Error

[Petar2015]

The signala I'm seeing online during data acquisition in LV, are a good looking signals, for input, I command the function generator with LV, and it generated the desired frequency, for the viscosity which I measure at the end, is much higher the theory value.

[RavensFan]

Why don't you post some screenshots of your actual signal, both with how it is being acquired with a DAQ card, and also with another instrument like an oscilloscope.

[Petar2015]

Here is the screenshot of the acquired signal during the aquisition. The upper graph is output, and the lower one is input...Function generator...

[billko]

There doesn't seem to be anything in the signal that couldn't be captured with something on the order of 10x sampling - at this level of detail, anyway.  What about modifying the scale on your output so we can see what one (or just a few) cycles looks like so we can see if you have some high frequency stuff hiding in the signal?

[Petar2015]

between the photodiode which measures output, I use a "Bruel & Kjaer" conditioning amplifier, so it filters and amplified the output, I took screenshot at the end of the sweep, because during the sweep I can't scale the graph, It's a little noisy but I don't care, whatis important for us is the resonance, which I sent you the screenshot of that. and it's clear

[jamiva]

The statement "What's the highest frequency harmonic you care about?" is not entirely true with regards to Nyquist. It’s not what you want, it’s what the spectrum looks like. I don’t know the specifics of the “Bruel & Kjaer” conditioning amplifier/filter. If it is introducing high frequency noise into your signal, it will be aliased in your measurements at 70 Hz. As I see it, you either need to sample at the Nyquist frequency of the “Bruel & Kjaer” (70kHz?), or low pass filter the raw signal before the A/D and sample at twice that cut-off frequency.

[billko]

---

@jamiva wrote:

The statement "What's the highest frequency harmonic you care about?" is not entirely true with regards to Nyquist. It’s not what you want, it’s what the spectrum looks like. I don’t know the specifics of the “Bruel & Kjaer” conditioning amplifier/filter. If it is introducing high frequency noise into your signal, it will be aliased in your measurements at 70 Hz. As I see it, you either need to sample at the Nyquist frequency of the “Bruel & Kjaer” (70kHz?), or low pass filter the raw signal before the A/D and sample at twice that cut-off frequency.

---

I agree that you should clean up the signal as best you can physically before anything software happens.  Besides making it easier to take readings, your test equipment will thank you for not exposing it to potentially damaging spikes and stuff of that nature.

[Petar2015]

Thanks so much for your replies, I really appreciate your time. But unfortunately they didn't solve my problem. It has been 3 months which I'm  working on this experiment. I tried with different sampling rate, but it took 3 months for me to figure out the reason of not getting the precise result, which was lower sampling rate!  

[billko]

Honestly, I believe your time would've been better spent figuring out why a lower sampling rate wasn't working for you.  I think that proper conditioning of the signal before taking your measurement probably would solve your issues.

[Petar2015]

So could you please explain how I can adjust the amplifier? Right now it's on 1KV.