confirmed cause, and the fix is to live with the working port, or to weigh repair against replacement if the working port is inconvenient (a DP-only port on a monitor whose PC output is HDMI, for example). Note that the ports are independent — a monitor can have one dead input and three healthy ones, and nothing about the rest of the monitor's behaviour will hint at it.

Run the same logic on the source side too, symmetrically: if the PC has two video outputs available (GPU plus motherboard, or two outputs on the GPU itself), moving the cable to the other output is the identical test in reverse. On a desktop with a discrete graphics card, that symmetric test is the wrong-output check below — and it is the one that catches the most common desktop cause of all, so if you have not done it yet, do it before anything more exotic.

This is the test that clears or condemns the monitor, and it is the most valuable test in the whole tutorial because it is free: connect a different device to it. A laptop, a game console, a streaming stick, a neighbour's PC — anything that produces a video signal. If the monitor displays that other device, the monitor is fully cleared: its power, its panel, and its input chain all demonstrably work, and the fault lives in your PC or in your PC's cabling. You have just converted “something is broken” into “the PC side is broken,” which is half the diagnosis done with zero cost.

Do it with a small method note: use the same connection type as the failing one where you can (if the suspect path is HDMI, test over HDMI), because the monitor's ports are independent and one can be dead while another works — testing a suspect HDMI input with a DisplayPort device proves the monitor works but does not prove the HDMI port works. Pair this test with its mirror image (the PC on another display, in the isolation section) and between the two you can usually pin the fault to one of exactly three places — monitor, cable, or PC — without opening the case or installing anything.

The other outcome is equally useful: if the monitor shows no signal with every device you try while its power LED behaves normally, the monitor's input electronics or its panel is the suspect, and the monitor-failure-indicators article below is what to read before spending a cent. If the monitor is a few years old and the failure followed a power event — a surge, a storm, a cheap strip that failed — a monitor mainboard failure is a realistic possibility, and it is a repair-versus-replace decision about the monitor, not a problem with your computer at all.

The classic desktop trap, and the single most common cause of “no signal” on a PC that was recently built, upgraded, or moved: the video cable is plugged into the motherboard's video ports — the vertical cluster near the USB ports — while a discrete graphics card is installed. On most systems, installing a discrete GPU disables the motherboard's outputs entirely; and on systems whose CPU has no integrated graphics at all (all AMD “F”-series processors and many others), the motherboard ports produce no signal under any circumstances. So a cable in the motherboard ports gives you a perfectly honest “no signal” even when the GPU is brand new and perfectly healthy — the cable is simply on a port that is not connected to anything that can display.

The fix takes thirty seconds: move the cable to the graphics card's ports — the horizontal ports lower on the back of the case, physically attached to the card rather than to the I/O shield cluster. If you have both a motherboard and a card, the card's ports are the ones to use, full stop. If you do not have a discrete card (a laptop, an APU system, a typical office PC), then the motherboard or laptop port you are using is the correct one, and the cause is elsewhere on the path — the good news is that this check is a two-second visual inspection, so it is cheap to rule out on every desktop you touch.

A second variant of the same idea, for completeness: on multi-port GPUs, the outputs are not always equal — some cards group ports by internal engine, and in certain configurations (a second monitor in use, an eGPU, a specific driver mode) one output can be disabled while the others work. If one output on your card is dead but another gives a picture, that is a port-level matter; if every output on the card is dead, that is a card-level one — seating, power, or the card itself — and the GPU articles below are where that branch goes.

One PC can contain two graphics engines — the CPU's integrated graphics (the iGPU, part of the processor) and a separate graphics card (the dGPU) — and a large share of “no signal” confusion comes from not knowing which engine is actually driving the display. The working rules: with a dGPU installed, most systems route the display through the dGPU and disable the iGPU's outputs (some BIOSes let you change or dual-enable this); with no dGPU, the iGPU on the motherboard ports is the only display path that exists; and on CPUs that have no integrated graphics at all, there is no iGPU path to fall back on — a system built around one of those is completely dependent on its card, which is why a card failure on such a system looks like a dead PC.

Why this matters diagnostically, in both directions. Direction one: if your cable is in the motherboard ports and the iGPU is disabled — or does not exist — you get “no signal” from a perfectly working GPU, which is the wrong-output case above. Direction two, the more powerful one: a working iGPU is a built-in diagnostic tool. When the dGPU's outputs are all dead, moving the cable to the motherboard ports and getting a picture is direct evidence that the graphics card (or its seating, or its power) is the confirmed cause — you have proven the rest of the system can display without buying, borrowing, or guessing anything. That iGPU test is the backbone of the GPU-failure diagnosis below.

Laptops add a third arrangement worth knowing: many run both engines — the dGPU for games and the iGPU for everything else — and switch between them, sometimes per-application. A failed switch, or a dGPU that has stopped initializing, can present as “external monitor no signal” while the built-in screen (driven by the iGPU) works perfectly, which makes laptop display problems feel inconsistent. That is the laptop article's territory. For desktops, the takeaway is simple: before any deeper diagnosis, establish the facts — is a card installed, does the CPU have integrated graphics, and which ports does each engine own? Knowing that turns half of the remaining sections from puzzles into checklists.

Reseating the GPU is the primary physical check inside the PC, and it is the one you reach for when the software-side testing has pointed at the card: no display from any of its outputs, after the cable and port tests have been done, especially after the system has been moved, dropped, or upgraded. What it addresses is mundane and common: a card that has shifted in its slot (the PCIe connection is a friction fit, and heavy cards sag on their retention bracket over months), a PCIe power connector — the 6- or 8-pin feed from the power supply — that has come partially unseated, or a slot contact that wants a clean break and remake. A card that is 95% seated can power its fan and still not initialize, which is exactly the “PC turns on, monitor says no signal” state.

The procedure, briefly: remove the side panel; unclip the PCIe power connector(s) from the card; release the slot's retention clip at the rear of the case; lift the card straight out; and reinsert it with firm, even pressure until the clip clicks. Reconnect the power connector(s) until they click. While the card is out, look — inspect the slot for debris, look at the card's gold contacts, and check the power connectors for bent pins, which is a visible cause of no-initialize that a reseat alone will not fix. The variants that extend the test: if the board has a second PCIe x16 slot, try it (rules the slot in or out); and if the CPU has integrated graphics, test with the card removed and the cable on the motherboard ports (rules the card in or out, definitively).

Two notes for recent builds, because they change the likely cause: a brand-new card in a brand-new build that never displayed is more often a power problem than a seating one — check that every PCIe connector the card's documentation lists is actually connected, because a card that needs an 8+8 pin feed on a system where only one feed is plugged in will not initialize; and a card that worked, stopped, and now needs reseating to work is telling you the slot or the bracket retention has worn, which is worth addressing (a bracket support or a stiffer case) rather than relearning every few months.

The scenario that this article exists for: the PC's fans spin, the case lights are on, the monitor says no signal, and nothing — not a BIOS screen, not an error, not Windows — ever appears. That is a system failing before it has initialized a display, a POST failure, and the single most common POST failure with that exact symptom is memory. A system with no working RAM does not produce a video signal at all: it is powered, it looks alive, and it has no picture, which is precisely “my PC turns on but the monitor has no signal.” The reason RAM matters is that the processor validates memory very early in the boot sequence, and if validation fails, the sequence stops — with the fans still running, because the fan power does not depend on it.

The evidence, when the board gives it to you: many modern boards carry four small diagnostic LEDs (commonly power, CPU, memory/VGA, and boot) that stay lit on the stage where the boot is stuck — a steady light on the memory indicator is the board telling you directly, in hardware, what this article is describing. Older boards beep instead (repeated short beeps or a long beep are classic memory errors — the board's manual maps the exact pattern to the cause, and that mapping is the diagnosis). If your board has neither LEDs nor a speaker, the manual's troubleshooting table is your map, and the reseat procedure below is your test either way. The same power-safety warning as the GPU article applies: shut down, unplug, drain, then work.

The test sequence, cheapest first: reseat every stick — RAM is a friction fit with clips on both ends, and a stick that is not fully clipped home is, without exaggeration, the most common RAM problem in the world; then run one stick at a time in the board's preferred slot (usually the second slot from the CPU, per the manual), moving sticks between slots as you go. The results read cleanly: a system that POSTs with stick A in slot 2 but not stick B has a failing stick; one that POSTs in slot 2 but not slot 1 has a failing slot; one that POSTs with neither stick alone but does with both (rare, but real) has an interaction problem worth a firmware update. This is the cheapest hardware diagnosis in computing, and on a “PC on, no display” case it is the answer more often than the GPU is.

Graphics cards rarely fail cleanly, and the failure usually arrives in stages — which is what makes the diagnosis possible, because each stage is evidence. The early signs are behavioural: visual artifacts under load (colored blocks, lines, corruption that follows the GPU's output), driver crashes (the screen flickers black and returns), and the operating system reporting that the display driver has stopped responding and recovered. The late signs are the absence: no output on any of the card's ports with any cable, output only at low resolutions, or the system no longer detecting the card at all. A card that has progressed from the first set to the second set over weeks is a card that is failing, and the no-signal state is its endpoint, not a random event.

The diagnostic evidence that points at the card specifically — and only at the card — is a short list, and each item on it is a test you can run: no display on any of the card's outputs, with different cables, after a full power cycle (which rules out the stuck-state and link-training explanations); the CPU's integrated graphics displays fine when the cable is moved to the motherboard ports (which rules out the monitor, and proves the system can display); the failure happens before Windows, so the BIOS screen itself is absent on the card's outputs (which rules out the driver); and, where Windows does load, Device Manager lists the GPU with a warning or not at all. Any one of those is a strong signal; the set together is a confirmation, and it is the difference between “probably the card” and “the card, with evidence.”

What not to do: do not declare the card dead on a single no-signal event — a seating problem, a missing power connector, or a failed link all mimic a dead card, and the reseat plus iGPU test is what separates them, which is why those articles come first. On the repair question: a failed GPU is a replace decision, or an RMA if it is under warranty — third-party GPU repair exists but is rarely worth it outside warranty, and if the CPU's iGPU is available it can carry the system in the meantime. If the card is covered, the evidence gathered above — which ports, which cables, the iGPU test result, when the behavioural stage started — is exactly what the warranty form will ask for, and having it in hand is what turns a multi-week RMA into a two-week one.

The symmetric question to the GPU one: when is the monitor itself the confirmed cause? The clean version of the evidence: the monitor shows “no signal” — or nothing — with every device you connect to it, a laptop, a console, a streaming stick, while its power LED behaves normally. At that point the monitor's input chain or its panel is the fault, your PC is cleared, and no amount of PC-side troubleshooting will change the outcome. That is the monitor-failure state, and it is reached only through the every-device test, which is why that test is the keystone of the isolation section.

The failure modes worth naming, so they are not misdiagnosed as PC problems: a monitor that shows its messages but accepts no input (a failed input board — the “no signal with everything” case); a monitor with a picture that carries a dead zone, a line, or a flicker (panel or internal connection trouble — a different symptom from no-signal, but the same family); a monitor that only wakes after a cold power-on (capacitor aging, which behaves exactly like a “power cycle fixes it” problem, which is a clue in itself); and the backlight case — a picture that exists but is so dim you can only see it by shining a torch at the screen, which is a failed backlight and, on some panels, a repairable fault. Of these, only the first is the no-signal case this tutorial addresses; the others are monitor-repair symptoms, and knowing their names is what stops a panel problem from being “solved” by replacing a healthy GPU.

The decision when the every-device test has condemned the monitor is a repair-versus-replace calculation, and the honest version is simple: repair is worth considering for a recent, expensive display and for specific fault types (the backlight case, a single dead port on a multi-port unit); for a monitor several years old, replacement is usually the correct arithmetic, and a mid-range panel will outlive a repair. Before you spend anything, though, run the every-device test properly — with at least two different devices and, where possible, a different connection type — because the most expensive misdiagnosis in this tutorial is a monitor “failure” that was actually one bad HDMI port, and every device you tried happened to use it.

The first software question is the one that sorts the remaining cases in half: does Windows even see the monitor? The test is Windows + P, which opens the projection menu, and Settings → System → Display, which lists the connected displays. If the monitor appears in that list but shows no signal, the video path is physically working end to end and the problem is a display mode or a configuration — read the resolution-and-refresh article, because that is where this branch goes. If Windows does not list the monitor at all, the operating system is not receiving a signal from the hardware, and the fault is below the OS: the cable, the port, the GPU, or a system that never reached a state where a display exists at all. That single check — “listed or not” — is worth doing early, because it tells you which half of the tutorial applies.

The projection states matter more than most people realize, and they are a genuine source of no-signal on laptops and dual setups. “Second screen only” turns the built-in display off — a laptop that wakes into that state looks like a dead laptop to anyone glancing at the lid, and an external monitor that is supposed to be primary can appear to have dropped out. “PC screen only” disconnects the external — the most common “second monitor no signal” that follows a laptop waking from sleep, and the fix is one keystroke. “Duplicate” forces both displays into a shared mode, and a display that cannot do the requested mode shows no signal while the other one is fine — which is the mode-conflict case in the multi-monitor article. If your no-signal is intermittent and tied to wakes, lock the projection state you actually want and watch whether it stops happening.

The “Detect” button in Display Settings re-queries for newly connected displays; it is worth one click after plugging in a display, though a healthy connection does not need it — if Detect finds nothing, the signal is not arriving and the button is not going to change that. And the reverse case is instructive: a monitor that Windows does list but that still shows no signal is almost always a mode problem (resolution or refresh beyond what the path delivers) or a marginal cable that carries the display's identity data but not the picture — both of which have their articles above, and both are cheap to fix, which is the point of having sorted the case first.

There is a two-second fix for a wedged graphics driver, and it is the fastest thing in this tutorial: press Windows + Ctrl + Shift + B. The screen goes black for a moment and you hear a beep — that is the reset completing, not a failure, and it is the intended sound. What it does is tell Windows to restart the display driver in place: the driver state, the current display mode, and the connection's software side are torn down and rebuilt, while nothing permanent is touched. No settings change, no files are written, nothing is at risk — which is exactly why it belongs at the top of the software half of the diagnosis, before Safe Mode and before any uninstall.

When to reach for it: the moment a screen freezes while the PC is still responsive (the cursor moves, other windows work, the machine is clearly alive); the moment a monitor goes black after a resolution change and then reports no signal; and as the first software test whenever a no-signal appeared out of nowhere on a system that was working minutes ago. If the reset brings the display back, note what preceded it — a driver that wedges regularly is a driver that needs a proper reinstall (the Safe Mode article is where that happens), and the pattern of the wedges — under load, after a wake, on a specific display, at a specific resolution — is the clue to which of those it is.

Equally important is what the reset does not fix, so the result is read correctly: a cable, a port, a dead GPU, or a system that never POSTs are all below the driver, and if Windows + Ctrl + Shift + B beeps and the no-signal remains, the driver state was not the problem. That is not a wasted two seconds — it has removed an entire category from the suspect list — but it means the path now runs toward the hardware side: ports, cable, output, card, and the POST checks. The reset is a sort, not a cure, and a good diagnostic uses it for exactly that.

A monitor shows “no signal” for display modes it cannot do, and this is more common than it should be because the failure is silent: when you — or Windows, or a freshly installed driver — request 2560×1440 at 144 Hz over a cable that is rated for less, or 4K at 120 Hz on a port that is wired for 60, the monitor simply does not show a picture. There is usually no “out of range” warning, no error, just the no-signal message, which is why it gets misdiagnosed as a connection problem so often. The tell that identifies this case: the display works at a lower resolution or a lower refresh rate and dies precisely when you push the setting up. If your no-signal is reproducible at one setting and absent at another, you are here.

The variables are three, and the fix is finding the highest mode all three of them agree on: the monitor's maximum (its specification, or its on-screen menu), the port and cable's bandwidth (HDMI 1.4 versus 2.0 versus 2.1, DisplayPort 1.2 versus 1.4 — each generation has a hard ceiling), and the GPU's output capability. The working method is deliberately conservative: set a known-safe mode first — 1920×1080 at 60 Hz is the universal safe start — confirm a picture, then step one variable up at a time (resolution first, then refresh) until the display drops. The step at which it drops names the limiting link: a cable that cannot do the mode gets replaced, a port that is capped gets worked around with a different port or connection type, and a monitor that is genuinely limited at that setting is simply accepted, with the setting backed off one notch.

Two related cases worth finishing here. First, the multi-monitor variant: when displays are duplicated, Windows forces them into a common mode, and a display that cannot do the requested mode goes to no signal while its neighbour is fine — the fix is extended mode with independent resolutions, or a mode both can do, and it is covered in the multi-monitor article. Second, the post-install case: a fresh graphics driver sometimes applies a mode the new hardware cannot deliver on a given port, and the clean reset is to boot into a basic-driver state, set the resolution manually to a safe value, and only then let the full driver take over. Both are the same underlying rule: the mode must fit every link in the path, and the no-signal message is the path's way of saying it does not.

Safe Mode is the driver problem's detector, and it works by subtraction: it boots Windows with a basic display driver and minimal services, so the difference between what displays in Safe Mode and what displays normally tells you whether the fault lives in the driver stack. Getting there: Settings → System → Recovery → Advanced startup → Restart now, then Troubleshoot → Advanced options → Startup settings → Restart, and press 4 (or 5, if you need networking) when the options list appears — or hold Shift while clicking Restart in the Start menu, which takes the same road faster. Then answer the one question this test exists to answer: does the display appear in Safe Mode?

The two outcomes, and what each one means. A picture appears in Safe Mode but not in normal mode: the hardware path — cable, port, GPU, panel — is proven fine, because it is carrying a display right now, and the installed graphics driver (or something it loads) is the confirmed cause. The fix is to uninstall the driver while in Safe Mode: Device Manager → Display adapters → your GPU → Uninstall, ticking the option to delete the driver software where it is offered, then let Windows fall back to the basic display adapter, and install a clean driver from the GPU manufacturer's site — or use a driver-cleaner utility if the uninstall leaves residue. No picture in Safe Mode either: the problem is below the driver — hardware (cable, port, card, or a system that never POSTs) or firmware — and the Safe Mode detour has still paid for itself, because it has ruled the entire software half of the tutorial out in one boot.

Two practical notes. First, the update-driven no-signal has a direct road through here: when a driver update broke the display, Safe Mode is where the uninstall happens and the previous driver version is where the fix comes from — the after-update article below covers that sequence end to end. Second, some systems with a failed discrete GPU cannot reach Safe Mode over the dead outputs at all, because there is no display to see the boot menu on; in that case the iGPU test (move the cable to the motherboard ports) is how you get a display in the first place, and it doubles as the GPU diagnosis. Safe Mode is not a fix — it is the instrument that tells you which half of the machine to work on.

The boundary question this article answers: is there a display before Windows? If the system shows the BIOS screen — or at least the manufacturer's splash — and then fails at Windows, the problem is in the operating system's half of the machine and the BIOS is not involved. If the system powers on and never shows anything before Windows, the problem is a POST failure, and the BIOS layer is where the evidence lives. The BIOS does two display-relevant jobs: it selects the primary display engine (integrated versus discrete graphics, on boards that offer the choice), and it is the first thing in the boot sequence that should produce a picture. A no-signal that stops at POST is a hardware-state problem — RAM, CPU, GPU, or the primary-display setting — and it is categorically not a driver problem, because no driver has loaded yet.

The evidence to look for, in order of usefulness: the board's diagnostic indicators — the four-LED debug display or a two-digit code, with the memory stage being the most common no-POST cause (the RAM article covers that test); the primary-display setting, where a BIOS configured for an engine that is not driving your cable (PEG versus iGPU versus Auto, on boards that expose it) produces a no-display with everything else healthy, and the fix is the setting or the cable on the matching output; and the CMOS clear — a jumper, a button, or removing the coin cell for a minute, per the board's manual — which resets the BIOS to defaults and is a legitimate step for a wedged configuration. The honest note on the CMOS clear: it resets everything the BIOS stores (boot order, memory timings, sometimes the clock), so you re-apply those afterwards, and a clear that changes nothing has told you the BIOS was not the cause, which is a useful result in its own right.

What this article is not: a general “try changing BIOS settings” step. The BIOS is where you go when the evidence points there — a no-display that stops at POST, a primary-display mismatch, a configuration that wedged after a change — and for the majority of no-signal cases it is the boundary marker between “this is a computer problem” and “this is a PC-inside problem.” The LEDs and codes are what tell you which side you are on, and reading them is a five-minute investment in the manual that saves an hour of guessing. If your board has no indicators at all and the system simply never POSTs after a change you made, the change is the suspect list: revert it (the after-upgrade article's method) before you go anywhere else.

A laptop's external display carries two variables a desktop never has. The first is port capability: laptops vary wildly in which of their USB-C ports actually carry video — some are data-and-charge only, and only the Thunderbolt-marked or “DP alt mode” ports drive a display, and the difference is invisible to the eye and fatal to the connection. The second is the projection state: Windows + P does more work on a laptop than people realize, because “second screen only” turns the built-in display off (a closed-lid scenario that looks like a dead laptop) and “PC screen only” drops the external (the most common “my external monitor went to no signal after the laptop woke”). Both variables are configuration, both are cheap to check, and both are checked before anything is suspected of being broken.

The working order for a laptop external that has gone to no signal: the built-in screen first — if the laptop's own display is also dark, that is a laptop power or display problem and the external's no-signal is a side effect, so stop and solve that one. The port second — try the laptop's other USB-C or Thunderbolt port, or its HDMI output if it has one, because a port that has stopped carrying video is a real thing and a different port clears it. Direct connection third — bypass any dock, because a dock is a separate failure point with its own article below, and a docked no-signal that is fixed by going direct is the dock, not the monitor. The projection state fourth — Windows + P, choose Duplicate, confirm both displays, and leave it there until the problem is solved, because an intermittent state change is a confounding variable. And the driver last — a laptop's graphics stack is usually hybrid (an Intel or AMD iGPU plus an NVIDIA dGPU), and a wedged switch between the two after an update is a classic “external monitor no signal” with a perfectly healthy built-in screen; the driver reset and the Safe Mode articles apply to laptops exactly as they do to desktops.

One situation to name because it produces a very specific signature: the lid. If closing the lid is set to “sleep” (the default on many systems), the external monitor goes to no signal the moment the lid closes — not because anything failed, but because the machine went to sleep and stopped driving it. The setting lives in the power options under “when I close the lid,” and setting it to “do nothing” is what makes a laptop-plus-external desk setup behave. If your external “stops working” precisely when you close the lid, that is the answer, and it takes ten seconds to fix.

A docking station or a KVM switch inserts a middleman between the PC and the monitor, and a middleman is a new failure point with its own power, its own firmware, its own cables, and its own ways of wedging. A no-signal that appears through a dock has a specific diagnostic signature, and it is the one that makes docks worth an article of their own: the direct connection works, the docked one does not. When you have that signature, the problem is not “the display” — it is one specific link in the chain (the PC-to-dock cable, the dock's power, the dock-to-monitor cable, or the dock itself), and the job is to find which one, which is the same isolation logic as the rest of the tutorial applied to a shorter chain.

The procedure, in the order that catches the most cases: power-cycle the dock — unplug both its power and its upstream cable, wait thirty seconds, reconnect; a dock is a small computer and it wedges the same way computers do, and the cold restart clears a share of “dock no signal” cases outright. Check the upstream cable second — the PC-to-dock cable must be one that carries video (a Thunderbolt or DP-alt-mode cable; a charge-only USB-C cable to a dock is a silent no-signal, and the dock's other features may still work, which makes it easy to miss). Check the dock's power third — some docks need their own supply and run degraded, or with no video, when it is missing or underpowered. And run the direct test — bypass the dock entirely; if direct works, the dock chain owns the fault and the three checks above find the link, and if direct fails too, the dock was an innocent bystander and the problem is upstream of it.

KVM switches add a dimension docks do not have: mode. A KVM that is switched to the wrong PC shows no signal on the display you are watching (the signal is going to the other machine's output); a KVM that needs a re-plug sequence — power off the KVM, switch, power on — will not pick up a new source by button press alone; and a KVM whose video specification is below what you are asking will fail exactly when you raise the resolution or refresh, which is the bandwidth case wearing a KVM costume. The isolation principle is unchanged and worth repeating because it is the whole method: remove one link at a time and watch whether the problem follows it — the link it follows is the link that is failing.

“Second monitor no signal” is a category of its own because it has causes that a single-monitor setup never encounters. The first is the mode conflict: in duplicate mode, both displays are forced into a common resolution and refresh, and a display that cannot do the requested mode shows no signal while its neighbour carries on — the fix is extended mode with independent settings per display, or a common mode both can do, and this is the single most common second-monitor cause. The second is the GPU's display budget: some cards — and many laptop GPUs in particular — support only a specific number of displays or a specific combination of ports (two at a time, or “one of this port plus any other,” per the card's documentation), and a third display, or an unlucky pair, goes to no signal with every cable healthy. The third is daisy-chaining attempted on hardware that does not support it: DP daisy-chaining works only from a port explicitly marked DP-Out on the upstream monitor, and a regular monitor port will not chain a second display no matter what cable you use.

The working order for a second monitor that refuses: the mode first — switch to extended and give each display its own resolution; if the second one appears, the mode conflict was the cause and you are done, which happens more often than the other causes combined. The port second — move the second monitor to a different output on the GPU, because the display budget is often per engine and a different port clears it. The cable third — a fresh cable for the second display, because a marginal cable that carries one display fine can fail to carry two, and a daisy-chain (if that is the intended topology) needs the right cable and the right marked port. The dock fourth, if one is in the chain — a docked second display adds the dock's display budget on top of the GPU's, and the dock's limit is often the tighter one. The results read cleanly: if the second monitor works alone but not alongside the first, the cause is almost always the budget or the mode, not a failing part — and that distinction matters, because it is the most common reason people replace a perfectly healthy monitor on a false diagnosis.

The confirmation, when you suspect the display budget: check the GPU's specification for its maximum display count and port combinations (it is in the card's datasheet, and the manufacturer's support pages list the combinations for laptop GPUs specifically), and count what you are asking it to do. If you are asking for three on a two-display card, the no-signal is the card being honest, and the fix is a different combination — or a second output source — not a repair. Multi-monitor problems are configuration problems wearing hardware costumes far more often than the reverse, and the order above is arranged to prove that before any hardware is suspected.

A no-signal that appeared after a hardware upgrade is, in the majority of cases, a connection problem rather than a failed part — and the order of suspicion follows whatever was touched. A new GPU: is it fully seated, are all of its PCIe power connectors actually connected (a card that needs an 8+8 pin feed on a system where one feed is missing will not initialize, and will present as no-signal with the fans running), and is the video cable on the card's ports rather than the motherboard's? A new or added RAM: is every stick fully clipped home, and does the board's manual list the combination as supported — because an unsupported memory combination is a POST failure, and a POST failure is a no-signal? A new monitor or cable: is the connection type and the requested mode within both devices' limits? Each of those is a two-minute check, and each of them is more common than the “the new part arrived dead” case, which is why they come first.

The diagnostic shortcut for upgrade cases is revert, one component at a time, and it is the cleanest binary test in the tutorial: put the old GPU back — if the old one works and the new one does not, the cause is the new card's connection, its power, or the card itself, and the reseat-and-power checks are next; pull the added RAM — if the system POSTs, the cause is the new stick or the combination; return to the old cable or monitor — if the display returns, the cause is the new cable or the new monitor's port or mode. Each revert changes exactly one variable, so the revert that changes the outcome names the cause, and that is worth far more than replacing a working part on a guess. The method is the same one the isolation section uses, just applied in reverse chronological order, because in upgrade cases the change is the most likely cause and the newest component is the most likely link.

Two upgrade traps worth naming because they masquerade as dead parts. The first is the power supply: a PSU that is under-rated for the new GPU lets the system power on, spin its fans, and never initialize the card — a no-signal with no other clue, no error, and no fix short of a larger supply or a smaller card, and the specification check (the card's power recommendation against the PSU's rated output and connectors) is the five-minute test. The second is the firmware: a new CPU on an older board sometimes needs a BIOS update before the board will recognize it, and a no-POST on a new CPU is a firmware problem, not a dead board — the board's support page lists the firmware requirement for each CPU generation, and a bridging method (a BIOS flash USB, where the board supports it) exists for exactly this situation. Both are “it worked until I upgraded” cases that look like a dead component and are not one.

A no-signal that appeared after a Windows update or a graphics driver update is, in most cases, a driver problem wearing a hardware disguise — and that is the good news, because driver problems are the cheapest kind to fix and the most reversible. The pattern to recognize: the system was fine, an update installed (a Windows Update, a GPU driver push, a “recommended” update from a manufacturer utility), and the display fails — often at boot, often only at a specific resolution, sometimes with the two-second driver reset temporarily bringing it back. The temporary fix is the clue: if Windows + Ctrl + Shift + B restores the display and it fails again later, the driver state is involved, and the driver stack is the suspect until proven otherwise.

The sequence, in the order that keeps it reversible: the driver reset first, because it is free and instant and tells you whether the driver state is the trigger. Safe Mode second (the Safe Mode article), to get a display running on the basic driver while the broken one is out of the picture. Uninstall third: Device Manager → Display adapters → the GPU → Uninstall, with the option to delete the driver software ticked where offered, which removes the updated driver rather than just disabling it. And the restore fourth — and this is the step that distinguishes an update case from a generic driver case: install the previous driver version from the GPU manufacturer's archive rather than the one the update just pushed, because reinstalling the same broken version is not a fix. The manufacturers keep their version histories on their support pages, and the version that predates the update is the one to install; if the problem clears on the old version, you have both the fix and the evidence for a support ticket about the new one.

The Windows-side variant is worth separating: a feature update or a Windows Update can change the display stack or “correct” a display setting (a resolution, a refresh rate) to one the hardware path cannot deliver, which presents as a mode problem rather than a driver-crash problem — the resolution-and-refresh article covers that branch, and the reset is a known-safe resolution set from a basic-driver state. And the honest framing for the whole situation: if a no-signal appeared after an update and the hardware tests (another cable, another port, another display) all pass, the update owns the cause until proven otherwise. The driver path above is not just the fix — it is the proof, and a no-signal that survives the full hardware isolation and clears on a driver rollback has been diagnosed end to end.

The mirror image of the monitor test, and the one that clears or condemns the PC's graphics side: connect the PC to a different display. Another monitor is ideal, but an HDMI TV is a universal test display and is usually closer than a second monitor — a TV's own “no signal” message is as informative as a monitor's, and its input selection works the same way. If the PC shows a picture on the other display, the PC's graphics side is cleared: the GPU, the output, the driver, the POST — everything between the power supply and the cable is proven fine — and the fault is in the original monitor, the original cable, or the original port. That single test converts “something is broken” into “the display A chain is broken,” which is the entire diagnosis in one sentence.

Run it with one method note that keeps the result clean: use a different cable if you can. A cable shared between the two tests is a shared variable — a bad cable fails on both displays and muddies the result, so the clean version of the test is a different display and a different cable, which isolates the PC from the entire original chain rather than just the original monitor. Give the new display a moment to initialize (DisplayPort re-trains, and TVs can take a few seconds to switch inputs), and account for the TV's own input selection, because a TV set to the wrong HDMI number shows the same message for the same reason as a monitor. The result table below maps each outcome to its cause.

The pair, done together — PC on another display, monitor with another device, both over a known-good cable — isolates the fault to one of exactly three places (monitor, cable, or PC) without opening the case or installing anything. That is the payoff of the entire isolation section, and it is the reason the tests are worth the ten minutes they take: they replace a guess with a location.

The whole tutorial reduces to one question, and it is worth stating plainly: which component does the problem follow? The isolation tests — another cable, another port, another display, another device — are designed so that the answer is always one of a small number of places, and once you have it, the fix is obvious. The logic is the same as every swap test in the guide: a fault that moves with a component when that component is swapped is that component; a fault that stays behind when a component is swapped is not. There is no interpretation step — the result either follows the swap or it does not.

The five outcomes, and what each one names: (1) another cable fixes it — the cable was the cause, and the case is closed with a replacement; (2) another port on the monitor fixes it — the monitor's port was the cause, and the fix is the working port (or a repair/replacement decision if that port is the inconvenient one); (3) another display fixes it — the original monitor's chain is the cause, and the every-device test decides between its port and the monitor itself; (4) another device works on the monitor — the monitor is cleared and the PC owns the fault, which sends you to the PC half of the guide; (5) nothing fixes it — the fault sits in the component all the tests share, which for a desktop is the GPU or a system that never POSTs (the RAM article), and for the monitor side is the panel or input board (the monitor-failure article). Every no-signal case ends in one of those five rows.

The discipline that makes the method work has two rules, and both are about test quality rather than test quantity. Change one variable at a time — a test that swaps the cable and the port simultaneously has two suspects and therefore no answer. And use known-good replacements — a “spare” cable that has never demonstrably worked is not a control, and a test against it proves nothing, which is why the guide keeps returning to “a cable proven on another device.” Do the isolation before you replace anything, because a no-signal that turns out to be a fifteen-dollar cable is the most expensive misdiagnosis in this tutorial, and the isolation sequence is specifically how you avoid it. The checklist at the end of the page is this exact sequence, in order, ready to work through.

Knowing when to stop is the same skill as knowing how to continue, and the rule is simple: you stop troubleshooting when the isolation tests have named the failing component and the fix for that component is a repair or a replacement rather than a setting. The signals that you are at that line are specific, and each one corresponds to a test result earlier in the guide: a GPU that fails every output test and is confirmed by the iGPU test (a replace or RMA decision, not a settings hunt); a monitor that fails with every device while its power behaves normally (a repair-versus-replace decision about the monitor); a board that does not POST even with known-good RAM and after a CMOS clear (a board or power-supply problem — a technician's domain, not a home-repair one); and a power supply that cannot carry the system's load (a replacement, full stop). In each case, the troubleshooting has done its job: it has converted “no signal” into a component and a category of fix.

What a technician will ask you — and why doing it first is worth an hour of their time: which cables and ports have you tried; does the system display in Safe Mode; does it display on another display; does the integrated graphics display work; what changed immediately before the problem started; and do the board's diagnostic LEDs show a code. A no-signal case that reaches a technician with those answers in hand is diagnosed in minutes; one that arrives as “my monitor doesn't work” is diagnosed in hours, and the difference is entirely the work this guide exists to get done beforehand. The sequence above is the list of answers, in the order a technician would ask them, which is not a coincidence — it is the order the evidence accumulates most efficiently.

The cost sanity check, stated once and meant to be remembered: do not replace a part to cure a no-signal until the isolation tests have pointed at it. A two-hundred-dollar GPU bought to cure a fifteen-dollar cable is the specific failure mode this entire page is organized to prevent, and every test in the guide is cheaper than that mistake. The warranty note completes the picture: a component that the tests have isolated and that is under warranty is an RMA, not a repair shop — and the evidence you have gathered, which ports and which cables and the iGPU result and when the behavioural stage began, is exactly what the warranty form asks for. When in doubt between “one more test” and “call it in,” the one more test is almost always cheaper; when in doubt between two parts, the tests are how you avoid buying both.