[hpsdr] ADC Question

[Glenn Thomas]

Hi Larry.

I appreciate your thought that a better ADC is likely to come with a 
DoD-sized price tag. Today at least. Would you agree that if something 
akin to Moore's law applies to ADC's, in 18 months or so we should have 
an ADC that is some linear combination of twice as fast, one bit wider 
and half the price of a now-current 3308-family ADC?

73 de Glenn wb6w

On 10/28/2012 4:11 PM, Larry Gadallah wrote:

***** High Performance Software Defined Radio Discussion List *****



(This is a long overdue response to Andrew's question)

My casual and somewhat disorganized research and readings regarding this 
lead me to a few conclusions:

1. People want the maximum dynamic range possible, to allow operation in 
crowded and/or high-signal environments
2. People want low noise and maximum sensitivity to allow detection and 
demodulation of very weak signals
3. The want all of the conveniences and "bells and whistles" that SDR allows

There is little doubt that current amateur radio SDR systems are already 
close to meeting items 1 and 3 above. I have my doubts about item 2, for 
the reason that Phil mentions: The ADCs that our DDC receivers use have 
a relatively high noise figure. Sure, you can add preamplification ahead 
of the ADC to help mitigate this, but that would degrade item 1 above.

What I have not found, which might put this question to bed, is an 
explanation of the limits of how good the noise figure of a real 100 
Msps ADC implementation can be. If we are close to the limit already, 
then perhaps this is a moot point. I suspect that there are some fab and 
materials issues that cause these kinds of limitations, so perhaps if 
some enterprising person designed a high speed ADC in GaAs or something 
like that, we could get better performance, but I'd imagine that it 
would come with a government/military price tag, not the cellular phone 
price tag that we are used to now :-)

As to the difference in performance between SDRs and analog receivers, 
my estimation is that the best HF analog receivers have a noise figure 
in the range of ~10 dB, and a dynamic range of about 90 dB. Our SDRs 
have much higher noise figures, in the 25-30 dB range, and a dynamic 
range of ~100 dB. I think the SDRs probably would outperform many analog 
receivers in high-signal situations, but I think in many cases the 
analog receivers would be better in low signal/low noise environments. 
It is a gross oversimplification, but the SDRs are better in the high 
end of the operating range, and the analog radios are better in the low 
end of the operating range.

I have also read a lot of comments from people who seem to look at the 
ITU atmospheric noise curves for suburban or rural areas and conclude 
that you don't need a lot of sensitivity or a low noise figure, 
especially at lower frequencies. However, I'd argue that top-band 
operators, MW DXers, and anyone who has experienced a beverage antenna 
might disagree and there are situations where an exceptionally low noise 
figure is a big benefit.


> Larry Gadallah, VE6VQ/W7 


 1351540278.0