As well as the stencil printing process described above, a printer must be able to carry out several other operations to complete the print cycle, which comprises:
The complete operation ‘cycle time’ is important in determining the throughput and capacity of a surface mount line. To avoid inspection adding to the cycle time, it is often carried out as an off-line operation, or using a separate work-station.
The description given below is of a printer integrated into a production line, but the comments made about board mounting, alignment and cleaning requirements apply generally. Note that there are many variations in the way mounting, alignment and cleaning are implemented, and that stand-alone printers will also handle boards manually rather than automatically.
The board is normally transported into and out of the working area by conveyor belts. Once in the working area, the board is stopped in the desired position either mechanically or by using an optical sensor. The board then has to be both clamped rigidly in position to prevent lateral movement, and supported to resist the downward forces during the print stroke, which would otherwise lead to warping and solder bleed under the stencil.
Fixturing can take several forms, including:
The solution chosen will depend on the application. Providing adequate support during second-side printing can be problematic, particularly when the first side assembly is densely packed with components, or the board is thin and/or flexible. A common approach is to use a dedicated tooling plate which is machined to accommodate the components. This gives better support than a bed-of-nails fixture, especially around the board periphery.
Clamping may be mechanical, with edge clamps which are thin enough not to impede squeegee travel (and which are correspondingly sharp!). Alternatively, the board may be held against the tooling plate with vacuum assistance, which has the apparent advantage that it will hold warped boards flat (although this task is adequately carried out by the downward pressure of stencil and squeegee).
Accuracy of alignment is critical. The aim is to position the board accurately and repeatably to align the solder lands on the board pattern with the stencil apertures. This requires three adjustments (X, Y, q): as the range of angular adjustment is small, alignment is often implemented using X1, X2, and Y adjusters as shown in Figure 15.
Manual adjustment has now largely been replaced by vision systems using a CCD camera to image fiducial marks on the board. A number of different shapes of fiducial have been used in the past, although the industry tends now to use a solid filled circle between 1mm and 3mm in diameter.
In many systems, the camera is inserted between the stencil and the PCB and looks both down to the board and up to the stencil, either using pneumatic shutters to determine the direction in which the camera points or viewing both targets simultaneously (Figure 16). In operation, the camera traverses diagonally across the board to take in both fiducials, and returns to a rest position outside the range of vertical motion of the board fixturing.
The sequence of operations is as follows:
Other printer manufacturers do not site the camera between stencil and board, and therefore require a precision camera movement.
Fiducials may be referenced either to light background (the stencil) or dark background (the board). Modern software systems work on the basis of fiducial scoring against target/accept scores. In the DEK system, a stored outline would yield a score of 999 for a perfect match. Practical values are 950 on a board and 900 on a stencil, with a target of 700 and an accept score of 500.
The accuracy of alignment depends on:
Differences in temperature and humidity during production can lead to small differences in dimensions, particularly on large boards, so there is no guarantee that the stencil will exactly match the board even though these have been generated from the same CAD artwork.
or this reason, the best any alignment system can do is to provide perfect alignment between stencil and the board at one point, and to minimise the errors at all others.
A vision alignment system looks at two fiducials, and can correctly align the fiducials to the line joining them and average the error along this (Figure 17).
The one point of perfect match is somewhere between the two fiducials. Often this point is midway between them, but some boards may have non-central areas which are particularly difficult to print. In such cases, it is possible to ‘weight’ the fiducial correction, so that the optimum alignment occurs at the desired board location. In all cases, however, exact alignment in X,Y and q, is only achieved at one place, and the inaccuracy increases with distance from this point.
A colleague complains that the PCBs you have printed are misaligned, because only the pads in the centre of the board have been accurately printed, and suggests that the vision alignment system is faulty. Explain why this is not likely to be the case, and what the probable cause is.
If possible, temperature should be kept constant within close limits: 21°C ±2°C is a typical specification. This is primarily in order to avoid any effect on the process resulting from changes with temperature in the viscosity of the solder paste, but also helps keep alignment consistent by maintaining constant stencil dimensions.
Stencil cleaning is a low-technology part of the printing process but is nevertheless vital. Cleaning is required both at the end of a run when the stencil is removed from the printer, and it is necessary to remove all accumulations of solder paste from the apertures and the stencil surfaces before the deposits harden, and also in-process.
The underside of the stencil gradually acquires solder paste through:
The rate at which this happens will depend on print parameters and stencil condition.
The underside of the stencil should therefore be cleaned periodically. This is particularly important for fine-pitch applications, because a small degree of contamination of the board by solder paste degrades the print through smearing. If not cleaned off, the resulting print smearing causes an increased likelihood of solder shorts and solder balling.
In both cases, cleaning may be by hand, or automatic. Programmable in-process stencil cleaning may be built-in to an automated stencil printer, while separate automatic spray-cleaning tanks may be used for stencils after a print run.
Automatic stencil cleaners are designed to allow unassisted cleaning of the underside of the stencil at user-programmable intervals (typically every few prints). Typical systems for cleaning the stencil underside use a lint-free wipe, running between supply and take-up rolls, so that an unused area of the paper is pressed against the underside of the stencil. Contaminants and paste removed from the stencil are trapped on the roll of material.
On some machines, this operation can be run dry or wet, using a cleaning fluid ‘sprinkled’ onto the absorbent paper by a solvent bar wetting system. Machines may be programmed for different combinations of wet and dry wiping, using wet cleaning to loosen dried solder paste residues.
Depending on the manufacturer, the cleaning process may also be assisted by vacuum, which is claimed to help remove solder paste from stencil openings and improve the clearance of partially-clogged apertures. The vacuum system operates in conjunction with twin blades on the under-stencil wiper.
Why does solder paste gradually accumulate on the underside of the stencil, and what problems are likely to occur if this is not regularly cleaned?
Paste needs to be stored correctly and allowed to reach operating temperature before the container is exposed to (potentially moist) air. Preparation for use should also include a degree of ‘paste conditioning’ to ensure that:
With a conventional stencil printer, the third of these is generally carried out by ‘kneading’ the paste, using the squeegees to work it backwards and forwards over a blank section of the stencil. This operation can be programmed on the printer and linked with automatic paste replenishment.
However, straightforward mixing to achieve homogeneity before application to the stencil is only possible when paste is supplied in pots, and even then it is easy for over-zealous stirring to incorporate bubbles of air into the paste. With cartridges, because direct working of the paste is not viable, any homogenisation necessary must be built into the dispense routine. For an automated system, one approach is to incorporate a circuitous paste flow route within the dispenser.
Visual assessment of paste deposits offers a quick and helpful guide to whether the process is under control. Best results are obtained with relatively low magnification (×4 – ×10) using a magnifier or projection microscope, as these allow the whole area to be scanned relatively quickly.
What are you looking for? A working definition of an acceptable print is one that has good definition and registration without any defects (such as slumping, scavenging, bridging and peaking):
More detailed standards are available, such as those originally developed by Martin Marietta, covering topics such as:
Be aware that, because they have not yet been formalised by IPC or similar international regulatory bodies, workmanship standards may differ slightly between assembly houses.
The measurement of the paste deposit is crucial to quality control. There are two significant aspects:
The methods of evaluating a print vary between fully automated inspection, both for coverage and paste height, and occasional operator visual checks. There is a trend towards building in automatic checks at the end of the print cycle, either within the printer or as a separate feature. As always, there are variations in practice between different manufacturers and users, depending to some extent on the required cycle time and the sophistication and speed of the optical arrangements. This is an area where significant progress has been made in recent years.
Optical inspection for coverage has in the past relied on there being a visual difference between a pasted and bare pad: this is very easy when printing onto a nickel-gold finish, but very much more difficult when printing onto solder surfaces.
The so called ‘Z-check’ for paste height can be carried out easily with a light-section microscope or laser equivalent. As shown in Figure 19, the height of the print, and some information on the topography of the surface, can be gained using oblique illumination through a slit, and viewing from above.
Both visual measurement systems are ‘non-contact’, eliminating any need to touch the solder paste or any surface on the board substrate, and are suitable for measuring wet solder paste as soon as it is printed. However, both approaches ignore the real aim, which is to deposit a known volume of solder paste, and are particularly susceptible to error due to thickness variations from effects such as scooping.
There are two general methods available for measuring solder volume.
While this second method is of limited utility, particularly in detecting stencil blockages and other localised defects, the approach has been used to monitor the amount of paste which is actually used for making the product, giving a guide to the scrap levels. These are surprisingly high in many companies, despite the fact that solder pastes may cost £50–£140 per kg. Figures of under 50% utilisation are not uncommon.
Think about the process of printing, and identify the ways in which solder paste might be consumed but not turned into good product.
[ back to top ]