The relationship and difference between FG, CHAR, PAT, SBA and ED in AEC Q100 automotive chip certification

After completing the two contents of AEC Q100 F Group, several contents about test methods and statistical guidance in AEC documents have been completed. These contents may be difficult to understand for friends who are new to AEC certification and automotive regulation systems, so a summary and comparison are made here.

In fact, looking at the entire AEC-related annexes and the content of AEC Q100, in addition to strict requirements for certified sample testing, there are also strict requirements and definitions for production test data statistics and monitoring. However, the dimension lacks requirements for the stability of test equipment. Although this is mentioned in the 5th ED power distribution of Group E, there is no specific test equipment acceptance guidance plan. However, I believe that all automotive component manufacturers have a 16949 quality system, so naturally they know how to release test equipment. I have also written an article about MSA for an in-depth introduction and interpretation, which I will not elaborate on here. Let's go back to the AEC Q100 document itself.

First, let’s recall the location of these verification items in AEC Q100 Figure 2.

As shown in the figure, these five items are distributed in Group E and Group F, belonging to E5-ED, E6-FG, E7-CHAR of the electrical performance verification test of Group E, and F1-PAT and F2-SBA of the defect screening test of Group F. In fact, PAT, SBA, and CHAR correspond to the three annex files of AEC Q001, 002, and 003, which are the general guidance documents for AEC certification. FG corresponds to AEC Q100-007, and ED corresponds to AEC Q100-009, two annex files of AEC Q100. The amount of information contained is very large.

During the reading process, you will find that the contents of the five files mention and quote each other, which makes it difficult for friends to understand the content of any one of them without reading all five files.

Let's first interpret the corresponding content item by item and introduce them in a more understandable order:

E6 - FG Fault Classification

The title of FG and the translators are not easy to understand. In fact, the FG project is the evaluation of the chip planning test method and the corresponding test coverage, that is, the chip manufacturing process test program. Whether it can simulate the various states of the chip in actual operation, whether it can completely test the chip input and output. FG is only for digital signals and digital parts of mixed digital and analog signals. (In principle, the analog output part requires 100% detection, so it does not need to pass this classification)

FG mainly studies the possibility that input and output are stuck and the signal cannot respond normally but cannot be detected.

E7 - CHAR characteristic properties

The purpose of CHAR is to tell you how to determine the specification parameters and upper and lower limits of a product. Under the premise of ensuring the stability of the product process (Cpk>1.67), an appropriate specification parameter and range should be given based on the actual test results of multiple batches of products. (Some friends may have doubts that if the upper and lower limits are enlarged, Cpk will naturally be good. This statement is indeed correct, but if the upper and lower limits are set too large, customers will also doubt the accuracy of the product, which will affect the application of this product, so it is necessary to choose the most appropriate parameter according to the actual situation.)

When a new product is launched or a product is changed, the CHAR process must be executed to redefine the specification.

E5 - ED Electrical Performance Distribution

The ED test content also counts the Cpk values ​​of the test results of multiple batches of products, and stipulates that Cpk should be above 1.67. However, the ED project must use the specification parameters as a benchmark, and the test equipment deviation must be corrected to obtain the results. Therefore, the ED project must be implemented after CHAR, and the ED project cannot be implemented without the specification parameters.

CHAR sets the specification parameters according to the actual situation of product production; after ED verifies the proposed specification parameters, it checks whether the stability of the remaining batch production process can meet the specification parameters.

F1 - PAT Process Average Test

The content of PAT is to constantly modify the upper and lower limits defined in the specification according to the actual production process, so as to ensure the reliable and consistent quality of the products produced. Here we need to understand the differences between several upper and lower limits. LSL/USL is the upper limit defined in the specification, LTL/UTL is the upper limit defined in the test procedure, and PAT is the upper limit defined according to the Robust mean and Robust standard deviation. Theoretically, the range of LSL/USL>LTL/UTL>PAT Limit, and some test contents defined in the AEC file must use the PAT limits as LTL and UTL, so PAT makes the test requirements more stringent.

PAT practice actually avoids the problem mentioned earlier in the CHAR process where the upper and lower limits are simply enlarged to ensure Cpk while ignoring the improvement of production process stability.

After reading this article, you will understand how strict automotive chip certification is. In the actual production process, I am not sure how many companies can do this, but this is indeed a process requirement that must be implemented.

F2 - SBA Statistical Yield/Binning Analysis

Compared with the above-mentioned intertwined concepts, SBA is much easier to understand. SBA actually includes two contents: SYA Statistical Yield Analysis and SBA Statistical Bin Analysis. The contents are to be based on the yield and bin statistics of at least 6 batches in the production process. Any wafer or batch below SYL1 or above SBL1 should be marked for engineering review. In addition, any batch below SYL2 or above SBL2 is required to be isolated and improvement measures are given.

It should be noted that the SBA content requires the use of the upper and lower limits of PAT for statistics, which also increases the requirements for the stability of the production process.

The above is a summary of the five items. Let’s briefly summarize them:

First, the test coverage must be defined according to the FG requirements;

Then the specification parameters should be determined according to CHAR requirements;

After determining the specification parameters, ED should be done to determine whether the products can meet the standard of Cpk>1.67;

Even if the sample passes the ED Cpk, a more stringent PAT test limit correction is required for long-term production;

Based on the test limits of PAT, batches are statistically monitored for SBA/SYA.

This is the relationship and difference between the five items FG, CHAR, PAT, SBA, and ED in AEC Q100. Do you understand?