A fast, built-in test system for oscilloscope manufacturing - HP 54600 oscilloscope - includes related articles on a verification strategy an diagnosing stimulus/response defects in production - Technical

Hewlett-Packard Journal, Feb, 1992 by Stuart O. Hall, Jay A. Alexander

[NOTE: SOME FORMULA'S HAVE BEEN OMITTED]

Following a verification strategy instead of a screening or characterization strategy, a special module was designed to replace the computer Input/output option module of the HP 54600 Series oscilloscopes. The resulting test system has reduced both equipment costs and test times to one tenth those of previous test systems.

The HP 54600A and 54601A general-purpose oscilloscopes have the lowest manufacturing cost of any oscilloscopes in HP's history. A large contributor to this is a test strategy that minimizes the cost of calibration and verification.

Traditionally, digital storage oscilloscopes have been tested with a system that evolved from a prototype characterization system or an environmental qualification system. The R&D team described a set of parameters that they wanted to see tested in changing environmental conditions, and the test engineer developed a system that tested those parameters as accurately as possible. This process resulted in a long list of parameters and a set of test algorithms and equipment optimized for accuracy. This test implementation is required when attempting to understand every aspect of a new product design. The test strategy for the HP 54600 family of oscilloscopes was developed to minimize the test process cycle time, using the least expensive set of equipment possible (see "Verification Strategy," page 22). Developing this strategy required a complete understanding of the hardware of the product, a commitment on the part of the design team to design margins into the product, and a willingness to recognize the difference between testing to characterize a design and testing to verify a product. Here we will describe this test strategy, from prototype characterization to production testing, and discuss the benefits of the resulting system.

Prototype Characterization

Like most new products, the HP 54600 Series oscilloscopes required extensive performance characterization at various points in the design cycle. The characterization test system developed for this was used both for characterizing prototype units and for qualifying the product to HP environmental specifications. Data from this system was used to influence the design team to increase margins on some specifications. In other cases the system verified that a circuit had more margin than the designer expected.

A goal was to characterize prototype instruments early in the design process. This had several positive influences on the success of the project. Because the characterization system was developed as an HP-IB-controlled automated test system, it was necessary to install the HP-IB parser software at an earlier stage than usual. This caused the parser code to be developed as an integral part of the instrument, rather than as an additional feature. The characterization process uncovered hardware and software problems that might otherwise have been missed until a point in the project when it would have been much more expensive to correct them. The system was also used to test new prototype test algorithms and to corroborate the techniques used in the production test process described later in this article. It was also recognized that the characterization test system would be required as an auditing system in production for ongoing verification of key assumptions regarding parameter relationships.

The standard tests performed with this system were developed using the traditional process described above. The system uses the following equipment:

HP 8131A programmable pulse generator

HP 8656B synthesized signal generator

HP 3458A programmable system multimeter

HP 437B programmable power meter

HP 4262A LCR meter

HP 3235A programmable multiplexer

HP 9000 Model 360 computer.

Thirty-one different types of tests are performed, generating more than 200 test parameters for the oscilloscope. (The term "parameter", as used in this article, means the result of a given test under one specific set of test conditions. Some tests have many different sets of conditions. For example, a trigger test that measures performance for two different input signal frequencies, on eight different vertical ranges, on two channels, will generate 2 x 2 x 8 or 32 separate result parameters.) By the completion of the production prototype phase, the characterization system had been used to verify that each of the more than 200 test parameters had specification margins of at least four standard deviations 4a).

Production Test

The production test system was developed according to three fundamental concepts. First, the system would be controlled by the device under test, or DUT (the oscilloscope currently being tested) to alleviate the cost and throughput disadvantages of an HP-IB controller system. Second, the system would use only two signal sources: a high-frequency source and a precision low-frequency source. Third, the test process would be designed to verify that the product is statistically normal. It is important to recognize the distinction between this last concept and the process of ensuring that all specifications are met. Of course, for this method to be sufficient to verify production instruments, all normal instruments must meet specification limits. This is the fundamental reason to include margins in the design of a product.