Browse Prior Art Database

Method for Controlling Test Execution in a Build-to-Order Test Environment

IP.com Disclosure Number: IPCOM000106630D
Original Publication Date: 1993-Dec-01
Included in the Prior Art Database: 2005-Mar-21
Document File: 2 page(s) / 94K

Publishing Venue

IBM

Related People

Faver, DL: AUTHOR [+2]

Abstract

The manufacturing test process for the RISC System 6000* consists of three key test steps. These steps are commonly known as screen, run-in, and final. The test software for each device is made up of a set of test units which test different portions of each device. For example, a hardfile might have one test unit which does a random seek test, another test unit which does a write/read/compare test, a test unit which does a media surface scan, etc. By providing a flexible method for rapidly changing at which test step each test unit is executed, test efficiency is maximized.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 44% of the total text.

Method for Controlling Test Execution in a Build-to-Order Test Environment

      The manufacturing test process for the RISC System 6000*
consists of three key test steps.  These steps are commonly known as
screen, run-in, and final.  The test software for each device is made
up of a set of test units which test different portions of each
device.  For example, a hardfile might have one test unit which does
a random seek test, another test unit which does a write/read/compare
test, a test unit which does a media surface scan, etc.  By providing
a flexible method for rapidly changing at which test step each test
unit is executed, test efficiency is maximized.

      Screen test is the first step in the manufacturing test
process.  It takes place immediately after the assembly of the
system.  Screen test is a manned operation, and devices are tested
one at a time.  A test operator is present during the entire screen
test.  Screen test stops as soon as an error is detected.  The main
purpose of screen test is:

1.  to ensure minimum functionality prior to preceding with the rest
    of test.

2.  to perform any manual test steps.  For example, display adapter
    output is visually checked at screen test

3.  prepare the system for the extended run-in test.

      Determining which test units to execute during screen involves
a trade-off.  It is efficient to capture as many failures as possible
early in the test cycle because this shortens cycle time.  Moving
systems rapidly through the manufacturing line is important because
it ensures the customer receives their machine when expected, and it
reduces inventory costs.  On the other hand, it is expensive to pay
an operator to run a screen station during the execution of long
tests which require no intervention.

      Run-in test takes place after screen test.  Run-in does not
require a test operator to be present during test execution.  During
run-in test, test units are executed over and over to verify system
reliability.  Unlike screen test, run-in will continue to execute and
log errors during the entire run-in period, regardless of earlier
failures.  Run-in test  supports concurrent test of devices, unlike
screen test which is sequential.  Therefore, more devices can be
tested during a given period of time as compared to screen test.
Run-in length varies from 3 to 24 hours depending on machine
complexity and previous run-in history of the same model.  If a
machine fails run-in, it must go back through the entire test
process.  Thus, a good deal of time is lost when a system fails
run-in test.

      Final test can be automated or manual.  The primary purpose of
final is to interrogate the run-in log and determine if the system
passed or failed.  Test units can be executed one last time at final
if necessary.

      An additional concern for RISC 6000 manufacturing test is the
build-to-order nature of the process.  The system is built and...