Error

[billko]

---

@Petar2015 wrote:

It's sweep, the frequency changes from 40-70, so you mean I adjust the sampling rate to 140.?

Okay, I'm going to try it, though I've tried 1000 before. 

---

Well, 140 Hz would be the minimum.  I like the idea of 10x instead of 2x.  Are you trying to capture soemthing out of band?  Like a harmonic?

[crossrulz]

---

@Petar2015 wrote:

It's sweep, the frequency changes from 40-70, so you mean I adjust the sampling rate to 140.?

---

Like I said, 140 is on the low side.  I would recommend at least 500Hz.

Here's a better question: What are you trying to measure?  If you are just trying to measure that 70Hz signal, then you should just need to sample at 140Hz (as a minimum).  If you are trying to measure something else, we need to know that.

[Petar2015]

Both of the input and output are harmonic. Now I'm acquiring with 140, then I'll try 700. I'll let you you how the result is.

[billko]

What's the frequency of the highest harmonic you care about?  You can start with twice that frequency and adjust upward from htere if you need to.

[Petar2015]

After recording the time response, I take FFT in MATLAB, so when the number of the data points were higher, I get more accurate result.

[billko]

I think you might need to think about the Law of Diminshing Returns.

[Petar2015]

Even for 70000 sampling rate, I can't obtain acceptable result, I tried 140..but nothing.

Would you please explain more about that law?

[crossrulz]

What is considered an "acceptable result".  In all data processing I have done, Nyquist works.  I'm thinking you have an issue with your math if you can't get proper results at 70kHz.  At this point, you are allowing  a lot of noise to enter the system.  So again, what EXACTLY are you trying to measure?

[Petar2015]

I measure the viscosity, after I take time response, I take FFT in MATLAB to obtain FRF, the width of the plot will determin damping and viscosity, it's so susceptible to the acquired data.

[Hooovahh]

If you are sampling a 70hz signal with a rate higher than 700Hz, and not getting a good looking signal, then your input to the system is not what you think it is.  Maybe the frequency is not 70Hz, or maybe it is with lots of high frequency noise on it.

Also we haven't seen any code yet, but I suspect you are just using a modified example for a finite or continuous measurement.