Keeping on level with 3D
07 March 2007
BGA and other high pin-count packages are particularly susceptible to coplanarity errors, but many manufacturers are failing to use adequate inspection techniques to detect these. Adaptsys’ Dan Eames outlines the options.
A two-dimensional optical inspection may be sufficient to assess simple dual-in-line packages for manufacturing defects, but complex, high pin-count packages such as Ball Grid Arrays (BGAs) need more careful examination. Unfortunately, despite the relatively high value of systems being built in Western Europe, most manufacturers are still relying on two-dimensional optical inspection to assess what is, effectively, a three-dimensional (3D) problem.
Coplanarity is the measurement of the deviation of distances of points from a plane. In the case of BGA and other high pin-count packages, this is the measurement of the difference in the height of the balls, or length of the lead terminations, compared to the seating plane or printed circuit board.
Even with modern manufacturing techniques, the dimensions of BGA terminations can vary to such an extent that one or more terminations do not make contact when mounted on a printed circuit board The ball terminations of BGAs can be squashed, sliced, sheared or simply be too small. For leaded components, the problems are with bent, or badly-formed terminations, or where a shard of metal, or burr, is present on the lead. Repeated handling of components during programming, test, and assembly can cause additional damage to terminations which can increase coplanarity errors and adversely affect manufacturing yield and failure rates.
Conventional two-dimensional (2D) optical inspection can check for defects, cracks, chips, marking and alignment as well as for low-end, implied or inferred coplanarity. This method relies on two images from the component, for example from the side and underneath, to give an approximate position in 3D space. Technically, however, 2D inspection is checking for colinearity, whilst accurate verification of true coplanarity literally requires the additional dimension provided by 3D inspection. Research by Adaptsys using their 3D optical inspection camera system has shown that, in a typical manufacturing environment, 1% of inspected BGAs have coplanarity errors which could make the component susceptible to failure.
There are two methods for defining the reference plane to which component terminations should be aligned: the seating plane method and the Least Mean Square (LMS) method. In the seating plane method, the three balls or lead terminations with the greatest distance from the seating plane are used, in combination with the centre of gravity, to define the reference plane.
In the LMS method, the least-mean-squares distances of all the balls or lead terminations are used to define the best-fit plane, with the maximum distance between the plane and a termination providing the figure for the maximum coplanarity error.
3D automated optical inspection equipment uses multiple cameras to create images of the target components. Whilst most 3D systems use the seating plane method to calculate the reference plane, Adaptsys uses an LMS algorithm to precisely locate the surfaces of leads and solder balls in 3D space to accurately measure coplanarity as well as other parameters such as lead pitch and width, foot angle, span, mark quality and part orientation.
A typical coplanarity inspection will be set to provide a rejection overkill (i.e. the percentage of good parts being rejected) of between 0.1% to 1%. This provides a margin of safety on the inspection and ensures a 0% pass-rate for parts which exceed the figure for coplanarity error.
The rejection ratio is not set directly, but should be the result of Statistical Process Control (SPC). An analysis of the rejected parts can identify where problems exist and allow those problems to be corrected to reduce the number of faulty parts. SPC can also be used to tweak the specifications of the inspection to ensure that optimal setup is maintained and to reflect the actual manufacturing flow for a particular PCB.
Some automotive standards already demand 100% coplanarity inspection prior to pick-and-place but there are benefits for other applications, particularly where component failure can have significant cost implications. Coplanarity inspection can reduce cost in two ways; it helps to minimise field failures and can identify faulty components so that they can be removed from the manufacturing process before additional cost is incurred by continuing to add value to an inherently faulty board. After inspection, components can be categorised as passed, rejected or repairable, if the overall cost of the system justifies the cost of repair.
Coplanarity inspection can be carried out throughout the manufacturing process. Adaptsys tape-and-reel systems and programmers can be fitted with a 3D camera system so that coplanarity inspection can be integrated with the taping and reeling process or as part of the programming cycle. Components can also be inspected as they are packaged into trays or tubes, or prior to pick-and-place.
Wherever it is applied, the key factors in effective inspection are accuracy, speed and repeatability. For all three of these factors, automated 3D inspection offers considerable benefits compared with manual optical inspection. The latest Adaptsys 3D optical cameras can, for example, support inspection rates of up to 45,000 Units Per Hour (UPH), with lead-width measurement to a resolution of 3 microns and coplanarity accuracy down to 10 microns.
Given the high value of systems being manufactured in Western Europe, automated 3D optical has a growing role to play in helping to increase manufacturing yields and adds a vital new dimension to the accurate placement of BGAs and other high pin-count packages.
Contact Details and Archive...
Related Articles...
Most Viewed Articles...