[Elecraft] K3 & ADAT ADT-200A by HB9CBU

[Kok Chen]

On Dec 4, 2008, at 1:06 PM, Bill Tippett wrote:

> This looks like a much better system than Flex's QSD.

For single signal use, I agree completely Bill.  By the time you've  
matched the sampling rate of a direct SDR to a signal with 500 Hz  
bandwidth, you have adequate dynamic range. (And not need the  
expensive Asahi codec that is used in the Flex 5000, to boot.)

However, the QSD based SDRs still has a dynamic range advantage when  
you want to demodulate (think CW skimmer for RTTY) multiple signals in  
a wider bandwidth, say 50 kHz.

With added complexity, you can of course implement multi-rate filter  
banks with direct SDR and stitch the result back together to create a  
wider bandpass.  Toby had earlier suggested multiple direct SDRs, and  
polyphase and multi-rate decimation is just one way to do it (i.e.,  
use the same RF chip, the rest is "just software" HI).  We can leap  
frog one another until the cows come home.

Moreover, that is not to say that the QSD guys can't put a "roofing"  
filter between their quadrature mixer and the A/D converters (just  
that they are not doing it today) to get even better BDR than they are  
getting today.

This QSD "roofing filter" would just be a matched low-IMD LPF pair for  
the I and Q channels.  A 500 Hz "roofing filter" would reduce unwanted  
large signals from clipping the A/D converter and can raise the 2 kc  
separation BDR of the 500 Hz receiving system to something that is  
just a function of the dynamic range of the QSD itself.

It is funny to imagine a "roofing filter" that is centered at DC, but  
its function is no different from the crystal roofing filters in  
superhets.  Instead of preventing out of band signals from reaching  
the I.F. amplifiers, you prevent them from reaching the A/D  
converter :-).

73
Chen, W7AY