## Theory of the Softrock

Expand Messages
• Ok, so I ve been trying to work through the math of the softrock receiver in my quest to build what I discussed in a previous post (multi-band
Message 1 of 2 , Oct 1, 2006
• 0 Attachment
Ok, so I've been trying to work through the math of the softrock
receiver in my quest to build what I discussed in a previous post
document "Software defined radio for the masses" a couple of times,
(admitting that I have stopped short of getting the pencil to the
paper, probably the best option to completely understand this) and
just need a few things clarified. My understanding is that its the
system impedance (50ohms) in conjunction with the shunt cap(s) that
determines the bandwidth of the reciever, and the xtal determines
your center frequency. Can you then, theoretically, open up the BPF
on the front end as large as you want? There has to be some
limiting factor or else the softrock design wouldn't have included
the BPF in the design. The ultimate bandwidth of the receiver is
limited by your soundcard. So, my question is, what is the limiting
factor of covering a wider bandwidth with a frequency synthesizer of
some sort? Please let me know if any of my thoughts on the
innerworkings of the softrock are incorrect. I'm just trying to
figure out if I need to have a mixer stage with IF or if I can just
use a wider front end filter with a synthesizer to realize my design.

Thanks, and sorry for my inexperience
Kyle
KC9JJL
• ... You have to have a filter because the detector will respond to odd harmonics of the LO, that is why for some frequencies we have been discussing dividing
Message 2 of 2 , Oct 1, 2006
• 0 Attachment
racerxgt2 wrote:
> Ok, so I've been trying to work through the math of the softrock
> receiver in my quest to build what I discussed in a previous post
> document "Software defined radio for the masses" a couple of times,
> (admitting that I have stopped short of getting the pencil to the
> paper, probably the best option to completely understand this) and
> just need a few things clarified. My understanding is that its the
> system impedance (50ohms) in conjunction with the shunt cap(s) that
> determines the bandwidth of the reciever, and the xtal determines
> your center frequency. Can you then, theoretically, open up the BPF
> on the front end as large as you want? There has to be some
> limiting factor or else the softrock design wouldn't have included
> the BPF in the design. The ultimate bandwidth of the receiver is
> limited by your soundcard. So, my question is, what is the limiting
> factor of covering a wider bandwidth with a frequency synthesizer of
> some sort? Please let me know if any of my thoughts on the
> innerworkings of the softrock are incorrect. I'm just trying to
> figure out if I need to have a mixer stage with IF or if I can just
> use a wider front end filter with a synthesizer to realize my design.
>
> Thanks, and sorry for my inexperience
> Kyle
> KC9JJL
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
You have to have a filter because the detector will respond to odd
harmonics of the LO, that is why for some frequencies we have been
discussing dividing the crystal frequency by 3 to use a crystal that is
easier to find, it's generally called sub-harmonic sampling.

The bandwidth is determined by the sound card sampling rate, within
reason, the cap size, the op-amp that follows need to have the proper
frequency response. People have successfully used the SoftRock with
48KHz, 96KHz, and 192KHz. The circuitry is capable of it, the limiting
factor is the sound card, of course you must have a PC capable of
handling such a flood of data.

A AD9851 DDS is capable with a 180MHz reference of outputting 60MHz and
it's inexpensive.

AT the higher frequencies the limiting factor can become the switches
due to leakage and unsymmetrical turn on/off times, that is seen as
increased noise, at some point the noise is the limit, that problem can
be helped with a pre-amp.

However you are not limited to just using switches as the mixer, there
are a multitude of mixer technologies that will also work, for example
you could use a Diode Ring Mixer, specially the balanced ones can be
good into the GHz range and are not expensive, the cost is that they
need a lot more power than a FET switch mixer. Gilbert Cell mixers can
be used as well as many others.

Hope it helps.
--

Cecil
KD5NWA