[Elecraft] K3: 2.1 vs 1.8

[Ron D'Eau Claire]

Doug, KR2Q, presented a superb explanation of why one needn't sweat too much
over the "roofing" (first I.F.) filter in the K3.  

I'm enough of an O.T. to remember the famous 1950's articles in QST "What's
Wrong With Our Present Receivers?" by Byron Goodman, W1DX, who pointed out
the astonishing idea that double and triple conversion receivers were not
necessarily better than a simple, single-conversion designs. Virtually all
"high end" receivers of the day used double conversion formats (like the
K3). Some even used triple conversion designs. In spite of what 'everyone
knew', those behemoths of the airwaves weren't as good as a simple
single-conversion design, thanks to the limitations of the circuits and
active devices available in those days. 

Those very expensive, top-end receivers offered astonishing features and
flexibility, but at a huge cost in raw ability to hear signals. 

Today engineers have vastly better devices to use in designing receivers,
and they allow high-performance multiple conversion designs with many
features and flexibility, if the receivers are designed carefully with
attention to having as few "birdies" and other spurs as possible and with
very careful management of the gain distribution throughout the design. 

However, multiple conversion receivers are far more complicated than a
single-conversion design. The more complicated something becomes, the fewer
people really understand the trade-offs involved. It gets especially
difficult to give up some basic "truths" we've lived with for many years
that have stood the test of time. One of those is that all the selectivity
in a receiver is best taken as close to the antenna as possible, so the
first i.f. (so-called 'roofing') filter should set the overall selectivity. 

In principle, that is true: we should strive to take all the selectivity
possible as close to the antenna as possible. But, in practice, it's not
really needed in a well-designed receiver using modern devices under almost
any real-world operating conditions. Taking the selectivity in stages, such
as with a roofing filter followed by a second i.f. variable filter (such as
the DSP filter in the K3), offers the chance to have features and
flexibility we can't get otherwise. 

Like all designs, it's compromise. There is *no* perfect design, no perfect
receiver. 

Engineering is *always* about compromises. 

Ron AC7AC