News
The short answer
A GOB-coated module can sometimes be repaired, but it should never be treated as a routine repair. The coating that protects the LEDs also locks them in place, so opening it is a second manufacturing operation, not a quick fix.
Repair makes sense only for a small number of confirmed LED or joint faults, on a module where the fault has been isolated, the coating chemistry is known, the reseal material comes from the same batch, and the module can be fully retested afterwards. In every other case — a widespread fault, an unconfirmed cause, an unknown coating, or a module already installed in a screen that must look uniform — replacing the whole module is the safer decision, even though it feels wasteful. The cheapest GOB rework, by a wide margin, is the one prevented by testing the module before it is coated.
On a bare SMD module, a dead LED is accessible. A technician can see it, heat its joints, replace it, and inspect the result directly. On a GOB module, that same LED sits under a continuous protective layer that covers the LEDs, the driver components, and the solder joints. Nothing can be reached without first removing part of that layer.
That single difference changes the whole repair. Removing coating is a mechanical operation next to fragile, closely spaced LEDs. It can crack a neighbouring package, lift a solder joint that was fine, or leave a rough edge that shows on the finished screen. Resealing then has to restore two things at once: protection and appearance. A repair that works electrically but leaves a visible seam, a dull patch, or a colour step has still failed on a display product, because the screen is judged by eye.
This is why published comparisons that label GOB repairability as simply “moderate” are not much help on the factory floor. “Moderate” does not tell a production manager whether this module, with this fault, should be opened at all. That decision needs a fault diagnosis first, and a set of gates that a repair must pass before anyone picks up a tool.
GOB dispensing machine with its module fixture. Rework must locally undo the coating this machine applied, without disturbing the rest of the module.
Most expensive GOB rework starts with the wrong diagnosis. A dark pixel is a symptom, not a cause. The fault may sit in the LED itself, in its solder joint, in a driver component, in a connector, or outside the module entirely — in the cabinet, the cable, or the receiving card. Only the first two cases can possibly be fixed by opening the coating. The others cannot, and opening a good coating to chase them only adds damage.
Work through this order before any coating is touched:
1. Confirm the fault is on this module. Swap the suspect module with a known-good module in the same cabinet position, or move the suspect module to a known-good position. If the fault moves with the module, it belongs to the module. If it stays with the position, the cause is outside the module and coating work is pointless.
2. Map the fault pattern. A single dark pixel, a whole row or column, a block, or a flickering pixel each point to a different layer of the circuit. Record the exact pixel coordinates and the pattern before anything is disturbed, with a photograph of the powered module if possible.
3. Check the easy causes first. Inspect connectors, edge contacts, and mounting for contamination or mechanical damage, and check the power and data path. These checks cost nothing and rule out faults that no amount of coating removal would fix.
4. Decide what “fixed” must look like. Agree in advance whether the repaired module must pass the same electrical test, the same optical check, and the same appearance check as a new module — and whether it will ship as a first-grade module, be downgraded, or be kept as a spare. If that agreement does not exist, stop. A repair without an acceptance rule is how disputed modules end up in a customer’s screen.
| What you observe | What it most likely points to | Should you open the coating? |
|---|---|---|
| One or a few dark pixels, fault follows the module | Individual LED or its joint | Possibly — if the gates in Section 3 all pass |
| Whole row, column, or block out | Driver, circuit trace, or data path | Usually not — opening the coating is unlikely to reach the cause, and the damage risk is higher |
| Fault stays with the cabinet position after a module swap | Cabinet, cable, power, or control hardware | No — the module is not the problem |
| Many scattered faults across the module | Process or batch-level problem | No — local repair cannot fix a widespread cause; quarantine the batch and investigate the process |
This table is a decision aid, not a diagnostic instrument. It does not replace electrical testing, and it does not prove a cause. Its purpose is narrower: to stop coating removal from being the first step, when it should be close to the last.
A GOB module should be opened for repair only if it passes all five gates below. Failing any one of them is enough to choose module replacement instead. This strictness is deliberate: each gate corresponds to a failure mode that turns a repaired module into a field return.
Gate 1 — The fault is small and confirmed. Only a small number of pixels, with the cause isolated to an LED or its joint by the checks in Section 2. A widespread or unconfirmed fault fails this gate.
Gate 2 — The coating chemistry is known. The encapsulant type must be identified from the supplier’s datasheet or the production record. Different chemistries behave very differently when removed: industry practice reports that silicone-type materials can generally be peeled or softened for removal, while cured epoxy is cross-linked and much harder to remove. If the chemistry is unknown, the removal method cannot be chosen safely, and the gate fails. No solvent or temperature should be guessed.
Gate 3 — The removal method is proven on scrap first. Whatever removal approach is planned must first be demonstrated on a scrap or rejected module of the same construction — same coating, same pitch, same board. If the trial damages neighbouring LEDs, leaves residue, or roughens the surface, the method is not ready for a shippable module.
Gate 4 — The reseal material matches. Reseal material should come from the same batch as the original coating wherever possible. Industry reports warn that a different batch of encapsulant can show a visible colour or transparency step on the finished screen, even when the difference looks small in the container. If matching material is unavailable, the repaired area must be judged against the appearance standard agreed in Section 2 — and if it cannot meet it, the module should not ship as first grade.
Gate 5 — Full retest is possible. The repaired module must be able to go through the same post-coating electrical and optical checks the line normally applies, plus a visual check of the resealed area under the same lighting the customer will use. A module that cannot be retested to the normal standard should not be shipped on the basis that it “looks fine.”
A note on field repairs
The gates above are written for a factory or workshop setting. For a module already installed in a customer’s screen, the balance shifts further toward module replacement: the working environment, the lighting for an appearance check, and the control over reseal curing are all worse on site, while the cost of a visible repair in front of the customer is higher.
Keeping calibrated spare modules from the same production batch is, in practice, what makes this decision painless — a point worth settling at order stage, not after a fault appears.
The sequence below describes the logic of a local repair. It deliberately gives no temperatures, solvents, or cure profiles, because those values belong to the specific encapsulant’s datasheet and must be taken from it — not from a general article. Any guide that prints a single temperature or chemical for “GOB removal” should be treated with suspicion.
Step 1 — Protect everything around the fault. Mask or shield neighbouring LEDs, keep-out zones, connectors, and edge contacts before any coating is disturbed. Coating debris and reseal material in a connector cause the same kind of field failure the repair was meant to prevent.
Step 2 — Open only the faulty area. Remove coating locally, over the faulty LED and no further than the repair requires. A wide opening increases the reseal area, the optical risk, and the chance of collateral damage, for no benefit.
Step 3 — Replace the part, then prove the joint. Replace the LED with the same part, or rework the joint if the joint is the confirmed cause. Inspect the joint before resealing — once the area is sealed again, it cannot be inspected optically.
Step 4 — Clean and prepare the opening. Remove all coating debris and residue from the opened area so the reseal material bonds to a clean surface. Contamination trapped under a reseal becomes a permanent inclusion, exactly as contamination trapped under the original coating would.
Step 5 — Reseal in a controlled pattern. Re-apply encapsulant with a controlled dispensing process, not by hand, so the resealed area matches the surrounding coating in coverage and edge definition. The reseal must be cured exactly to the encapsulant supplier’s datasheet profile — under-cured material stays soft, over-cured material can stress the board, and either problem may only appear weeks later.
Step 6 — Retest, including appearance. Run the normal post-coating electrical and optical checks, and inspect the resealed area at normal viewing distance under normal lighting. Record the module, the fault, the repair, the reseal batch, and the test result, so a later field question can be traced back to this repair rather than to the whole production batch.
A desktop CCD jet dispensing machine. The reseal step needs this kind of controlled dispensing — a hand-applied reseal cannot match the surrounding coating edge.
Neighbour damage during opening. Fine-pitch LEDs sit close together. A tool that removes coating over one pixel can easily nick the package next to it. The damage may not show electrically at first, which is why the post-repair optical and electrical retest matters more than the repair itself.
A visible reseal seam. Even a technically sound reseal can fail the appearance test: a step in surface height, a change in gloss, trapped bubbles, or a colour difference from a different material batch. On a video wall, modules sit side by side, so a seam that is invisible on a bench can be obvious on the wall.
Sealing a fault that was never the LED. If the real cause was a driver, a trace, or a connector, a perfect LED replacement changes nothing. The module returns to the field with its original fault plus a reseal seam. This is the failure that Section 2 exists to prevent.
Rushing the cure. A reseal that is handled, packed, or installed before it is fully cured can deform, collect dust, or lose adhesion at the edge. The cure profile comes from the material supplier, and the repaired module should not skip it because a shipment is waiting.
The most reliable way to reduce GOB rework is to move detection earlier. A dead pixel found before coating is an ordinary SMD repair: accessible, inspectable, and cheap to verify. The same pixel found after coating becomes the operation described in this article. That is why the test sequence around the coating step deserves as much attention as the coating step itself — a final functional and optical check immediately before coating, and a full post-coating check before packing, catch faults at the cheapest point available.
Two habits support that sequence. First, keep the production record for each batch — coating type, batch, and cure record — so a later fault can be judged against known material, rather than guessed. Second, agree the rework policy with the customer before production: which faults are repaired, which modules are replaced, what a repaired module may ship as, and how many spare modules from the same batch will be held. Settling those questions after a fault appears, under shipment pressure, is when bad repairs get approved.
Rework quality ultimately depends on how well the original coating was applied and how precisely the reseal can be placed. The GOB-LXS330 glue dispensing machine for LED displays is JHIMS’s dedicated machine for this coating step: a teaching-program machine with a 400 × 400 mm fixture, a running accuracy of ±0.05 mm, a 30 L glue capacity, an input pressure range of 0.5–0.8 MPa, a Panasonic PLC with touchscreen control, parameter storage, and an anti-drip valve.
Those specifications are taken from the JHIMS product page for the machine; JHIMS does not currently list a dedicated GOB machine page on its commercial site, so the product page is the single source for this machine’s specifications and should be read as such.
For the reseal step itself, dispensing precision is the controlling factor, and it is worth checking against a precision reference: the JH-D900 precision dispensing machine lists a repeat accuracy of ±0.02 mm with a CCD vision system and a dispensing area of 860 × 400 mm. The D900 is a precision dispensing machine, not a GOB machine, and it is cited here only as a precision reference for controlled reseal dispensing.
Readers comparing a GOB line should also look at the JHIMS GOB LED dispensing solution page and the process guide to glue dispensing in fine-pitch LED module production, which covers the dispensing variables that decide coating quality before any rework question arises.
Can a single dead pixel in a GOB module always be repaired?
No. It can be repaired only if the fault is confirmed at that LED or its joint, the coating chemistry and removal method are known and proven on scrap, matching reseal material is available, and the module can be fully retested. If any of those conditions fails, module replacement is the correct outcome.
Is it true that GOB modules cannot be repaired at all?
No — that overstates the case in the other direction. Industry reports consistently describe GOB repair as difficult and specialised, requiring local coating removal and resealing, not impossible. The honest position sits between the two slogans: repair is possible in limited, well-controlled cases, and uneconomic or risky in most others.
Why does the reseal material batch matter so much?
Because a display is judged as a wall, not as a single module. A small difference in colour or transparency between encapsulant batches can become a visible patch when the repaired module sits next to its neighbours. That is why the article insists on same-batch reseal material, or an appearance check that the repaired module must pass before it ships.
Should a repaired module be sold as new?
That is a commercial and warranty decision that must be agreed in writing before the module ships — not a technical fact this article can settle. What the engineering side can insist on is narrower: the repaired module must pass the same electrical, optical, and appearance checks as a new module, and its repair must be recorded so it remains traceable.
Bottom line
Treat GOB rework as a gated exception, not a standard operation. Diagnose first, apply the five gates, prove the method on scrap, reseal with matching material under a controlled process, and retest to the normal standard. When a gate fails, replace the module and keep the record — a replaced module costs one module, while a bad repair can cost a customer’s wall.
For process questions about GOB coating or reseal dispensing, contact JHIMS at [email protected] with your module size, pixel pitch, coating type, and the fault pattern you are seeing.
Teams
whatsapp
telegram
