Method for the speedup of test vector generation for digital circuits

• US 5,418,792 A
• Filed: 12/02/1992
• Issued: 05/23/1995
• Est. Priority Date: 12/02/1992
• Status: Expired due to Term

• Current Assignee: AT&T IPM Corp. (AT&T, Inc.)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Maamari, Fadi
• Primary Examiner(s): Ramirez, Ellis B.
• Assistant Examiner(s): PEESO, THOMAS R
• Application Number:
  US07/984,651
• Time in Patent Office: 902 Days
• Field of Search: 371/27, 371/22.1, 371/23, 371/16.5, 371/19
• US Class Current: 714/738
• CPC Class Codes: G01R 31/318392 for sequential circuits G01R31/318544 takes precedence

Abstract

The generation of a test for detecting faults in a circuit (10) can be speeded up by first selecting a successive one of a first set of faults for targeting and thereafter determining the effort required to detect a predetermined number of the first set of faults. Each of the remaining faults is then successively targeted, with the amount of effort spent to detect each of the remaining faults being adjusted in accordance with the amount of effort spent detecting the previously targeted fault. In each test cycle, faults that are untestable, or too difficult to detect during that cycle, are eliminated from consideration to improve the efficiency and speed of the test generation process.

Claims

1. A method for speeding up the generation of a test for detecting a set of faults in a digital circuit by dynamically varying how much effort is expended to generate the test, comprising:

(a) establishing an allocated test generation effort;

(b) selecting one of the set of faults for detection;

(c) generating a test for the purpose of successfully detecting the fault while monitoring the effort spent at generating the test;

(d) recording the amount of effort actually-expended to generate the test, if the actually-expended effort is not greater than the allocated test generation effort amount, when a test that successfully detects the selected fault is found;

(e) repeating the steps of (b)-(d) until a test has been found to successfully test a separate one of a prescribed number of faults in the set;

(f) dynamically adjusting the allocated test generation effort in accordance with the highest of the recorded efforts by increasing the test generation effort when the highest of the recorded efforts is above a prescribed value and by decreasing the test generation effort when the highest of the recorded efforts is below the prescribed value;

(g) selecting one of the faults not previously successfully detected;

(h) generating a test for the selected fault while expending the now-adjusted allocated test generation effort, and if the test does successfully test the fault, then recording the actual test generation effort expended to generate the test;

(i) dynamically adjusting the allocated amount of effort expended to generate a test in accordance with the recorded amount of actually-expended effort;

(j) repeating the steps of (g)-(h) until all of the remaining faults in the set have been selected; and

(k) increasing the test generation effort by a prescribed amount and repeating the steps of (e)-(j) until at least a predetermined number of faults have been successfully detected.

2. The method according to claim 1 wherein the test generation effort corresponds to a maximum allowable time interval for generating a test.

3. The method according to claim 1 wherein the allocated test generation effort is adjusted during step (f) in accordance with a first prescribed multiple of the highest actual test effort.

4. The method according to claim 3 wherein the first prescribed multiple equals 1.5.

5. The method according to claim 1 wherein the test generation effort is increased during step (h) in accordance with the greater of:

(1) the test generation effort amount, as last adjusted, and (2) the product of a second prescribed multiple and the highest recorded test generation effort.

6. The method according to claim 5 wherein the second prescribed multiple equals 1.2.

7. The method according to claim 1 further including the steps of:

(l) initially establishing a difficulty threshold, representing a maximum level difficulty threshold for a fault above which a test will not be generated therefor, prior to step (b);

(m) initializing a recorded unsuccessful fault difficulty value to zero prior to step (b);

(n) comparing, after selecting the next fault during step (b), the difficulty threshold to a difficulty value associated with how much effort is required to successfully test said selected fault;

(o) re-executing step (b) when the difficulty of the currently selected fault exceeds the initially established difficulty threshold;

otherwise

(p) comparing, for said selected fault when no test is generated therefor, whether the difficulty of the fault exceeds a fourth predetermined multiple of a highest one of the recorded unsuccessful difficulty value and, if not, then re-executing step (b);

otherwise

(q) recording the difficulty value of the selected fault as unsuccessful;

(r) readjusting the difficulty threshold value to a lowest recorded unsuccessful fault difficulty value if a predetermined number of unsuccessful difficulties have been recorded thus far before re-executing step (b).

8. The method according to claim 7 wherein the fourth predetermined multiple equals 1.2.

9. The method according to claim 7 further including the steps of:

(s) de-recording the recorded unsuccessful fault difficulty value after adjusting the amount of test generation effort and before re-executing steps (e)-(j); and

(t) re-setting the difficulty threshold to its initial value after adjusting the amount of test generation effort and before re-executing steps (e)-(j).

10. The method according to claim 1 further including the steps of:

performing a fault simulation for each test found to successfully detect a fault to determine which other faults (if any) will be detected by the test; and

eliminating those faults determined by fault simulation to be detectable from further consideration.