[R-390] Receiver noise floor needs

[Perry Sandeen]

List,
Wrote: "Noise floor measurements seem to me to be of uncertain value in an RFenvironment where atmospheric noise in "S-6" and above, and anyreceiver induced noise is well below."
That's probably true. And there are some urban areas where HF SW is a total bust.
Now for the BUT (A debate that will never end)
I spent 18 months on 2nd (Swing 3 pm to 11 PM and 3rd shift (Mids 11 PM to 7 AM
Karmusel Air Station was an intercept base about 100 miles west of Istanbul, Turkey. Besides radio interception the base was a HF radio link between the eastern boarder of Turkey and ENT AFB (USAFSS) in Colorado Springs, CO. Out send power was DISB with three 30KW SSB linear amps 24x7. 
Our part of the link was to receive from eastern Turkey to Croughten (SP) AFB in England.  They intern re-broadcast HF to Ft. George Mead in Maryland,USA.
The 2nd and 3rd shift job was standby maintenance (hardly any) and to maintain reception from the 3 bay Hugh's receiver and the R390A's. 
Now to my point - at last.
Mother nature does what she wants, when she wants, and as long as she wants.
Most nites, we easily received rock and roll AM pirate radio ships off the English coast and would feed the audio to the intercept operators and of course our receiver site.
We had a spare "A" and a rtty converter and teletype printer set up so mainly on the nite shifts we would copy AP, UPI, Reuters, what ever.  We went through cases and cases of yellow copy paper.
On several occasions I could not hear the audio shift side tone of the receiver but the rtty printer kept pumping out plain text. 
Many nites the bands were so quiet we would easily pick up WWV on 10 and 15 MHz (17KM sky-wave I was told) as well as foreign SW broadcasts all over the place.
For this, having a low noise figure was useful.
Noise floor numbers aren't the end-all for receivers but they are far easier to measure than IP3, receiver stability drift and audio distortion to name a few.  That's why most people will try to measure it.
The problem is that unless one has a HP, Fluke (I have both) or other high end signal generators and-even-then, true sub-micro-volt numbers are hard to measure with repeated accuracy.
The noise floor "numbers" are averages but omit the fact that ionosphere can and at times much, much quieter.
An indirect proof.  Look at all the receiver specs of the modern transceivers on the market. Making the receivers that good costs money.  If it didn't add value they wouldn't do it.
So one makes choices that are case specific or "How good can I make this sucker, cost be dammed".  My personal choice is the latter.
Whatever floats your boat.
Regards,
Perrier