Thread: What's the physical meaning of voltage noise unit V/(Hz)^1/2

Hello experts,
Could anyone explain what's the physical meaning of voltage noise unit V/vHz?
If V/vHz is relatively large what that implies?

Best regards,
bittware

Noise is really power and proportional to bandwidth. The density is noise power/bandwidth. To get voltage units you have to take the square root of both. The square root of power is voltage, the square root of Hz is the square root of Hz.

Originally Posted by flatulent

Noise is really power and proportional to bandwidth. The density is noise power/bandwidth. To get voltage units you have to take the square root of both. The square root of power is voltage, the square root of Hz is the square root of Hz.

So how to measure this parameter at the output of a power supply? I think it should be a derived parameter. If I am right what are the V/vHz calculation inputs?

Your best bet is to put a resistive load on the supply and ac couple to your measuring device. The frequency range for the noise measurement should be specified. Use a spectrum analyzer to see what shape the output noise spectrum takes. You can read the screen when set to voltage scale and use a narrow resolution bandwidth. The manual should tell you the correction factor.

If all you have is a true RMS meter you can put a simple RC low pass filter after the series capacitor and get a crude measurement of the RMS voltage over the bandwidth which is actually 1.57 times wider than the -3 dB frequency of the filter.

[added later] I forgot to be clear and say you divide the RMS meter reading by the square root of the noise bandwidth which is 1.57 times the 3 dB bandwidth.

Originally Posted by bittware

Hello experts,
Could anyone explain what's the physical meaning of voltage noise unit V/vHz?
If V/vHz is relatively large what that implies?

Best regards,
bittware

it gives power spectral density of noise...(PSD)

A very good explanation:

http://sound.westhost.com/noise.htm

Originally Posted by flatulent

Your best bet is to put a resistive load on the supply and ac couple to your measuring device. The frequency range for the noise measurement should be specified. Use a spectrum analyzer to see what shape the output noise spectrum takes. You can read the screen when set to voltage scale and use a narrow resolution bandwidth. The manual should tell you the correction factor.

If all you have is a true RMS meter you can put a simple RC low pass filter after the series capacitor and get a crude measurement of the RMS voltage over the bandwidth which is actually 1.57 times wider than the -3 dB frequency of the filter.

[added later] I forgot to be clear and say you divide the RMS meter reading by the square root of the noise bandwidth which is 1.57 times the 3 dB bandwidth.

Hello flatulent,
What is meant by "after the series capacitor"? Where does the "series capacitor" come in? How do you derive out the "1.57" factor? If I am interested in band pass noise energy, should the low pass filter be replaced by a band pass filter? If yes, what's sort of band pass filter you would recommend?
In fact, I have a true RMS meter, i.e. FLUKE187. (Its specification is attached.)Is it possible to use it alone to find out whether the power supply output noise is below 20uV, rms(10Hz to 1MHz) level(convert to 20nV/√Hz, am I right?)? After all, for me, the spectrum analyizer is too expensive to afford. If the precision is concerned, what's your advice on chosing proper RC and making layout?
Other than the multimeter, I also have a Tek TDS3054B(500 MHz, 5.0 GS/s) oscilloscope which is what I have the most advanced instrument. Would the oscilloscope's FFT module make any help? If the answer is unfortunately no, will it be able to do any help on other aspect for identifying the noise level?
Waiting for your savvy advices.
Thanks in advance.

Best regards,
bittware

After the series capacitor means that you AC couple your measuring device to the supply output so that the DC value does not get to the measuring device.

The VOM should do the trick on the AC true RMS measuring mode.

A loose definition of noise is the output power, or power spectral density (power in a given bandwidth) with no input signal. The advice on measurement is good, but don't forget to have a calibrated resistor or metal film resistor for the measurement so the noise is ideal thermal noise i.e. 4ktR(df).