Re: [linrad] A simple front end

[Edson Pereira <edson@xxxxxxxxxxxxxxxxx]

Hi Ron,

I have been experimenting with a few very simple circuits looking for the
easiest way to get 40 kHz from the 20m band into the Linux PC (and trying
to educate myself a little about RF mixers, low noise amplifiers, and
SDR). My intention is to find a reasonable simple and cheap quadrature
receiver for HF frequencies. 

My first receiver used a pair of SA602A and a NE5532 Op Amp. The circuit
was a verbatim copy of the circuit described in the Phillips AN1981
Application Notes. I knew the limitations of the SA602A and just wanted to
get started. The next circuit I built used two single balanced mixers
using two diodes on each mixer. The next one used two sub-harmonic mixers
using four diodes each. The sub-harmonic mixer circuit is similar to the
two-diode balanced mixer, but with two more diodes in an anti-parallel
configuration. This mixer worked much better than the SA602A and slightly
better than the single balanced diode mixer. One advantage of the
sub-harmonic mixer is that the LO is at half of the received frequency.
This is an advantage since the irradiated LO energy is smaller. Also, the
noise figure seems to be relatively lower (comparing the averaged FFT
noise floor of the three circuits.) The lower signal I can hear with the
sub-harmonic mixer is about -135 dBm. I could not measure the 1 dB
compression point since my RF generator stops at -10 dBm. At -10 dBm there
is no distortion. The SA602A mixer is useless for signals stronger than
-35 dBm.

In all circuits, I used a simple crystal LO. I have a DDS, based on the
Wellpine TC170C030AF chip (it is a kit sold in stores here in Tokyo), but
the output is not very clean. For the quadrature phase shift I used a
+45/-45 degrees RC network.

The next circuit I will build is based on the Sampling Detector used on
the SDR-1000 article. I will also use a simple (and clean) crystal LO. The
only thing I need to get started is to get the PI5V331 chip soldered in an
universal SMT circuit board. The leads of the chip are very difficult to
handle without special SMT tools. From the literature I've read, the
performance of the quadrature sampling detector should be better than the
circuits I have already tried. If true, I will be very pleased.

On the software side, I have linrad and two SDR prototypes (for pure
experimentation). One uses a conventional FIR/Hilbert Transformer filter
and the other uses FFTs (based on the SDR-1000 article.)

I have been enjoying a lot reading about and playing with SDR circuits and
software. I will be glad to share information about the successes and 
mistakes.

73,

-- Edson, pu1jte, n1vtn, 7n4ncl

On Thu, 7 Aug 2003, W6WO wrote:

> I am keen to experiment with the "very easy way to get 40 kHz of 
> radio spectrum into a PC" that Leif described in Fig 2 of the Jan/Feb 
> 2003 QEX. My interest is only the cw end of 20 meters
> For an RF stage I plan to use a Mini-Circuits ERA-5 amplifier with a 
> band-pass filter ahead of it. Instead of the quarter wave of coax the 
> SPH-16 voltage controlled phase shifter looks like the way to go 
> unless I can find a cheaper solution. I looked for low noise audio 
> ICs that were not surface mount and the best I can find is the LM833. 
> Initially I will use a signal generator for the LO but a crystal osc 
> would make it a self-contained front end that could be packaged in 
> quite a small box.
> I would like to share work of this kind with others, either via the 
> list or direct.
> 



LINRADDARNIL