[hpsdr] Star-10 Transceiver article in QEX

[John Miles]

Hi, Constantin --

>I do now if you remember me or not, I bought your
>board for your sysnt published in QEX to test it.
>However I wrote to you about another topic

Sure!  Good to hear from you.  I'll cc: your message to the list since that
appears to have been your intent.  (You'll need to subscribe if you'd like
subsequent traffic to appear on the reflector without being copied by
someone else.)

>. I reed your comment about Star 10 radio on internet.
>I am sorry to inform that you are wrong in your assumptions.
>First I build and tested the boards and radio in my lab
>working with Cornel. In fact are 2 star 10 , one is
>Cornel's, and another is mine with slights differences.
>There are  not spurs to be heard in AM radio from xtal
>of CPU as you suggested . You apparently do not know how radio is build.

RFI doesn't come from CPU crystal, but from the CPU and its I/O connections
themselves.  Great to hear it's clean in the "AM pocket radio test", but
that has nothing to do with the 32 MHz clock frequency that Cornell
mentioned as the reason.  Not everything in a CPU happens at one edge per
clock.

Cornell's use of a microcontroller, rather than a CPU with external memory
and peripherals, is a major advantage at avoiding RFI.  The use of a part
that happens to be clocked above the HF band is not meaningful at all.

>The dynamic range is as published. I invite you to measure it .
>The test was done using last ARRL procedure -110 dbm on ch
>of interes and a strong signal @2,5,10,20,100khz  ,
>using FFT S/A to monitor spurs, -110 dbm sig.

It's unlikely that a +40 dBm signal at those offsets would fail to yield
spurs and increased noise due to reciprocal mixing (not to mention a smoked
front end).  Yet that is what you imply when you speak of 150 dB of dynamic
range.

I will confess that I am not familiar with the ARRL's officially-sanctioned
measurement procedure.  It's not a useful procedure if it yields
dynamic-range figures like 150 dB.  An error in measurement or communication
seems more likely.

>For you info I sent you a few picture with my lab, my
>self assembling radio, the plot of sysnt used as lo.
>The synt plot consist of 2 picture , one only synt
>500 hz - 20 khz @-137dbc/hz and second photo wiyh
>radio  synt lower trace on S/A and upper trace is
>pll noise of a new Agilent sig gen .Freq. of plots are
>89.2 MHZ, , on the dial 14,2 MHz. I am very familiar
>with your test of sysnt noise published on internet

Can you provide a few details on how your phase-noise measurement setup
works?  Obviously, you can't measure noise at levels anywhere near -137
dBc/Hz on a conventional spectrum analyzer, without the use of additional
hardware and calibration steps to suppress the carrier and amplify the
residual noise.

>So making comments without knowing what is about
>is not in Ham spirit...

Very true.  I'm sure that many of the points being debated come down to
differences in test methodologies.  It certainly does look like a very nice
HB rig, at any rate.

-- john, KE5FX