Fault Coverage :

A parameter used in quality assessment of digital circuits is fault coverage. Fault coverage is related to another term known as fault grading. The measurement of fault coverage is done with reference to a set of test vectors. The vectors detect number of faults out of all possible faults in the design. The mathematical equation for fault coverage is,


(Total Detected Faults) / (Total Fault Population)

However different tools report different fault coverage numbers. The absolute total number of faults in the design is not used as the denominator of the measurement process. The errors in fault coverage occurs when different tools identify faults differently. Tools classify stuck-at 1’s and 0’s on gate inputs and outputs as faults, whereas other tools classify stuck-at 1’s and 0’s on gate inputs, gate outputs, and the interconnecting wire nets as faults.

 

The way in which the fault coverage is calculated is as follows,

  1. Each circuit node is taken in sequence and held to 0 and circuit is simulated,
  2. When a discrepancy is detected between the “faulty machine” and the good machine, the faulty is marked as detected and the stimulation is stopped.
  3. This is repeated for setting the node to 1. In turn, every node is stuck at 1 and 0.