In the electrical design process, there are instances where special integrated circuits are required for a specific purpose, where the choice is often between Field Programmable Gate Arrays (FPGAs) or Application Specific Integrated Circuits (ASICs). Usually, in digital applications, the electronic engineer may prefer a FPGA during the prototyping phase for the flexibility of programming and immediate availability of components. However, once the functionality has been demonstrated and the component needs to be utilised in production, a FPGA may well be unattractive from a cost point of view. ASICs offer the advantage of highly optimised logic and dedicated metal interconnect, which can match or better the functionality and performance of most FPGAs, with significant cost reduction.

For the unit costs, there is a single point at which the financial attractiveness of FPGAs and ASICs cross over. This is illustrated in the graph, which depicts a real example (Altera EPF10K50R versus GE Aviation 1.2micron standard cell ASIC). In this example, the cross over point, taking the NRE conversion cost into account, occurs at 1000 units.

The crossover point for a particular application will depend on several factors including:

• Gate count requirement

• Speed

• RAM and Macros

• Pad count

• Package type

• Operating environment

These factors mainly determine the geometry, process and die size, which leads to the selection of the most appropriate semiconductor technology for the application.