[hpsdr] Proposal for a 'low-end' RX/TX board for HPSDR

[Philip Covington]

On 7/26/06, Ahti Aintila <oh2rz.sdr at gmail.com> wrote:
> On 27/07/06, Philip Covington <p.covington at gmail.com> wrote:
> > Here is another one to look at:
> >
> > <http://www.philcovington.com/HPSDR/TEMP_STUFF/isd_07262006-1.pdf>
> >
> > I went slightly different and used a SN74AUC2G53.
> >
> > 73 de Phil N8VB
> >
>
> Phil,
>
> Good idea to use that SPDT switch. No need for inverted pulses. I may
> modify my schematic for that component, but not want to go to the
> active integrators. Too many components!
>
> 73, Ahti OH2RZ

The active integrators are an attempt to provide a reasonably well
defined range of impedance to terminate the BPF filters both in and
out of band.  I also wanted to remove the relatively large capacitive
load on the switch by the sampling capacitors in the QSD.

In the SDR-1000 QSD the sampling capacitors look like a short well
outside the bandwidth of the QSD.  The series resistors before the
switches are an attempt to provide a defined impedance looking into
the QSD.  This terminates the preceding BFP properly but only out of
band.  Within the bandwidth of the QSD, the sampling capacitor's
reactance increases to infinity the closer you get to the sampling
frequency (or DC).  The BPF does not have a well defined impedance
looking into the QSD in band - it is actually very very high due to
the input impedance of the instrumentation amp ( I think it is around
60 Mohm).

I did some experimentation with various analog switches with various
capacitive loads to ground on one side of the switch.  Each different
switch and even the same switch within the same family or part number
behaved differently in the tests.  I could find no predictable
response when the switches were operated at > 10 MHz and the
capacitive loads were varied within the values needed for a 48-192 kHz
BW QSD... it depended on the particular (physical) switch used.  This
is why I wanted to avoid the sampling capacitors hanging off of the
switches.

The fully differential amp in my circuit can be replaced by the OPA
part the Bob uses... the AD8139 is probably overkill.

73 de Phil N8VB

 1154006536.0