Whether to install SPI is usually decided by gut feel. This evaluation takes a different approach: first run a pFMEA to score each failure mode in the printing step and calculate the RPN, and let the numbers decide what to fix first; then test on an actual machine to confirm the equipment’s cycle time and repeatability can keep up with the line.
1. Current state: quality held up by FAI
Solder paste printing has three things to check: whether the paste volume is correct, whether the print position has shifted, and whether there is foreign matter contamination. Specifications were in place, but in practice quality was confirmed mainly by FAI (first article inspection), with only sampling checks during production.
The difficulty lies in the paste’s physical nature. It is a non-solid material that deforms very easily during handling and measurement, so it is inherently hard to measure; existing inspection methods weren’t accurate enough to give precise results. Assurance of quality and yield was therefore concentrated too heavily on FAI, with no real-time, comprehensive monitoring: once the first article passed, nobody knew about the thousands of boards in between.
At the same time, pin count kept rising: 800K → 1KK. More points mean sampling covers a smaller share, and the limits of human inspection become more obvious.
2. pFMEA: RPN sets the priorities
Following process failure mode and effects analysis, each of the three risk items was scored for severity (S), occurrence (O) and detection (D), and the three were multiplied to get the risk priority number (RPN):
| Failure mode | S | O | D | RPN |
|---|---|---|---|---|
| Solder paste volume within specification | 8 | 7 | 8 | 448 |
| Print position shift (X-Y) | 6 | 5 | 8 | 240 |
| Foreign matter contamination (fibers, dust, etc.) | 5 | 7 | 6 | 210 |
All three have high detection scores (6 to 8), meaning current methods have a hard time catching them in real time. That is the main reason the RPNs are high, and exactly what SPI is meant to solve. Solder paste volume ranks first at 448 and is handled first.
3. Machine verification: cycle time and repeatability
An evaluation can’t rely on the catalog alone. Test conditions: CKD VP5200 equipment with 5 µm / 2.5 µm lenses, test ratio 1:10, 34,243 pads inspected in total.
| Item | Result |
|---|---|
| Single inspection time | 15 seconds (including loading and unloading) |
| Full strip estimate | 15 seconds × 10 = 150 seconds, meets target |
| Repeatability test | One sample inspected 30 times in a row; standard deviation (STDEV) extremely low |
Pay special attention to repeatability. An extremely low standard deviation means the equipment gives highly consistent results when inspecting the same object repeatedly. The measurement system itself must not be a source of variation, or every judgment downstream is meaningless. Only when both efficiency and stability pass is implementation recommended.
4. Estimating payback
The benefit of SPI isn’t “how many bad boards it catches” but yield. The estimate is simple and needs only three numbers:
- First get annual output value (annual volume × unit value).
- Set scenarios for yield improvement, for example +0.5% and +1%, and multiply each by annual output value to get annual savings.
- Total equipment investment ÷ annual savings = payback years.
In practice, each 0.5% gain in yield roughly halves the payback period. The value of this method is that it moves the discussion from “the equipment is expensive” back to “how much do we lose each year without it,” which is the cost actually being lost. Actual figures vary with each plant’s volume and unit price; plug your own numbers into this framework.
5. Conclusions and next steps
- Abnormal solder paste volume is the higher-RPN risk item and must be addressed first.
- In mass production, the inspection cycle must be kept strictly within 150 seconds.
- SPI reduces defects upstream rather than sorting them out afterward.
- Next, a high-spec industrial PC will be added to handle the data volume, moving toward 100% inspection and steadily reducing reliance on manual work.
About this document
This page is compiled from the JIAHON Engineering Department’s SPI equipment evaluation record. The customer’s capacity, output value and commercial terms have been removed; only the evaluation method, risk analysis framework and machine test data are kept.
Standards cited
Refer to the current version from each issuing body for the standards cited here.
- IPC J-STD-001 Requirements for Soldered Electrical and Electronic Assemblies. Issued by the Global Electronics Association (formerly IPC).
- IPC-A-610 Acceptability of Electronic Assemblies. Issued by the Global Electronics Association (formerly IPC).
Further reading: for selection, acceptance and line cases explained more clearly than a spec sheet: troubleshooting solder paste printing defects (working back from defect shape to the consumable to change); a full guide to how an SMT line works (printing, placement, reflow, SPI / AOI).
CKD is a trademark of its respective owner; JIAHON is not the authorized distributor of CKD in Taiwan.
Request a quote
Send your board type, pin count and target cycle time for an SPI evaluation. Request a quote →