Automated Process Control Example
(click on elements for more information)

Here we are depicting a more complex equipment cell that can perform Advanced Process Control (APC) functions.  The core of the system is efficient data collection.  An Equipment Interface and multiple Sensor Interfaces gather process data from the equipment and sensor instrumentation.  These components can use the Datahub and directly write SQL statements to a Hub process, saving the data into tables.  Because the tables are partially keyed by the observation time, just writing the data into these tables creates a properly ordered, combined set of observations.  In other words, the potentially difficult issue of time synchronization becomes a snap. Also, the data can be keyed by criteria that will be used to subset the data for analysis such as process step.  Fourth generation SQL statements can then be used to efficiently manipulate the data as subsets. 

Not all process runs are alike, and the analyses need to be varied depending on the process program, material identification, or other context criteria such as process chamber.  The power of the Datahub's dynamically interpreted SQL selection statement can be harnessed as a context selection mechanism.  For example, the Fab user can configure certain analyses to be performed when the process program matches specified values.  During execution, the context data for a process run is represented as a data row in an SQL table.  The context criteria associated with an analysis is used in an SQL selection statement on the table.  If the process run is selected by the criteria, the corresponding analysis is performed.

Hume Integration can also supply a Statistical Process Control (SPC) server component designed for automated use which is capable of all of the classic Shewhart control chart analyses, as well as cusum and EWMA.  The SPC component can be a key building block for an APC system since the SPC analyses are used for both traditional univariate methods, and for analyzing the output statistics of multivariate methods.

The analysis processing and creation of output data is potentially time consuming.  Therefore a separate Datahub process is being used, leaving the Distributed Message Manager responsive to other clients.  Similarly, notification of SPC alarms or other conditions via e-mail is potentially time consuming, so a separate process is depicted.

  About Us                   Contact Us                 News and Views             Integrators                Documentation                Support 

Serving the semiconductor manufacturing industry since 1992                          Copyright Hume Integration Software 2021