This section describes how to use the basic functions of the Diagnostic Profiler to assist in testability analysis. The testability analysis metrics produced by the Diagnostic Profiler are based on the definition of tests to be applied to the design. Testability analysis parameters are usually only meaningful in the context of a defined test or diagnostic strategy, which involves an engineering trade-off and analysis process. The development of a test/diagnostic strategy is described in under Specify Tests & Repairs.

   Testability Analysis

As designs become more complex, testing these designs becomes more difficult. It is no longer feasible to consider test of these complex designs after the design has been completed. Test must be developed concurrent with design. This requires tools to allow the designer to concentrate on his key goal - getting the design to perform - while allowing the test engineer to influence the design without undue delay or impact to the designer.

The Diagnostic Profiler allows the test engineer to design the test profile in a concurrent mode. The designer requires constructive advice on the design in terms of its inherent fault detection and isolation capabilities. This advice must be presented in a timely and focused manner.

The Diagnostic Profiler creates and uses a model of the design directly off the CAD/CAE data that the designer is creating. This allows the test engineer to work concurrent with the design engineer. The Profiler provides the test engineer with automatic calculation of testability characteristics. This allows the test engineer to quickly perform engineering analysis of the design, calculate testability characteristics and feed the resulting conclusions to the designer.

For the designer, the testability characteristics can help guide him or her in improving the inherent testability of the design. This may include adding or modifying built-in test or adding accessibility via additional test points or breaking up design loops. For the test engineer, the Profiler can provide insight as to what available measurement points should be used to maximize fault detection and isolation capability, within predefined test constraints. It can also verify that the test program will meet test requirements.

Calculating Testability Parameters

1. Fault Detection percentage
2. Fault Isolation percentages to the part level
3. Fault Isolation percentages to the repair item level
4. Ambiguity Group identification, composition and probability
 

    To Calculate Testability Parameters

2. Highlight the Testability Analysis tool in the tool list. Press Use Tool, or double-click on the Testability Analysis tool in the tool list.

NOTE: make sure that you have the appropriate project and candidate loaded.

The Testability Analyzer screen will appear. If you have not made any test or repair item specification changes since the last time you performed testability analysis, and testability analysis calculations for the design remain valid, or unchanged, then the testability analysis screen will display the testability analysis metrics. If changes have occurred, you will be prompted to re-calculate testability.

3. Press Run Calculations to initiate calculations. Press Cancel to exit the Testability Analyzer. Press Continue Anyway to display the last valid testability analysis calculations.

Testability calculations take several minutes to perform. The time to calculate testability is related to the size of the design, and the number of tests specified.

4. When calculations are complete, testability parameters are displayed on the screen.

Fault Detection refers to the capability of recognizing that a fault has occurred. Any part in a design can fail. Detection of a failure is critical for confident operation of the design. The higher the fault detection percentage, the better the capability. Fault detection calculation is based upon tests that have been identified, and signal flowing the design making its way out to observation points to tests. Initially, fault detection calculation is a good check of the integrity of your design model (see Edge Candidate).

The main Testability Analysis tool screen displays the percent fault detection. If fault detection is less than 100%, you can view additional details related to those fault locations not detectable, by pressing the Non-Detectable option for the ambiguity groups display.

When viewing Non-detectable ambiguity groups, you can retrieve specific identification of either the parts, repair items, or faults that are non-detectable. Normally, viewing this information at the fault level is most meaningful. To view detailed identification of non-detectable faults, click on the box under the faults column in the ambiguity groups table. A drop-down list box will display the faults in the non-detectable ambiguity group. A scroll bar is available for you to scroll through the list when the list contains more than six items.

Fault locations can be detectable, non-detectable, or excluded from the model. Items/faults may be excluded from the model by setting their failure rate to zero. Items excluded from the model are excluded from testability calculations by virtue of setting their weighing factor (failure probability) to zero. Items excluded from the model are not identified on the testability analysis screen, but are identified in the testability analysis reports.

See individual sections for specific algorithms used for fault detection, fault isolation and ambiguity groups.

Fault Isolation

The Testability Analysis screen displays the percentage of fault isolation capability at both the part level and the repair item level. If all parts and repair items are the same, as is the default, then these percentages will be the same. If any parts have been grouped into a single repair item, then these numbers will be different, and will reflect the appropriate percentages. To group parts into a repair item, use the Specify Repair Items tool available from the Specify Tests and Repairs tab.
 
Fault isolation percentages are weighted by failure rate of the parts and the distribution of the part failure rate among the faults. By default, all parts have an equivalent failure rate, and all fault locations have an equivalent distribution of the part's failure rate. To modify this default, use the Specify Failure Rates tool available from the Specify Tests and Repairs tab.
 
The Testability Analysis screen displays the percentage of fault isolation percentages up to the maximum size of the largest ambiguity group (i.e., to one, two, three or 'X' number of parts).
 
Additional details on the composition of the ambiguity groups is available at the bottom portion of the screen. This includes, for each ambiguity group, ambiguity group probability, repair items in the ambiguity group, parts included in the ambiguity group and faults included in the ambiguity group.

Ambiguity Groups

Identification and composition of Ambiguity Groups is important in testability analysis since it identifies specific areas where additional test coverage (i.e., observability) is required.

The Testability Analysis tool displays details on each ambiguity group in the lower half of the screen. These additional details include, for each ambiguity group, ambiguity group probability, repair items in the ambiguity group, parts included in the ambiguity group and faults included in the ambiguity group. You can use the scroll bar at the right side of the table to scroll through the ambiguity groups present in the design.
 
Ambiguity group probability is the probability that that ambiguity group will occur in a maintenance session. The probability is based upon the failure rate attributable to the faults within the ambiguity group relative to the overall design's failure rate (i.e., percentage of the total failure rate). The ambiguity group probability is displayed in the ambiguity group table. The ambiguity group table is ordered by probability.

NOTE: Since the ambiguity group table is ordered by probability, ambiguity groups of more than one repair item or more than one part may be low in the list. Normally, you will be interested in which ambiguity groups contain more than one repair item or part. Therefore, remember to scroll through the ambiguity group list to spot ambiguity groups containing more than one repair item or part.

Repair Items in the Ambiguity Group displays the number of repair items in each ambiguity group. To identify which repair items are included in that ambiguity group, click in the box associated with that ambiguity group under the "Repairs" column. A pull-down arrow will appear. Press on that arrow, and a list box will identify the repair items in the ambiguity group.

Parts in the Ambiguity Group displays the number of parts in each ambiguity group. To identify which parts are included in that ambiguity group, click in the box associated with that ambiguity group under the "Parts" column. A pull-down arrow will appear. Press on that arrow, and a list box will identify the parts in the ambiguity group.

Faults in the Ambiguity Group displays the number of faults in each ambiguity group. To identify which faults are included in that ambiguity group, click in the box associated with that ambiguity group under the "Faults" column. A pull-down arrow will appear. Press on that arrow, and a list box will identify the faults in the ambiguity group. Note that this list will include both fault location names and failure modes, if applicable.

All of the detailed information on ambiguity groups is also available in the testability analysis reports.

Testability Analysis Algorithms

- Percentage of detectable fault locations
- Percentage of detectable parts.

Observable test locations are those locations associated with a measurement. Using all the observable measurement locations, each fault location is checked for whether or not that fault location is detectable by that measurement. The failure rate value for all detectable fault locations is summed together. This summed value is divided by the total of all failure rate values. This results in the percentage of fault detection. This is the number that is displayed on the main screen as the fault detection percentage.
 
In addition to this number, the testability reports provide detail on the percentage of detectable parts. If a part is only partially detectable (i.e. only some of the fault locations that make up that part are detectable), it is assumed to be a non-detectable part for the purpose of this calculation. In this case, the sum of all the failure rate values for the detectable parts are divided by the sum of all failure rate values.
 
Fault Isolation Algorithm

For each detectable fault location, if that fault is not distinguishable from another fault, then they are grouped together. In other words, using only accessible test locations, if two fault locations are equivalent from a fault propagation perspective, then they are grouped into a single group. The total number of these groups that contains more than one part is the total number of ambiguity groups. The maximum group size is equal to the number of parts in the largest group.
 

Ambiguity Groups Algorithm

The detailed list of ambiguity groups gives the total number of parts, the total number of fault locations, and a list of fault locations grouped by part for each ambiguity group.
 
On the main screen, the fault isolation to one, two and three parts is displayed. On the detailed screen the fault isolation to x parts is given where x is the number of parts to 100% isolation. This number represents the percentage of faults that can be isolated to x parts or less and is weighted using failure rate numbers. The percentage of parts that can be isolated to x parts is derived by summing the failure rate numbers for each group whose group size is equal to or less than x parts. This number is divided by the sum of all failure rate numbers for all groups.

Sample Testability Calculations

Test Descriptions:
Test Name Description
Edge Default Edge Test
 
Test Information:
 
Test Name     Category     Cost     Order     Datafile
Edge             Edge                 1         1         Edge.out

Disabled Tests: None
 
Fault Coverage:
Test: Edge
Measurement: CALCTA.IN_1
CALCTA.IN_1

Measurement: CALCTA.IN_2
CALCTA.IN_2

Measurement: CALCTA.OUT_1
CALCTA.IN_1, CALCTA.IN_2, CALCTA.OUT_1, U1.1
U1.2, U1.3, U2.1, U2.2, U2.3, U4.1, U4.2, U5.1 ,U5.2

Measurement: CALCTA.OUT_2
CALCTA.IN_1, CALCTA.IN_2, CALCTA.OUT_2, U1.1
U1.2, U1.3, U3.1, U3.2, U3.3, U4.1, U4.3
 

Testability Analysis reports generated by the Diagnostic Profiler will reflect the testability characteristics for a specific candidate. The testability calculations must be run prior to report generation.

Generating Reports

    A tabbed window will appear, displaying report categories available.
 
2. Select the Testability Analysis tab by clicking on it. A list of testability reports is shown in the report list. Select a report by clicking on the title.

    A small box will appear to the left of the reports selected.
 
3. Press the "View" pushbutton at the left of the window.

After a few moments, a display window will appear containing the report. From this screen, you can view the report, search the report for a text string, save the report, or print the report. When saving the report, the report is saved in the standard Rich Text Format (RTF), a format which can be imported into most word processors. Standard window interface functions are used.
 
Detailed information on report functions.
 
4. Press Done when you have completed viewing the report.
 
Testability Analysis Reports

The testability analysis reports contain more detailed information than is available on the Testability Analysis tool screen. The following is a list of the testability reports:

- Testability Analysis Summary
- Testability Analysis Details
 
Testability Analysis Summary

Date: 3/15/97
Work Area: Samples
Project: A11
Candidate: WORKING CANDIDATE

Fault Detection Summary:

Total Number of Items in Design Parts=42 Faults=289
Items Excluded from Model Parts=1 Faults=57
Items Included in Model Parts=41 Faults=232
Detectable Failures Parts=41 Faults=232
Non-Detectable Failures Parts=0 Faults=0

Fault Detection Probability (Percentage): 100.000%
 
Fault Isolation Summary:

Parts
85.582% to 1 Parts or Fewer (85.582%)
99.302% to 2 Parts or Fewer (13.720%)
99.302% to 3 Parts or Fewer (0.000%)
99.302% to 4 Parts or Fewer (0.000%)
99.302% to 5 Parts or Fewer (0.000%)
100.000% to 6 Parts or Fewer (0.698%)
 
Detectable Ambiguity Groups Summary:
Total Number of Fault Groups 63
Groups Containing More than One Part 15
Maximum Number of Parts in a Group 6
Weighted Average Number of Parts in Ambiguity Groups 1.172111

Detectable Ambiguity Groups:
Group #     Probability     # Repairs     # Parts     # Faults
1                     52.906%           1                         1           40
2                         6.613%         1                         1             5
3                         5.299%         2                         2             6
4                         2.650%         2                         2             3
5                         2.650%         2                         2             3
6                         2.646%         2                         2             3
7                         2.645%         1                         1             2
8                         2.645%         1                         1             2
9                         2.645%         1                         1             2
10                       2.645%         1                         1             2
11                       2.645%         1                          1              2
12                       2.645%         1                          1              2
13                        2.645%         1                          1              2
14                        2.645%         1                          1              2
15                        2.645%         1                          1              2
16                        0.819%         1                          1              5
17                        0.661%         1                          1              1
18                        0.661%         1                          1              1
19                        0.116%         1                          6              12
20                        0.116%         1                          6              12
21                        0.116%         1                          6              12
22                        0.116%         1                          6              12
23                        0.116%         1                          6              12
24                        0.116%         1                          6              12
25                        0.116%         1                          2              4
26                        0.116%         1                          2              4
27                        0.116%         1                          2              4
28                        0.116%         1                          2              4
29                        0.010%         2                          2              4
30                        0.008%         1                          1              2
31                        0.008%         1                          1              2
32                        0.008%         1                          1              2
33                        0.008%         1                          1              2
34                        0.005%         1                          1              1
35                        0.005%         1                          1              1
36                        0.005%         1                          1              1
37                        0.005%         1                          1              1
38                        0.005%         1                          1              1
39                        0.005%         1                          1              2
40                        0.005%         1                          1              1
41                        0.005%         1                          1              1
42                        0.005%         1                          1              1
43                        0.005%         1                          1              1
44                        0.005%         1                          1              2
45                        0.005%         1                          1              1
46                        0.004%         1                          1              1
47                        0.004%         1                          1              2
48                        0.004%         1                          1              1
49                        0.004%         1                          1              2
50                        0.002%         1                          1              2
51                        0.000%         1                          1              1
52                        0.000%         1                          1              2
53                        0.000%         1                          1              2
54                        0.000%         1                          1              2
55                        0.000%         1                          1              2
56                        0.000%         1                          1              2
57                        0.000%         1                          1              2
58                        0.000%         1                          1              2
59                        0.000%         1                          1              2
60                        0.000%         1                          1              2
61                        0.000%         1                          1              2
62                        0.000%         1                          1              2
63                        0.000%         1                          1              2

Non-Detectable Ambiguity Group: None
 
Testability Analysis Details Report