Fault detection in non-BS clusters (groups of non-BS components that are present in boards with BS) might be done exclusively through the surrounding BS components, but smaller test times and higher fault coverage may be achieved if external test resources are available as well. The BS infrastructure and external test channels may be combined mainly as represented in the three cases shown in figure 1.
The symbol is used in figure 1 to represent an external test channel with direct physical access to the PCB node. The three situations shown correspond to different trade off solutions in terms of test speed (and eventually fault coverage as well) versus cost (in resources of test equipment and physical access provisions), which may be summarised as follows:
Test vector generation for the non-BS cluster is done by an automatic test pattern generation tool according to a specific fault model and to the test resources that are assumed to be available (functional test of the cluster or in-circuit access to internal cluster nodes). For the case shown in figure 1.c, post-processing of the functional test vector set is then employed to serialise and insert the test stimuli in the proper locations in the BS chain(s).
As complex packaging and mounting technologies become more and more common, the situation shown in figure 1.c becomes the preferred choice. It should be pointed out that the need to scan in / out each test vector / responses may lead to unacceptably low test speed, for those cases when the BS chain(s) include hundreds / thousands of cells and the number of test vectors is very large. For certain types of circuits (for example, most combinational logic clusters), alternative test techniques may however be applied, such as pseudo-random test generation and response compaction by signature analysis, which enable test vector application at the rate of one test vector per clock cycle.