Half the Screen Normal, Half Abnormal? Panel-Drive, Signal-Chain & Backlight Fault Diagnosis Guide

| HK GPU Fix

Half the Screen Normal, Half Abnormal? Panel-Drive, Signal-Chain & Backlight Fault Diagnosis Guide

Key Takeaways

When exactly half the screen fails while the other half stays normal, the straight boundary is the clue: screenshot, on-screen-menu and torch tests separate a panel-drive fault from a signal fault and a backlight fault before any part is blamed or replaced.

Why a Fault Can Split the Picture Exactly in Half

A modern flat panel does not draw the picture as one indivisible sheet. The timing controller takes the incoming video signal and fans it out to source-driver circuits, usually chip-on-film assemblies bonded to the panel edge, that drive columns of pixels, while gate drivers drive rows. In many panels the left and right halves, or the upper and lower halves, are fed by separate driver groups, separate flex connections, or separate halves of the timing-controller output. If one group loses its drive, its data, or its power, exactly the half it serves goes wrong while the other half keeps working.

That clean geometric split is the signature of this topic. Scattered graphics-card rendering errors usually appear as random speckles, colour blocks, or tearing across the whole picture, and they do not obediently follow one straight line down the middle of the screen. A razor-sharp line at mid-screen that stays put whatever the image content changes is behaving like a panel-drive or signal-processing boundary, not like random rendering corruption. Drawing that line at the start matters, because readers who confuse a half-screen split with full-screen noise will use the wrong tests and reach the wrong conclusion.

Three Families of Cause, Kept Apart

A half-screen fault has three main families behind it, and their diagnostic directions and repair implications are entirely different. Mixing them up sends the reader down the wrong path.

The first family is panel-side drive failure. A source-driver group, its flex bond, or its supply rail fails, and the affected half goes black, white, a solid colour, frozen, or filled with vertical stripes confined to that half. The boundary is sharp and does not move when the image content changes. This family is a panel repair or replacement question, not a settings or driver question, and the article must say so plainly rather than implying that a few clicks will fix it.

The second family is a signal-chain fault before the panel. The graphics card, cable, port, or scaler delivers a damaged or half-decoded picture. Typical pictures include one half showing the wrong content: torn, shifted, duplicated from the other half, or compressed. This family has one telling feature: the split may appear only at one resolution or refresh rate, and changing the mode changes the symptom. If the split follows the computer or the cable, the panel is largely exonerated.

The third family is a backlight half-fault, which looks like a half fault but is not a picture fault. Edge-lit panels use light-emitting diode strings along one or more edges; if one string or its driver output fails, one half of the screen goes dim or dark while the picture is still there, faintly, underneath. The distinction from the drive family is critical: a backlight fault keeps the picture, a drive fault loses or corrupts it. Confusing the two turns a backlight power problem into a condemned panel, or the other way around.

Beyond these three, a scaler fault can also split the picture, for example by processing one half at the wrong scaling. That is why the resolution-change test below matters: scaler problems usually move with the mode, while drive problems do not.

The Diagnostic Order: Six Tests, Cheapest First

Order is part of the diagnosis. Start with the tests that cost nothing and split the largest families first, and only then move to tests that need a different cable or a different display. Each test below is stated as an if-then: if the result is this, it points there; if it is that, it points elsewhere. A list of things to try, with no branching logic, is not a diagnosis.

Test 1: The Screenshot Test

Take a screenshot, or share the screen to another display, and view it elsewhere. If the screenshot shows the split too, the fault is at or before the framebuffer: graphics-card rendering, the driver, or the application output, and the panel is probably innocent. If the screenshot is clean while the eye sees the split, the fault is after the framebuffer: cable, port, scaler, timing controller, panel drive, or backlight. This single test kills half of all wrong guesses, which is why it comes first.

One caveat must be stated. A protected-content window can make a capture look different from what the eye sees, for reasons unrelated to the fault. Use a plain desktop or white-window screenshot for this test, not a video player.

Test 2: The On-Screen Display Menu Test

Open the monitor's own menu and place it so it straddles the boundary, if the menu allows positioning, or simply observe whether the menu itself is corrupted on the bad half. If the menu is clean and sharp on both halves, the panel's own drawing path works, and the incoming signal is suspect: cable, port, graphics card, or scaler. If the menu itself is split, striped, or missing on the bad half, the fault is in the panel drive chain: timing controller, source driver, or flex bond. The logic is simple: the menu is generated inside the monitor and never travels through the cable, so a broken menu cannot be blamed on the cable.

Test 3: The Monitor-Swap and Cable-Swap Test

Move the computer to another display, and move another source, such as a laptop, to the suspect display, changing one thing at a time. If the split follows the computer, the graphics card, driver, or cable side is suspect. If the split stays with the monitor on any source, the monitor side is suspect. If the split disappears on a different cable or port, the cable or port side is suspect, but re-test the original cable alone before blaming the port. Changing two things at once and declaring a winner destroys the evidence.

Test 4: The Resolution and Refresh-Rate Change Test

Drop to a lower resolution and a standard refresh rate, for example the panel's native resolution at 60 Hz, or one step below it. If the split vanishes at lower settings, suspect bandwidth, display-stream compression, scaler, or cable-margin behaviour at the original mode. If the split is identical at every mode, settings are not the cause, and the drive or backlight family becomes more likely. The point of this test is the comparison after the change, not the mere fact that the picture looks different at a new mode.

Test 5: The Torch Test, for the Dark-Half Case Only

Only when one half is dark rather than corrupted should this test be used. Shine a torch at an angle on the dark half. If the image is faintly visible, the panel is still drawing and the backlight string, its driver, or its power feed is the suspect. If no image is visible at all, the pixels are not being told what to draw, which is a drive or signal fault, not a backlight fault. Used at the wrong time, on a half that is distorted or striped rather than dark, the torch test misleads, so its scope must be stated before it is used.

Test 6: Reading the Boundary Itself

Look at the boundary. A razor-sharp straight line at mid-screen, stable across content, points to drive-group or scaler-split behaviour. A soft gradient, vignette, or blotch fading across the boundary points to a backlight uniformity problem, not a drive split. A boundary that moves, flickers, or follows a window points to a rendering, compositor, or signal problem, not a fixed panel drive. The shape of the boundary is itself evidence, not decoration.

TestResult pointing to the signal or graphics-card sideResult pointing to the panel-drive or backlight side
Screenshot testScreenshot shows the split tooScreenshot clean, split visible only to the eye
On-screen display menu testMenu clean and sharp on both halvesMenu itself split, striped, or missing on the bad half
Monitor-swap and cable-swap testSplit follows the computer, or vanishes on a different cableSplit stays with the monitor on any source
Resolution and refresh-rate testSplit vanishes at lower settingsSplit identical at every mode
Torch testNot applicable: distortion or stripes are not a backlight problemFaint image visible means backlight; no image at all means drive or signal
Boundary readingBoundary moves, flickers, or follows a windowSharp straight line fixed in place, or soft gradient meaning backlight uniformity

Boundary Cases Where the Conclusion Changes

If only one application or one game splits while system-wide tests are clean, suspect that application's own rendering or scaling, or its fullscreen mode, not the panel. Cross-check with a second application at the same resolution before blaming hardware.

If the picture is fine from a cold boot but splits after sleep and resume, suspect link re-training or display re-enumeration after wake before suspecting panel hardware. A full power cycle of the monitor, switched off at the switch rather than left in standby, is a legitimate elimination step, distinct from rebooting the computer.

If the split appears only at high refresh rate or high resolution, it points to cable margin, display-stream compression, or scaler limits, as covered by test 4. This article does not expand into a general discussion of high refresh-rate problems, which belongs to the bandwidth and display-identification topic.

A half fault that comes and goes with warm-up narrows the family but does not by itself prove which component is at fault. Cold-versus-hot fault timing has its own dedicated discussion, and this article must not promise that warming or cooling fixes a drive fault, nor present temperature change as a repair method.

Vertical stripes confined to one half must be read differently from full-screen stripes. Full-screen regular stripes point to interference, cable shielding, or graphics-card artifact patterns. Stripes confined to one half with a sharp boundary point to that half's drive group. The line drawn at the start of this article exists to prevent exactly this confusion.

Common Misconceptions: Wrong Direction Is Worse Than No Direction

First, the claim that half the screen is broken so the graphics card must be dying usually has the direction backwards. A clean half split with a clean screenshot and a corrupted on-screen menu points to the monitor, while a graphics-card fault typically produces scattered artifacts, crashes, or whole-screen failure, not a neat half. Declaring the graphics card guilty before the screenshot and menu tests is the most common misdiagnosis in this topic.

Second, reinstalling the driver will fix it, but only within a narrow condition. Only if the screenshot test implicates the framebuffer side does a driver reinstall make sense. If the menu itself is split, no driver reinstall can help, because the menu never passes through the driver. That precondition must be stated before any driver step is listed.

Third, the cable is always the culprit, which is ritual rather than diagnosis. The swap test exists precisely because cable, port, and panel-drive faults can look identical. Buying a new cable before running the first three tests spends money instead of diagnosing, and very often ends with the fault still sitting in the monitor.

Fourth, a dark half means the screen is dead, which is a misjudgement the torch test exists to prevent. A backlight-string fault is a different repair from a panel-drive fault, with different parts and a different prognosis. Before writing off a whole display for a dark half, run the torch test at least once.

Fifth, pressing or flexing the screen edge fixes it, which must never be recommended as a repair. At most, the article may note as an observation that a fault which changes when the panel is moved is consistent with a flex-bond or contact problem and belongs in professional assessment. Turning that observation into a do-it-yourself instruction risks turning an assessable fault into further damage.

Stop Points and What to Record Before Seeking Inspection

Stop during checking if the panel is cracked, if there is an unusual smell or scorch mark, if the monitor power section makes an abnormal sound, or if the thought of opening the case arises. Half-screen diagnosis should stay within settings, screenshots, the monitor menu, cable swaps, and display swaps. Panel drive and backlight power work belong in professional assessment.

Before seeking inspection, gathering the facts makes the inspection much faster. Record the monitor model, which half is affected, whether the boundary is vertical or horizontal, whether it is a sharp line or a gradient, the screenshot test result, the menu test result, the cable-swap and monitor-swap results, the resolutions and refresh rates at which the fault appears, the torch test result where applicable, and whether the fault is constant or intermittent. The more complete the record, the less the inspection has to guess from scratch.

For inspection in Hong Kong, the shop is Room 1209, Dali Plaza, Shek Mun. The telephone number is 2186 8010 and the WhatsApp number is 6558 6806.

Frequently Asked Questions

1. Half the screen is normal and half is abnormal. What should I do first?

Run the screenshot test, then the monitor menu test. The screenshot separates the framebuffer side from everything after it, and the menu separates the incoming signal from the panel drive. Both tests cost nothing, need no parts, and split the largest families first, so later cable and setting tests have a direction.

2. The screenshot shows the split too. What does that mean?

It means the fault is at or before the framebuffer: graphics-card rendering, the driver, or the application output, and the panel is largely exonerated. The next steps are driver, application-setting, and single-application checks, not a replacement display.

3. The screenshot is clean but my eyes see the split. What does that mean?

It means the fault is after the framebuffer: cable, port, scaler, timing controller, panel drive, or backlight. Run the menu test next. If the menu is broken too, the panel drive is the direction; if the menu is clean, the incoming signal is the direction.

4. The monitor menu itself is split. Does that mean the display cannot be saved?

Not necessarily, but it means the problem is in the panel drive chain, not in settings or drivers. Reinstalling drivers, changing cables, or changing resolution will not change what the monitor's own menu looks like. The value of this result is that it stops effort in the wrong direction and moves the question to panel assessment.

5. One half is dark. How do I tell a backlight fault from a drive fault?

Shine a torch at an angle on the dark half. A faintly visible image means the panel is still drawing and the suspect is the backlight string, its driver, or its power feed. No image at all means the pixels are receiving no instruction, which is a drive or signal fault. The torch test applies only to a dark half, not to a distorted or striped half.

6. The boundary is a soft gradient, not a sharp line. Is it the same fault?

No. A soft gradient, vignette, or blotch points to a backlight uniformity problem, not a drive split. That class of light-uniformity defect sits closer to the dead-pixel, bleed, and glow topics, and its diagnostic direction differs from a sharp-boundary drive fault. Do not apply the same conclusion to both.

7. Only one game splits, and everything else is normal. Is the panel failing?

Very probably not. A single-application split with clean system tests points to that application's rendering, scaling, or fullscreen mode. Compare with a second application at the same resolution. If the second application is normal, the panel is unlikely to be the cause, and the game's display settings are the place to check.

8. The split appears only after waking from sleep, but a cold boot is fine. How should I read that?

Suspect link re-training or display re-enumeration after wake before suspecting panel hardware. Switch the monitor fully off at its power switch and back on as a separate elimination step, distinct from rebooting the computer, and see whether the result changes before drawing a conclusion.

9. The split vanished after I changed the cable. Does that prove the cable was faulty?

Not by itself. Changing the cable also changes the connector, can renegotiate the signal mode, or can temporarily improve contact through reseating alone. Re-test the original cable on its own and confirm that the split stably follows that cable before declaring it the cause.

10. One half has vertical stripes and the other half is normal. How is that different from a full-screen stripe problem?

The range and the boundary are the key. Stripes confined to one half with a sharp boundary point to that half's drive group. Regular stripes across the whole screen point to interference, shielding, or graphics-card artifact patterns. The two directions are entirely different, which is why the range must be established before any conclusion.

Related articles

Related services

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