Automating the test

[developer help]

I have to test the functionality of DIO, ADC, PWM & CAN modules of a software,

I will write a VI to switch ON & OFF the digital ports & send analog voltage... According to the test case and read the status & voltages and display on the front panel. and also check whether the test is PASSED or FAILED. etc and the test will generate the test report as HTML file.

I have multiple test inputs

eg,.  SW1  SW2  SW3  SW4

TC1   ON   OFF    OFF  OFF

TC2   OFF  ON     OFF  OFF

TC3   ON    OFF    ON   ON ...

         Ai1   Ai2   Ai3

TC1    5v    0v    0v

TC2    0v    5v    5v... like this of too many combinations, so it will be difficult to test manually. i need to do Automate the test, it should get the test case from a file or from NI Test Stand and perform the test & the test report should be generated.

Can you pls give me some ideas how i can automate the test & how to get the test inputs etc.

Thanks..

[JÞB]

Developer,

I would have to recommend using a third party vendor.  While LabVIEW and TestStand both have incredible power, the info you seem to be looking for is the answer to the almighty question "What is a Test Executive?"  The answer to THAT question depends on your company's needs, local or Federal regulations, the point(s) in the business cycle that the test executive is designed for and a mean of the corporate talent pool.  

Sorry to say, but these questions are best answered on a case by case basis so they can address specific needs of the business they serve.

That being said, some general rules apply.

1) a generic "Test" must have some way of learning what the experiment consists of.  Some use sequences in Test stand, Some a path to a LabView type def and a configuration file and some are hard coded with constants or variables and other methods exist.

2) a generic "Test" must have some way of controlling stimulus package.  Temp, Humidity, Voltages, frequencies, lumins, amps, ohms, time etc.... or any combination of these.

3)a generic "Test" must have some way of measuring the Device Under Test's response with tracebility to an NIST standard so the measurement is reproducable.

4)a generic "Test" must have some way of recording all of the above.

So, as you see, an automated test is identicle to any other experiment.  However, you must teach a computer the steps to take (harder than it sounds but, "its a living!")

As for referances to good sources- well, click my name or the names of many others who post on the forum.  I'm hardly working!