Post-Analysis Referencing
Overview
There are a few nSpec analyses that must take a previously run analysis as input. This includes Custom Reporter, Custom Exporter, Cropped Images, Report Summary Image Export and Defect Classifier. One or multiple previous analyses can be input to the post-analysis.
Parameters
All of these analyses include the Base Analysis IDs Mode and Base Analysis IDs parameters, which are used to reference the input analysis.
Parameter Name | Parameter Value | Description |
Base Analysis IDs Mode | Absolute, Relative | Absolute points to the numeric integer reference of a previously run analysis. Relative points to the most recently run analysis. |
Base Analysis ID |
| The user must input a comma separated set of IDs. 1 or more IDs can be input. If the mode is Absolute, it is an integer or set of integers. For example,“1” or “1, 2”. If the mode is Relative, it is an input analysis name. See the Analysis Reference List below for the Analysis Reference ID for each analysis. For example, “DeviceInspection, Gen IV Analysis” means that the first input analysis will be the most recently run analysis of type Device Inspection, while the second input analysis will be the most recently run Gen IV Analysis. |
Using Relative Mode
When using Relative mode, you also have the option to reference any previously run analysis, not just the most recently run analysis. by adding a numeric indicator at the end of the analysis name. Let’s say an image database ran 3 analyses:
Gen IV AI Analysis (parameter set 1)
Gen IV AI Analysis (parameter set 2)
Device Inspection
To point at #2 and #3, you would input the following for the Base Analysis IDs field:
"DeviceInspection, Gen IV Analysis"
To point at #1 and #3, you would input the following for the Base Analysis IDs field:
"DeviceInspection, Gen IV Analysis.2"
The .2 at the end of Gen IV Analysis informs nSpec to reference the second most recently run Gen VI Analysis. In this way, you have complete flexibility to point at any previously run analysis using both absolute or relative modes.
If you wish to reference the most recent analysis, there is no need to add the “.1” after the analysis internal name reference. This means that “DeviceInspection” is the same as “DeviceInspection.1”
Internal Analysis Reference ID and User-Facing Analysis Names
The user-facing name of an analysis is not necessarily identical to the ID that should be used to reference that analysis. The following Analysis Reference ID values should be input to the Base Analysis ID parameter when using Relative mode.
Analysis Name (User facing) | Analysis Reference ID (IntName) |
|---|---|
AI Analysis | AIAnalysis |
Basic Selection (Contrast, Morphological Range) | nrad_rangeselect |
Basic Selection (Intensity, Exclusive) | nrad_intensityselectexclusive |
Basic Selection (Intensity, Inclusive) | nrad_intensityselectinclusive |
Crack Detection V2 | CrackDetectionV2 |
Custom Exporter | Custom Exporter |
Custom Reporter | Custom Reporter |
Device Inspection | DeviceInspection |
Die Yield | nrad_dieyield |
Dislocation Defect Analyzer | Dislocation |
Gen IV AI Analysis | Gen IV Analysis |
Gen V AI Analysis | GenVAnalysis |
High Resolution Mosaic Generation | High Resolution Mosaic Generation Analysis |
Surface Scattering Analysis | Surface Scattering Analysis |
Utility: Launch Executable | LaunchScript |
Using Absolute vs. Relative Mode
When running in production, we recommend always using relative mode.
When testing a new script, it might be useful to use absolute mode. However, in the majority of cases, we believe that relative mode is most likely to be the optimal mode, especially when running in production.
Consider a case where you scan a wafer and perform a Group Analysis on this scan. The first time, using absolute references to point at the 1st and 2nd analyses may work fine. But what happens if you run that group analysis a second time? Now, the post-analysis is still performed on the first and second analyses ever performed, rather than the ones most recently performed as part of that new analysis group.