How to Design an Efficient SMT NG Board Handling Process?

An efficient SMT NG board handling process is built around one rule: a board that fails SPI or AOI must never become an uncontrolled object in the normal production flow. The line should identify the result, route the board to a defined NG, review, or rework position, preserve its order and identity, and release it only after the required disposition is complete. A well-planned SMT conveyor solution for stable PCB handling treats this as a complete material-flow problem, not just a reject bin added beside an inspection machine.
The practical challenge is that NG boards are not all the same. One may need a quick operator confirmation, another may require component replacement, and a third may be held for engineering review. At the same time, good boards should keep moving when the downstream route is available. This guide explains how to design the process from inspection result to final disposition, including equipment layout, buffer capacity, traceability, operator work, controls, exception recovery, and acceptance testing.
Table Of Contents
- What An Efficient SMT NG Board Process Must Achieve
- Map The Full Route From Inspection Result To Disposition
- Separate Good, NG, Review, Rework, And Uncertain States
- Design The Physical Handling Equipment And Buffer Points
- Protect Board Identity, Order, And Traceability
- Balance Capacity, Operator Work, And Line Continuity
- Build Controls, Interlocks, And Recovery Logic
- Validate The Process Before Production Release
- Frequently Asked Questions
What An Efficient SMT NG Board Process Must Achieve
Define efficiency as controlled flow, not simply faster movement
In an SMT line, NG handling begins when a process decides that a board cannot be released as a normal good board. That decision may come from SPI, AOI, an operator review, an in-circuit test, or a downstream quality gate. Efficiency means the board reaches the correct next state with minimal ambiguity and without forcing every other board to follow the exception route. A conveyor that moves quickly but leaves operators guessing which board failed is not an efficient NG process.
The process should therefore make five outcomes explicit: pass and continue, NG and hold, review required, rework in progress, and rework returned for verification. Some factories also need an identity-uncertain state for a barcode failure, manual removal, or interrupted transfer. These states should be visible in the equipment interface and understandable to operators, quality staff, maintenance, and automation engineers.
Protect four production objectives at the same time
First, the good-board route must remain physically separate from boards that are not released. Second, the NG route must be accessible enough for inspection, repair, and disposition without unsafe reaching or repeated line stops. Third, board identity and order must remain recoverable after a stop, manual intervention, or rework return. Fourth, the process must provide a clear response when the NG area is full or the next destination is unavailable.
These objectives often compete. More buffer positions may protect continuity but consume floor space. A manual review table may be easy to access but weak for traceability. A tightly integrated automatic route may reduce handling but require better interfaces and fault recovery. Before selecting equipment, write down which trade-off matters most for the factory’s product mix and quality workflow. This prevents a local improvement from creating a larger line-level problem.
Pay particular attention to the transfer points where an NG board changes direction, stops for review, or returns to the main route. The guide to common PCB handling problems caused by poor conveyor design is useful for identifying skew, unstable support, uncontrolled stops, and difficult recovery before they become routine production events.

Map The Full Route From Inspection Result To Disposition
Start with the physical and digital event sequence
Draw the route as a sequence of events rather than a list of machines. A typical sequence includes board arrival, board detection, result receipt, result confirmation, route selection, transfer, slot or station confirmation, operator action, disposition entry, and release to the next process. For each event, identify what the board is doing physically and what the control system believes is happening. The most dangerous gaps appear when those two views can diverge.
For example, a board may leave AOI before the result arrives, or an operator may remove a board while the software still shows it in a buffer slot. The line needs a defined hold state for both situations. It also needs a recovery method that does not rely on an operator moving boards until the screen “looks right.” The result of how PCB conveyors improve SMT line capacity depends on this kind of controlled flow, because unplanned stops and manual reconciliation reduce practical capacity even when nominal conveyor movement is fast.
Include every route that can occur in real production
Do not map only the normal pass and reject paths. Include consecutive NG boards, a full buffer, a downstream machine not ready, a result timeout, barcode read failure, a board removed for engineering, a rework return, and an operator correction. Include the path for boards that are placed on hold for a quality decision and the path for boards that must be scrapped or transferred to a separate area. These cases determine whether the process remains stable during the moments when it is most likely to be tested.
A practical route map also records handoff ownership. Production may own the conveyor, quality may own disposition, engineering may own an uncertain defect, and maintenance may own the equipment fault. If ownership is unclear, a board can remain physically safe but operationally forgotten. Assign a next action and responsible role to each exception state before the line is installed.
Measure the decisions that create waiting
Record how long a board waits for an inspection result, operator review, repair, engineering release, and downstream acceptance. These waiting times determine whether the line needs a short controlled hold, a multi-position buffer, a separate rework route, or a process change. The purpose is not to promise a universal buffer size; it is to make the design traceable to the factory’s actual decision delays and board arrival pattern.

Separate Good, NG, Review, Rework, And Uncertain States
Do not reduce every failed board to one reject category
A board can be physically present at the inspection exit while its quality status is still pending. It can be confirmed NG but not yet assigned to a repair action. It can be repaired but waiting for verification. It can pass a second inspection but still require a controlled return to the main route. Treating all of these conditions as “reject” hides the next required action and makes it harder to design correct movement rules.
At minimum, define separate states for inspection pending, pass released, NG held, manual review, rework required, rework in progress, rework returned, and identity uncertain. The factory may add engineering hold, scrap confirmed, or quality release depending on its process. Each state should have a permitted location, a permitted movement, a required confirmation, and an owner. If the board cannot be safely classified, it should remain in a hold state rather than silently joining the good-board flow.
Make good-board protection physical as well as digital
Software status alone cannot prevent mixing if the board can be placed on the same open surface as good product. Use dedicated buffer positions, separated lanes, clear labels, physical stops, or controlled operator access where appropriate. The design guidance on why AOI needs a controlled NG buffer is especially relevant when the inspection decision and operator disposition do not finish at the same time.
Physical separation should also cover rework returns. A repaired board should not look identical to a newly inspected pass board until the required verification is complete. If the route passes through a common conveyor, the control system should use an explicit return event and a second identity or quality confirmation. This is more reliable than expecting operators to remember which board came from which table.
The same principle applies at SPI. A board that fails paste inspection may be sent for review, cleaning, or reprint-related action before it reaches placement. Use a dedicated state and location to prevent SPI NG boards from mixing with good boards, then carry that identity through later inspection and rework decisions.
Write operator actions in terms of decisions
Operators need instructions such as “confirm defect,” “send to rework,” “hold for quality,” “remove and record,” or “release after verification.” Messages such as “NG error” or “conveyor fault” are too broad when several actions are possible. An HMI should show the board location, reason for hold, permitted next action, and whether the line can continue. Clear wording reduces trial movements and makes training easier across shifts.

Design The Physical Handling Equipment And Buffer Points
Choose equipment by route function
A straight conveyor can connect compatible equipment, but an NG process often needs more than a straight handoff. A buffer provides controlled waiting positions. A turn conveyor changes line direction. A translation unit moves a board or carrier between parallel routes. An inspection conveyor provides a stable interface for review or positioning. These functions should be chosen from the route map rather than from a generic request for “an NG conveyor.”
When the exception route needs inspection access and stable positioning, review the I.C.T high-end SMT inspection conveyor as an equipment-role reference. When the line must turn into a review or repair area, a 90-degree SMT turn conveyor may be relevant. When two parallel paths must share a review or rework route, evaluate a PCB translation shuttle conveyor for board handling.
Check support and access across the complete PCB range
Board dimensions, panelization, thickness, edge clearance, component overhang, carrier use, and board orientation affect how a PCB behaves at every transfer. A design that works with a rigid single board may not behave the same way with a large panel or a flexible board. Check the transition between conveyor surfaces, side-rail adjustment, center support, stop position, sensor location, and the space available for a technician to recover a board.
Operator access is part of the equipment design. A buffer that stores boards efficiently but requires reaching across a moving path can create a handling risk. A review station should make the board, result, and next action visible. A rework route should prevent repaired and unreleased boards from being returned through an ambiguous entry point. Service doors, belts, sensors, rails, and electrical cabinets also need enough access for routine work.
Reserve space for exception movement
Draw the largest board, the longest carrier, and the complete movement envelope of a turn or shuttle. Add the operator standing area, safety clearance, maintenance opening, carts, and any doors or utilities. The nominal machine footprint is not the usable installation footprint. Many NG processes fail in practice because the equipment fits the line drawing but the operator cannot remove the board, open the guard, or inspect the sensor without stopping adjacent equipment.
Plan A Controlled NG Board Route
Share your inspection points, PCB range, rework path, buffer need, line direction, traceability events, and service constraints with an I.C.T engineer before selecting the handling equipment.
Protect Board Identity, Order, And Traceability
Keep physical position and quality status connected
Traceability is not created by a barcode reader alone. The process must connect board identity, inspection result, physical position, movement event, operator intervention, and final disposition. A barcode may be read before SPI, at AOI, at the NG entry, or at the rework return depending on the line architecture. What matters is that a board entering an exception route does not become anonymous while the system is waiting for a decision.
Keep transport status and quality status separate but related. “Board is in buffer position three” describes where the board is. “Board requires rework” describes what may happen next. A PLC can confirm physical events such as board present, transfer complete, blocked, and removed. A higher-level line or MES system may record inspection result, repair action, quality release, and production history. The interfaces and ownership should be agreed before installation.
Preserve order through queues, removals, and returns
Board order becomes difficult when several NG boards arrive together, when one board is removed for review, or when a repaired board returns after later boards have already passed. Define whether the process is first-in-first-out, result-priority, operator-selected, or controlled by a quality rule. If order can change, require an explicit event and record which board moved into which position.
Operators should not have to rebuild the sequence from memory. The station can show slot occupancy, board identity, result state, and permitted action. If a board is manually removed, the system should require confirmation and record the new physical state. A reset that only clears an alarm while leaving the old board record in the slot can create a hidden mismatch that appears much later at rework or final test.
Use visual and digital controls together
Labels, stack lights, HMI messages, slot indicators, barcode confirmation, and event logs should reinforce the same process state. A red indicator should not mean “machine fault” in one area and “NG board waiting” in another. The operator should be able to identify the board, the reason it is held, whether the line may continue, and what confirmation is required. This combination of physical and digital clarity is especially important during shift changes and maintenance recovery.

Balance Capacity, Operator Work, And Line Continuity
Size the buffer around decision delay and arrival pattern
NG buffer capacity should be derived from the time required to review, repair, release, or remove boards and from the way NG results arrive. Occasional isolated defects may need a short controlled hold. Consecutive failures, batch review, remote rework, or a quality team that works at intervals may require more positions. The design should also account for board pitch, slot spacing, carrier use, operator reach, and the consequences of a full buffer.
Use observed or agreed process assumptions for the design review: expected board arrival interval, time to receive the result, average and worst-case review delay, repair response, and maximum acceptable upstream hold. These assumptions are not universal performance claims. They are inputs that the factory must confirm and update when the product mix or inspection strategy changes.
Separate short variation from a persistent bottleneck
A buffer can absorb short differences between inspection, review, rework, and downstream release. It cannot solve a persistent defect source or an undersized repair process. If the buffer fills repeatedly, ask whether the root cause is solder-paste or placement quality, inspection tuning, repair staffing, engineering response, downstream capacity, or a missing route. Adding positions without reviewing the cause may move the visible queue while increasing WIP and traceability work.
Look at both empty and full behavior. An empty NG buffer should not delay a pass board. A partially occupied buffer should preserve the planned order and remain accessible. A nearly full buffer should warn the responsible team before it blocks the AOI exit. A full buffer should create a defined upstream hold or controlled stop, not an instruction to place more boards on an untracked surface.
Design the operator workload, not only the machine cycle
Count the actions required per NG board: identify, inspect, confirm, label if required, route, record, repair, return, and release. Review how many of these actions can occur while the main line continues and which actions require a safe stop. The best arrangement may be a small automatic buffer plus a dedicated review station rather than a large buffer that forces operators to walk, reach, and reconcile records.
Shift changes and breaks should be included in the workload model. A process that works when one experienced operator is present may become unstable when the area is handed over. Clear status display, explicit ownership, and recoverable records help the next operator understand why a board is held and what must happen before it can move.
Build Controls, Interlocks, And Recovery Logic
Define normal, blocked, and uncertain control states
The control sequence should include board detected, result pending, route ready, transfer active, destination occupied, transfer complete, NG hold, review required, rework return, blocked, and identity uncertain. The upstream equipment needs a release permission. The destination needs a ready signal. The transfer needs sensor confirmation. A timer can support a sequence, but it should not be the only proof that the board reached its position.
Interlocks should prevent a board from entering an occupied position, a shuttle from moving while the receiving side is unsafe, or a pass release from occurring while the inspection result is missing. When the NG area is full, the signal should identify a capacity condition rather than presenting it as a generic motor fault. When communication times out, the system should hold the board and show the required confirmation rather than silently defaulting to pass.
Make recovery a designed process
Test what happens after an emergency stop, power loss, sensor interruption, communication timeout, board skew, manual removal, or restart. The recovery sequence should establish the physical board position, compare it with the stored state, and require a clear operator confirmation when the two do not agree. It should not move every board automatically to “reset” the station.
Recovery also needs a rule for a board that cannot be identified. The safest process may be to move it to a quarantine or review position, mark it as uncertain, and require quality confirmation. The exact rule depends on the factory’s traceability system, but the principle is constant: uncertainty must be visible and must not be converted into a normal pass event by convenience.
Connect the controls to the complete line
Agree the handshakes with SPI, AOI, rework, downstream equipment, barcode readers, and any MES or line-management system. Define which system owns the quality result, which system owns the physical position, and which event permits release. This boundary avoids duplicated commands and unclear responsibility during faults.
Ask suppliers to provide an event list, state diagram, I/O or communication map, alarm list, manual intervention rules, and restart procedure. These documents are more useful for commissioning than a generic statement that the conveyor is “automatic.” They also give production, quality, maintenance, and automation teams a shared way to review the process before the equipment arrives.

Validate The Process Before Production Release
Test the hardest board and the complete exception route
Acceptance testing should use the largest panel, smallest PCB, most flexible board, most difficult orientation, and actual carrier or support method intended for production. Run pass, NG, review, rework, rework-return, and identity-uncertain cases through the complete route. An empty conveyor running smoothly does not prove that a populated board will transfer safely or that the result state will remain attached to the correct physical board.
Test normal flow and abnormal flow separately. Normal tests should confirm stable support, sensor detection, transfer timing, slot assignment, operator visibility, and downstream release. Abnormal tests should include consecutive NG boards, full buffer, blocked destination, missing result, barcode failure, board skew, manual removal, emergency stop, power recovery, communication timeout, and wrong-slot confirmation. For each case, record the expected state, actual state, alarm, operator action, and final disposition.
Use a buyer-focused release checklist
| Validation area | Question to confirm | Evidence to keep |
|---|---|---|
| Route separation | Are pass, NG, review, rework, and uncertain boards sent to distinct permitted states? | Route test records and state map |
| Board stability | Do support, rails, stops, and handoffs work across the full PCB and panel range? | Representative-board transfer test |
| Capacity | Does the process report available, occupied, nearly full, and full conditions before the upstream line is affected? | Buffer and blocked-flow test |
| Traceability | Does physical position remain aligned with identity, inspection result, and disposition after intervention? | Barcode, position, and event-log review |
| Recovery | Can the line restart without releasing an unidentified or unreleased board? | Fault, restart, and operator-recovery record |
| Serviceability | Can operators and technicians reach boards, sensors, belts, rails, and controls safely? | Maintenance and jam-recovery trial |
Release should require agreement from production, quality, maintenance, and automation owners. The sign-off should cover the board range, result logic, buffer capacity, operator work instruction, traceability events, safety behavior, maintenance access, and recovery rules. If a limitation remains, record it as a controlled operating condition rather than leaving it as an informal assumption.
Reviewed by I.C.T Engineer: an efficient NG board process is ready for production only when it controls abnormal boards without losing the identity, order, or recoverability of the good-board flow.
Build A Stable SMT NG Handling Process
Discuss inspection results, NG routing, board range, buffer capacity, traceability, controls, rework access, and FAT requirements with the I.C.T team for a complete SMT line solution.
Frequently Asked Questions
SMT NG Board Handling Questions
What Is The Most Important Rule In An SMT NG Board Handling Process?
The most important rule is to keep an NG or uncertain board out of the normal good-board route until its status is confirmed. The process should assign a defined physical location, quality state, next action, and responsible role. This protects good-board flow while giving quality and production teams a controlled way to review, repair, return, or quarantine the board.
When Does An SMT Line Need An NG Buffer?
An NG buffer is useful when inspection decisions, operator review, repair, or downstream release take longer than the immediate board handoff. It provides controlled positions for boards that cannot yet continue and can reduce improvised placement beside the line. Capacity should be based on arrival pattern, decision delay, rework response, access, and the consequence of a full buffer.
How Can A Factory Prevent NG Boards From Mixing With Good Boards?
Use separate physical positions or routes, clear state indicators, controlled release rules, and identity confirmation at manual intervention points. The control system should distinguish pass, NG, review, rework, and uncertain states instead of treating every board as ordinary conveyor waiting. The operator should always see where the board is, why it is held, and what action is allowed next.
What Equipment Can Be Used In An NG Board Handling Route?
The equipment depends on the route function. A buffer provides controlled waiting positions, a turn conveyor changes line direction, a translation unit connects parallel paths, and an inspection conveyor supports stable review or handoff. Select the function from the factory layout, PCB range, operator access, control interface, and rework route rather than choosing a generic reject conveyor by name.
What Should Be Tested Before Releasing An NG Handling Process?
Test pass, NG, review, rework return, missing result, barcode failure, full buffer, blocked destination, board skew, sensor fault, emergency stop, power recovery, communication timeout, and manual removal. Confirm board stability, state transitions, physical position, identity, alarm behavior, operator action, and restart rules. Use the hardest intended board and keep evidence for both normal flow and exception recovery.







