ADEXL: Making Dependent Tests

calcVal() is your friend here. The approach is as follows:

1. In your first test, have an output which calculates the result you want to pass to the second test. Make sure it has a name. Let's call that "result1" and the first test is called "test1".
2. For whatever you are hoping to control with the second test, ensure you have a design variable for that (and that the design variable is a global variable which is used by the second test but not the first test). Let's call that "variable2"
3. Drag the output "result1" over to the value of the global variable variable2. You'll see it creates it with an expression calcVal("result1" "test1") (you could type it in, but drag and drop will do it too).

That's it. Now it knows there's a dependency and test2 won't run until test1 is completed. You can also use calcVal in output expressions to calculate results which are a function of more than one test - again, this sets up the dependencies.

Regards,

Andrew.

Hi Andrew,

How about if I want to use "result1" to define a variable for "test2" in the Corners Setup?

Thanks

I don't see what benefit having this as part of the corners set up would be. If you have the dependency in the variables, then you get the result from the corresponding corner anyway.

Regards,

Andrew.

You won't get the corners though, and it is a pain since ADE-XL sorts out the results differently when you sweep a variable instead of using the Corners Setup.

I want to use "result1" and "result1a" from "test1" in "test2" to define min and max corner cases. I want the results to be part of one corner set, not 2 corner sets (one for min and one for max).

Unfortunately, I haven't managed to get the variable sweep to work with "result1" and "result1a" either. I used the syntax you mentioned above: calcVal("result1" "test1") calcVal("result1a" "test1"). ADE-XL only takes "result1" and ignores "result1a". I used an inclusion list with no luck either.

OK, so you're asking a different question from the one in the original post, and hadn't explained that what you were trying to do something different (this is precisely why the forum guidelines ask you not to post on old threads, especially when there's some difference in your question).

If you want to use a previous result to define points in a sweep set either via swept variables or via corners, you can't do that.

Andrew.

Thanks for your prompt response.

What I am after is very similar to this thread; the only difference is the sweep/corner option. General forum guidelines often urge you to ensure, before you post a new thread, that the answer you are after is not closely related to something already posted.

Anyway, I have now managed to sweep the dependent variable by using 2 inclusion lists (not both in one):

{Inclusion List}calcVal("result1" "test1"){Inclusion List}{Inclusion List}calcVal("result1a" "test1"){Inclusion List}

It'd be very beneficial if this worked in the corners. I know for a fact it used to work some time ago (not in the latest ADE-XL).

I wasn't talking about general forum guidelines (presumably for other forums), but the specific forum guidelines for this forum (pinned post at the top of each forum).

Anyway, if this worked before (I haven't tested in old and new versions) and doesn't any more, then it's a regression and you should contact customer support to report it. That said, I think this may be an untested feature and it just happens to work in some situations and so we need to formalise its use so that it's documented and tested behaviour.

Regards,

Andrew.