Making sense of KSA992 noise specification

Howdy

I am trying to make sense of the KSA992 Noise specification. How does the "Noise figure" in mv correspond to a noise figure specified in db?

http://www.mouser.com/ds/2/149/KSA992-88789.pdf

I am used to low noise transistors having charts like this, with the noise figure specified at different collector currents and source resistances.
http://pdf.datasheetcatalog.com/datasheet/toshiba/913.pdf
I am assuming I will have to manually calculate these for the KSA992 transistor.

I must say I can't make much sense either.
The equivalen input noise is 2.5uV (25mV -80dB)
The source impedance (100kohms) has a self-noise of 1.2uV over the 1kHz BW.
This must be deducted quadratically from 2.5uV, for a result of 2.2uV.
Since it's a BJT operating at 1mA collector current, one can assume noise is predominantly resulting from noise current.
So, 2.2uV into 100kohms is 0.22nA. Over the 1kHz BW, it computes at 6.9pA/sqrtHz.
That does not like like very low noise. Compare to an LT1115 (VLN opamp that operates at similar collector current), that achieves 1.2pA/sqrtHz.
So either my calculations are wrong or these specs are not convincing...
It's very unusual to pair a 100k source with a BJT operating at 1mA Ic.

iampoor1 said:

... How does the "Noise figure" in mv correspond to a noise figure specified in db?....

Click to expand...

http://www.ids-elektronik.de/421.pdf
page 3

moamps said:

http://www.ids-elektronik.de/421.pdf
page 3

Click to expand...

This datasheet is much more comprehensive than the previous one.
Graph shows In=0.75pA/sqrtHz at 1mA, resulting in 2.37uV into 100kohms with 1kHz BW. Not too far from my previously calculated 2.2uV.
Noise voltage contribution is negligible at ca. 63nV.
The combination of Ic=1mA and Rg=100k results in NF=ca. 6dB. I can't figure out why this has been chosen as typical data.
Noise figure map shows that optimum Ic for 100kohms source Z is around 20uA.

What Abbey said... the high beta (400-800 for E)  is a typical low noise device characteristic, but 100k termination is not typical for low noise bipolar device application.

JR