| Flying probe tester is a system to finish testing Printed Circuit Board in manufacturing environment. It is an improvement based on traditional machine tool. It makes use of probe to replace the machine tool and sets up 4-8 quick step-motor-driven probes in X-Y framework. It finishes test by making contact between probe and PCB. The probe is driven to move in Z direction. The testers convey the CAD files into useful files for testing. These files include the coordinate and value in network of the solder-points, which decide the movement of the probes. Generally, there are thousands of solder-point in a PCB. If the probes don't move in an optimum track or inferior optimum track, it will waste lots of time. To solve this problem, TSP is used in this system.Traveling Salesman Problem (TSP) is considered as an old and difficult problem in combinatorics. To solve it effectively has not only great theoretical function, but also very important pragmatic value in the filed of calculable theory.Firstly, the actual using value of TSP in flying probe tester is introduced based on the using background. Then Traveling Salesman Problem is described and its mathematics model is provided. Some correlative algorithms are introduced and their capability of solving TSP is compared.Secondly, Genetic Algorithms is used in TSP. A method for the improvement on crossover operator aiming at TSP is introduced. Furthermore, a method to solve a special type of TSP - the Matrix Type of TSP is introduced based on the thought of local optimizing.Finally, the exploitation process of the controlling system for flying probe tester, including hardware and software, is introduced. The solder-points is considered as cities in TSP, as well as, probe as salesman, the total length among all solder-points as the cost of travel. The TSP is used to optimize this cost. As is shown, the software for flying probe tester has the right testing function and the TSP has been rightly used in the path optimizing for the probe.
|
PDF Full Text Request