Express spectral measurement units?

[WSOCT]

What are the units of Express Spectral Measurements (Setting=power spectrum, results = dB, no window, no averaging...see attached)?

I am using the simulate signal express vi to simulate a 100 Hz signal with 1024 points (voltage = 1 V). (By the way), Is that Vpeak or Vrms?

I take the output of the simulate signal, run it into Spectral Measurements and then a graph. The amplitude is -3dB, which is what one would get if you scaled the FFT for # of points, didn't correct for real signals (i.e. multiply by 2), and assumed it is 1 Ohm resistance (why not 50, or why not make this selectable in the spectral measurements express vi).

So, given the above assumptions are true, how can I simulate a 1Vpeak sine wave of 100 Hz into 50 ohms and get the spectral measurement results of that?

Thanks.

[Kristen_K.]

Hi WSOCT,

The Voltage will be peak voltage because the express VI uses the Sine Waveform VI behind the scenes, which specifies that the amplitude is the peak voltage.

 Can you explain what you mean a bit more by didn't correct for real signals?

[WSOCT]

Hi Kristen,

In the attached (from my first posting) is the vi that demonstrates everything. Please run that and explain why the graph shows the peak at -3 dB given that the peak voltage of the sine wave input is set to 1V.

[JÞB]

---

WSOCT wrote:

Hi Kristen,

In the attached (from my first posting) is the vi that demonstrates everything. Please run that and explain why the graph shows the peak at -3 dB given that the peak voltage of the sine wave input is set to 1V.

---

It is because you have a very pure sine wave.  Exactly half the Total power is within 1 Hz of the selected frequency and one half of the Total power is in therest of the spectrum.  The math is quite involved but, The "Fourier series" is a good introduction.  The results are a logrhythmic ratio of the power at any one frequency to the total power.

[WSOCT]

My point is that the RF Spectral Measurement doesn't correct for resistance (50 Ohms).

1Vpeak sine wave's Fourier Xform is .5 at +fc and -fc. Convert to dBW and you get -3 just as you should (i.e. R=1 Ohm).

Or, to come at it differently:

P = Vrms^2/R.

Vrms = 1/sqrt(2).

P = .5/R

Would be nice to have the ability to set 'R' (in most cases, you'd want R=50) in the express vi's so that you don't have to correct for it if you are trying to simulate a 50 Ohm environment.

(convert 1Vp in a 50Ohm system to dBW and you'd get -20, not -3).

So, I guess my question/point is that both express vi's are using R=1.

[Kristen_K.]

Hi WSOCT,

You can always submit a product suggestion.  It sounds like you have a valid argument.  Here is the link to submit a product suggestion.  You'll need to click the feedback link.  This feedback is looked at by R&D when considering features to add in the next release of the product.  

[JÞB]

Yes,

You will need to apply a scale.  the VI returns a an array of ratios that add up to 100% (or nearly).  You could prove this by changing it to a linear ratio an integrating.

It sounds like you want the actual power out of the FFT yet scaling the power AFTER optaining the spectum would be the prefered method.  

 to say "1Vpeak sine wave's Fourier Xform is .5 at +fc and -fc. Convert to dBW and you get -3 just as you should (i.e. R=1 Ohm)."

is an incorrect assumption in that you are assuming a referance for the ratio that is incorrect.  The referance in this vi is TOTAL power not the watt

[WSOCT]

OK, now that we are all on the same page

How would you recommend I do what I am trying to do? I am simulating several things including a noise floor, so scaling "after" is nearly impossible (but preffered).

Should I stick the express vi's and scale, or is the some other way?

[JÞB]

---

WSOCT wrote:

OK, now that we are all on the same page

How would you recommend I do what I am trying to do? I am simulating several things including a noise floor, so scaling "after" is nearly impossible (but preffered).

Should I stick the express vi's and scale, or is the some other way?

---

OK the express vi's are nice but----- Not too professional- they involve a lot of overhead that you don't need but.... just saying.

So, yes, simulate your signal and all its components. (or acquire it any how you want).  Ring it through the FFT to convert to power vs Freq.

Now you need to know something else about the input waveform- you only had voltage vs time to start with and you only have percent of total power vs frequency now.  Apply what you know about the input signal to scale the spectrum into absolute power.  You've already demonstrated the math so I won't repeat it.  

Glad we're on the same page.

<Set Joke mode=True>

"there are 3dB kinds of people, those who understand exponents and, those who don't"