[hpsdr] preselector control

[Eric Ellison]

Hi Henry/Graham

I agree. Too early to call.  Graham you might consider calling a
'brainstoming session' on Teamspeak. Phil, Bill, Chris and others are there
on Friday's at 02:00. Would be fun to listen in! 

These were my thoughts, like Henry's in the Pandora thread with Ben.

> My feeling about Alex is it should be a 3 way.

> Put a DIN connector on it, to mount on Atlas, pick up power and control
from the buss.

> Also put external connections i2c connection, reset, ground, JTAG and
> whatever else is needed to control it from OZY. External power connection
from a standard PC power Molex 12v-G-G-5v.


> Also make it 'stand alone for any receiver project going on out there,
> outside SDR.


> Make it modular with plug in filter sections in any case.

******************************************************
That would be very hot!  Something that could be used outside HPSDR 
would certainly increase interest!
*******************************************************

Every time I see a project proposed in HPSDR my first thought is can we get
an equivalent 'off the shelf'. Second is can we make this project board
appeal to folks outside just Ham Radio. Third is is there any 'duplication'
of parts existing on other project boards (FPGA on Ozy for example). In the
case of Alex, almost any receiver/tranciever project in QST, RadCom, QEX or
other could benefit from an 'off the shelf' or easily modified high quality
filter board. Especially if it can take a few watts of exciter power. Then
instead of selling 300-500 highly customized (HPSDR) filter boards we are
into selling potentially thousands. I think ATLAS - OZY and possibly an EPI
board are in that category already if we hit the Robotics or FPGA groups.


Hi Henry:

Well, we are not far enough along to split the project yet.

Initial work will be focused on verifying that the topology we are
considering can be made to work over the frequency range, and whether we
can use solid state switches and still make the filter performance,
or whether we need to go to relays.

Packaging and control decisions will come later, after we know how big it
will be, and how sensitive the inductors are to picking up local fields.

By going from a ten section fixed tuned preselector to a continuously
tuned preselector, we have gone from something where a simple
SPI bus was fine, and a small CPLD was overkill, to needing
a small CPU or a large CPLD and a fairly large lookup table,
or some clever algorithms to train and manage it.

--- Graham




 1173913400.0