02-22-2010 02:17 PM
This is very difficult question. So only very expert can respond.
Suppose that U have periodic signal. Simple signal: square wave, but with dutyfactor very small.
U need only to sample one time in a high level, one time in a low level, but U must be sure not to lose a sample in the short high part of signal after same periods (maybe for accumulation of delay). It's obvious that U can't use simple loop because so U take more samples than necessary. HOW CAN U DO THIS?
If u have solution post VI for LW8.
02-22-2010 02:30 PM
First, please follow a few rules on the forum.
Don't spell out in ALL CAPS unless you are trying to yell.
Don't use text message abbreviations like "u". Use full English words.
It's not clear what you are talking about. Can you draw a sketch of your periodic signal and make notations on it as to what the different parts you are talking about are?
02-22-2010 02:35 PM
Michelle,
The answer to your question may depend on what you intend to do with the data. Two samples per period simply specify the amplitude of the signal (for a signal known to have only two levels). However, those two samples alone are insufficient to specify the frequency and duty cycle. If you also have the time of each sample, that put limits on the possible values of frequency and duty cycle. To completely specify such a signal you need at least two amplitude measurements and two timing measurements. The amplitude measurements must be made during the portions of the period when the amplitude is not changing. The timing measurements must capture the times at which the transitions from one level to the other occur.
Typically sampling is done at a rate sufficiently high to capture valid samples for each type of measurements. If it is only desired to save small amounts of data (such as amplitude of duty cycle), those reduced parameters are calculated and the raw data is discarded.
Please describe what you are actually trying to do so that someone may provide more specific help.
Some of your questions are very fundamental. If you are not familiar with LabVIEW, please look at the tutorials as a good way to get started.
Lynn
02-22-2010 04:04 PM
Maybe I'm missing something, but this sounds like the type of behavior usually handled by an Event. Nothing happens until the "signal" (e.g. whatever triggers the event) changes state. I don't know much about how LV handles these things interally, but it sure doesn't take up any processing time when the state is not changing (I have some VIs that use Event structures to perform operations based on the change in state of front panel controls).
Regards,
Michael Tracy
Synergy Microwave
02-22-2010 04:15 PM - edited 02-22-2010 04:16 PM
02-22-2010 04:28 PM
02-22-2010 04:30 PM - edited 02-22-2010 04:30 PM
Ravens Fan wrote:First, please follow a few rules on the forum.
Don't spell out in ALL CAPS unless you are trying to yell.
Don't use text message abbreviations like "u". Use full English words.
It's not clear what you are talking about. Can you draw a sketch of your periodic signal and make notations on it as to what the different parts you are talking about are?
^^ +1. ALL CAPS is considered poor netiquette to many.
02-22-2010 04:31 PM
Let's start at the beginning and you provide the hardware you have in order to capture the pulse. Also, is the pulse constantly being output or not? With the correct DAQ card, you can set a trigger and capture the entire pulse with a single DAQmx Read.
And please, what the heck does "I take many samples I take all zero" and "it's only a loss of memory" mean?
02-22-2010 04:31 PM
02-22-2010 04:35 PM
Part of me is thinking this is homework so I'll give a snippet so at least you'll have to create your own LV8 version. All you need is a simple integration. No memory worries, no waiting around. The final value of the integral is the amplitude of the impulse.