Paste Inspection Frequency: Sampling and Trend Analysis
Solder paste inspection measures the deposit after printing and compares it with the stencil aperture, so that a print problem is found before components are placed on top of it. How often that measurement is made is a decision about risk rather than about technology: too rarely and a drifting process is discovered by a wave of defects, too often and inspection consumes capacity without adding information. Paste inspection frequency should be set from the stability of the process and the cost of the defect.
What Paste Inspection Is For
The measurement reports volume, area and height for each deposit, and compares them with the aperture. A deposit that is short, that is displaced or that has slumped into its neighbour is a defect that will become a joint defect after reflow, and the inspection catches it at the point where the board can still be washed and reprinted.
Inspection also serves as a process control instrument. The mean volume across a board responds to squeegee speed, separation, paste condition and stencil wear, so the same measurements that find defects also describe the health of the printer.
Sampling Frequency and How It Is Chosen
Three factors set the frequency. The first is the stability of the process: a printer that has been running the same product for weeks with a flat trend needs less inspection than one that is changed over several times a day. The second is the criticality of the product, which is really a statement about the cost of a defect that escapes. The third is the capability of the process, since a process running close to its limits needs more frequent checks.
A common starting point is the first board of every lot, every board after a changeover, and then a sample at an interval through the run. Products with apertures near the limit of printability are often inspected more frequently, while a mature product on a stable line may be sampled once every few hours, provided the trend chart has been flat for long enough to justify the reduction.
Deposit Volume Limits and Tolerance Bands
Volume limits are set as a percentage of the aperture volume, often with a target band and a wider reject band. A typical specification accepts deposits between 60 and 150 percent of nominal, warns outside 80 to 120, and rejects outside the wider figure, with the exact numbers chosen for the product.
The band should follow the joint rather than a generic table. A power device with a large pad tolerates a wide volume variation, while a fine-pitch lead needs a narrow one because the same absolute variation is a much larger fraction of the deposit. Setting the band per aperture size rather than per board gives a more useful result.
Trend Analysis and Control Charts
A single measurement says whether the board in the machine is acceptable. A series of measurements says where the process is going, and that is where the value lies. Plotting the mean volume and the standard deviation for each lot on a control chart turns the inspection into paste inspection data that can be acted on before a limit is reached.
Trend analysis also separates the sources of variation. A mean that moves while the spread stays constant points to a setting; a spread that widens while the mean is stable points to a stencil that is wearing or to paste that is ageing. The two require different responses, and the chart distinguishes them without any additional measurement.
When to Move to Full Inspection
Some products justify inspecting every board. A board with fine-pitch devices, a high value per unit or a customer requirement for traceability may be inspected at 100 percent, and the inspection data then becomes part of the record for that board.
The decision should be revisited rather than left in place. A product that was moved to full inspection because of a temporary stencil problem can be returned to sampling once the trend is flat, and a product that has been sampled for years may need tighter control after a change of paste or stencil supplier.
Response to an Out-of-Tolerance Reading
The response is defined by the result rather than by the count. A single isolated deposit outside the reject limit suggests a stencil defect, a clogged aperture or a foreign particle, and the action is to inspect the stencil and clean it. A spread of deposits outside the warning band across a whole board suggests a process setting, and the action is to stop and check the printer.
Where the readings are marginal but acceptable, the useful action is to record them and continue while watching the paste bridging risk at reflow. A process running at one edge of its tolerance will eventually produce a defect, and the trend is the notice that the margin has gone.
Inspection Frequency by Product Type
A board with a single fine-pitch package and a handful of passives can be sampled lightly, because there is little to go wrong and the consequences are local. A board with many thousands of small deposits is the opposite: the probability of at least one marginal deposit per board is high, and the inspection frequency has to reflect that.
Coarse, high-volume products sit in the middle. Their apertures are forgiving, so the inspection is really monitoring paste condition and stencil wear rather than chasing defects, and a lower frequency with a good trend is sufficient.
Correlation With Visual and X-Ray Results
Paste inspection is one of three views of the same process. Visual inspection after reflow sees the consequences, automated optical inspection sees joint shape, and X-ray sees what is hidden under a package. Where paste inspection results are good and downstream defects still appear, the correlation between the three needs to be examined before the inspection frequency is increased.
Comparing the two sets of data is useful in the other direction as well. If a rise in paste volume variation is followed two hours later by a rise in solder balling, the correlation justifies intervening on the print rather than on the oven.
Records and Process Ownership
The inspection record should carry the board identity, the stencil identification, the printer settings, the measured means and the limits applied. With those fields, the frequency can be justified from the data rather than from a policy that nobody can explain.
Ownership matters more than frequency. Inspection performed because a procedure requires it produces numbers; inspection performed by somebody responsible for the print produces adjustments. Where the two differ, the frequency can be halved and the outcome improved, which is worth remembering before another measurement point is added to a line that is already capacity limited.
FAQ
How often should solder paste inspection be run? The first board of every lot, every board after a changeover and then samples at an interval through the run. Products with apertures near the limit of printability justify more frequent checks.
What deposit volume variation is acceptable? Commonly 60 to 150 percent of nominal with warning bands at 80 and 120 percent, set per aperture size rather than per board. The band should follow what the joint can tolerate.
Does inspecting every board improve yield? It catches defects that would otherwise be reworked or scrapped after reflow, but it does not improve the process. Without trend analysis and a response to the data, frequent inspection mostly generates records.
