Using a Pi SBC with Shortwave Radio (4)

birdiesimages2Figure 1:  Main signal is at about +30 KHz

OK, so I’ve been singing the praises for the Soft66-Lite.  Yes, I like it.  But – it’s not a $1000 dollar radio.  It’s an $18 dollar radio. It has its limitations. With the BPF on the front end, I can get rid of most artifacts that might interfere with my reception.  However; the BPF is 2 MHz wide (intentionally).  The local oscillator, IQ detector, and software is expected to take up the slack relative to the additional selectivity that is needed.  And, to a great extent, it does.

lessimages1Figure 2:  A signal that is alone (except for noise from a local PC computer and other sources).

We really want the situation we see in figure two, rather than the multiple signals shown in figure one.  That is true unless we really want to listen to the huge signal that is causing the images!  Obviously, if we don’t want to listen to the big signal, then the signal we want to hear may be covered up by the big one, and we wouldn’t be able to hear it.  Figure one looks almost as if the BPF was not actually installed in the circuit.  It was present, I know, but maybe it was not making a good connection.  I’ll have to do more testing.

Anyway, the BPF is coarse, by nature, so it isn’t used to drive the passband down to the 3 KHz we need for the weather fax reception.  In a $1000 dollar radio, we would see the heterodyning of signals, perhaps the use of multiple IF stages, isolated parts in certain instances, maybe including a mechanical filter (at least in the old days) and shielding between major receiver components.  The software does a great job with decoding, demodulation, and to a great extent the selectivity as well.  But, it can’t keep the receiver parts isolated.  A metal box would help, I suppose.  The big signal in figure one was a shortwave AM broadcast station on ~ 6 MHz, located about a thousand miles from here.  It probably was running 50 kilowatts of power, and we were experiencing good ionospheric conditions.  It was a killer signal.

IMO the faxes that started out great, but then faded, were often fading due to some de-sensing caused by the other signals.  As the other signals grew, the desired signal shrank. I am not a receiver design expert, but it seems the issue is not sensitivity, but instead it is selectivity.  The RF amp in the unit provides sufficient sensitivity, IMO.  It has some impressive specs, for those willing to check out the datasheet. But somewhere along the line, our signals are occasionally getting swamped.

What can be done to improve the little SDR in this area? I’m not sure, but I’d like to see a write-up by the author of the radio (maybe with suggestions) about how to improve it (without spending a lot of money).  Money can fix it, but we’re looking to make something of nothing (isn’t everybody?) but also have our cake, and eat it too! There are some upgraded Soft66 models on the zao site.  I may try one of them, to see if I get improved results 🙂  In particular, the Soft66-LC looks good.

So, I’m looking at the little “sidecar” SDR, and I like it.  It’s tiny (unlike some clunky shortwave receiver box I’d have to lug around).   It’s easily powered from the tablet, and sips only the tiniest bit of battery capacity.  It’s tunable from the tablet.  The tuning could be automated, so as to search for fax signals and print them, etc.  It’s easy to play with the unit, it pretty easy to understand, and there’s very little worry of damaging anything ($18?).  Note that in figure one, the plus/minus 80 KHz scale is what the software displays when the “audio scale” is selected. In reality, I can tune anywhere inside of the width of my Cirrus card’s capture sample rate (192 KHz), which is plus/minus 96 KHz from the center frequency (zero on the audio scale).

It’s much easier to grab faxes with a $500 – $1000 shortwave radio, and I’ll certainly always have one of those in my shack.  But, the fun factor of the Soft66-Lite far exceeds its cost.

I’ve been thinking about adding a DDS in order to get rid of the crystals. But, that adds about $50, plus my labor.  Hmmmm.  In the meantime, I’ve purchased a number of crystals that allow the bandwidth to include a few weather fax stations. The most bang-for-the-buck is 4.3 MHz. Here’s a rundown:

4.3 MHz crystal:

Multiplier                         CF    WX Station

  • x2 (sv1= M; sv2 = M)  4.30  4.344 New Orleans, LA
  • x2 (sv1= M; sv2 = M)  4.30  4.235 Boston, MA
  • x4 (sv1=2,3;sv2=2,3)   8.60  8.682 Point Reyes, CA
  • x3 (sv1= M; sv2=1,2)   6.45  6.4?? Nova Scotia

(CF = center frequency; note: last stage is a divide by 2)

3.2 MHz crystal:

Multiplier                               CF     WX Station

  • x2 ( sv1 = M; sv2 = M )     3.200 3.253 Resolute, Canada
  • x4 ( sv1 = 2,3; sv2=2,3)    6.400 6.341 Boston, MA
  • x5.3333 (sv1=M;sv2=2,3) 8.532 8.504 New Orleans, LA

(CF = center frequency; note: last stage is a divide by 2)

4.5 MHz crystal:

Multiplier                        CF      WX Station

  • x4 (sv1=(M; sv2=M )  4.500 4.416 Sydney Nova Scotia
  • x6 (sv1=2,3;sv2=1,2) 13.50 13.?? Halifax

(CF = center frequency; note: last stage is a divide by 2)

9.0 MHz crystal:

Multiplier                       CF      WX Station

  • x2 ( sv1=M; sv2=M )  9.000  9.110  Boston, MA (note 3)
  • x5 (sv1=1,2;sv2=2,3) 22.50 22.527 Point Reyes, CA
  • x4 (sv1=2,3;sv2=2,3) 18.00 18.068 Ham radio band


  1. Unless stated, jumper sv3=(1,2), which makes multiplier active
  2. (IIRC – Halifax Nova Scotia inactive)
  3. Might be possible by “pulling the crystal” see text
  4. Frequency list is not exhaustive, but what I’ve found for Xtals

Read More …

Note: the author does not have a recent, applicable background in circuit building, or battery related issues, so this is presented as the work of a hobbyist, and is not meant for duplication by others. Readers should look elsewhere for design advice and info.

Note: the soft66-lite hardware is a product that is sold on a Japanese website ( – and is not affiliated with this site or author in any way.