Solder Paste Inspection (SPI): Measurements, Defects and Limits Before Reflow
Solder paste inspection (SPI) checks paste deposits after stencil printing and before components are placed. It can flag a missing, insufficient, excessive or misplaced deposit while the board is still easy to clean and reprint. A 3D SPI system measures deposit height and derived volume in addition to the two-dimensional footprint. It does not prove that a solder joint will be acceptable after reflow, because placement, wetting and thermal behavior occur later.
That stage boundary is the main reason to use SPI. It gives the assembly team an early observation of a controllable process, rather than waiting for a bridge or open joint to appear at final inspection. Philifast’s assembly-defects guide places paste printing among the causes of downstream defects; this article examines the paste checkpoint itself.
Where SPI sits on the line
The normal sequence is stencil print → inspect paste → place components → reflow → inspect joints and test the assembly. A suspect paste image should trigger review of the print, stencil, board support, paste condition or alignment. Reprinting a board after an agreed cleaning process may be possible at this stage. Once components are placed and reflowed, corrective work is different and can be more disruptive.
The printer first pushes paste through stencil apertures onto exposed pads. The stencil lifts away, and the paste must release with enough shape and volume to support the later joint. SPI compares each programmed deposit with its expected location and geometry. A low reading is therefore a symptom of the print, not yet a diagnosis: the aperture may be partly blocked, paste may not have released cleanly, the board may lack support, or the measurement recipe may be wrong. The operator needs the image and the process history to distinguish those possibilities.
Illustration: SPI is an early process checkpoint; the actual inspection and rework plan is specific to the product.
What the measurements mean
| SPI output | What it describes | Why it matters |
|---|---|---|
| Area or coverage | The deposit’s visible footprint on or near its target pad | A narrow or missing footprint can indicate poor stencil release; spread toward another pad can signal a bridge risk. |
| Position or offset | Difference between deposit location and programmed pad location | A shifted print can reduce overlap with the component termination or connect neighboring deposits. |
| Height | Vertical profile of the deposited paste, available in 3D measurement | A low or uneven deposit can reveal release or support problems. |
| Volume | Estimated three-dimensional amount of paste | It helps compare deposits with the programmed target and trend printer performance. |
| Shape / bridge indication | Irregularity or connection between deposits | It can signal a print defect requiring review before placement. |
These are paste measurements, not final solder-joint dimensions. RichPCBA’s 3D SPI explanation identifies the same core parameters, but equipment-specific accuracy and speed claims on any supplier page should not be used as universal acceptance limits. A 2D system cannot directly measure deposit height or volume in the way a 3D system does.
Koh Young’s inspection explanation identifies volume, height, area and offset as the core outputs of 3D SPI. They are related, but not interchangeable. A deposit can cover the expected area yet be too low; another can have enough volume but be offset toward a neighboring pad. For a fine-pitch component, that offset may matter more than it does on a large isolated pad. The useful record is the measurement at a named pad on a named board, compared with a limit chosen for that feature.
Illustration: Typical patterns that prompt investigation; the shapes are schematic rather than pass/fail criteria.
How to set limits without a misleading universal percentage
There is no single paste-volume band that fits every component, stencil opening and board finish. Programmed targets and alarms depend on the stencil design, pad geometry, component package, paste, printer setup and measured process capability. The same deposit may be tolerable for one joint design and unacceptable for a fine-pitch neighbor. Define the critical pads and evaluation rules with the assembly engineer and the product’s acceptance requirements.
For a first article, review representative deposits and false alarms before treating a numeric alarm as an automatic reject. For repeated production, trend location and volume by pad and by print run. A rising pattern of low-volume deposits may point to stencil clogging or paste release; a common offset across the board may point to registration or support. The corrective action is a process investigation, not merely changing an alarm until it stops firing.
Read patterns before changing the printer
Different patterns suggest different investigations. One low deposit repeated at the same fine-pitch aperture points toward stencil opening, paste release or cleaning. Many pads shifted in the same direction suggest registration or board support. Excess paste on adjacent pads suggests a bridging risk and may call for a stencil, print-pressure or alignment review. Random isolated alarms may indicate unstable printing, contamination or an inspection-program issue. None of these patterns proves a root cause on its own.
The response should preserve the board and panel identity, pad location, measured deviation, operator disposition and any process adjustment. If the same defect recurs after cleaning, record the recurrence and escalate it; repeated reprints without diagnosis can hide an unstable process. If the alarm is a false positive, document why the deposit is acceptable under the agreed criteria before adjusting the recipe. This makes SPI useful for process control rather than a machine that merely counts red flags.
Why one fixed percentage fails
An SPI recipe often compares measured volume or area with a programmed target, but the target comes from the stencil aperture and component land design. A large thermal pad may be intentionally divided into several paste apertures; treating each as though it should cover the whole copper pad would create meaningless rejects. A fine-pitch QFN or connector row may need tighter attention to lateral spread because a small bridge can join adjacent pads. The board finish, paste type, stencil thickness and release behavior also affect the print. Set limits by feature and validate them against first-article results and downstream defects, rather than importing a generic percentage from a supplier page.
What SPI cannot see
SPI takes place before placement and reflow. It cannot confirm component orientation, solder wetting, hidden BGA joint shape, electrical continuity or circuit function. Post-reflow AOI evaluates visible assembly features; X-ray can help examine certain hidden joints; electrical and functional tests answer different questions. Philifast’s PCBA testing guide outlines those later checks. A good test plan uses the methods together according to the failure modes that matter for the product.
Similarly, a good paste print does not guarantee a good joint. An incorrect part, placement offset, poor wetting or unsuitable reflow profile can still create a defect. Conversely, a flagged SPI measurement may require engineering review rather than automatic scrap if the alarm is poorly tuned.
SPI is also not the correct evidence for a component that is hand soldered later, a through-hole joint made in another process, or a hidden BGA joint after reflow. Those features belong to the inspection/test method for their own stage. If a final defect appears, retain the SPI record alongside placement, reflow and post-reflow evidence so the team can decide whether the printing stage contributed. The absence of an SPI alarm does not prove that the print was causally unrelated; it only states that the programmed measurements passed their limits.
Diagnose the spatial pattern before changing the printer
An isolated alarm and a repeated line of alarms are different process signals. A single low deposit can come from a local aperture, paste transfer or measurement issue. A row of low deposits may point toward a stencil region, board support, wipe condition or print setup. A consistent offset across one board area deserves a different review from an offset that follows one package type across the panel. Treat these patterns as hypotheses to check against the print, not as automatic diagnoses.
Start with the board and panel map. Compare the flagged pads by reference designator, package, board position and print side. Check whether the same location was abnormal on the previous print or whether the pattern appeared after a setup change. Then inspect the actual print and relevant process records: stencil and aperture revision, paste identity and condition, print recipe, support and clamping, wipe history and measurement recipe. A machine alarm that appears only after a recipe edit may reflect the measurement setup rather than a suddenly changed deposit.
The purpose of this review is to preserve evidence while finding the cause. If the physical deposits are consistently shifted, adjust the responsible alignment or support condition and verify the next print. If volume changes broadly, review paste transfer and stencil condition before narrowing limits. If a few small apertures repeatedly flag while the joints later perform acceptably, engineering may decide that pad-specific limits need validation. Record the reason for every process or threshold change and compare the resulting print with the same set of features.
Control the recipe and escalation decision
An SPI result depends on how the board is registered, which pads are measured, what the software calls nominal, how height or volume is calculated, and which limits trigger review. Keep the inspection program tied to the PCB revision, panel layout, stencil revision, side and paste process. A design change that moves a pad, changes the panel, or changes a stencil aperture can invalidate an old recipe even when the product name remains unchanged.
For a new product or a changed process, agree on the sample used to tune the program and who approves its limits. Keep representative images or measurement exports for accepted, borderline and rejected deposits where the process requires them. When an alarm occurs, define who may stop the line, clean and reprint, accept under an approved disposition, or request engineering review. Do not use “the software passed” as a substitute for a disposition record when a downstream defect has been found.
The scope of a cited standard matters. IPC-7527 is a guideline for evaluating solder-paste printing and supporting process improvement; its own scope distinguishes that purpose from solder-paste material requirements and stencil design. J-STD-005B covers solder-paste material characterization and testing, not the process acceptance limit for every printed pad. Neither reference creates a universal SPI percentage limit for every pad and assembly. Check the IPC revision table, name the contractually selected revision, and validate product-specific limits with the process owner.
Trend confirmed defects separately from false calls. A rising alarm count can mean the process moved, the recipe changed, or more marginal deposits are being reviewed; the count alone cannot distinguish them. Tie SPI records to the same board identity and revision used by downstream inspection or test. The BGA X-ray guide i PCBA test-plan guide explain what later evidence can and cannot establish about the finished assembly.
When several pad sizes share a board, avoid treating one average score as the health of the entire print. Small apertures, large thermal pads and fine-pitch leads respond differently to stencil transfer and can have different functional consequences. Review representative feature groups and any known high-risk components separately. If an inspection system reports a board-level summary, request enough location-level data to identify whether a single local exception is being hidden by otherwise normal deposits. Keep the summary for trending, but make the disposition from the relevant features and the agreed process limits.
Questions to settle with the assembler
Ask which pads are inspected, whether the method is 2D or 3D, what measurement and review criteria apply, how the first article is approved, and what record will be available for a rejected print. Also agree on cleaning/reprint limits and who can authorize a process change. Those answers are more useful than a generic claim that “SPI is used,” because they define what evidence the inspection will produce for your assembly.
For a board with several cavities in one panel, ask whether reports identify the individual board position. For two-sided SMT, ask whether both print stages are covered and how underside components are supported during the second print. Request the link between an alarm, its disposition and any later AOI or X-ray finding. This creates a traceable answer to “what happened on this board?” rather than a batch-level claim that SPI was present on the line.
