[Premium-Rx] The new R&S EM510 Direct-Sampling HF Receiver

[Pieter Ibelings]

This is the biggest misconception on ADCs. When your ADC is 16 bits, the 
96dB number means that if you insert a signal at full scale (0dBfs), the 
integrated noise over half the sample rate will be 96 dB down. In other 
words, if you inject a signal of say 10dBm (assuming full scale is 10dBm) 
and you are sampling at 100 MHz, then the noise will be broadband in nature 
and will be -86dBm (96dB down) spread over the band of 0-50 MHz. If you set 
the display to have a 1Hz resolution bandwidth, you will see the noise floor 
at -86dBm-10*log(50MHz) = -163 dBm. This is equivalent to a noise figure of 
the ADC of 11 dB. In this setup, I would be able to insert a carrier and see 
it from about -163 dBm all the way to 10dBm on the display. So dynamic range 
is not equal to S/N of an ADC which is calculated from the Effective Number 
of bits (ENOB) multiplied by 6.02dB +1.76dB. The signal to noise numbers 
above assumes that all of the noice is random in nature and caused by 
quantization noise in the ADC.

Pieter, N4IP

Here is a link to an explanation:
http://www.analog.com/en/content/0,2886,760%255F%255F88014,00.html







>
>You made an excellent point. The basic problem with the direct-sampling
>approach is the ADC headroom. A 16-bit ADC, no matter how fast, has 16*6 =
>96 dB theoretical dynamic range. (I understand that some device
>manufacturers are claiming SFDR > 96 dB, but am not sure how that is
>accomplished.)

_________________________________________________________________
Find a local pizza place, movie theater, and more….then map the best route! 
http://maps.live.com/?icid=hmtag1&FORM=MGAC01