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

[Ahti Aintila]

On 27/07/06, Philip Covington <p.covington at gmail.com> wrote:

> 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
>
Phil,

You are talking so nicely about the active integrator, so I have no
alternative  but admit. Being a simple and silly person myself, I love
KISS. Here is my latest simple schematic
diagram:<http://kotisivu.dnainternet.net/ahti/sdr-1000/newmix4.pdf>

I figured out that the OPA1632 is a combination of two inverting
amplifiers, so why not they could not work as two integrators.
Unfortunately, I am still waiting for the ordered components to come,
so no tests made yet. This is just a paper exercise. If this idea
works the QSD (and QSE likewisw) can be put to a really small area.

By the way, I like isolating tranformers to separate the dirty digital
world from the sensitive analog circuits.

What do you say, and what am I missing this time?

73, Ahti OH2RZ

 1154016588.0