[Elecraft] Analog vs. Digital Front Ends

[Joe Subich, W4TV]

On 9/15/2015 9:13 PM, Eric Swartz - WA6HHQ, Elecraft wrote:
> I think what everyone is missing here is that Adam has clearly
> stated that his tests were not designed to be tests to directly
> compare analog and direct sampling radios, and certainly they were
> not intended by Adam for direct comparisons to the tests run by the
> ARRL and Sherwood. They were originally intended as -additional- lab
> tools for technical people and those like us here at Elecraft and
> Flex, ICOM, Yaesu etc to use for evaluation of our designs going
> forward.

When one makes the same tests and reports the data in the same format
the reader is naturally going to compare the numbers of one type of
radio (upconversion or downconersion superhet with analog product
detector or DSP demodulator) with others (direct digital conversion
DSP).  It is up to the test designer to make those tests directly
comparable *or* very clearly state the difference and why the two can
not be compared directly.

There is certainly no indication of the differences/incompatibilities
in Adam's "Multi-Band NPR Test Results".  The difference in test
methodologies is documented only in "Noise Power Ratio (NPR) Testing
of HF Receivers" and even that document fails to clearly show the
differences and *why* the test results of traditional receivers can
not be compared directly with those of DDC SDRs.

Simply providing MDS for each test condition in each test would go a
long way to establishing comparability.  Identifying "Ptot dBm" and
"Noise Loading dBm" consistently across all receiver topologies
*along with* MDS would be even better to establishing a common set
of test parameters that would allow the various topologies to be
*compared directly* for the ability to copy a signal at the [internal]
noise floor in the presence of strong adjacent channel interference
which is really what *all* of the receiver comparison tests are about.

Real world performance is not about one parameter in a vacuum.

 > One other note - the 2.7 kHz SSB filter he used on his test of the K3
 > (without new synth) does not have as good of a stop band and shape
 > factor as our 2.8 kHz and narrower 8-pole SSB filters.

Actually the tests of K3 S/N 5812 are indicated as using the KSYN3A.

However, even comparisons of "analog" receivers are murky ... one needs
to look carefully to see than Noise Loading of the Icom 7800 and 7851
*drops* by almost the same amount as the preamp gain (just like a DDC)
where the noise loading of the K3 varies by only a small part of the
added gain.  I won't even point out that the 7800 beats the 7851 on
160 meters - another anomalous result.

73,

    ... Joe, W4TV


On 9/15/2015 9:13 PM, Eric Swartz - WA6HHQ, Elecraft wrote:
> I think what everyone is missing here is that Adam has clearly stated
> that his tests were not designed to be tests to directly compare analog
> and direct sampling radios, and certainly they were not intended by Adam
> for direct comparisons to the tests run by the ARRL and Sherwood. They
> were originally intended as -additional- lab tools for technical people
> and those like us here at Elecraft and Flex, ICOM, Yaesu etc to use for
> evaluation of our designs going forward. We can certainly take
> additional data in the lab to make these tests even more useful as
> comparisons between different technologies etc.
>
> Your comment about including the mds of the rig and the A/D limit level
> at the settings used for each of his tests will help a lot though in
> comparing the results. I'd also like to see that.
>
> One other note - the 2.7 kHz SSB filter he used on his test of the K3
> (without new synth) does not have as good of a stop band and shape
> factor as our 2.8 kHz and narrower 8-pole SSB filters. Thereis a big
> difference between thge two for this type of testing.  I'd like to see
> the noise power test data for a K3 (or K3S) with new synth and an 8 pole
> 2.8 kHz or narrower filter. We did surprisingly well with the 2.7 kHz
> 5-pole in his tests. I know what he will see with the 8-pole filter and
> new low noise Synth. :-)
>
> 73,
> Eric
> /elecraft.com/
>