At 07:23 PM 12/22/2005, David Queen wrote:
Part of the reason the SDR1000 is so good with strong close in signals has
got to be the very low phase noise and that fact you only have on RF
source of phase noise.
That got me to thinking, has anyone compared the difference close in phase
noise at the Tayloe Detector using the AD9854 with its 12 bit ADC and the
AD9951 (14 bit adc) AD9951 14 bit adc and a Johnson counter? It would seem
that there might be some advatage to a single filter chain and conversion
back to digital. I fully admit this is just a guess having no equipment to
check it.
The whole sidebands, phase noise, spur issue with DDSes, particularly into
a mixer, is pretty complex. There's a good analysis of the DDS spur issue
from Martin Pechanec
http://www.geocities.com/CapeCanaveral/5611/dds.html
However, Martin doesn't consider the effects of driving the QSD/QSE, nor
the effects of the low pass filter in the SDR1000 (which, being something
other than a gaussian, has a "non linear" phase vs frequency
characteristic). With these sorts of detectors/modulators, the timing of
the zerocrossings is moderately important. (otherwise, you could just feed
the MSB into the logic and be done with it.. leaving aside the DAC and
filters).
Jim W6RMK