Browse Prior Art Database

Smart retest methodology for semiconductor chips to increase tester productivity

IP.com Disclosure Number: IPCOM000236898D
Publication Date: 2014-May-21
Document File: 3 page(s) / 179K

Publishing Venue

The IP.com Prior Art Database

Abstract

When a device fails a production test, the device often is subjected to retesting. Due to running high parallelism at testing, the existing production retest method is to re-run all tests no matter how many tests have been completed and passed by the rejected units. This causes the production retest test time to be the same as normal production test time (same number of tests performed), which means that the retested units have a high cycle time. This paper presents a method to reduce the production retest test time for high parallelism testing device by skipping the tests already passed by the rejected unit and performing only the failed block tests. This method will cover all the tests, as it only separates the tests into either production testing (passed blk) or the production retest (failed blk) to achieve the retest test time reduction.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 51% of the total text.

Smart retest methodology for semiconductor chips to increase tester productivity

Abstract

When a device fails a production test, the device often is subjected to retesting.  Due to running high parallelism at testing, the existing production retest method is to re-run all tests no matter how many tests have been completed and passed by the rejected units.  This causes the production retest test time to be the same as normal production test time (same number of tests performed), which means that the retested units have a high cycle time. This paper  presents a method to reduce the production retest test time for high parallelism testing device by skipping the tests already passed by the rejected unit and performing only the failed block tests. This method will cover all the tests, as it only separates the tests into either production testing (passed blk) or the production retest (failed blk) to achieve the retest test time reduction.

Introduction

Tester capacity constraint is the main problem for all the testing industry. The issue can be resolved by either purchasing more testers or reducing the test cycle time. Reducing the test cycle time is a better option as it will save the millions of dollar needed to purchase a new tester. Reducing cycle time can be achieved by either reducing the production test time or reducing the production retest test time. The Smart Retest methodology described in this paper reduces the production retest test time for high parallelism testing by skipping the passing block tests that have done in the production testing and performing only failing block tests during production retest. The Smart Retest methodology begins by analyzing the device's top failures and groups the failures into five (5) blocks. The 1st block, called 'Start up block' includes setup integrity tests like continuity test. The 2nd block, called 'High fallout block' includes the Top5 failures for the device upon which retest will be performed. The 3rd block, called 'Mid fallout block' includes the Top 6-10 failures for the device undergoing retest. The 4th block, called 'Low fallout block' includes the remaining tests for the device. The last bock, called 'Close up block' includes the ending test for the device, like JVT test. The methodology arranges the test flow start with the Start up block -> Low fallout block -> Mid fallout block -> High fallout block -> Close up block. The purpose of arranging the test flow as per above is to cover as many tests as possible during the production testing. Due to testing at high parallelism, rejects units will be sitting on the test site waiting for the remaining units to complete test before being binned out as rejects so th...