Create signal at time of first sample of failed signal from a Limit Test

[JockeR]

Using the Limit Test I'd like to pick out the time of the first sample of the signal that fails the test.

Then I'd like to create a unit step signal (from 0 to 1) at this specific time.

I'm using SignalExpress 2009.

Is this possible? In that case, how is it done?

Thank you for any help.

/Jocke  🙂

[JockeR]

I've solved a part of the problem.

The step signal is created using the "Create Analog Signal" with the "Signal type" option set to "Formula".

The formula is entered as  a*step(t-T)  where T is the time of the step.

The remaining problem is how to extract the time of the first sample, T in the above formula, of the signal that fails the Limit Test.

[Davidek]

Hi Jocke

I tried something with the results from the Limit Test. If you look at the components (press the + sign in front of the signal) of the limit test results you have one thats called "Failed signal" which gives you the part of the signal that is outside of the boundry. What I did was that I used this conmponent as an inout to a Formula step and used the formula step(x0), where x0 is the variable assigned to the input signal. This produced a step that that goes to 1 when the signal breaks the limit and then goes back to 0 when the signal returns inside the limit again. Would that work for you?

If not something else you could try is to do it with an alarm. Go to the recording options tab and choose the Alarms option. Here you can add an alarm that triggers at certain conditions, among them when your signal breaks limits (above, below single limit and between ot outside two limits). You can then set an alarm action. Remember that these alarms only work when you are recording and the signal in the alarms is among the signal that is being recorded.

Best Regards

David

NISW

[JockeR]

Hi David, thanks for your reply.

Your suggested solution is one step in the right direction.

But as you said, the step returns to zero when the tested signal later passes the test and that's not what I want.

The test should trigger the step signal and then remain in that state no matter what happens to the tested signal. 

I'm trying to do this in post-recording processing so using an alarm is not an option.