Even if your noise source is not so flat (say it's got some 1/f component or something) you could still measure the noise source un-filtered, then the noise source filtered and take the ratio between them to get the filter response, this is how we would measure optical filters with poor light sources that showed peaks and all sort of irregularities. As long as the power spectral density doesn't drift with time you'll be good but even then you could fight drift with averaging. This way you could get good nuff results without expensive noise sources. Edit: A minor problem with this is the ratio is usually super noisy outside of the noise bandwidth, the reason being if your noise drops too much you don't have resolution to observe how the filter is affecting it
The obvious thing that you left out is showing the response without the noise source (just background noise)! You can also show the response without the amplifier, too.
Hi Sebastian, first thanks for making this series of noise measurement video, it really help me understand how spectrum analyzer is related to noise level measurement, and I have a question, before turn on noise source and amplifier power, noise level is at about, -110dbm to -115dbm, after power on amplifier, level goes up to about -96dbm (shouldn't be -110+30=-80dbm?) and after power up noise source, become -80 dbm (noise source has a +15db ratio, so it looks correct), so amplifier only amplify by 20dB?( not 30dB?)
This is very interesting, specially for testing Surge Arrester EMI/RFI Filtering capabilities, that may range from 20kHz to more than 10MHz. I am in need to conduct such tests in several Surge Arresters.
This is very informative but you left out the part about how to do it with fft. I tried to get it to work on my Agilent 54622A and had no meaningful results at all. I’m using a Rigol function generator wide band noise function for my source on a low pass audio filter and see no response trace, just the noise. Can you give me some hints on how to do this right?
great video here. instead of a pricey scope, also can use cheap SDR dongles. i use these as my spectrum display and shows the filter response very nice. keep up the great videos, see my channel for more radio fun. take care.
![Noise Figure Measurement [Gain Method]](http://i.ytimg.com/vi/IgXKS0yJ1-E/mqdefault.jpg)
![Noise Figure Measurement [Gain Method]](/img/tr.png)
Thanks very much for the informative video. A video on how to build a noise source would be great too.
Even if your noise source is not so flat (say it's got some 1/f component or something) you could still measure the noise source un-filtered, then the noise source filtered and take the ratio between them to get the filter response, this is how we would measure optical filters with poor light sources that showed peaks and all sort of irregularities. As long as the power spectral density doesn't drift with time you'll be good but even then you could fight drift with averaging. This way you could get good nuff results without expensive noise sources.
Edit: A minor problem with this is the ratio is usually super noisy outside of the noise bandwidth, the reason being if your noise drops too much you don't have resolution to observe how the filter is affecting it
I relay like these "How to" videos with real equipment on the table! keep them going :)
Great extension of the last video, thanks for sharing your tips and techniques. 73 - Dino KL0S.
The obvious thing that you left out is showing the response without the noise source (just background noise)! You can also show the response without the amplifier, too.
Hi Sebastian, first thanks for making this series of noise measurement video, it really help me understand how spectrum analyzer is related to noise level measurement, and I have a question, before turn on noise source and amplifier power, noise level is at about, -110dbm to -115dbm, after power on amplifier, level goes up to about -96dbm (shouldn't be -110+30=-80dbm?) and after power up noise source, become -80 dbm (noise source has a +15db ratio, so it looks correct), so amplifier only amplify by 20dB?( not 30dB?)
This is very interesting, specially for testing Surge Arrester EMI/RFI Filtering capabilities, that may range from 20kHz to more than 10MHz. I am in need to conduct such tests in several Surge Arresters.
Glad you liked the video.
I found out that the FFT function of my scope is not good for measuring frequency responses. Using an SDR pen is actually better.
I would be very interested in building a noise source :-)
This is very informative but you left out the part about how to do it with fft. I tried to get it to work on my Agilent 54622A and had no meaningful results at all. I’m using a Rigol function generator wide band noise function for my source on a low pass audio filter and see no response trace, just the noise. Can you give me some hints on how to do this right?
Thanks for the video
great video here. instead of a pricey scope, also can use cheap SDR dongles. i use these as my spectrum display and shows the filter response very nice. keep up the great videos, see my channel for more radio fun. take care.
SDR's are usually limited to a few MHz bandwidth I believe