Every assembly line already has an inspection step. It is called final test, or worse, the customer. The real question is not whether your defects get found, but where, and how much value you have added to the board by the time they are.
Defects get more expensive the longer they live
A solder paste deposit that is 40% short on volume costs almost nothing to fix at the printer: wipe the board, reprint it, and move on. The same deposit after placement means removing parts. After reflow it is a rework station, a skilled operator and a thermal cycle the board was never meant to see. At functional test it is diagnosis time as well. In the field it is a return, a failure analysis and a difficult phone call.
The industry rule of thumb is that the cost of a defect rises about tenfold at each of those stages. The exact multiplier varies by product, but the direction never does.
Most solder defects start at the printer
It is widely reported in the industry that the majority of surface-mount solder defects trace back to the paste printing step, with figures of 60 to 70 percent often quoted. The precise share depends on the study and the product, but the reasoning is simple: printing has the most variables. Stencil condition, aperture design, paste rheology, squeegee pressure and speed, board support, separation speed, temperature and humidity all interact, and they drift during a shift.
That makes the printer the highest-leverage place to measure. If the paste is right, most of what follows goes right.
Human inspection does not scale
A modern board can carry thousands of solder joints. An operator with a microscope can sample a few and judge them by eye. They cannot measure volume, they tire, and two operators will disagree about the same deposit. Sampling also means that most of what you ship was never looked at.
Automated inspection measures the same way every time, records a number instead of an opinion, and does it at line speed.
Inspection is process control, not just sorting
The lowest-value use of an inspection system is as a gate that separates good boards from bad. The highest-value use is as an instrument. When paste height on one corner of the board begins to trend down over twenty prints, the system is telling you about board support or stencil wear before a single board fails. Fixing that cause removes a whole family of future defects.
- Sorting catches the defect you made.
- Process control stops you making the next thousand.
- Data gives you evidence for customers, auditors and your own engineering decisions.
Where to inspect
| Stage | Method | What it catches |
|---|---|---|
| After printing | Solder paste inspection (SPI) | Insufficient, excess, offset and bridged paste |
| After placement | Automated optical inspection (AOI) | Missing, wrong, shifted or reversed parts |
| After reflow | AOI | Solder joint defects, tombstones, bridges, lifted leads |
| End of line | Electrical and functional test | Whether the board works |
You do not need all of them on day one. If you can only add one step, add it where your defects originate. For most lines that is the printer.
What inspection will not do
Inspection does not fix a bad stencil design or a paste that has been out of the refrigerator for two days. It tells you those things are hurting you, quickly and with numbers. The improvement comes from acting on what it shows. Lines that buy a system and never look at the trend charts get the sorting benefit and leave most of the value on the table.
The short version
Inspect as early as you can, measure rather than judge, and use the data to fix causes. The savings come from the boards you never have to rework.