Monitor Keeps Disconnecting and Reconnecting? Display Disappearance, Hot-Plug Detection & Link Retraining Diagnosis Guide

2026-11-13 | GPU Repair | 30 minutes | HK GPU Fix Editorial Team

Monitor Keeps Disconnecting and Reconnecting? Display Disappearance, Hot-Plug Detection & Link Retraining Diagnosis Guide

Key Takeaway

A monitor that disconnects and reconnects by itself, with a device connection chime and windows moved, has disappeared from Windows — the system briefly believes it was removed. That is different from a black screen, where the monitor stays listed. This guide separates five layers and walks through seven steps to find why the disappearance keeps repeating.

What This Problem Is — and What It Is Not

Picture the scene: the computer is in normal use, nobody has touched a cable, nothing has been plugged in or unplugged, the machine has not been woken from sleep and has not been restarted — and suddenly the screen goes dark, Windows plays its device connection or disconnection chime, the desktop shrinks to one monitor, and the windows that were on the second monitor are moved onto the main screen. A moment later the chime sounds again, the monitor reappears, and the desktop stretches back out. Then, minutes later, it happens again. And again.

That scene is the problem this guide covers, and three features define it. First, the monitor disappears from Windows: during the event, the monitor is no longer listed in Windows display settings, and the desktop contracts as though the monitor had been unplugged. Second, there is evidence of a removal and a reconnection: the device connection or disconnection chime, and a record in Event Viewer under Kernel-PnP, the Windows component that logs device enumeration. Third, the repetition is spontaneous: it happens during normal use, with no plug, unplug, wake, restart, or monitor change to trigger it. All three features must be present. If any one of them is missing, the problem is probably a neighbouring one, and the boundaries section below names which.

The decisive fork in this whole guide is the first feature, and it is a check the reader can do in seconds: when the problem happens, open Windows display settings and look at whether the monitor is still there. If the monitor is still listed but the picture is black, the problem is a signal problem — the monitor is present, the picture is interrupted — and the neighbouring guide about brief black screens applies. If the monitor is not listed at all, the problem is a disappearance: the system has decided the monitor is gone, and the desktop has been rebuilt without it. The two look identical from the chair — the screen goes dark either way — but they are different failures with different causes and different tests. Every step below depends on getting this fork right first, which is why it is Step 1.

Four conditions must all be true before the steps below apply:

  1. The disappearance happens during normal use, with no cable touched, no monitor plugged in or unplugged, no wake from sleep, and no restart.
  2. The monitor vanishes from Windows display settings during the event, and the desktop contracts. Windows may play a device connection or disconnection chime.
  3. The monitor comes back by itself, or after a short wait, and the event repeats — it is not a one-off.
  4. The graphics card is otherwise able to show the Windows desktop. If the whole display is broken, the problem is not a repeated disappearance — it is a display problem, and a different guide applies.

Several nearby behaviours look similar but are not this problem. A picture that goes black for a second or two and comes back, while the game keeps running, the sound keeps playing, and the windows stay where they were, is a brief black screen: the monitor never disappeared, and only the signal was renegotiated. Windows and icons in the wrong place after a monitor wakes up are a layout problem. A monitor that misbehaves only at the moment a second monitor is plugged in is a hot-plug event at the moment of plugging in, not a spontaneous disappearance. A monitor that fails only after waking from sleep, only after a monitor has been changed, or only at startup or restart, has a trigger — and this guide is about disappearances with no trigger at all.

How a Monitor Tells the Computer It Exists: Hot-Plug Detection and the Four-Step Handshake

A modern display connection is not a one-way pipe that simply carries a picture. It is a two-way conversation, and the conversation has a strict order. Understanding that order is what makes the diagnosis possible, because the steps below test one part of the conversation at a time.

Step one of the handshake: hot-plug detection, or HPD. The monitor provides a 3.3V signal on the hot-plug detect pin, telling the graphics card that a monitor is present. High means the monitor is there; low means it is gone. The timing matters, and the rules are precise. If the HPD signal stays low for more than 2ms, the receiving end treats the monitor as disconnected and closes the link. If it dips low for only 0.5 to 1ms, that is an interrupt event — a short pulse asking the source to read the monitor's status registers again. The source is required to sample the HPD signal at intervals greater than 250µs, so a dip shorter than that may not even be seen. These numbers are the official rules, and they are the reason a brief dip and a long drop are different events with different consequences: a short pulse triggers a re-read, while a drop longer than 2ms triggers a removal.

Step two: EDID. Once the monitor's presence is confirmed, the graphics card reads the monitor's EDID over the AUX channel — a small data block in which the monitor describes itself: its model, its supported resolutions, and its refresh rates. If the EDID cannot be read, the handshake stalls.

Step three: link training. The card and the monitor then train the link in two stages: clock recovery, in which the monitor locks onto the card's signal clock, followed by channel equalization, in which the signal quality on each lane is tuned. The AUX channel carries the back-and-forth for this stage, and each AUX transaction must complete inside a hard timeout window of 300 to 400 microseconds, or the graphics card marks it as an AUX timeout. After a set number of retries — usually three to five consecutive failures — the driver gives up the handshake entirely and assumes the receiving end has disconnected or failed. Link training is where a connection running at the edge of its bandwidth first shows its weakness: a high resolution combined with a high refresh rate and HDR demands more from the link than a modest setting does, and a link that trains cleanly at a lower setting may fail repeatedly at a higher one.

Step four: HDCP. Finally, the two ends complete HDCP authentication, the content-protection handshake. Only then does the picture appear.

When everything works, this four-step handshake runs once, at connection, and then stays quiet. When a monitor keeps disconnecting and reconnecting, the handshake is running over and over again, and the question the diagnosis must answer is which part of it is failing, and why it fails repeatedly with nothing being touched.

The Five Suspicion Layers

A repeated monitor disappearance can come from five layers, and they must be kept separate. Mixing them together is how a deep-sleep problem gets treated with a driver reinstall, or a dock problem gets treated by replacing the monitor, and how the real culprit survives because the wrong layer was treated. The five layers below follow the handshake from the physical signal outward to the system, and each one has its own deciding test.

Layer One: The Hot-Plug Detection Layer — the 3.3V Signal Keeps Falling

The first layer is the HPD signal itself. The monitor provides a 3.3V signal telling the graphics card it exists. If that signal drops below its threshold, or fails to deliver a clean 0.5 to 1ms interrupt pulse, the card can misread the situation: an unstable HPD connection that drops and returns makes the card believe the monitor has been removed and then reinserted, over and over. Intermittent poor contact, oxidized pins, a bent cable, dust in the connector, or a damaged HPD pin can all make the signal fall repeatedly. The key property of this layer is that the fault follows the physical connection: it is the same monitor on the same port, disappearing and reappearing, and the deciding test is whether changing the cable or the port changes the behaviour. A damaged HPD pin is the cleanest form of this fault — a complete failure — but this guide is about the intermittent form, where the pin works most of the time and drops out without warning.

Layer Two: The Link Training Layer — AUX Timeouts and Bandwidth Edges

The second layer is link training. DisplayPort link training has two stages, clock recovery and channel equalization, and the AUX channel that carries the negotiation must complete each transaction inside a 300 to 400 microsecond window. Most intermittent link drops stem directly from AUX handshake failures during link training: after three to five consecutive failures, the driver abandons the handshake and assumes the receiving end has disconnected. A link that is running at the edge of its bandwidth — a high resolution combined with a high refresh rate and HDR — is the first to show this weakness, because the training has less margin for error. The deciding test is whether lowering the resolution or the refresh rate by one step makes the disappearances stop. If it does, the link was at its edge, and the fault is in the training layer, not in the monitor.

Layer Three: The Monitor Power and Deep Sleep Layer — the Monitor Drops Its Own Signal

The third layer is the monitor itself, and specifically its power state. A monitor that has fallen into deep sleep drops its HPD signal, and the graphics card then believes there is no monitor. This state is called DisplayPort Deep Sleep, and it is a property of the monitor's power management, not of the graphics card. A monitor's power-saving setting can also cut the connection signal entirely after a period with no signal, and the monitor then has to be woken before the card sees it again. The deciding test is the 60-second power clear: pull the monitor's AC power cord for 60 seconds to clear a stuck DisplayPort hot-plug detect state, then reconnect and watch whether the disappearances stop. A second test is to turn off the monitor's deep sleep or power-saving setting in its on-screen display, and to lock the input source manually instead of relying on automatic input scanning. If the disappearances stop after either change, the fault was in this layer. This layer is also the reason the trigger matters so much: a monitor that only disappears after a period of inactivity, and never while it is actively in use, is behaving differently from a monitor that disappears in the middle of active use — and the two point at different layers.

Layer Four: The Intermediate Device and USB-C Layer — Docks, Switches, and Lane Sharing

The fourth layer is everything between the graphics card and the monitor: docks, hubs, switches, splitters, and adapters. Each of these re-enumerates the connection on its own schedule, and a re-enumeration that takes slightly too long, or that stutters, can look exactly like a monitor disappearing and reappearing. A dock's DisplayPort output that keeps disconnecting and reconnecting is a documented pattern, and the cause can sit in the laptop, the cable, the monitor, or the dock itself — which is why the comparison in Step 3 exists. USB-C adds a further complication: a single USB-C cable must split its high-speed lanes between DisplayPort video, USB data, and charging, and a change between 2-lane and 4-lane mode changes how much connection the video side can sustain. A USB controller or driver problem can also make a monitor disconnect and reconnect repeatedly even when the monitor itself is healthy, sometimes appearing after a firmware update and accompanied by charging interruptions and a connection chime every few seconds. The deciding test for this whole layer is the direct-connection comparison: bypass every intermediate device with a single compliant cable, straight from the graphics card to the monitor. If the disappearances stop, the intermediate device — or the cable feeding it — is the suspect, and the search narrows to which one.

Layer Five: The Driver and System Enumeration Layer — Enumeration Failures and Power Management

The fifth layer is the driver and the system. The driver can fail to enumerate the display device, or time out while doing so, and when enumeration fails, the system's behaviour is that the device is not there — not that the device shows an error message. That distinction is the signature of this layer: a device that is missing from Device Manager is different from a device that is present with a warning. Windows power management can also force re-enumeration: PCIe link-state power management and USB selective suspend can put a device to sleep and wake it again, and each wake is a new enumeration event. A monitor firmware handshake problem belongs here too. The deciding evidence for this layer is in Event Viewer: the Kernel-PnP enumeration records, matched in time against the disappearances. If a Kernel-PnP record appears at exactly the moment the monitor vanished, the system saw a removal event — though the cause may still sit in a lower layer, and the records must be read together with the direct-connection and cable comparisons, not on their own.

The Seven Diagnostic Steps

The seven steps below are ordered so that each one rules out a layer before the next one is tried. Work through them in order, change only one thing at a time, and write down the result of each step before moving on. A note that says what was tried and what happened is worth more than a memory of having tried something. The steps are the same in both language versions of this guide, and the numbers stay the same throughout. Do not rush to change settings before the pattern is recorded: changing settings before the pattern is known only creates false conclusions.

Step 1: Qualify the Event — Is the Monitor Gone, or Only Black?

When the problem happens, open Windows display settings and look at whether the monitor is still there. If it is still listed but the picture is black, the problem is a signal problem, and the neighbouring guide about brief black screens applies instead. If it is not listed, the problem is a disappearance, and this guide applies. At the same moment, note whether Windows played a device connection or disconnection chime. The chime is evidence of an enumeration event, and its timing — at the disappearance, at the return, or both — should be written down. A monitor that shows "No Signal" on its own screen is different again from a purely black picture: "No Signal" means the monitor is powered and waiting for a signal it is not receiving, while a black picture with the monitor still listed means the signal path is interrupted somewhere else. The distinction between those two on-screen states is worth recording too, because it separates a monitor that has lost its input from a picture that has simply gone dark.

Step 2: Record the Pattern

Write down how often the disappearance happens, how long the monitor is gone each time, which monitor it is, what was being done when it happened, whether one monitor or two are connected, which port is in use, and which cable. A pattern that only appears with two monitors connected points somewhere different from a pattern that appears with one monitor alone. A pattern that only appears at a high refresh rate points at the link training layer. A pattern with no visible rhythm at all is still information: it rules out scheduled causes such as a sleep timer. Changing a setting before the pattern is recorded means the reader can never tell whether the setting helped, because there was no baseline to compare against.

Step 3: The Direct-Connection Comparison

Bypass the dock, the switch, the splitter, and the adapter. Connect the graphics card straight to the monitor with a single cable that meets the specification for the resolution and refresh rate in use, and watch whether the disappearances stop. If they stop, the intermediate device — or the cable feeding it — is the suspect, and the search narrows to which one. The comparison must be given time: a disappearance that happens once an hour will not be ruled out by ten minutes of direct connection. Run the direct connection for as long as the pattern in Step 2 says is needed to see an event, and only then conclude. If the disappearances continue on a direct connection, the intermediate layer is ruled out. This step is the single most informative comparison in the whole guide, and it should never be skipped in favour of a driver reinstall.

Step 4: The Cable and Port Comparison

Try a different cable — shorter, and carrying a certification mark for the standard in use. Try a different output port on the graphics card, and a different input port on the monitor. Then lower the resolution or the refresh rate by one step, and watch whether the disappearances stop. If a shorter certified cable stops them, the old cable was at the edge of its capability. If a different port stops them, the old port's HPD pin or connector is the suspect. If lowering the setting stops them, the link was at its bandwidth edge. Each of these is a separate comparison, and they should be tried one at a time, with a launch-and-watch period after each one, so that the reader knows which change made the difference. A note on cables: the important property of a cable is its specification and its certification, not its price. An expensive cable that does not meet the specification for the resolution and refresh rate in use will fail in exactly the same way as a cheap one.

Step 5: The Monitor Power Comparison

In the monitor's on-screen display, lock the input source manually instead of relying on automatic input scanning, so the monitor is not re-scanning and re-negotiating on its own. Then pull the monitor's AC power cord for 60 seconds to clear a stuck DisplayPort hot-plug detect state, reconnect it, and watch. Check the monitor's deep sleep and power-saving settings, turn them off, and observe again. If the disappearances stop after any of these changes, the fault was in the monitor's power layer: the monitor was dropping its own HPD signal, and the card was correctly reporting that the monitor was gone. A monitor that only disappears after a period of inactivity deserves special attention here: deep sleep is the first suspect when the disappearances follow idleness, and the power comparison is the test that confirms or rules it out.

Step 6: The System Evidence Comparison

When the problem happens, open Device Manager and look at the Monitors section: has the monitor's model changed to a generic monitor, or has the entry vanished entirely? A model that has changed to generic means the system lost the EDID and re-enumerated the monitor as an unknown device; an entry that has vanished means the enumeration failed altogether. Then open Event Viewer and look at the Kernel-PnP enumeration records, matching their timestamps against the times the monitor disappeared and returned. A Kernel-PnP record at exactly the moment of a disappearance is first-hand evidence that the system recorded a removal event. Record the times, the record details, and the Device Manager state together, because the three pieces of evidence only make sense as a set.

Step 7: The Stop Line — Cross-Test with Another Monitor, Then Stop

Swap in a different monitor and run the same setup — same cable, same port, same settings — and watch whether the disappearances follow the monitor or stay with the computer. If they follow the monitor, the monitor is the suspect. If they stay with the computer on a different monitor, the computer side is the suspect. Before any step that involves opening the graphics card, bending a connector pin, or updating monitor firmware, stop. Those steps are beyond home diagnosis: a bent pin can be broken permanently, a firmware update that is interrupted can leave a monitor unusable, and opening a card can void a warranty. Bring the machine in with the records from Steps 1, 2, and 6: the disappearance-versus-black-screen result, the pattern log, the direct-connection and cable comparison results, the monitor power comparison results, and the Kernel-PnP timestamps. Inspection is charged at the site's published three-tier fees of HK$400–2,800, quoted after inspection.

A Quick Comparison: The Seven Diagnostic Steps at a Glance

StepWhat to doWhat a positive result meansLayer tested
1. Qualify the eventCheck display settings during the event; note the chimeMonitor missing means disappearance; monitor listed but black means a signal problemDisappearance vs black screen
2. Record the patternLog frequency, duration, monitor, activity, ports, cableA pattern points at its layer; no pattern rules out scheduled causesAll layers
3. Direct connectionBypass dock, switch, splitter, adapter; one compliant cableDisappearances stopping means the intermediate device is suspectIntermediate and USB-C
4. Cable and portShorter certified cable; different ports; lower resolution or refresh rateStopping means the cable, port, or bandwidth edge was at faultHPD, link training
5. Monitor powerLock input source; pull AC power 60 seconds; turn off deep sleepStopping means the monitor was dropping its own HPD signalMonitor power, Deep Sleep
6. System evidenceDevice Manager monitor entry; Event Viewer Kernel-PnP records matched in timeA Kernel-PnP record at the event time confirms a removal eventDriver and enumeration
7. Stop lineCross-test with another monitor; stop before opening the card or updating firmwareFault following the monitor means the monitor; staying means the computerAll layers

A second comparison is worth keeping beside the first, because it stops the most common wrong turns: replacing a monitor that was never at fault, or reinstalling a driver that was never the cause.

SymptomMost likely layerFirst test, not first fix
Monitor missing from display settings, chime soundsHot-plug detection or link trainingQualify the event, then cable and port comparison (Steps 1 and 4)
Disappearances stop on a direct connectionIntermediate device or USB-CDirect-connection comparison (Step 3)
Disappearances only after a period of inactivityMonitor power, Deep SleepMonitor power comparison (Step 5)
Kernel-PnP record at the exact moment of the eventDriver and system enumerationSystem evidence comparison (Step 6)
Disappearances follow the monitor to another computerMonitor power or firmwareCross-test with another monitor (Step 7)
Picture black for a second, monitor still listed, windows unmovedSignal renegotiation, not disappearanceNeighbouring brief-black-screen guide

Boundaries: Problems That Look Similar but Are Not This One

Several problems look like a repeated disappearance but belong to other guides, and naming them here saves the reader from following the wrong diagnosis.

A picture that goes black for a second or two and comes back on its own, while the game keeps running, the sound keeps playing, and the windows stay where they were, is a brief black screen. During that black screen the system still believes the monitor is there, and the picture returns without any re-enumeration. The neighbouring guide on that subject states plainly that a purely signal-renegotiation black screen leaves the game and the sound running, and everything is normal after it recovers. This guide is about the opposite case: the monitor is gone from the system, the desktop has contracted, and the windows have been moved.

Windows and icons in the wrong place after a monitor wakes up are a layout problem, not a disappearance problem. That guide's subject is what happens after the signal returns. This guide investigates why the disappearance event itself keeps repeating; it does not investigate how to restore the window layout afterwards. The layout guide itself admits that as long as the disappearance events repeat, the moving and resetting will repeat too, and changing graphics settings or reinstalling a driver will not make the disappearance events themselves go away.

A black flash at the moment a second monitor is plugged in is a hot-plug event at the moment of plugging in: the enumeration and handshake process is interrupted repeatedly, and the main monitor flickers with it. This guide's premise is the opposite: no plug or unplug action at all, during normal use. The trigger is what separates the two.

A monitor that misbehaves only after waking from sleep belongs to the sleep guides: waking black screens can be divided into three layers — the channel link failing to retrain, the display device failing to re-enumerate, and the screen itself not waking — but the trigger is the wake, and this guide is about normal use with no wake. A monitor that misbehaves only after a different monitor has been connected belongs to the monitor-change guide. A monitor that takes a while to show a picture at startup, or that misbehaves only on restart and not on a cold start, has a startup or restart trigger. A single output port that is permanently dead while the other ports work is a single-port fault, where the fault follows the port; this guide is about the same monitor on the same port repeatedly disappearing and reappearing. And USB devices behind a monitor's hub that stop working after the monitor sleeps are a USB enumeration problem: the picture and the USB are two separate paths.

Six Common Misconceptions

Misconception 1: A black picture means the monitor disappeared

No. A brief black screen is a signal renegotiation while the HPD signal is still present, and the system still believes the monitor exists. A disappearance is the HPD signal falling, and the system treating the monitor as removed. The way to tell them apart is to look at Windows display settings while the problem is happening: if the monitor is still listed, it has not disappeared.

Misconception 2: Windows moving means the graphics card is broken

No. When a monitor briefly goes offline, the system understands the event as the monitor no longer existing, and the desktop contracts. The windows are moved because there is nowhere else for them to go. That movement is a protective behaviour, not a fault in the graphics card's computing — and if the disappearance repeats, the moving repeats with it.

Misconception 3: Reinstalling the driver will fix it

No. HPD, Deep Sleep, and cable-layer problems will not stop a monitor from briefly going offline just because the driver was reinstalled. The driver check belongs later in the order, after the direct-connection, cable, and monitor power comparisons. Reinstalling the driver first, and watching the disappearances continue, proves nothing except that the driver was never the cause.

Misconception 4: A more expensive cable will definitely fix it

No. The important property of a cable is its specification and its certification, not its price. An expensive cable that does not meet the specification for the resolution and refresh rate in use will fail in exactly the same way as a cheap one. The comparison to run is a shorter cable with a certification mark, not a dearer one.

Misconception 5: A monitor missing from settings means the monitor is broken

No. An enumeration failure can sit in the cable, the port, the dock, or the driver layer. Declaring the monitor dead before the direct-connection and cable comparisons have been run is a misdiagnosis. The cross-test in Step 7 — a different monitor on the same setup — is the test that separates a monitor fault from a computer-side fault, and it comes last for a reason.

Misconception 6: Unplugging and plugging back in fixed it, so it must have been a loose contact

No. Unplugging and plugging back in triggers a fresh round of enumeration and handshake, and it cannot by itself prove a loose contact. The fault must be followed: if it follows the cable, the cable is suspect; if it follows the port, the port is suspect; if it follows the monitor, the monitor is suspect. A single recovery after a re-plug is evidence that re-enumeration works, not evidence about what failed.

When to Stop and Bring the Machine In

Stop when all seven steps have been worked through, one variable at a time, and the disappearances still repeat. Stop sooner if the monitor disappears and never comes back, if the graphics card disappears from Device Manager, if the machine crashes, or if any step involves opening the graphics card, bending a connector pin, or updating monitor firmware. Those are signs that the fault is beyond home diagnosis, and further changes risk making the evidence harder to read.

Bring the machine in with the notes from Step 1 and the results of each step tried: the disappearance-versus-black-screen result, the pattern log, the direct-connection and cable comparison results, the monitor power comparison results, the Kernel-PnP timestamps, and the cross-test result with another monitor. In-store inspection takes place at Room 1209, Dali Plaza, Shek Mun. You can call 2186 8010, or reach the shop on WhatsApp at 6558 6806. Inspection is charged at the site's published three-tier fees of HK$400–2,800, quoted after inspection. There is no home-visit service and no Mong Kok shop; inspection is carried out in the shop.

A diagnosis that ends at the stop line is not a failed diagnosis. A note that says "Monitor disappears from display settings three times an hour during normal use, with a device chime each time; direct connection does not help; a shorter certified cable does not help; deep sleep off does not help; Kernel-PnP records match each disappearance" is a precise description of where the fault lives, and it is exactly what an in-store inspection needs to start from the right place instead of from the beginning.

FAQ

Q1: The picture goes black for a second. Does that mean my monitor disappeared?

No. A brief black screen is a signal renegotiation while the HPD signal is still present, and the system still believes the monitor exists. A disappearance is the HPD signal falling, and the system treating the monitor as removed. The way to tell them apart is to look at Windows display settings while the problem is happening: if the monitor is still listed, it has not disappeared.

Q2: My windows keep moving to the main screen. Does that mean my graphics card is broken?

No. When a monitor briefly goes offline, the system understands the event as the monitor no longer existing, and the desktop contracts. The windows are moved because there is nowhere else for them to go. That movement is a protective behaviour, not a fault in the graphics card's computing. If the disappearance repeats, the moving repeats with it.

Q3: Will reinstalling the driver fix a monitor that keeps disconnecting and reconnecting?

No. HPD, Deep Sleep, and cable-layer problems will not stop a monitor from briefly going offline just because the driver was reinstalled. The driver check belongs later in the order, after the direct-connection, cable, and monitor power comparisons. Reinstalling the driver first proves nothing except that the driver was never the cause.

Q4: Will buying a more expensive cable fix it?

No. The important property of a cable is its specification and its certification, not its price. An expensive cable that does not meet the specification for the resolution and refresh rate in use will fail in exactly the same way as a cheap one. Try a shorter cable with a certification mark, and a different port, before spending more on a cable.

Q5: My monitor is missing from display settings. Does that mean the monitor is broken?

No. An enumeration failure can sit in the cable, the port, the dock, or the driver layer. Declaring the monitor dead before the direct-connection and cable comparisons have been run is a misdiagnosis. The cross-test in Step 7 — a different monitor on the same setup — is the test that separates a monitor fault from a computer-side fault.

Q6: Unplugging the cable and plugging it back in fixed it. Does that mean it was a loose contact?

No. Unplugging and plugging back in triggers a fresh round of enumeration and handshake, and it cannot by itself prove a loose contact. The fault must be followed: if it follows the cable, the cable is suspect; if it follows the port, the port is suspect; if it follows the monitor, the monitor is suspect. A single recovery after a re-plug is evidence that re-enumeration works, not evidence about what failed.

Q7: How do I use the connection chime and the Kernel-PnP records together?

The device connection or disconnection chime is Windows announcing a device enumeration event: write down the time of each chime. In Event Viewer, look at the Kernel-PnP enumeration records and match their timestamps against the times the monitor disappeared and returned. Matching times confirm that the system recorded an enumeration event, not merely that the picture went black. Also check Device Manager, under Monitors, for whether the monitor's model has changed to a generic monitor.

Q8: How do I tell an HPD drop from a link training failure?

An HPD signal that stays low for more than 2ms means the monitor has been removed, and the system takes it out of display settings; a brief dip of 0.5 to 1ms is only an interrupt event asking the source to read the monitor's status registers again. In a link training failure, an AUX transaction cannot complete inside its 300 to 400 microsecond window, and after three to five consecutive failures the driver assumes the receiving end has disconnected. The two look alike, so tell them apart with the downgrade comparison: lower the resolution or the refresh rate by one step — if the disappearances stop, the link was at its bandwidth edge; if they continue, the HPD layer is the likelier suspect.

Q9: How should I handle DisplayPort Deep Sleep?

A monitor that has fallen into deep sleep drops its HPD signal, and the graphics card then believes there is no monitor. Lock the input source manually in the monitor's on-screen display instead of relying on automatic input scanning, pull the monitor's AC power cord for 60 seconds to clear the stuck deep-sleep state, and check the monitor's deep sleep and power-saving settings — turn them off and observe again. If the disappearances stop after the 60-second power clear, the monitor power layer is the suspect.

Q10: When should I stop trying things myself and bring the machine in for inspection?

Stop when all seven steps have been worked through, one variable at a time, and the disappearances still repeat; or when the monitor disappears and never comes back; or before any step that involves opening the graphics card, bending a connector pin, or updating monitor firmware. Bring your notes: the disappearance-versus-black-screen result, the pattern log, the comparison results, the Kernel-PnP timestamps, and the cross-test result. In-store inspection is at Room 1209, Dali Plaza, Shek Mun, charged at the site's published three-tier fees of HK$400–2,800, quoted after inspection. You can call 2186 8010 or WhatsApp 6558 6806.

Related articles

Related services

If you need Chip-level GPU repair, see these service pages: