How to troubleshoot dead pixels, modules, and flickering

LED faults are noisy: one black diode, a dark rectangle, a flickering column, a rainbow module, or a wall that only misbehaves on camera. The mistake is to start by remapping the entire wall or replacing the processor. The method is to change one variable at a time and decide which layer failed.

This tutorial is a field isolation guide. It assumes you can run test patterns from the processor or control software.

Name the symptom precisely

Write one sentence:

• Single pixel or small cluster — likely a failed LED or a tiny module defect.
• One module (often 250×250 mm or similar) — module, ribbon, or hub-board connection.
• One cabinet — receiving card, cabinet PSU, cabinet file, or both data ports.
• From cabinet N to the end of a chain — jumper, port budget, or sending-card output.
• Whole wall — processor, input signal, house power, or brightness/blackout.
• Only on camera — refresh, scan, shutter, or rolling-shutter beat. The wall may be innocent.

Take a photo of the fault and of the cabinet ID test pattern. You will rotate modules and forget which hole was cursed.

Dead pixels versus dead modules

A dead pixel is one RGB site. Many indoor products allow a few. If it sits in a face on IMAG, replace the module when you have a spare; do not spend an hour on one diode with a soldering iron unless you are a depot tech.

A dead module is a rectangle. Typical causes:

1. Ribbon or flat-flex not seated (the most common “failed LED” in rental).
2. Hub board or module connector damage from a rushed front-service pull.
3. Module logic failure.
4. Wrong module in a mixed-pitch spare kit.

Isolation:

1. Reseat the ribbons. Power down if the manual says so.
2. Swap the suspect module with a neighbor. If the fault moves with the module, replace the module. If the fault stays in the hole, the cabinet (hub, ribbon, receiving card) is the problem.
3. Swap in a known-good spare from the same pitch and scan type.

Do not mix GOB/COB/SMD cosmetics in the center of a wall unless you like a visible patch.

Cabinet-level failures

If a whole cabinet is black, dim, or the wrong slice of video:

1. Confirm the cabinet has power (fan, PSU LED, or a voltage check). A dark cabinet with no PSU light is power, not mapping.
2. Confirm data in and data out LEDs on the receiving card. No link on “in” means the problem is upstream. Link on “in” but not “out” can be a card or a setting.
3. Reseat the two Ethernet jumpers. Touring tabs break.
4. Swap the receiving card with a preloaded spare. If the cabinet returns, program the spare to the house file and retire the bad card.
5. If the cabinet shows a scrambled internal map, reload the cabinet file to that card only.

A cabinet that shows a repeated or shifted image often has the right power and the wrong map slot. Check software before you condemn hardware.

Flicker you can see with your eyes

Visible flicker or shimmer on the wall itself (not only in a viewfinder) usually means:

• Unstable power (voltage sag, shared motor circuit, loose powerCON)
• Data errors (CRC, exceeded port pixels, bad cable, EMI from dimmers)
• Receiving-card or sending-card overheating
• A module in a death rattle (often one module strobes)

Isolation:

1. Drop brightness by half. If flicker vanishes, you may be hitting PSU limits or a low-grayscale bug. Note it and test circuits.
2. Take content to a static test pattern. If flicker remains, it is not the media server.
3. Unplug the last half of a data chain. If the remaining wall is stable, the fault is in the dropped section or you were over the pixel budget.
4. Replace the first jumper after the last good cabinet.
5. Feel (carefully) for a receiving card that is much hotter than its neighbors.

Packet errors often show as sparkles, tearing, or a cabinet that “rains.” That is data. Solid rhythmic flicker on the whole wall is more often power or a processor output.

Flicker that only appears on camera

This is the IMAG special. The wall looks stable, the director sees bands or strobing.

Causes:

• LED scan / multiplex beating with the camera shutter
• PWM brightness method beating with rolling shutter
• Frame rate mismatch (29.97 camera vs 60 wall, etc.)
• Auto shutter or auto flicker-reduction on a camcorder fighting the wall

Fixes to try in order:

1. Match wall frame rate to the house (59.94 with 59.94 cameras is a common US church standard).
2. Change camera shutter to a multiple or factor that the LED vendor suggests (often around 1/60, 1/120, or a specified angle). Turn off auto shutter.
3. Raise or lower LED refresh / scan settings in the cabinet file if you know what you are changing — and test on a spare cabinet first. Wrong scan settings can damage the look or the hardware.
4. Slightly change brightness; some PWM bands move.
5. Change camera position or lens. Moiré and scan lines are angle-dependent.

Do not “fix” camera flicker by turning the wall into a blurry mess unless the producer accepts that trade.

A 10-minute triage order

When the room is filling and someone is shouting:

1. Is program still in the processor? (Input lock)
2. Did a breaker open? (Which columns)
3. Did a data port drop? (Link lights)
4. Is it one module? (Swap or ignore until after)
5. Is it camera-only? (Tell IMAG to park shutter; do not rebuild the map)

After the event, do the slow isolation. Heroes who remap live create two problems.

Parts to keep in the house kit

• Two modules of each pitch in service
• One receiving card preloaded with the house cabinet file
• Power and data jumpers
• A small screwdriver / module puller the product wants
• A printed map of ports and circuits
• Photos of a known-good test pattern on this wall

If you cannot swap a module in five minutes, practice on a spare cabinet in the shop. Front-service tabs teach humility.

Common mistakes

Replacing the processor first. Processors rarely kill a single module.

Firmwaring the whole wall during a fault. Now every card is an unknown.

Using a spare module from a different scan type. It may light. It will look wrong and may not be safe for that driver.

Ignoring heat. A rack in a closet will flicker at minute 40 of the concert.

Declaring “dead pixels” on a dirty wall. Smoke-juice residue and pollen look like failed diodes until you clean them (see the maintenance tutorial).

Log the fault

Write: date, cabinet ID, module position, what you swapped, what fixed it. Patterns appear — a wet outdoor connector, a bad batch, a column that always loses a jumper when the wall is trimmed in.

Troubleshooting is not intuition. It is moving the fault or making it stay, then replacing the layer that moved.