[hpsdr] Preselectors

[Graham Haddock]

Frank:
Comments below...

FRANCIS CARCIA wrote:
> The Cubic claims it can change frequency in 10 ms. The problem 
> with relays is transients generated as the frequency changes due to 
> relays making and breaking. Even solid state switches could cause 
> transients during wide frequency sweeps.
Yes, any tuning modification to a tuned circuit will generate some kind 
of transient.
What are you trying to do? 
What is it that you are worried about?
Is a ten ms. transient when you switch filters, or settings inside of a 
filter, unacceptable?

If you are using Mercury for a spectrum analyzer, then I recommend that 
you turn
off the preselector and the Mercury's preamp.  In that configuration, 
the Mercury
has an IP3 of +50 dBm and a noise figure of somewhere around 20 dB.

If you are doing spectral analysis requiring the preselector and preamp, the
program will need to pause and throw away data for a few milliseconds, every
time you cross an internal tuning boundary inside the preselector.  I 
think the
preselector can be designed so as to have no tuning boundaries
inside a given ham band, other than perhaps 80 meters.

> Discrete filters,
>  A standard board could be built with two different applications 
> depending on parts installed. One configuration for Ham band only and 
> the other for the whole spectrum.
> I've seen a number of SDR posts in different places where guys build 
> brick wall preselectors to improve performance.
> I think the preselector should be in a shielded box away from digital 
> noise or I'm about to leard some cool PCB layout skills from a master. fc
It might be a shielded card that plugs into the Atlas bus, or a shielded 
box that
mounts elsewhere.  We are not that far along yet.  I am still trying to 
fit it into
the allotted space for an Atlas bus card.  The larger inductors we are 
considering
are not self shielding like a Toko can.  Toroids still have some 
external field.
No PCB layout will fix that.

--- Graham

>  
>   


 1173803389.0