07-29-2016 09:19 AM - edited 07-29-2016 09:21 AM
@RavensFan wrote:
What are the units for the X-axis? Is that 30 seconds, 30 milliseconds, 30 microseconds?
Let me repeat that last question.
Additional question. Are you saying that the waveform you show contains the entire waveform when the generator goes from 0 to 500 Hz?
What would really be helpful is if you ran your VI, stopped it, then goto File >> Make current Values default. Save your VI and attach it here.
Then we can play with your actual data.
07-29-2016 09:20 AM
The units for X axis is seconds. I samlped 30 seconds' signals.
The acquisition rate is 1000Hz.
Sorry , I do not konw the meaning about "the specific frequency of generator". Can you explain it again?
07-29-2016 09:22 AM
You say that you are sweeping from 0 to 500 Hz. But what is the generation frequency at the time you show your waveform?
Is it showing a snapshot in time when the generator is at let's say 100 Hz?
Or is that the entire sweep and it is 0 Hz at the 0 time, and at 500 Hz at time 30?
07-29-2016 09:24 AM
Thank you, Bob
This is the response of a cantilever, the first three modes mean the first three natural frequences of it.
We can easily see the first three modes in the graph, at where the first three peak values of magnitudes are.
07-29-2016 09:32 AM
Here is my vi :
You'll see I generate a sweep sine signal, and many other measure vis
Also some filters ,while I set their cutoff frequences high to give the original signal.
Sorry my English is not very fluent.
07-29-2016 09:35 AM
Oh, I got you !
Yes, the sweep signal is 0 Hz at the 0 time, and at 500 Hz at time 30.
07-29-2016 09:38 AM
Can you also attach a picture of the stimulus?
Oh, of course. I just noticed that your Timer loop (what does this run on, by the way, a cRIO or something like that? Surely not on a PC ...) has what I assume is a D/A write and an A/D read (very strange, but if it's FPGA stuff, that's just outside my Experience zone).
You do know, I presume, that you must low-pass filter your signal to remove frequencies above the Nyquist frequency (500 Hz) before you sample it, don't you? Doing a low-pass filtering after the sample doesn't remove any aliased input. For obvious reasons, the pre-sampling Filter, called an "Anti-aliasing Filter", is usually an Analog circuit to (you guessed it!) avoid Aliasing issues.
Bob Schor
07-29-2016 09:44 AM
Another ture is that when I use this sweep sine signal to sitimulate my cantilever, I heard some periodic vibrations. You know ,the nosies of the vibrations changed from high to low then high again, just periodically!
When I set the sweep range from 0 to 500Hz, I may heard five or six periods' change, every period may last about 6s to 5s , while when I set the sweep signal range from 0 to 50Hz, I may only heard two turns' change, each lasts for about 15s.
It seems that the graph to be reasonable!
07-29-2016 09:45 AM
OK, aliasing is probably not your problem.
That timewave does not look like a sweep response to me.
Blog for (mostly LabVIEW) programmers: Tips And Tricks
07-29-2016 09:54 AM
Right Bob,it runs on a cRIO.
As I told upstairs,the stimulate is a sine that the frequency is 0Hz when time is 0s, and 500Hz in 30s time.
Yes, I do konw the filter is not used in this vi , I just set it at 500Hz, to not affect the results.
But in my experiments it can be used to filter high frequences, I can set the parameters. I use the filter to remove the high modes so I can only use the first three modes to do system parameters' estimation.
The result is the same if I remove the filter, trust me.