RTL-2832 and Raspberry Pi for ARPT and FUNcube

RTL-2832 and Raspberry Pi for ARPT and FUNcube

Last year I was playing with a cheap USB dongle for weather satellite reception. Eventually though I gave up on it, as although I could receive reasonable pictures with it, the device kept causing my machine to freeze.

Anyway, I recently dropped a raspberry Pi in the shack to act as a router between my shack and my wireless network. It got me wondering as to the possibilities of running the dongle via it. Turns out, it was as easy as… pi! (Sorry… ūüôā )

It really only takes a few simple steps:

  1. Download the rtl tools source code via git from OsmoSDR
  2. Build up the tools (make sure you’ve installed the build-tools package on your pi)
  3. Make sure you unload an DVB kernel modules with rmmod (after checking with lsmod)
  4. Run rtl-tcp on the raspberry pi
  5. And then connect via TCP from something like SDRSharp

Doing this I’ve had no problems with anything freezing, and have captured images with as much ease as I did before – although I have a dodgy coax cable I need to fix.

Further, I’ve gone on to receive beacons (6m, 2m and 70cm) as well as the FUNcube telemetry. Actually, to my surprise I’m the only VK7 to have recorded reception of any telemetry (according to the FUNcube Upload Ranking page).

To finish, here’s a couple of images received via this setup (one from NOAA-15 and one from NOAA-19). And the title picture of New Zealand is from NOAA-19.


WSJT-X Going the Distance – 16,900 km

JT65 QSO with AB1J at some 16,917 km.

JT65 QSO with AB1J at some 16,917 km.

After successfully getting WSJT-X 1.3 up and running last night, I decided to throw out a quick CQ on JT65. Even though my waterfall was showing some interference I still managed to pick up a contact way out over on the West Coast of America some 16,900 kms away!

Other than being my longest QSO to date, what really surprised me was:

  • The fact the QSO was achieved with such high level of interference with¬†the signal barely visible on the waterfall (see below); and
  • That this was with JT65, so imagine what could be achieved with JT9 – shame the majority of activity of the two is JT65!

Looking forward to doing more with WSJT-X!!

Waterfall showing QSO with  AB1J. Note the interference and how weak (visually) the successfully received signal was.

Waterfall showing QSO with AB1J. Note the interference and how weak (visually) the successfully received signal was.

WSJT-X and Windows XP

As mentioned in my post about my first experiences with WSJT-X and JT65/JT9, I had issues getting any version other than 1.1 to successfully decode. A bit more searching and some playing around last night and I now have successfully decoded with 1.2.1 and 1.3 on Windows XP.

My solution in the end was to simply open up wsjtx.ini in the install directory, and then add the following lines at the very end:


This (I understand) has the effect of returning the¬†sampling methods back to what was in version 1.1 – which essentially uses 12kHz sampling rather than 48kHz. There is another option however this is sound card driver specific, and that’s to set the default sampling rate (via the control panel, etc.) to 48kHz. However, for my sound card (Audigy 4) I was unable to find this setting so went with the above solution.

I hope this helps others who may have been stumbling on this. Albeit, I do now realise it’s documented in two places:

  • The above solution is in the manual’s FAQ section – however currently the formatting is a bit askew in the manual; and
  • The details about making sure your sound cards default sampling is set to 48kHz is specified in the Windows install instructions in the manual.

Good luck!

An easy 15,800 with JT9

An easy 15,800 with JT9
JT9 QSO with K8TLC at some 15,813 km.

JT9 QSO with K8TLC at some 15,813 km.

I was passing the shack’s computer last night before dinner and I noticed JT9 showing a very clean waterfall. I saw someone calling but it was really on the edge of my filters passband. So in case someone else was about I threw a CQ out up in a stronger section of my passband.

To my surprise the other caller (K8TLC) saw it and immediately replied. This resulted in a full QSO at some 15,813 km. Full exchange can be seen in the screenshot.

I’ve also found some information on getting WSJT-X versions 1.2 and up working on Windows XP. Seems in my haste I missed the pertinent information in the manual. If all works out I’ll be sure to share.

WSJT-X and DX with JT9 and JT65

Last week I came across JT9 as a mode for DX on HF. I’d not come across it before (possibly being side tracked with WSPR), but before I knew it I was downloading WSJT-X, reading the tutorial and running it up.

However, first attempts to receive were a complete fail. I was sure of success as it was when there was no interference (for once) and I had heaps of JT65 signals roaring in and system time was in sync to ~100 milliseconds. But not a single decode.

I then proceeded to just give it time and let it run some 24 hours. But still nothing…

A bit of searching around the net revealed I wasn’t the only person who had troubles decoding with WSJT-X. Further, it seemed maybe that for those who had troubles all used to work with version 1.1 of WSJT-X.

I proceeded to download WSJT-X 1.1 and lo’ an’ behold suddenly I was decoding signals left right and centre. Both in single JT65 mode (for which there seems greater activity) and also in split mode decoding both JT65 and JT9 transmissions.

Since then I’ve tried again with version 1.3 and also 1.2.1 (a couple of times), but for the sake of me I can not get decodes with either – oh, except a single decode once… And looking on pskreporter (a¬†great site BTW) it would seem very few are using the latest version of WSJT-X. Further, looking at a¬†WSJT-X changelog I note there were changes starting with 1.2 about how the sound hardware was accessed. So I wonder if maybe that’s had a flow on effect to those running on Windows XP (such as my shack PC).

Anyway, at least I have a version that works (just a shame some nice waterfall improvements came in in version 1.2) so I proceeded to try some contacts. My first was with VK5HAM but I made a fair mess of things as I wasn’t as comfortable with the UI as I thought I was. Anyway, the one I was most impressed with was with HA1AD. Details were:

Their grid square: JN87
My grid square: QE37
Mode: JT9
Freq: 14.078850 MHz
Report sent: -06
Report received: -09
Distance: ~12,995
My power: 5W

That is indeed my first true (ie. bi-directional) DX contact.

After that, I was hooked and tried a few more, but often shooting too far (15,000+ kms). It’s amazing how quickly the time passes by with these modes.

So, if you’re yet to try these¬†modes then I highly recommend them. They take WSPR a step further, and with reverse beacon networks such as PSK Reporter you’ve got everything you need (ie. WSPR beaconing but with option for proper QSO contacts)!

One last thing though before I go. Later on my interference returned, but to my surprise WSJT-X kept on decoding. The following two images show WSJT-X picking signals out of the noise. If you take a look at the waterfall @ 23:23 and then just near 1200Hz you can see the signal (just) from ZL3HAM, but at the same time you can see the wonderful interference I enjoy.

Although an extreme example, I have actually had QSOs in amongst lesser interference with WSJT-X this week-end. Something I would have struggled with voice and many other digital modes.

Interference Continues

Example interference around 14.070 Mhz

Example interference around 14.070 MHz

I’m still looking into my interference issue and have found a few more tidbits about it. One is that it seems to be seriously broadband in that I tracked it from about 7.9MHz to 23.0MHz repeating about every 5.04kHz. (Keeping in mind that my antenna is tuned for the 20m band so at those ranges any reception is pretty weak.)

Further, I started to wonder if it was my radio, so I swapped from my FT-817 to my IC-718 and the issue was still there.


The interference differs too. Around the main frequency of my antenna I see squiggly lines on the waterfall (as above), however once I get further away I see what I primarily consider straight lines as seen on the right.

Anyway, to visually see things a bit more I also put together two time lapse videos of the last two 24 hour periods. The first being on the Yaesu¬†and the second on the Icom.¬†Unfortunately to show them here on WordPress, they’d like me to upgrade my account – so I’ll have to get back to you as to how I’ll share them.

Another thing I tried, tonight I pulled everything out and now only have the Icom connected directly to the antenna via the wall socket and have it running on a battery. But still the interference persists.

I’ve also done things like disconnected the antenna – waterfall went quiet as expected; turned off the radio – waterfall went quiet as expected.

But the waterfall capture is starting to show me is two things:

  1. There seems to be good digital activity on 20m an hour each side of 1300hrs each day – hopefully I can enjoy it at some point; and
  2. A pattern might be emerging – at least I think I see the interference start around 7pm and stop around 10:30pm – but need many more days yet (and there’s also a possible morning start and stop too. And that makes me further think it’s a person doing something and maybe in the form of heating.

Anyway, hopefully a bit more monitoring might reveal some patterns as I’m kind of out of ideas to try – I’m fairly comfortable to say the source is not from my block.

20m Interference – Starts and Stops

So I just sat and watched the waterfall for a bit in Spectrum Lab on 20m (14.070MHz). To my surprise I observed the noise suddenly start and stop (see below).

That’s making me think it’s more likely some kind of electronic device, and just maybe something like a reverse cycle heater (or heat pump as they seem to call them in VK7). Anyway, I’ve now setup spectran to go nice and slow and do 15 minutes captures – so I’ll see what that yields.

But still, if anyone has any ideas, please drop me a comment. ūüôā

Maidenhead Converter Library

Click to github project

Click to github project

I’ve been putting together a little groovy/java library for doing maidenhead conversions – mh-converter. It’s pretty early on at the moment, but is basically documented, has a test suite and is known to work. So effectively I’ve put it out there as a 0.1 type release. As well as providing a JAR for use in other projects, the JAR also provides a command line interface.

Anyway, if you want to have a play then feel free to grab a copy from github at:


You can then build yourself (as long as you have a JDK installed) and do with it as you like.

I will be making further changes (maybe) as I’m yet to actually use it as the intended library I was after for something else I’m working on. So we’ll just see where that takes it.

Attempt to setup for 20m – interference

Example of interference on 20m - 14070

Example of interference on 20m – 14.070 MHz

My long term goal here is to have an inverted V for 40m. However the other week I was looking at my 10m vertical in the garage and started wondering what I could throw up in the short term for HF. It then dawned on me that between the eave support for my 2m/70cm antenna and a pole in my back yard for an old arbour I’m pulling down,¬†there’d¬†be just enough space for a 20m dipole!

So last week-end I finally got a basic dipole up, tuned it and the SWR was looking good! I attempted some WSPR and I was happily receiving stations from America and Europe (albeit, my shack computer currently has no network connection Рso I was unable to post spots). In return, I could see I was also making the same distance.

However shortly after that I realised this was causing TVI. I thought this was surprising considering it was low power and at 20m, however whenever I transmitted I immediately trigged a ‘No or weak signal’ message on my TV.

Example of interference on 20m - 14070

Example of interference on 20m – 14070

Easy fixed, I wont transmit until the week-end (when I’ve time to investigate and address) and will just attempt to receive some SSTV. But this turned into a rather mediocre result and led me to look closer at what I was receiving. I then realised¬†there was little point fixing the TVI issue as I was going to have trouble even receiving!! The image on the right is what I saw (which is also reflected in the S meter sitting up at a solid 9).

The image was taken at 14.070 MHz as one of the main things that attracted me to 20m was the hope of lots of digital work – and that’s focused between 14.070 and 14.090. But up at 14.230 (the SSTV calling frequency) the image was similar albeit the lines are more straight/constant.

Anyway, I did the normal thing and turned off¬†the power to my¬†house and ran on battery supplies. But the issue persisted – indeed, it didn’t seem to change at all. So I guess there’s some localised interference, but I’m not entirely sure what this pattern might be caused by or indeed what my next choices might be. I’m kind of just taking it as a sign that 20m ¬†is out… But that said, if you recognise this pattern (or have suggestions) please jump on in and leave a comment.

10mComparisonAs a comparison, I thought I’d see if I could tune the dipole up on 10m an see what that looked like. The image to the left shows the result. Further to this though, I was able to happily send and receive WSPR to America as well as the reception of a couple of SSTV images also from America.

So whether that means I cut the thing down to size for 10m and just go with that until I get my proper 40m setup done… I don’t know. Anyway, it’s the end of the week-end so I guess I’ll just have to ponder it a bit.

Maidenhead GIS Layers


So I’ve been a bit quiet on the radio front lately as I’ve been pretty busy with non radio things. However, that said I’ve still been playing with some code with a radio focus. Specifically, I’ve been doing some analysis of WSPR data that I hope soon I’ll be able to post up.

But most recently I’ve written a small groovy script (can be seen¬†on¬†gist.github.com) to generate GIS layers of a maidenhead grid to help with some visualisation in¬†QGIS. I thought I’d share it in case others find it useful, as well as the resulting GIS files – as both the original GeoJSON files and the Shape files.


ESRI Shape file:

The script also generates a 6 character grid, but I’m sure you can appreciate the size of that it comes out as – not entirely practical.

I really do love how easy it was to build these in Groovy with no external dependencies and a minimal amount of code and fuss. If you’ve not¬†checked¬†out¬†Groovy¬†for your JVM platform programming – I can’t recommend it enough.