I suppose it would be easy to put an analog loop around the front end that
reduces the gain as the saturation point of the A/D is approached. A variable
gain amplifier or passive pin attenuator would work well. Another approach
would to use the computer to generate a D/A output based on the noise floor of
the frequency of interest and maximum signal coming into the A/D. Flex does
this with an amplifier and a pad or none to get 3 input levels so the dynamic
range can be shifted in 3 10 dB steps by the operator. Frank WA1GFZ
Andreas Troschka <[EMAIL PROTECTED]> wrote: I agree with you, JR,
please note that the characteristics of a RF direct sampling ADC scheme circuit
is actually far from being comparable to a commercial high-end (narrow-band)
architecture with say typical 120dB dynamic range or better.
Else the best ADC actually obtainable (on the consumer market) cannot compete
with the more traditional architectures. Yet!
For this reason some AGC must be introduced.
There are at least a couple of solutions to look at.
The actual tendency of ADCs manufacturers is to include in the top-level ADCs
(e.g. the Linear Technology LTC2208/09) a PGA (Programmable Gain Amplifier) as
input amp.
This is useful to give a degree of two step AGC due to the fact there are two
selectable amplification factors available.
Another possibility is the use of an external ADC driver or a VGA (Variable
Gain Amplifier) which are mostly working in the frequency range of the ADCs
used.
Luckily, high performance instrumentation amplifiers are offered by the same
manufacturers and being perfectly matching to their ADCs inputs.
Of course there is a third solution using a more classical amp/attenuator
configuration but this may require more effort to be correctly realised for a
wide bandwidth architecture like this.
The interesting thing is that in all the solutions both the AGC detection as
the generation of the control signals for the ADCs input amp/driver can be done
by the use of software at a following stage using DSP (instead of leaving it in
the hands of some traditional static hardware based algorithms).
The DSP solution offers the opportunity to develop adaptative and dynamic
algorithms based on the particular setup of the SDR's filters, of the signal,
the unwanted-signal, the mode and the actual band noise.
In other words, applying such software autoadaptative algorithms, very high
performances can be reached, where the AGC strongly depends from the received
signals characteristics, process unthinkable with traditional architectures.
One critical point to have really care of is the regulation of the RF input
level to let the ADC to work beneith its critical upper limit.
Therefore, with use of a RF front-end (either if a PBF as if more complex
circuits), adaptation of the source impedance to the ADC's input requirements
has to be studied scrupulously as it is not as simple as it may seem.
Variable or even wrong impedance adaptation results easily in significantly
higher even order harmonics (as degraded SFDR, Spurious Free Dynamic Range).
Actually I'm studying a working solution using the ADC direct sampling
architecture else if I'm aware that this way brings no astonishing results in
terms of final dynamic range (so long).
The reason I'm working anyway on this is the fact that in some way it makes
sense waiting the industry to develop more performant ADCs and the future will
than payback the effort. ;-)
There are more very interesting things to discuss about the engineering of the
scheme mentioned but this may be a subject for another thread.
vy 73s de ik2wqi - Andreas
jr_dakota wrote:
>As long as your front end and AD converter have enough dynamic range,
> there is really no need for AGC and simple amp somewhere in the
> frontend chain can increase the front end range enough .... I had
> thought of using a good audio compressor/limiter circuit after the QSD
> but when you start running the numbers you'll find you screw up the
> blocking dynamic range. Large signals in the 48K bandwidth will
> decrease the gain and render weak signals inaudible ... traditionally
> this is handled by a strong front end and a narrow IF filter to block
> out the strong station but that defeats our purpose here of grabbing a
> fairly wide chunk of the RF spectrum
>
> Better to go with a 2 tier (Maybe 3) gain structure using software AGC
> and let the software filters get rid of the strong signal (Both are
> done better in software than hardware today)
>
> JR
[Non-text portions of this message have been removed]
------------------------ Yahoo! Groups Sponsor --------------------~-->
Yahoo! Groups gets a make over. See the new email design.
http://us.click.yahoo.com/XISQkA/lOaOAA/yQLSAA/ELTolB/TM
--------------------------------------------------------------------~->
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/soft_radio/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
