Dictionary based Data Compression for SOC with Multiple Scan Chains

碩士 === 元智大學 === 資訊工程學系 === 94 === Reducing test application time and test data volume are major challenges in SoC design. In this thesis we propose a new dictionary-based test data encoding technique for multiscan-based designs. To reduce the data volume transferred from ATE to the core under test,...

Full description

Bibliographic Details
Main Authors: Kuo-En Tseng, 曾國恩
Other Authors: 曾王道
Format: Others
Language:zh-TW
Published: 2006
Online Access:http://ndltd.ncl.edu.tw/handle/95880643828728291356
Description
Summary:碩士 === 元智大學 === 資訊工程學系 === 94 === Reducing test application time and test data volume are major challenges in SoC design. In this thesis we propose a new dictionary-based test data encoding technique for multiscan-based designs. To reduce the data volume transferred from ATE to the core under test, we classify the data slices of the scan chain into high using frequency and low using frequency. The high using frequency slices are arranged in the low address portion of the dictionary and the low using frequency slices are arranged in the high address portion. Two different lengths of indices are then assigned for these two classes of slices to point to the entries of the dictionary. When sending a high using frequency of slice, only low portion of address is needed and the data volume can then be reduced. The decompression logic in the core under test is very simply and we can get high compression ratio by using variable encoding technique without increasing the complexity of decompression logic. Furthermore, to reduce the size of the dictionary, we use a scan cell reorder method to increase the compression ratio. Dictionary-based is based on the use of a small number of ATE channels to deliver compressed test patterns from the tester to the chip and to drive a large number of internal scan chains in the circuit under test and it is especially suitable for a reduced pin-count and low-cost DFT test environment. Experimental results show the proposed approach indeed can get a high compression ratio of the test data and can as well reduce the test times.