[Elecraft] K3 Harmonic Distortion

[Jim Brown]

On Tue, 02 Sep 2008 17:06:29 +0100, Brendan Minish wrote:

>I used an old copy of SMAART pro that I have from my Sound 
engineering
>days. 

I have that software, and have long used it to analyze ham filters. 

>the sound card is a delta44 

I don't know this card, but it's own noise floor (or input stage) may 
be the limiting factor. 

Look at the first data set in 

http://audiosystemsgroup.com/K3FilterStudy.pdf

The black curve (the Headphone output) and the red curve (Lin Out) are 
set to produce the SAME output level, but the harmonic distortion is 
55 dB stronger at the Line Output!  The difference between the LF 
traces is IM, and is also significantly greater at the Headphone 
output. Notice also the dynamic range of this measurement, which is 90 
dB. The excitation is band noise, so there's another 10 dB or so 
between the top of the screen and digital clip. 

My measurement system is EASERA SysTune (considerably more advanced 
than Smaart Pro), and the input device is the EASRA Gateway, which has 
a 24-bit A/D and Firewire interface. I'm fairly confident that the 
noise floor in my measurements is the radio, not the measurement 
system!  

That first data screen shows 3rd harmonic at -55 dB for a gain setting 
of 1 (and that is the noise floor of the radio). At a gain setting of 
2, it's -54 dB. At a gain setting of 3, it's -49 dB. At a gain setting 
of 5, it's -40 dB. For a setting of 10, it's -28 dB, and at 20, it's -
29 dB. 

Note also the very significant difference in excitation. My test 
signal is not a CW signal, but broadband noise from an antenna, band-
limited by the IF filters. Your test signal is a sine wave. My test 
signal is somewhat representative of trying to copy a signal in QRN, 
or pick one signal out of many in a contest. Your signal is 
representative of a test bench or a signal with little or no noise on 
it. As I'm sure you know, those of us working in pro audio find pink 
noise far more useful than sine waves help us hear real world problems 
in systems.  

73,

Jim Brown K9YC