AnsweredAssumed Answered

N=kTB = -174dBm/Hz (@290K)

Question asked by xmo on Jan 27, 2013
Latest reply on Jul 18, 2013 by n3go

Like • Show 0 Likes0
Comment • 3

N=kTB = -174dBm/Hz (@290K)

So says every reference and application note I have ever read. NIST says this is wrong and it should be -177dBm/Hz.

http://tf.nist.gov/phase/noisemeas.html

I am curious as to the ramifications of this change on test instrumentation such as spectrum and noise figure analyzers, calibration of noise sources, etc.

Does Agilent have an application note or document that clarifies this discrepancy?

No one else has this question

Outcomes

3 Replies

tabbott Jan 29, 2013 2:16 PM
Mark Correct
Correct Answer
Hi -

All Agilent application notes regarding noise figure measurements all list kTB at -174 dBm/Hz. It seems every Google search on the web also agrees. There is some good information on Thermal Noise in Agilent application note 57-1 in the Glossary on page 26. In order for this value to change from -174 dBm/Hz, one of the three variables in kTB must change. Was it Boltzmann’s constant, temperature or bandwidth, which is referenced to 1 Hz?

The ramifications around test equipment would be huge.
Like • Show 0 Likes0
Actions
Dr_joel Feb 2, 2013 8:24 PM
Mark Correct
Correct Answer
> {quote:title=xmo wrote:}{quote}
> N=kTB = -174dBm/Hz (@290K)
>
> So says every reference and application note I have ever read. NIST says this is wrong and it should be -177dBm/Hz.
> >
No document that I know of, but NIST says "It has been verified both theoretically and experimentally that the single-sideband PM (and AM) noise floor due to thermal noise is -177 dBc/Hz, relative to a carrier input power of 0 dBm; this result differs from that found in early literature"

This refers to Carrier-to-noise and not ktB noise floor. To dig deeper, we know that adding noise can add some AM and some PM noise to a carrer. I would guess that this refers to the fact that the noise power is split equally between AM and PM when added to a carrier signal.
Like • Show 0 Likes0
Actions
n3go Jul 18, 2013 9:54 PM
Mark Correct
Correct Answer
Both are correct. NIST states ktb = -177dBm as an SSB noise floor. Excess noise sources (e.g. diodes) produce white random, and therefore incoherent noise. An SSB demodulator will thus see the noise content of both sidebands, in the post detection spectrum. As such, the noise contribution from the undesired (opposite) sideband folds over and adds to the noise in the desired sideband, effectively doubling the detected output noise power. Hence a NF measurement of an SSB detector measured without filtering (suppressing) the opposite sideband's spectrum during the noise figure measurement, will measure 3dB higher because it is in reality, a DSB noise measurement. Thus; in a true SSB measurement, ktb = -177 dBm, but most measurements are made as DSB measurements because wideband AM detectors are commonly employed which perform DSB detection by design, and ktb=-174 dBm becomes the assumed noise floor.

A SSB system front end measured post detection, is a filtered system with opposite sideband rejection, and ktb = -177 dBm applies... even though the excess noise measured post detection is itself a DSB measurement.
Like • Show 0 Likes0
Actions

N=kTB = -174dBm/Hz (@290K)

Outcomes

Related Content

Recommended Content