[hpsdr] VHF/UHF downconverter project for HPSDR

[Ante Vukorepa]

Hi, folks!

I've been lurking about for quite a while now and finally took the plunge last week and ordered a basic RX HPSDR set of boards from Gerd (Atlas, Mercury-EU and Ozy_mini, which i hope to expand with Metis later on). Now, i've always been mostly a listener (even on boards and mailing lists such as this - yes, i know, pretty passive and lazy of me), so receivers are my primary interest...

The thing is, i'd like to expand my HPSDR at some point to cover VHF and UHF as well (up to 70cm), but, alas, there are no convenient, broad-coverage and - most importantly - elegant enough solutions that would enable me to do so. So i've been thinking...

(bear with me, all of this will probably sound naive, but as far as HAM is concerned, i am pretty naive and still starting out)

*Option A* is building a preselector and utilising undersampling. There are, furthermore, two sub-options:

1) designing a selection of narrow band-pass filters, a la Alex, for 2m, 70cm and perhaps a few other bands such as FM broadcast, air band, perhaps even public service frequencies in the 170 MHz area
2) designing a selection of broad band-pass filters, each covering one Nyquist zone, up to 8*61,44=491,52 MHz.

Option A1 is a possibility, although it somehow doesn't seem elegant enough (whatever that means - it's just a gut feeling). Option A2 would require too many filters, even if zones without much interest were thrown out. Then there are the technical obstacles... 70cm band lies very near one of the zone limits, i.e. it would fold back extremely close to the Nyquist frequency, resulting in atrocious sensitivity (if any). Synthesizing filters with enough rejection and enough Q would require 7th, 9th and 11th order chebyshev and elliptic designs. The capacitances and inductances involved in anything above 2m would be difficult to manufacture consistently (if at all). Yadda yadda, etc.

So, i've rejected the idea of a VHF/UHF and opted for *Option B*. Which is down conversion. I see three possible (probably naive) ways of building a suitable and flexible enough down converter:

1) using Si570 (high grade LVDS version, if possible) as the LO to feed a high IP3 Mini Circuits frequency mixer, then filtering out the higher "image" with a 500MHz-ish low pass filter
2) same as 2, but with DDS (Analogue Devices AD99xx)
3) same as 2, but with a bank of relay switchable fixed frequency oscillators

B1 has the advantage of being cheap and simple, with a good amount of existing information, performance data and control methods. Both B1 and B2 have the advantage of continuously adjustable LO frequency, meaning the part of the spectrum that's of interest could always be centered w/ respect to Mercury's bandwidth. B2 is more expensive, but probably more stable. B3 would probably offer the highest quality, as long as the bank's frequencies are set up so zones of interest are down converted inside Mercury's bandwidth "comfortably".

B1 is probably the best option for initial experimentation and my own (not very demanding) purposes. So that's the route i think i'll go, as soon as i catch some free time and start putting all this down on paper. My plan is to try and design a down converter consisting of fore-mentioned elements (Si570, frequency mixer, filter) plus a microcontroller, that would fit an Atlas card and would be controlled through the bus, via I2C. The vision is to have the ability to bypass and turn all the elements, including the microcontroller, off when not in use, so the impact on the existing HF signal chain is minimal.

My experiences with PCB routing and design are strictly "digital" and (relatively) low frequency, so - if this project sounds viable at all - i'll quite probably need help with RF-related issues like stray capacitance, impedance matching etc.

So... what say ye? :) 


-- 
Ante Vukorepa
Sent with Sparrow (http://bit.ly/sigsprw)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20110922/3094360e/attachment-0002.htm>