No Signal When Launching a Game but Desktop Is Fine? Mode Switch, Bandwidth, DSC and System-State Diagnosis Guide

Key takeaway: If the desktop is fine but the monitor reports no signal on launching a game, while sound keeps playing, the fault sits in the mode switch, not in the machine. Set the desktop to the game's mode, lower the game's mode one step, and check whether the system is alive. That separates the display chain, the mode, and the driver.
The desktop is fine. Windows boots, the login screen appears, documents open, videos play, and the picture is steady at any resolution and refresh rate. Then one game is launched, or the display switches to one particular mode, and the monitor immediately reports "No Signal". The picture does not come back on its own. The game sound keeps playing. The Caps Lock light still answers. The machine is still alive, and the screen is not.
The instinct in this situation is usually to blame the graphics card, or to buy a more expensive cable. The card is assumed to have failed because the fault appears when the card is asked to work hardest, or the cable is replaced because cables are the cheapest part of the chain to change. Both instincts are premature. A desktop that works normally proves that the card can output a picture. A fault that appears only at the moment of a mode switch points at the switch itself: the new mode, the re-handshake that the new mode demands, or the driver state that the new mode exposes. The first job is therefore not to replace anything, but to find out whether the display chain can hold the game's mode at all, and whether the fault follows the mode or follows the switch.
This guide covers only a narrow shape. Four conditions must all be present. First, the desktop is normal at any resolution and refresh rate: booting, logging in, documents, and video playback all work. Second, launching one particular game, or switching to one particular mode, makes the monitor report "No Signal" immediately, and the picture does not return by itself. Third, the system is still alive: the game sound continues, the Caps Lock light responds, or pressing Win+Ctrl+Shift+B produces a beep or a flicker. Fourth, the loss of signal lasts until the mode changes: quitting the game, lowering the game mode, or returning to the desktop brings the picture back. If any of these conditions is missing, the diagnosis changes. A machine with no signal from the moment it is switched on is a static signal-chain problem. A picture that goes black for a second or two and returns by itself is a transient switching problem. A machine that restarts, freezes, or stops responding entirely is a different shape, at the power or system layer. A picture that goes black part-way through a game, after the mode has been stable for a while, is a mid-game failure, not a launch failure.
A gap left by five guides
This combination — no signal at the moment a game starts, a persistent loss of signal, a system that is still alive, and no restart — is not covered by any single neighbouring guide. Five guides between them leave it uncovered.
One guide hands the shape over in its opening section, writing that a particular mode with no picture at all, where the monitor reports that it receives no signal, belongs to the category of signal interruption, and is different from the picture-corruption subject of that guide — and it names no guide that receives it. Another guide, on no signal at boot, contains the words "game", "gaming", and "load" zero times in its whole text: it covers a static no-signal chain at power-on, with its cable, port, EDID, and DSC layers, and it has no way to divide a fault by the trigger of launching a game. A third guide, on a persistent black screen during play, states in its opening section that a black screen at the moment a game is launched, with the whole machine restarting by itself, points to the power-connector side, and that its own subject is a fault that appears only after a game has been running for a while. A game that fails at launch is therefore excluded from it. A fourth guide, on the power connector, covers only the case where the machine restarts — the black screen and restart ending, with a Kernel-Power 41 event. This guide's shape has no restart, so it is not that guide either. And the guide on a black screen during a switch covers only a blackout of a second or two that returns by itself, not a persistent loss of signal.
Put together, the position is this: one guide hands the launch-time no-signal shape away, the boot-time guide never mentions gaming, the mid-game guide excludes launch-time failures, the power guide takes only restarts, and the switching guide takes only brief blackouts. The combination of launching a game, a persistent loss of signal, a system that is still alive, and no restart is claimed by no guide at all. That gap is the reason this guide exists, and it fills that gap and nothing else.
A mode switch is a re-handshake, and this guide is the failed side
Every change of resolution, refresh rate, colour format, or signal timing forces the monitor to read the new signal again, lock onto the new synchronisation, and adjust its internal processing. The graphics card and the monitor must establish the signal link again — a re-handshake. The switching guide already states this mechanism, and this guide uses it without repeating the whole of that guide: where the switching guide covers the case in which the re-handshake succeeds but the picture goes black for a moment, this guide covers the case in which the re-handshake fails and the monitor reports no signal. The trigger is the same. The outcome is different. The two must not be merged, because their tests are different: a brief blackout that returns by itself needs no further division, while a persistent loss of signal needs the mode, the cable, and the driver to be tested separately.
A game that switches to exclusive fullscreen at a different resolution or refresh rate is the most common trigger, because it changes the mode in one step rather than gradually. A game that switches colour mode, turns HDR on, or changes the colour depth does the same thing in a different dimension. The desktop, running at its own steady mode, never asks for the switch, which is why the desktop can be perfectly stable while the same machine fails the moment a game asks for a different mode.
Bandwidth: the same axis as artefacts, a different outcome
Higher resolution, higher refresh rate, and higher colour depth multiply together: the amount of data that must be sent from the graphics card to the monitor in each second rises sharply. The guide on artefacts at high modes states this mechanism, and this guide shares the axis but not the outcome. When the required bandwidth passes what the chain can carry, the artefacts guide describes data arriving but corrupted — speckles, blocks, and colour errors. This guide describes the other outcome: the data never establishes a usable link at all, and the monitor reports no signal. Artefacts mean the data arrived but wrong; no signal means the link could not be established. The two are the two ends of the same bandwidth axis, and the test that divides them is the same test that divides this guide from the artefacts guide: whether the monitor shows a corrupted picture or reports no signal at all.
The ktcplay guide states the mechanism in its own words: "A high-refresh connection that depends on the right cable standard is where many game-specific failures become easier to explain. A desktop at idle may sit right on the edge of signal integrity, then a game switches to exclusive fullscreen at 144 Hz or uses a different color mode and the link drops." A desktop at idle sits close to the edge without crossing it. The game pushes the chain over the edge, and the link drops. That is why the same cable and the same monitor can be perfectly stable on the desktop and fail the moment a game starts.
DSC: a compression negotiation that can fail
Display Stream Compression, or DSC, is a visually near-lossless compression technique used to fit a high-specification picture into limited bandwidth. A combination that does not support DSC must carry the full picture on raw bandwidth alone. When a game switches to a high mode, the DSC negotiation can fail, or the monitor or cable may not support DSC for that mode at all, and the result is a loss of signal. The DSC switch itself is a comparison test in this guide, not a setting recommendation: turning DSC off and testing again divides a DSC-negotiation failure from a raw-bandwidth failure. If the picture returns with DSC off, the fault points to the DSC negotiation layer. If the picture is still absent, the fault points elsewhere.
The driver can reset under load: two causes in parallel, no ranking
TDR is Windows' driver timeout reset mechanism, with a default limit of 2 seconds: if the graphics processor cannot finish its work within 2 seconds, Windows judges the driver to have stalled and resets the display driver. A successful reset brings the picture back; a failed reset can become a blue screen. The techdows guide places two causes side by side in its own words: "The graphics driver resets under load, or the game switches to a mode the monitor rejects." Both are possible, and this guide tests both. It does not rank them, because no outside source ranks them, and ranking them would be inventing an order the sources do not support.
Win+Ctrl+Shift+B is the display-driver reset shortcut recorded in Microsoft's official documentation, and it is used in this guide as a comparison test at the driver layer, not as a repair. Pressing it and hearing a beep or seeing a flicker shows that the driver reset ran; whether the picture returns divides the driver layer from the mode layer. The livemonitor.online guide states the same observation in its own words: a signal that drops for a few seconds only in games looks like a driver reset. Event Viewer provides the other half of the driver comparison: event 4101 from the Display source, "the display driver stopped responding and has successfully recovered", or a LiveKernelEvent 141 entry. Those codes are reported here only as they are described — a driver reset happened — and they must not be extended into a claim about which part is faulty. The presence of event 4101 does not by itself prove that any particular component is broken.
Power is an exclusion branch, not the main subject
The power-connector guide and the power-supply guide describe a typical ending of a black screen plus a restart, not no signal with a system that is still alive. This guide therefore treats power as an exclusion branch only: if the machine restarts when the fault happens, the diagnosis leaves this guide and goes to the power-connector or power-supply side. The claim that a power-supply transient spike causes "no signal with the system still alive" has no confirmation from any outside source consulted for this guide, and it is not written here as a confirmed cause. It is preserved as an unconfirmed idea, and the exclusion branch is the only place it appears: a restart during the fault is the signal to leave this guide, and the absence of a restart is the signal to stay in it.
System state: Dynamic Refresh Rate, leftover modes, and windowed play
The mode is not only the numbers in the game's settings. Windows itself can change the mode. Windows 11 Dynamic Refresh Rate (DRR) switches the refresh rate automatically between different states of use, and a leftover display mode from an earlier setting can still be in effect when the game asks for its own mode. When a game switches to exclusive fullscreen, Windows and the game change the mode together, and the comparison tests have to divide which side the fault sits on.
Switching the game to borderless windowed mode changes the switch itself: the game no longer takes exclusive control of the fullscreen mode, so a picture that returns in borderless windowed mode points to the switching layer. Turning DRR off and testing again, switching to borderless windowed mode and testing again, and resetting the display settings and testing again are the three comparisons for this layer. Change one thing at a time, and write down which change first brought the picture back.
The five suspect layers
| Layer | What it means | Decisive comparison |
|---|---|---|
| Mode and bandwidth layer | The game switches to a mode the display chain cannot carry (resolution, refresh rate, and colour depth multiplied past the bandwidth limit) | Desktop same-mode comparison, plus in-game lower-mode comparison |
| Cable, port, and DSC layer | The cable has too little bandwidth headroom, the port version is wrong, or the DSC negotiation fails | Cable and port change comparison, plus DSC on-and-off comparison |
| Driver reset layer | A driver timeout reset under load (TDR), or a driver-version problem | Win+Ctrl+Shift+B comparison, plus Event Viewer event 4101 or LiveKernelEvent 141 comparison, plus driver-version comparison |
| System-state layer | Windows 11 Dynamic Refresh Rate (DRR), a leftover display mode, or exclusive fullscreen versus borderless window | DRR on-and-off comparison, plus windowed-mode comparison, plus resetting the display settings |
| Power exclusion layer | A loose power connector or a power supply with too little headroom (exclusion only: a restart during the fault means this side, not this guide) | Restart-or-not comparison, plus a visual inspection of the connectors after the power is off |
Each layer has its own comparison, and no single comparison covers all five. The desktop same-mode comparison separates "the chain cannot carry this mode" from "something goes wrong at the moment of switching". The in-game lower-mode comparison separates the mode and bandwidth layer from the driver layer. The system-alive comparison separates this guide from the power guides. The order in the seven steps below is the order that wastes the least time.
Seven diagnostic steps
Step 1 — Define the shape
Confirm the four shape conditions: the desktop is normal, launching one particular game or switching to one particular mode produces no signal immediately, the system is still alive, and the loss of signal lasts until the mode changes. Write down the baseline before changing anything: which game, which mode (resolution, refresh rate, colour depth, and whether HDR is on), what the desktop mode is, which cable is in use (identify its version, and do not judge it by its price), the monitor model, and the driver version. Ask two further questions and write down the answers: does the fault happen every time this game is launched, or only sometimes, and does it happen only in this game, while other games are normal. A diagnosis built on a vague description such as "the screen goes black when gaming" cannot be checked later, because it does not say which game, which mode, or whether the system was still alive.
Step 2 — Compare whether the system is still alive
While the monitor reports no signal, check three things and write down all three: does the game sound continue to play, does pressing Caps Lock change the light, and does pressing Win+Ctrl+Shift+B produce a beep or a flicker, or bring the picture back. Sound continuing plus a Caps Lock light that answers means the system is still alive, and the diagnosis continues in this guide. A machine with no response at all, or a machine that restarts by itself, means stop: the diagnosis leaves this guide for the power-connector, power-supply, or mid-game black-screen side. This step is the first fork in the whole guide and must not be skipped. The guide on a persistent black screen during play states in its own words that continuing sound is its most important evidence, and this guide turns that same evidence into the first step of a launch-time failure.
Step 3 — Compare the desktop with the same mode
Quit the game and return to the desktop, then set the desktop to the game's resolution and refresh rate (Settings, System, Display, Advanced display) and leave it there for a while. If the desktop holds the mode steadily, the chain can carry that mode, the fault is not a simple bandwidth problem, and the diagnosis moves to Step 4 to examine the switching and driver layers. If the desktop also loses signal, the fault points to the bandwidth or cable layer, and the diagnosis moves to Step 5. This step is the most important single comparison in the whole guide: the same mode, steady on the desktop and failing in the game, is what divides "the chain cannot carry this mode" from "something goes wrong at the moment of switching". Without it, the two are indistinguishable.
Step 4 — Compare the game at a lower mode
Open the game's settings, lower the resolution by one step (for example from 4K to 1440p) or lower the refresh rate by one step (from 144 Hz to 120 Hz or 60 Hz), and launch the game again. If the picture returns at the lower mode, the fault points to the first layer, the mode, bandwidth, or DSC layer. If the lower mode still has no signal, test the system state before moving to Step 6: turn off Windows 11 Dynamic Refresh Rate in the advanced display settings and test again; switch the game to borderless windowed mode and test again; and reset the display settings and test again. Change one thing at a time, and write down which change first brought the picture back. If none of those brings the picture back, the fault points to the third layer, the driver or load layer, and the diagnosis moves to Step 6. Change only one number at a time, and write down which number is the dividing line. The ktcplay guide mentions a forum case in which lowering 144 Hz to 120 Hz stopped the black screen; that is a forum case, not a verified conclusion, and it must not be written as a guarantee. A lower mode is a comparison test that proves the fault follows the mode, not a repair.
Step 5 — Compare the cable, the port, and the DSC setting
Change to a cable that is known to support the required bandwidth, identifying its version rather than judging it by its price. Change to another output port on the graphics card, or another input port on the monitor. In the monitor's on-screen menu or in the driver, turn DSC off and test again, if there is such an option. Change one thing at a time and test after each change, writing down which change first brought the picture back. Remember the lesson of the direct-connection audio guide when changing cables: the cable is only one link in the chain, and a mode beyond the monitor's capability, a DSC negotiation failure, or a driver reset will not be fixed by any cable, however expensive.
Step 6 — Compare the driver reset
Check whether Win+Ctrl+Shift+B produces any response: a beep or a flicker means the driver reset ran. Open Event Viewer and look for event 4101 from the Display source ("the display driver stopped responding and has successfully recovered") or a LiveKernelEvent 141 entry, and compare the driver version before and after an update. The presence of event 4101 or LiveKernelEvent 141 points to the driver reset layer; their absence points back to the mode and bandwidth layer, and the diagnosis returns to Steps 3 to 5. These numbers and codes may only be stated as they are described here. They must not be extended into a claim that event 4101 proves which part is faulty: the guide on a persistent black screen during play states this plainly.
Step 7 — Stop if the fault is still undivided
If the fault still cannot be divided after these steps, or if the machine restarts when the fault happens, or if there is a smell of burning, or if the plastic casing is seen to be deformed, or if a power connector is suspected of having melted, stop and bring the written record to Room 1209, Dali Plaza, Shek Mun for inspection. A power-connector inspection must only be done after the power is off: never wiggle a power cable while the machine is live. The telephone number is 2186 8010 and WhatsApp is 6558 6806. The fee for inspection is given in the closing section below.
Boundary cases
A machine with no signal from the moment it is switched on is a static signal-chain problem, at the boot-time no-signal layer, not a launch-time mode problem. A picture that goes black part-way through a game, with sound continuing, and needs a restart to return, is a mid-game failure, at the persistent black-screen layer: the mode was already stable and the fault appeared later, which is the opposite trigger from this guide. A black screen at the moment a game starts, with the whole machine restarting, is a power problem, at the power-connector or power-supply layer: the ending is a restart, not no signal with a system that is still alive. A switch that goes black for a second or two and returns by itself is a transient switching problem, at the switching-blackout layer, not a persistent loss of signal. Artefacts that appear only at high resolution or high refresh rate are a threshold-mode problem: artefacts mean the data arrived but wrong, while no signal means the link could not be established at all. A black screen that appears after logging into Windows is a login-time black-screen problem, at the driver-loading and desktop-composition layer, not a game mode-switch problem. An intermittent black screen of a second or two that returns by itself, with no fixed trigger, is an intermittent black-screen problem, not a launch-time loss of signal. A second monitor that goes black while the main monitor is fine, returning when the desktop is restored, is a secondary-monitor problem, not a main-monitor launch-time loss of signal.
Six common misconceptions
- "No signal the moment a game starts means the graphics card is broken." Wrong. A desktop that works normally proves that the card can output a picture. The fault sits in the mode-switch layer, not in the card itself.
- "No signal means the whole machine is dead." Wrong. The game sound continues and the Caps Lock light answers, so the system is still alive. No signal means only that the display chain is broken, not that the whole machine has died. The guide on a persistent black screen during play states that continuing sound is its most important evidence.
- "A more expensive cable will definitely fix it." Wrong. The cable is only one link in the chain. A mode beyond the monitor's capability, a DSC negotiation failure, or a driver reset will not be fixed by any cable, however expensive. The desktop same-mode comparison must be done first.
- "Lowering the resolution fixed it." Wrong. Lowering the mode is a comparison test that proves the fault follows the mode and bandwidth layer. Until the real cause — the cable, the DSC negotiation, or the driver — is found, it must not be written as a repair.
- "The power supply is too weak, so that is why there is no signal." Wrong. A power supply with too little headroom, or a loose power connector, typically ends in a black screen plus a restart, at the power-connector or power-supply layer. No signal with a system that is still alive cannot be blamed directly on the power supply. Power is an exclusion branch here, not a confirmed cause.
- "Pressing Win+Ctrl+Shift+B brought the picture back, so the driver is broken." Wrong. That shortcut only resets the display driver, and it is a comparison test. A picture that returns after pressing it points to the driver or composition state, but it does not mean the driver is broken, and it does not mean the fault will not happen again.
Sources and what they do and do not establish
The techdows guide "Computer turns on but monitor says no signal (9 ways to fix)" states in its own words: "Why does my monitor say no signal when I play a game? The graphics driver resets under load, or the game switches to a mode the monitor rejects." It gives the comparison methods used in this guide: pressing Win+Ctrl+Shift+B, setting the game's resolution and refresh rate to values listed in the settings, and turning off Windows 11 Dynamic Refresh Rate in the advanced display settings. The two causes are placed side by side, with no ranking, and this guide preserves that: both are tested, and neither is written as the main cause.
The ktcplay guide "Monitor No Signal in Games: Causes & Easy Fixes" states that when no signal appears only in certain games, the problem is usually in the display chain rather than in the monitor panel itself; that if audio continues, Windows may still be running while the graphics card, cable, port handshake, or driver fails under gaming load; and that a game switching to exclusive fullscreen at 144 Hz or using a different colour mode can make the link drop. It also states that if one game fails at launch, another fails after ten minutes, and a third never fails at all, the pattern does not automatically prove a bad graphics card. The 144 Hz to 120 Hz case it mentions is a forum case, not a verified conclusion, and it is used here only as that.
The livemonitor.online guide "Monitor says 'No signal' while the computer is running" states that a signal dropping for a few seconds only in games looks like a driver reset, and that pressing Win+Ctrl+Shift+B resets the display driver. It also states that if the signal drops only in games, it looks like a graphics driver issue. It is used here for the driver-reset observation, not as proof of a driver fault.
The box.co.uk guide "Gaming Monitor Showing 'No Signal' / Blank Screen" states that many no-signal issues appear right when a game or operating system forces an unsupported setting, and that no signal and a purely black picture are two different observations. It also states that a picture in the BIOS with no picture in Windows points to a settings layer. It is used here for the distinction between no signal and a black picture, and for the unsupported-setting trigger.
The wikigamingpc.com guide "Gaming PC No Signal Monitor: Cable and Port Fixes" is a general no-signal guide covering the input source, the cable, the port, and the choice between the graphics card output and the motherboard output. It describes no signal at boot, not a launch-time trigger, and it is used here only as general background. Its methods must not be applied directly to the launch-time shape.
There is a limit in the sources that this guide states openly. No outside source consulted for this guide separates the two mechanisms — a mode rejected by the monitor versus a driver reset under load — with a controlled comparison. The desktop same-mode comparison and the in-game lower-mode comparison in this guide are the division method, not a quotation. The power-supply transient spike as a cause of no signal with a system that is still alive has no confirmation from any outside source, and it is written here only as an unconfirmed idea in the exclusion branch. No percentage, case count, success rate, or fee beyond the site's published three-tier fee appears in this guide, because none is supported by the sources used. The fee itself is given in the closing section below.
FAQ
The desktop is fine, but the monitor says no signal the moment I launch a game — is my graphics card broken?
Not necessarily. A desktop that works normally proves that the card can output a picture. The fault sits in the mode switch: the game asks for a different resolution, refresh rate, or colour mode, and the display chain fails to establish that new mode. First do the system-alive comparison and the desktop same-mode comparison, and do not blame the card directly — if the desktop holds the game's mode, the chain can carry it, and the fault is in the switching or driver layer, not in the card itself.
While the monitor says no signal, how do I tell whether the system is still alive or the whole machine is dead?
Check three things: whether the game sound keeps playing, whether pressing Caps Lock changes the light, and whether pressing Win+Ctrl+Shift+B produces a beep or a flicker. Game sound continuing plus a Caps Lock light that answers means the system is still alive, and the diagnosis stays in this guide. A machine with no response at all, or a machine that restarts by itself, is a different shape, at the power or system layer. No signal means only that the display chain is broken, not that the whole machine has died.
How do I do the desktop same-mode comparison, and why is it the most important test?
Quit the game and return to the desktop, then set the desktop to the game's resolution and refresh rate (Settings, System, Display, Advanced display) and leave it there for a while. If the desktop holds the mode steadily, the chain can carry that mode, and the fault is not a simple bandwidth problem — it is in the switching or driver layer. If the desktop also loses signal, the fault points to the bandwidth or cable layer. The same mode, steady on the desktop and failing in the game, is what divides "the chain cannot carry this mode" from "something goes wrong at the moment of switching". And if the desktop holds the mode, the cable is already carrying that mode, so the fault is elsewhere: the cable is only one link in the chain, and a mode beyond the monitor's capability, a DSC negotiation failure, or a driver reset will not be fixed by any cable, however expensive.
Lowering the game's resolution or refresh rate brought the picture back — is the problem fixed?
No. Lowering the mode is a comparison test, not a repair. It proves that the fault follows the mode and bandwidth layer, because the picture returns when the demand on the chain is reduced. If the picture returns at the lower mode, the fault points to the mode, bandwidth, or DSC layer; if the lower mode still has no signal, the fault points to the driver or load layer. Until the real cause — the cable, the DSC negotiation, or the driver — is found, the fault will return at the higher mode. Change only one number at a time, and write down which number is the dividing line.
Pressing Win+Ctrl+Shift+B brought the picture back — does that mean the driver is broken?
No. That shortcut only resets the display driver, and it is a comparison test, not a repair. A picture that returns after pressing it points to the driver or composition state, but it does not mean the driver is broken, and it does not mean the fault will not happen again. Check Event Viewer as well: event 4101 from the Display source ("the display driver stopped responding and has successfully recovered") or a LiveKernelEvent 141 entry points to the driver reset layer. Those codes show only that a driver reset happened; they do not by themselves prove which part is faulty.
The machine restarts when the fault happens — is that the same problem?
No. A restart during the fault is a different shape, at the power-connector or power-supply layer, and it leaves this guide. This guide covers only no signal with a system that is still alive and no restart. If the machine restarts, stop, and do not keep testing the mode layers. A power-connector inspection must only be done after the power is off, and a power cable must never be wiggled while the machine is live. The claim that a power-supply transient spike causes no signal with a system that is still alive has no confirmation from any outside source, and it is not written here as a confirmed cause.
If the fault still cannot be divided
Bring the written record described in Step 7 to Room 1209, Dali Plaza, Shek Mun for inspection. The telephone number is 2186 8010 and WhatsApp is 6558 6806. There is no Mong Kok shop, and there is no on-site service. The site's published three-tier fee is HK$400–780, HK$480–1,380, and HK$680–2,800, an overall range of HK$400–2,800, with the actual fee quoted after inspection. Do not delete the original settings; they are the comparison material.