Is there away to run parallel executions of different test sequences?

[lc_garza]

The new process models of TS 2.0 (ParallelModel.seq and BatchModel.seq) allow to run the same test sequence on multiple UUTs at the same time. Is there a way to use these process models to execute different test sequences on multiple UUTs at the same time? Is there a way to select which test sequence will be run in each parallel execution? If not, any clue of where should I modify these process models to achieve this functionality?

[RayFarmer]

Hi,

Using the ParallelModel.seq as the base

The new process model in the zip file has a new step in the Initialize TestSocket sequence which generates a new Client Sequence File path based on the Computer Motherboard Test Sequence.seq string.

This new path name is passed as a parameter to the Create Test Socket Execution for each of the test socket executions.


Unzip the file and place the new Process Model into \Components\User\Model\TestStand and place the test sequences into the folder \Examples\Demo\LabVIEW\Computer Motherboard Test.

Launch the seq editor and set the Station Model to use the new Pocess model ( you can set it back when you have finished). Open one of the new test sequences and select the Test UUTs to start the sequences. The
four new sequences will be launched.

I can see that the sequence that you initially open wont be your test sequence but a sequence that allows your test selection or sets up the sequences to run, setups your process model to run etc.

(This new Process Model sequence is from Teststand 2.01 beta, so if you have any problems then cut the new step and paste it into your ParallelModel sequence)

Hope this is a start
Regards
Ray Farmer

[mhk]

I tried your demo. But for the first socket the file with the ending 3 is loaded not the file with the zero. When I modify the model to use the endings 1 to 4 it works loading the right files.

[RayFarmer]

Hi,

Thanks for that.
I guess a better solution would have been to pass an array of sequence files would could have been selected via some dialog panel.

Regards
Ray Farmer

[activextp]

Hi,
it seems to be good, but it don't works on teststand 2.0 and its a little bit complicated do you have a simple sequence with different execution on the same time

[RachaelBrown]

Hi Ray

Is it possible to change the sequence file once the test socket is initialised?

i.e. if I wanted to run sequence0 for the first UUT in Socket0 and then sequence1 in socket0 without terminating the other sockets.

[RayFarmer]

Hi,

Sorry for delay - Holiday

It would be difficult to run another sequence file in the the same TestSocket. You would have to close that execution down and then launch a new execution for the next Sequence File. You would have to do some major changes to the process model to handle this.

Could you not just call your Sequence1 from your Sequence0 at the end via a SequenceCall adapter.

Regards
Ray Farmer

[RachaelBrown]

Hi Ray

Thanks, I thought it would be difficult to do.

What I really want is the user to be able to select different "UUT Models" with out having to reload all the sequences. Some of our tests can take 5/10mins to load and our production runs very low batches (often a batch of 1).
Since posting the original comment, I've considered creating a �Super Sequence� that will call the correct �UUT Model� sequence or dynamically loading the sequences as and when they are required.

Cheers

Willow

[Ray.R]

Hi Ray,

Will your model work with completely different sequences? As in running two different products in parallel, sharing resources, but having different tests??

And running TS-3.1..

R.

Message Edited by JoeLabView on 03-23-2005 02:45 PM

[RayFarmer]

Hi JLV,


Interesting idea.

It's been awhile since I looked at this but I surpose its possible. Although the original example used the Computer Mother Board sequence file and called four versions of it, they were infact four seperate files. So I guess it could equally have been four totally different sequence files.

Of course my example didn't really do anything, other than show that the mechanism was feasible. Also, there were a few teething problems, which I didn't really get around to sorting out, what with holidays etc...

>>>> Maybe some of the other guys (willow, activextp, mhk, lc_garza) have achieve a working solution? <<<<


Can you give a bit more information?

Regards
Ray Farmer