Understanding Function Generator Noise for Calibrating Noise Measurements

[Dele]

I want to calibrate my measurement setup for noise measurements. I have taken some measurements but I need to confirm that my measurement correlates with the Johnson noise theory.

To do this, I have tried measuring resistors but I seem not to get the right results. My supervisor then recommended using a function generator to do the calibration. I have measured certain noise voltages (0.1Vpp, 0.5Vpp and 1.0Vpp). However, I don't seem to understand how these relate.

My question is this. what does the noise spectral density measured from the function generator noise output translate to? A resistance? How do I compare and what might I compare with.

P.S. I intended to compare the measured spectral densities to certain resistances as calculated from the Johnson noise equation. Is this correct??

Thank you

 

[f95toli]

It will depend on the specific function generator model you are using. I assume you are using its "noise" output? Function generators generally assume a 50 ohm load
and ssuming the noise is reasonably white (i.e Johnson-like) you should (check the manual!) be able to just calculate the total noise power directly (1Vpp into 50 ohm).

Note that it should be fairly straightforward to measure the noise power from a high value resistor at room temperature and I would trust that over the output of a function generator. There isn't really much that can go wrong if you do this; especially if you are using an instrument that can measure the PSD directly such as a FFT analyzer.

 

[Dele]

It will depend on the specific function generator model you are using. I assume you are using its "noise" output? Function generators generally assume a 50 ohm load
and ssuming the noise is reasonably white (i.e Johnson-like) you should (check the manual!) be able to just calculate the total noise power directly (1Vpp into 50 ohm).

Note that it should be fairly straightforward to measure the noise power from a high value resistor at room temperature and I would trust that over the output of a function generator. There isn't really much that can go wrong if you do this; especially if you are using an instrument that can measure the PSD directly such as a FFT analyzer.

Thanks so much. I am strongly tending towards measuring the noise power in high value resistors ATM.

While at it, might I ask (might look stupid). The spectrum analyzer I use give a whole range of the spectrum different PSDs. In fact I see something similar to the 1/f in the result. I am therefore not sure whether to trust that also. I expect to see white noise of same power over the frequency range (100Hz) in my case.

Thanks for your response

 

[Henryk]

How high is your resistor value? Are you using a coax cable connection to the analyzer?
If you, you might be shunting the signal by the coax cable capacitance.