Topband: Norton preamp

[Terry Conboy]

At 08:49 AM 2007-01-22, Ford Peterson wrote:
>Just to be clear, when I evaluate a given preamp using my spectrum analyzer
>instead of a slide rule, I can quickly identify the gain function of the
>amp.  For discussion, let's say that gain is 12dB.  As I pump more RF into
>the amp, I can see that the output is always 12dB higher, until I reach the
>1dB compression point.  This is the point at which I don't get 12dB, I only
>get 11dB.  Simultaneously, I can visually see distortion products (harmonics
>and intermod) developing.  As I push even more RF into the front end, I can
>see the IMD develop to the point where any additional RF in produces no
>additional desired output, and serves to only develop yet more IMD.  This is
>the point I have always concluded to be the TOI of the amp.

Ford,

This may seem like an intuitive way to estimate TOI, but it most 
likely isn't related to the TOI at all.

For example, an amplifier that will produce no more than 0 dBm output 
no matter how hard you drive it, could easily have a TOI of +40 dBm, 
if it has very high levels of negative feedback.  Remember that over 
the linear range of an amplifier, the 3rd-order IM products increase 
3 dB for each 1 dB increase in input level, so all a high TOI implies 
is that the IM levels with normal signals are extremely low.

When I was a green engineer, with a freshly printed BS, my first real 
job was designing mixers and amplifiers for analog microwave systems 
for telephony applications.  At the time, these systems were the 
workhorse of the long distance carriers, but since there was no 
regeneration as in current digital systems, it required extreme 
control of IM products to prevent them from growing to unmanageable 
levels over a multitude of radio "hops".  Some of the amps had 
bandwidths to carry up to 1200 or more channels, each 4 kHz wide.  To 
simulate traffic loads, we used wideband noise that covered the 
entire band then looked at the IM products that fell into slots 
notched into the noise at the top, middle and bottom of the 
band.  The power levels were modest but the linearity requirements 
were very hard to meet.

73, Terry N6RY