online signal conditioning

[godfather]

I'm searching a VI that is able to analyse a periodic pulse-signal with known amplitude and frequency and give out a warning as soon as one pulse is missing. The acquired signal don't need to be saved.

Bye

Christian

[Michael Aivaliotis]

A late response here but I would recommend that you create a VI that counts a fixed number of transitions (low to high crossing a given threshold). Do this for a certain sample set. Let's say 1000 samples. Then I would compare the count of transitions to the total number of samples collected. If you have fewer transitions for the given samples, you have a problem.

[didierj]

Depending on how you get your signal and how high the frequency is, there are different possibilities. But there is no fully wired vi (except maybe somewhere in the examples).

1. Method of Michael for fast signals,

2. for slow signals calculate the time difference between two transitions. If the time is longer than "1/f+delta" (where delta is e.g. 1/10 of your period time) then a transition is missing.

3. If you get your signal from a copper wire, then you can supply it to a timer card. Set the timer card so, that its counter counts downward at a certain speed. If the counter reaches zero an event is produced to LV. Now a transition on your input signal triggers the timer card and refills the counter with a certain value. As long as the input signal produces transitions, the counter never reaches zero. If a transition is missing, then the counter reaches zero and produces an event to LV.

(the whole thing can also be done with a resistor, condensator and comparator).

Didier

[godfather]

A late response here but I would recommend that you create a VI that counts a fixed number of transitions (low to high crossing a given threshold). Do this for a certain sample set. Let's say 1000 samples. Then I would compare the count of transitions to the total number of samples collected. If you have fewer transitions for the given samples, you have a problem.

Hi Michael,

I solved my problem in a similar way. The VI aquires data over about 100 pulses and measures the time from rising to falling edges and from falling to rising edges. Because it is not pwm signal, these times have to be almost equal. If not, there is an error.