Screen Streaks and Line Defects: Source and Detection

A thin coloured line down a display looks like a single fault and is almost never one. The same appearance can come from a driver output, a bond joint, a crack in the glass, a connector that is not seated or a cable that has been flexed too often. Knowing which of those is present decides whether the unit is scrap, repairable or a warranty argument, and the difference can be found in minutes without opening the panel.

Part of the Quality, Testing & Compliance guide — 33 articles on defect classes, inspection, qualification and certification.

What causes a vertical line on a screen?

Usually a broken connection to one column or row.

A line appears when one column or row of pixels stops receiving its correct drive. The interruption can be inside the source driver, at the bond between the driver and the glass, within the glass’s own wiring, in the flexible cable or at the system connector. Because the pixel itself is usually healthy, the line is a connection problem rather than an image problem, and that is why the same symptom can be permanent, intermittent or temperature dependent.

The failure families at a glance

Four families, four signatures.

Driver failures affect one output channel and appear as a single line or a small group. Bond failures affect a block of lines and often change with pressure or temperature. Glass failures show as a crack, a branch or a corner defect and may spread. Interface failures affect the whole image as well as the line, and they change when the cable moves. The first diagnostic step is to classify by that signature, because each family has a different owner.

Source driver and gate driver failures

One channel, one line, stable.

A source driver drives the columns and a gate driver drives the rows, so a source fault usually produces a vertical line and a gate fault a horizontal one. A single dead output produces a line in one colour because the other two subpixels still respond; a group of outputs produces a band. Driver damage is often electrostatic in origin, which is why the handling controls on the assembly line are part of the defect history and why the practice published by the EOS/ESD Association applies to the factory floor as much as to the design. Output-level damage is generally not field repairable.

Bonding and bonding-related defects

Pressure sensitivity points at the bond.

Chip-on-glass and tape-automated bonding joints are fine-pitch and mechanical, so they fail from thermal cycling, moisture ingress, flexing or a contamination particle trapped during bonding. The signature is a block of lines that changes when the module is pressed or warmed, and sometimes recovers when it cools. A bond fault found at incoming inspection is a supplier issue and should be rejected with the temperature and pressure conditions recorded; a bond fault that appears after a year of field use usually indicates a mechanical or thermal design problem in the enclosure.

Panel internal shorts and cracks

Mechanical, and usually progressive.

A crack in the glass can sever column wiring and produce a line that branches near the edge, and an internal short between adjacent lines can create a pair of lines or a bright column. Cracks often start at a mounting point, a screw boss or an edge that was chipped during assembly, which is why a line defect that appears shortly after a mechanical change is a design warning rather than a component failure. These faults are not repairable and the correct response is to find the mechanical cause, because replacement panels will fail the same way.

Cable, connector and interface causes

If the whole image is unstable, suspect the path.

A partly inserted connector, a damaged FPC tail, a cracked solder joint or a wrong timing setting can all produce a line by corrupting one data lane. The distinguishing evidence is that the rest of the image is also disturbed: flicker, wrong colours elsewhere or a shifting image. Reseating the connector and flexing the tail gently while watching the screen is a legitimate diagnostic, and the effect of cable design on this class of fault is covered in the discussion of display specification terminology and in the comparison of industrial and consumer monitors.

Streaks and mura: a different family

Broad, soft and usually a uniformity defect.

A hard line has a sharp edge and a single position; a streak is broad, soft and often changes with viewing angle. Streaks come from backlight uniformity, light-guide scattering, bonding bubbles or cell-gap variation, and they belong to the mura family rather than the line family. The distinction matters because the acceptance criteria differ: a line is either present or not, while a streak is judged against a uniformity limit under defined conditions, as covered in mura and luminance uniformity.

How to localise the fault quickly

Six steps, before any disassembly.

• Photograph the line against a uniform mid-grey and a full white field.
• Change the input to a solid colour and to a test pattern to separate data from panel faults.
• Flex the cable and reseat the connector while watching for a change.
• Warm and cool the affected edge to test bond sensitivity.
• Press lightly at the driver edge to test for a bond or crack response.
• Repeat at the second unit from the same batch to see whether the fault is systematic.

These steps cost nothing and usually identify the owner of the defect before a supplier conversation begins. Where the fault is a colour judgement, the traceability expectations of NIST and the published guidance from the CIE describe how the measurement should be made.

What is repairable and what is not

Half of these faults are terminal.

A loose connector, a damaged cable or a firmware timing error is repairable. A failed source output, a cracked glass or a burn in the driver is not, and attempts to repair a bonded module usually destroy it. The useful distinction for a service organisation is what can be corrected in the field with standard tools, because a repairable-looking fault that is not repairable generates a return, a re-test and a customer complaint for nothing.

Preventing line defects in production

Three controls remove most of the risk.

The first is electrostatic control at every station, because driver damage is invisible until the line appears. The second is a defined flex and mate limit for the cable and connector, with the tail clamped so that it cannot be pulled. The third is mechanical review of every mounting point that touches the glass, since a line defect that appears after a redesign is usually a stress path rather than a component. Together they belong in the assembly documentation that IPC practice describes and in the product’s quality assurance and factory compliance records.

What to record for a supplier claim

Evidence decides the outcome of the argument.

A usable claim states when the fault appeared, whether it changes with temperature or pressure, whether the whole image is disturbed, what the input pattern was, and which batch and panel identification the unit carries. Photographs at two backlight levels and a short video of the flex test are more persuasive than a description. Where the display is used in regulated equipment, the lifetime and safety context set by IEC 62368-1 and the EU EMC Directive also shape how a failure has to be documented, and panel trend data from houses such as TrendForce can show whether a defect pattern is industry-wide. The same evidence pack serves a monitor product such as a patient monitor display, whose interface and certification requirements are set out in patient monitor displays, specs, interfaces and certifications.