vendredi 1 novembre 2013

Increasing testability

For this project, my client asked me to increase the capacity of their test suites for your after-sales service: going from 500 units daily to 2500 units per day (units here are modems ADSL / VDSL telephone, HPNA, WIFI , WAN, Ethernet and USB.) Using benchmarks aftermarket in this case is to analyse and give a diagnosis to replace defective components and then verify that the repair has solved the problem.
To optimize the test suites in the best way, the analysis is essential to decide what to do I found several factors that limited the ability.:
  • too many manual actions: the test is unfair, well, whenever operator intervention The can not perform actions on another bank.
  • b ank usability testing: the operator loses time to place the unit, connect and disconnect.
  • software usability testing: test results are not clearly identifiable, the operator loses time to read the diagnosis, or worse, think again.
  • false failures: if false failures, operator definitely test and the test is restarted, we lose time. False failures They can be caused by hardware (measurement instability due to instrumentation wiring, connectors or other components) and software (processing measurement, analysis, tolerance).
  • the use of hardware resources: all tests are run one after the other, while some can run in parallel on the same drive or multiple drives.
  • process: having a single bank, to execute all actions and testing is not always the best solution. The number and functions of banks depend on the product, the tests to be performed, the environment and why we test.
  • optimization of the line: the placement of the test suites may penalize the flow, forcing the operator to waste time in their movement and increase the risk of handling errors that may damage the product.
  • traceability: having the product history is important to optimize the repair, fluidize (not try a product that has been tested failed previous tester without any repair), diagnose problems l test Provide maintenance actions, but also to analyze production.
After presenting an analysis of the identification of problems, I worked with my client on several points:
  • Review of the line and testers to improve the flow, ergonomics and movement of operators and products.
  • use of the material now available, replacing and adding new, rewiring and connector type for better reliability and speed.
  • l migration software testing from test scripts executed from the Linux command line to test sequences of TestStand on Windows by creating a simple screen (test in progress, good or bad result, pending test) to test parallelization to reduce the time.
  • test report generation, data centralization in the network and development of data analysis software.
  • operator training and actions optimization.

Aucun commentaire:

Enregistrer un commentaire