[Elecraft] K3 Design

[Don Wilhelm]

Don,

You bring up some valid considerations.

When I first got into ham radio back in the mid-50s, the big selling 
points for any receiver were sensitivity and selectivity.  Not much 
mention of IMD, dynamic range, signal to noise ratio, etc.  The 
evolution of ham receivers regarding the number of IF conversions is 
quite interesting to me.

As of the mid-1950s, the best selling point for a receiver seemed to me 
to be the number of IF conversions.  If it was not a triple conversion 
superhet, it did not get much attention unless the brand was Colins - 
that was the only practical way to provide selectivity at a reasonable 
cost.  Yes, the Collins mechanical filter provided selectivity at an IF 
of 455 kHz, but many receivers converted down to 50 to 100 kHz because 
sharp LC tuned IF stages were possible at that low frequency.  I myself 
bought into that camp and being a ham without funds to buy the latest 
and greatest, I built my own version of the HBR-16 receiver using 85 kHz 
IF transformers salvaged from the BC-453 surplus receivers.  It used 
plug-in coils for band changes (and yes I built general coverage coils), 
but it did provide a good selective receiver - I even had a Q-Multiplier 
in it operating at 85 kHz for super selectivity when needed - it worked 
great, but I have no idea what the IMD and Dynamic Range numbers were 
for it - they were likely poor by today's standards.

Then there was an article in QST that caught my attention - asking 'What 
is Wrong with Our Present Receivers.  That article (among other things) 
seemed to indicate that the multi-conversion receivers were NOT the best 
way to go.  In fact, if I can generalize, the receiver with the fewest 
number of mixers will perform better - given all other conditions remain 
the same - simply because of the reduced number of mixer products.

While many ham transceivers used the technique of converting to a high 
IF to enable general coverage and then down-converting from there, the 
K2 (and a few others) came along with a single conversion design.  The 
K2 performance showed that a single conversion design could perform very 
well in a ham-band only receiver.  The stage by stage gain and noise is 
controlled to provide a good dynamic range and IMD measurements.

The K3 adds two things to the K2 philosophy (in terms of IF conversions) 
and that is to convert to 15 kHz rather than direct to audio to allow 
the DSP to 'do its thing' at that 15 kHz input; and general coverage has 
been added, but not by upconverting to 70 MHz and then down again, but 
by providing a local oscillator that can tune almost continuously along 
with a set of 'general coverage' bandpass filters.  Certainly the 
performance numbers outside the ham bands are not likely to be as good 
as the ham-band only numbers (the bandpass filters are wider), but it is 
quite possible to achieve general coverage without adding conversion 
(mixer) stages - of course the receive capability at the IF frequency is 
a slight problem, so there is a gap in general coverage near the IF 
frequency.

All in all, I believe this is the 'best of both worlds' - the number of 
IF conversions is minimized and general coverage is available with only 
a small band where full coverage is not possible.

73,
Don W3FPR

Don Rasmussen wrote:

> 
> The numbers seemed to prove that K2, OMNI VI+, Drake
> R4C, Corsair II, all ham band only have the best
> numbers as compared to the very best broadband designs
> - most often from Japan.
>