[hpsdr] Selectivity design/measurement

[Marco IK1ODO]

At 20.10 19/02/2007, Ahti wrote:

> > With no post detector amplification needed, it implies that the detector
> > should have very, very short audio connections to the A/D converter.
> >
> > - Dan, N7VE
>
>Yes Dan,
>
>and possibly the sampling switches and capacitors should be assembled
>as a small daughter board on the Janus module!
>
>73, Ahti OH2RZ

Good idea, but take care to stop QSD switching transients before the A/D input.

About the blocking dynamic range: isn't it simpler to take the A/D 
noise reading in dB referred to full scale, as Winrad does?
Then you take the Winrad's filter equivalent noise bandwidth (1.4 
times the displayed "Resolution" field") and calculate the S/N on the 
desired bandwidth.

I did that way to compare different sound cards, and it worked well. 
Easy math, and no external generator needed - more, S/N measurements 
matched that done with an HP 3400A (true RMS meter). I used Winrad 
with averaging set to 24 and 2dB/div vertical scale to take readings. 
I don't know if Winrad already works with the current HPSDR board 
set, but other programs may do the same, I think.

I would be VERY interested to see an audio spectrum at 192kHz 
sampling rate, vertical scale calibrated in dBfs.

73 - Marco IK1ODO


 1171921846.0