Topband: Norton preamp

[Tom Rauch]

> I've been following this thread and took a look at Dallas 
> Lankford's
> article.  It looks interesting.  And although "...he seems 
> to be very
> meticulous and conservative," he also boast of the magical 
> 2N5109 preamp as
> having an IP2 of +80dBm to +100dBm based on single versus 
> P-P design (top of
> page 7).  I can absolutely, positively, beyond any 
> reasonable doubt, assert
> that this is an absolute impossibility.  That's so 
> fantastic I had to
> re-read it an actually ponder how big a signal that really 
> is...
>
> 1W is +30dBm
> 1KW is +60dBm
> 1MW is +90dBm
>
> I can absolutely assert that anything looking like a 
> 'circuit' in the signal
> path of 10 MegaWatts of RF will absolutely vaporize in a 
> spectacular blue
> flame instantly.

Hi Ford,

IP measurement points are theoretical points where the input 
signal would generate an undesired product that is equal to 
the good signal. That can, of course, really never happen in 
real life. That power level is never obtained or required in 
a test!

But back to the real world is good.

I can obtain a repeatable reliable +45 to +47dBm OUTPUT 
third order intercept with a pair of 5109's (and nothing 
close to that with the other transistor mentioned). Since I 
can set the gain over a wide range with feedback or 
attenuation the input TOI can be anything up to that value 
(unity gain). Going above that requires some pretty hefty RF 
power transistors, but I have built things in the +50dBm 
output TOI range.

Second or even order intercept is greatly affected by 
balance in a push-pull amp since one transistor nulls the 
distortion by the other. You can get some very fantastic 2nd 
order intercept numbers in a carefully balanced circuit, but 
they don't really mean a thing to us. Even order problems 
can be easily filtered, it's the odd order products we are 
stuck with.

All this (and a ~1dB better noise figure from "noiseless 
feedback") is largely needless and useless on lower bands 
because even with 3000 feet of F-11 (RG-11 size) coax, noise 
propagated in by skywave from distant sources sets the 
system noise floor (even on 40 meters).....and that is with 
all the amplifiers located in the house in my quiet rural 
location.

I find the very high TOI  of 40-50dBm useful in my 
application but the rest just makes for meaningless 
conversation.  If you are in a noise limited urban 
environment even a 20dBm TOI might be good enough, with 
proper input filtering. Second order intercept, once a 
reasonable value, is pretty much meaningless beyond that 
point.

73 Tom