How to measure SNR using Spectrum Analyzer? (dBc or dB)

[nuvoletta]

Hi, I'd like to know how can i measure SNR from Sigma-Delta ADC using Spectrum Analyzer in Simulink.
When i looked into Spectrum Analyzer, these measurements were given.

- Signal Power 23.33 dBm
- SNR 58.75 dBc

I was little bit confused to see dBc because SNR is represented in dB and i read one article from this forum (https://www.edaboard.com/threads/how-to-convert-dbc-to-db.312112/#post-1334852) saying that there's no need to convert dBc to dB ?

But here's what i thought...

Since dBc means decibels relative to carrier, from the given Signal power(23.33 dBm) and SNR(58.75 dBc), you can derive Noise power which is -35.42 dBm.
Then you can convert dBm to Watt which means Signal power = 215.278 mW and Noise power = 287.078 nW.
Finally you can put these numbers to SNR = 20log(Signal power/Noise power) which leads to 117.5 dB.

I don't know which number (58.75 or 117.5) that i can trust for SNR. Can anyone help me on this?? Thanks!!

 

[FvM]

Finally you can put these numbers to SNR = 20log(Signal power/Noise power) which leads to 117.5 dB.

Please correct to SNR = 10log(Signal power/Noise power). Respectively, same value as given in the measurent.

 

[nuvoletta]

Oh my mistake! Thank you so much!!!

 

[kaz1]

SNR in dBc is SNR in dB relative to signal being carrier. so all those conversions back and forth are not needed. It is by itself SNR in your case.

 

[JohnBG]

Nuvoletta, to define carrier to let’s say a spurious nearby [dB] suffice, there's no need to append a 'c' to the unit tag.

[dBc] is used for one side (read IEEE 1139) carrier to noise ratio but together with the spectral distance, as [dBc/Hz].

[dBc/Hz] is the preferred unit to measure carrier (without modulation) sharpness at agreed, usually short spectral distance, from carrier.

Quiet oscillators show ‘deep’ (large absolute values) and test measurement manufacturers display phase noise values in [dBc/Hz] at more or less standard distances from carrier, allowing potential customers to compare according to phase noise in [dBc/Hz] as a figure of merit of spectrum analysers, signal generators, test receivers and in general any RF device attempting to receive or transmit carrier modulated signals.

So bear in mind:

1. [dB] units are enough just to rate power differences among spectral spikes.

2. For specific one side carrier sharpness, often measuring phase noise [dBc/Hz] is the industry standard, typically at distances like 100kHz, 200kHz or 1MHz.

3. When comparing phase noise figures make sure same SPAN applied: A cheaper equipment may show a larger [dBc/Hz] but at 1, 2, or 10MHz while expensive equipment may show a slightly [dBc/Hz] but at far closer to carrier.

4. In any case, ask for the graph, that should show a ‘down hill’ smooth slope: ups and downs denote noise regrowth caused by same equipment.

 

[FvM]

Nuvoletta, to define carrier to let’s say a spurious nearby [dB] suffice, there's no need to append a 'c' to the unit tag.

Necessary or not, dBc is the usual measure for supression of harmonics and spurious components in radio regulation standards.

 

[JohnBG]

I agree.
[dB] and [dBc] are dimensionless but often people also shorten [dBW] to just [dB] now Power, [dBW] being the ratio between a power measurement and 1W reference signal assumed reliable, now no longer dimensionless, and [dBmW] is also often shortened to [dBm] again Power measured in relation to 1mW source.

One also often encounters [dBV] [dBmV] and [dBuV] but here no one omits Volts, because of the far too often dB(W) dBm(W) omission.

 

[danadakk]

These might help -

https://cdn.rohde-schwarz.com/pws/dl_downloads/dl_application/application_notes/1ma201_1/1MA201_9e_spectrum_analyzers_meas.pdf

Here are some classic HP ap notes. Do a "Ctrl F" keys on page and search for "spectrum",
as yo scroll down page you will see table entries highlighted in the ap not listing which
are spectrum analyzer topics.

HP Application Notes

Regards, Dana.

 

[JohnBG]

Wonderful answer danadakk.

I agree with Nuvoletta that [dBc] is often used as a simple [dB] distance between spectral carriers.

But as danadakk points out with the R&S app : [dBc] is the standard way to measure whatever spectral level found at 1Hz from carrier, by default, or [dBc/Hz] when at another spectral distance different than 1Hz.

And it hasn't got to be anothet carrier, a harmonic spike, but at 1Hz, when using [dBc] alone, the chances are it's phase noise from same reference carrier.