ADS-B setups – Comparing internal and external aerials

Following on from my last update, I’ve spent a fair bit of time experimenting with two very different ADS-B aerial configurations.  I have two Raspberry Pi3s, both running identical software and exactly the same receivers, the FlightAware Pro Stick Plus which is a version of the RTL-SDR dongle with a built in preamplifier and bandpass filter.  This may (or may not) be as good as the other dongle I own combined with the Uputronics preamplifier/filter but I wanted both setups to be identical.

My first station is using a commercial aerial, a Moonraker “Radar-110” 1090 MHz collinear base antenna mounted on the side of the house, 38ft above the ground and in the clear.  Between the aerial and the receiver is approximately 15m of ExoFlex 15 coaxial cable which I calculate will give about 1.4dB loss at 1090 MHz.  The receiver is connected directly to the bottom of the coax with the Raspberry Pi attached directly to the receiver.

Moonraker Radar-110 base antenna

Moonraker Radar-110 base antenna at 38′ AGL

The second station is a homebrew aerial, a two element J-Pole collinear which I made using an SO-239 socket and a length of enamelled copper wire.  This is connected via a PL-259 to SMA adapter and then directly into the FlightAware receiver so there’s no loss whatsoever between the aerial and the receiver.  It’s mounted in my loft and as you can see, it’s quite heavily blocked in a south-southwest direction.

Homebrew two element J-Pole collinear

Homebrew two element J-Pole collinear in the loft

I’ve been running these two installations simultaneously for a few weeks and the results are interesting. Here you can see the figures reported for the total number of aircraft received daily and the total number of position reports received daily.  Here are the figures from the last two weeks.

ADS-B statistics - Averages at bottom

ADS-B statistics – Averages at bottom

The first surprise may be the actual numbers.  Yes, I really am seeing an average of nearly three thousand aircraft per day.  That’s a lot!

Perhaps more surprisingly, I’m seeing more aircraft on the internal aerial than the external aerial although the receiver attached to the external aerial is receiving more position reports per day.  I know this isn’t down to either system being swamped by signals as I’ve run gain optimisation on each setup and they’re adjusted appropriately.

Another thing to look at is the variation in distances.

Breakdown of distances received using the Moonraker aerial

Breakdown of distances received using the Moonraker aerial

Breakdown of distances received using the homebrew J-pole

Breakdown of distances received using the homebrew J-pole

Again, I think it’s surprising to note that I’m receiving a lot more signals from 150nm out using the internal aerial as compared to the external aerial.

Finally, it’s worth looking at the heat map for each setup.  This shows the maximum range received in each direction.

Heat map of aircraft received using the external Moonraker aerial

Heat map of aircraft received using the external Moonraker aerial

Heat map of aircraft received using the internal homebrew J-pole

Heat map of aircraft received using the internal homebrew J-pole

This is what I would expect to see.  The external aerial is in the clear and has a good view to the horizon with no obstructions whereas the internal aerial doesn’t have such a good takeoff and is clearly blocked in a couple of directions.

My conclusion to all this is that you don’t need a big external aerial, mounted really high in order to be able to see a lot of aircraft traffic.  Because all the aircraft traffic is up in the air that means that everything you’re receiving is direct line-of-sight and apart from local obstructions, there’s nothing in the way.  Of course, location is important and although I’m only 72′ above the sea, I happen to be in a prime location for spotting air traffic over the UK and north-west Europe.

Finally – If you’re interested, you can actually connect to my system here.  I have both receivers feeding a piece of software called Virtual Radar Server which combines the two inputs and plots them on a map.  It’s live aircraft data showing exactly what I’m receiving.  To access it click here.

Posted in ADS-B, Construction, Data, Raspberry Pi | Tagged , , , | Leave a comment

Building an ADS-B receiver setup

I realised recently that I’ve barely touched this blog for most of 2016 and I suppose that’s because I’ve not done a massive amount of radio this year.  My low powered QRSS/WSPR transmitter continues to run mostly 24/7 on various bands (on 30m at time of writing) and my radio high point of the year was being involved with the organisation and operation of the GB5RC special event station back in August.

My Hexbeam is currently down following a rotator failure and although I have a replacement rotator, I’ve not had the time or manpower available to get it back up again.

But I digress from the subject of this entry.

A few weeks ago I bought a Raspberry Pi3 to use for my own personal DX Cluster to replace the old computer I have running Windows XP.  I spent a while playing around with it, getting it working and then decided to look for other things to use it for.

I re-discovered aircraft tracking using the ADS-B system.  I had an old RTL-SDR dongle and so I bought myself a second Raspberry Pi3 and started experimenting.  I went through a few different builds and finally ended up using the instructions on the ADS-B Receiver Project site.  There are more detailed instructions on the Flightradar24 site showing how to set things up to ‘feed’ what you see to the internet.

I started off by connecting a 6m/2m/70cms collinear to the receiver and I was receiving just a few aircraft out to 50 miles or so but I wasn’t overly happy with that so I built a quarter wave groundplane aerial out of some copper wire and an SO239 socket.   By hanging that in my window, I was instantly receiving aircraft from over 150 miles away.  Absolutely incredible.

Quarter wave groundplane aerial for 1090MHz ADS-B reception

Quarter wave groundplane aerial for 1090MHz ADS-B reception

This wasn’t enough for me.  I did some research and ordered myself a combined pre-amp and bandpass filter by Uputronics which arrived a day later and my reception range increased out to over 200 miles.  I could have just set this up in the loft and forgotten about it, it would be a good setup.

Uputronics preamp and bandpass filter for 1090MHz ADS-B reception

Uputronics preamp and bandpass filter for 1090MHz ADS-B reception

But of course, anyone who knows me will be aware that I don’t like to compromise and an internal aerial is a compromise.  I could do better.  I was spotting over 1,000 aircraft/day using that internal aerial so what could I do with an outside aerial?  Two years ago, my 10m vertical broke in a storm and the pole it used to be mounted on was sitting outside.  I sourced an external aerial, fitted it on the pole using my old run of RG213 and switched it on.


Although I was receiving many more aircraft, they were all within 50 miles.  A quick calculation of the cable I was using suggested well over 4dB of loss, so over 60% of the signal was being lost in the cable.  It occurred to me that I had a run of EcoFlex 15 coiled up in the loft from my old VHF aerial installation so I fed that outside and ran it up the mast to the aerial.

Success (well, sort of).

I’m now receiving more aircraft from further away.  I’m still not 100% it’s working as it should be so I’ve ordered some test gear and once I’ve got that I can do some more checking.

Over 250 aircraft being tracked in the skies above G6NHU

Over 250 aircraft being tracked in the skies above G6NHU

A basic setup for ADS-B reception can be done cheaply with a small home made aerial, a Raspberry Pi and an RTL-SDR dongle.  I’ve gone a little further but this is a simple project which can be completed in an evening.

Posted in ADS-B, Construction, Data, Raspberry Pi | Tagged , , , , | Leave a comment

Video of the ten minute multi QRPp mode transmission

After my entry a few days ago, I’ve made a video of my Hans Summers Ultimate 3S transmitter going through all four modes in a single ten minute frame.  It’s annotated with comments along the way describing what’s happening at every stage.

Posted in Amateur radio, Construction, CW, Data, HF, QRSS, WSPR | Tagged , , , , | Leave a comment

Running four QRPp modes in a ten minute frame

Over the last month or so I’ve been doing some experimenting using a newly built Hans Summers Ultimate 3S transmitter fitted with the new oven controlled crystal oscillator and I’m now running four different modes in a single ten minute frame with space at the end for a calibration cycle.

You should note that to be able to run this setup, you need the v3.09 or later firmware.  There is a bug with the CW timing in all versions prior to v3.09 which means that once you go above 12wpm, the CW becomes progressively harder to decode and the RBN will struggle to receive it.  If you have v3.08 you may be able to do this by setting your CW speed to 12 and decreasing your calibrate time but in any versions prior to v3.08 you can’t have separate timings for the different modes and can’t configure multiple messages so this configuration isn’t possible.

Also, I would expect you to be locked to a GPS for calibration for this entire procedure.  It may work if you have a nice stable transmitter, all manually calibrated without a GPS but I’ve not tried it that way so I can’t guarantee it’ll work.  You do need your clock to be absolutely spot on for the WSPR cycle to start on time.

The four modes are WSPR, FSKCW, CW and slow Hell.

In order to do this, I’ve set up four transmission slots for the four modes as follows:

Slot Mode Power Frequency
0 WSPR 23 07.040.060
1 FSKCW 00 07.039.890
2 CW 01 07.027.300
3 Slow Hell 02 07.039.888

The numbers used in the power column are linked to the messages – For WSPR I’m running 200mW as indicated by the 23 but for the other three modes, the numbers call the messages as described below.
For slow Hell, I’ve selected a frequency a couple of Hz directly below my FSKCW so as just to centre it a little bit more with that mode.

My messages are configured as follows:
” #CS” | “CQ #CS #CS TEST” | “#CS”
The | is the delimiter character (this is a solid block on the U3S screen) and the ” are just for show – You don’t enter them. Please note that there’s a space in the first message before the #.
I’ve used the short messages here to save entering my callsign multiple times. When the U3 goes into transmit, the messages transmitted are as follows:

Message number Used by which mode Message text
0 FSKCW [space]G6NHU
2 slow Hell G6NHU

Speed settings are:

CW dit Hel
24 006 17

So that’s 24wpm on CW, six second dits on FSKCW and each character takes 17 seconds to send via slow Hell.

Calibration timings are set to 01 040 which means that the calibration cycle will run for 40 seconds and there’s just enough time for this to complete before the next frame starts. Depending on your callsign length, you may need to shorten the second parameter a little bit to allow the cycle to complete.  Start with 40 seconds, watch the frame complete and then keep an eye on the timing.  If the calibration routine doesn’t finish before the next frame starts, simply reduce the time.  If you find that you simply don’t have enough time in the ten minutes to complete the transmission cycle, never mind the calibration then you can try reducing your dit setting from 6 to 5 in the Speed menu page which will help.

Frame start settings are 10 08 which means that the frame will start at 08, 18, 28, 38, 48 and 58 minutes past the hour and run for ten minutes. I’ve chosen this start time because many grabbers will stack received images and this keeps the FSKCW/slow Hell within a single stack.

This configuration means that you can be spotted by a number of different systems. You’ll get the instant gratification of being reported on the WSPR network and also your CW will be picked up by the Reverse Beacon network but be aware that you won’t get as many CW spots on the RBN as you do via WSPR. You will also be spotted by any QRSS grabbers which happen to be on the same band.

I’m sure that many of the Hans Summers kits have been sold and are only working on WSPR, simply because the owners don’t know anything about the other modes.  If this article has raised your interest in the other modes that you can run from your transmitter, or even if you already know about them, I thoroughly suggest you take a look around the active grabbers to find a clear frequency before you just pick one and start transmitting.  In particular, 30m is already very busy.  For an up to date list of current WSPR/QRSS allocations, take a look at this page.

Here’s a frame from my own grabber on 40m showing WSPR, FSKCW and slow Hell.  You can’t see the CW because it was on a totally different frequency to the grabber and anyway, at 24wpm, it would be a mere blur.

WSPR, FSKCW and slow Hell in the same frame

WSPR, FSKCW and slow Hell in the same frame

To demonstrate the difference in signal between FSKCW and slow Hell, here is a single frame captured at the Pensacola Snapper grabber run by Bill, W4HBK.  You can see the FSKCW is much stronger than the slow Hell.  Bill is around 4,500 miles (7,300 kilometres) away from me.

G6NHU single frame at W4HBK

G6NHU single frame at W4HBK

To emphasise this even more, here’s a frame made up from three stacked images at the grabber operated by Pete, ZL2IK in New Zealand.  You can see the slow Hell is barely readable compared to the FSKCW.  Pete is located over 13,600 miles (21,900 kilometres) away by long path, which is the route this would have taken.

G6NHU on a stack of three frames at ZL2IK

G6NHU on a stack of three frames at ZL2IK

Finally the best image of the lot is another one made by Bill, W4HBK who stacked around sixty ten minute frames to produce this composite image.  It really is a testament to the stability of both the transmitter and the receiver that such a good image can be produced over such a long time.

G6NHU stacked at W4HBK

G6NHU stacked at W4HBK

Posted in Amateur radio, Construction, CW, Data, HF, QRSS, WSPR | Tagged , , , , | Leave a comment

Revisiting D-STAR and the GB7TE repeater

Back in March 2011 I wrote about three different modes of communications, EchoLink, D-STAR and QSONet. In that piece, I discounted all of them and although I said that D-STAR was the one which most interested me, I wasn’t planning on using it due to a lack of any D-STAR repeaters nearby.

Fast forward to earlier this year and I received a telephone call from Tony, G0MBA who had seen an article in the RSGB magazine, RadCom about Icom offering to donate D-STAR equipment to groups to help them set up repeaters.  We discussed it at length and decided to try and get involved with this.  We talked to the rest of the Martello Tower Group members and an application was made to Icom.

We were successful and on the 14th April 2015 Icom told us that they would provide us with a VHF Repeater Module and Controller for a D-STAR repeater.  Our plan was to remove the GB3TE repeater which had served the local area for over twenty five years and replace it with D-STAR.  We submitted an application for an NoV for GB7TE and that was issued on 28th May 2015, a very quick turnaround.

On the 4th June, a big box arrived from Icom containing all the repeater hardware.

A big box arrived, all wrapped in Icom tape

A big box arrived, all wrapped in Icom tape

We’d been doing some background work, looking at documentation and sorting out hardware for the server which was going to run GB7TE and on the 13th June, Tony G0MBA, Peter M1BRR and myself all met up to start assembling and setting up the repeater.

Tony with the new kit for GB7TE - The first time we've had brand new hardware for a repeater

Tony with the new kit for GB7TE – The first time we’ve had brand new hardware for a repeater

We had quite a few problems getting GB7TE actually built.  We went through three or four different PCs and servers and around six or seven different versions of Linux until we got the one which worked as intended. It was a major learning curve as none of us really knew anything about setting up a D-STAR repeater until we started this project.

Finally, after multiple false starts, quite a few bottles of beer, some cakes and even a curry, on the 5th of July we had a working system and on the 8th July, it was fully connected to the D-STAR network and on test at Tony’s workshop into a dummy load.

On the 17th July, the new GB7TE rack was installed on site at Holland-on-Sea, at the bottom of the radar tower.  We were almost ready to go.

GB7TE on site at Holland-on-Sea

GB7TE on site at Holland-on-Sea

The repeater ran for one day as a single, stand alone box before the ADSL connection went live and on Saturday the 18th July 2015 at 14:25z it was connected into the D-STAR network.  We were live!

G6NHU and G0MBA looking happy while testing the new GB7TE

G6NHU and G0MBA looking happy while testing the new GB7TE

It wasn’t all plain sailing though and the next few weeks saw repeated visits back to the tower to make tweaks to the setup.  We had quite a few problems which were eventually all tracked back to a faulty ADSL router and once that was finally replaced, everything has been working well since.  We also persuaded some local radio and TV personalities to record some voice announcements for us and they’ve been on air identifying the repeater every fifteen minutes since the 11th November 2015.

As a result of all this, I borrowed an Icom ID-51E handheld radio and after having used that for a couple of weeks, I decided to buy my own D-STAR wireless and I bought an Icom ID-5100E, removed the Kenwood TM-D710GE from my car and installed the 5100 instead.

Icom ID-5100E installed in G6NHU's car

Icom ID-5100E installed in G6NHU’s car

Despite the fact that Icom don’t supply a proper mobile mounting bracket for the control head, I bought their MBA-2 and paired it with a RAM Mount RAM-B-102U-A which I imported from the USA and that does a great job of fixing the head to the dashboard.  For what it’s worth, I think it’s pretty crazy that the only bracket Icom produce is a suction mount.

So how am I finding D-STAR?

Honestly, I still don’t really care for the internet part of it.  I’m more than happy to use D-STAR but my normal procedure when I get in the car is to unlink GB7TE from any reflector it’s connected to and have a shout through it.  If nobody replies, I’ll link it back to a reflector to listen for a few minutes and undoubtedly, after a while, I’ll repeat the procedure.  I don’t mind listening to other people chatting but as I said in my previous entry, I’m not interested in talking to someone thousands of miles away as I’m driving to and from work. Conversations through repeaters are normally just people talking fluff to pass the time as they’re driving around.  I don’t mean that offensively, I’m sure I’m just the same as I waffle on about nothing in particular to local friends!

As for the Icom ID-5100E though, that’s a different matter.  It’s an absolutely superb wireless.  Once programmed, it’s incredibly easy to use and the touch screen is a real joy.  I was originally going to set up the Kenwood I took out of the car in my shack but I’ve decided to get another ID-5100 to use indoors instead.

Perhaps my opinions on using D-STAR through the internet will change as I become more familiar with it.  If you have D-STAR and want to give me a shout to tell me I’m wrong or just for a quick chat (about nothing in general), I’m normally on in the mornings as I drive to work for half an hour or so at any time from 05:30 (UK time) onwards and listening to GB7TE which is generally linked to reflector REF001C.  You can check the status of GB7TE by clicking here and I’ll always transmit something to it when I get in the car so my callsign should show on the list of last heard stations.  I’m travelling home from work and in range normally from around 16:15 (UK time) on weekdays.

Posted in Amateur radio, D-STAR, FM | Tagged , , | Leave a comment

Building a QRP Labs OCXO

I’ve not posted much this year because I’ve not been doing much new stuff. I’m still messing around with the very low power modes and enjoying that a lot. I’m transmitting WSPR and QRSS all the time and tend to leave the transmitter running on one band/mode for a month or so at a time.

In the last week, I’ve built a QRP Labs OCXO (oven controlled crystal oscillator). I originally started building one of these quite some time ago but I made a big mistake and ended up just binning it and forgetting about it.  This time I was determined to do better.

I started off with this kit of parts.

Kit of parts for the OCXO

Kit of parts for the OCXO

After a while, I’d managed to assemble the two boxes for the oven and fit the oven components.  This is small and fiddly work, to give an idea of scale, I’ve added a 5p coin.

The OCXO partly assembled

The OCXO partly assembled

A few hours later, everything was done.  Here you can see the two halves of the oven, one each side of the main PCB.

Side view of the finished OCXO

Side view of the finished OCXO

The top of the completed OCXO

The top of the completed OCXO

To build this oscillator took around six hours and on top of that, once I got it into the final box and then into a transmitter, it took another couple of hours of gentle tweaking to get the temperature control set correctly.  It’s a case of adjusting it, waiting for ten to fifteen minutes for it to stabilise and then adjusting it again, repeating that process over and over until the minimum frequency is reached.  It’s a slow job but worth it.

I also built a replacement transmitter, a Hans Summers Ultimate 3S and fitted the OCXO to that.  I’ve been running it for a few days now and the stability is really quite incredible for such a low priced unit.  The previous transmitter used a DDS module which took a while to warm up and then tended to drift depending on room temperature and even just breathing on it could cause it to drift.  I’d sort of worked out the foibles of it and it wasn’t actually bad, I had it all built into a sealed case which was an effective oven anyway.  But this new OCXO is absolutely amazing.  It’s calibrated by GPS and once it’s reached operating temperature in around ten minutes, the stability is superb.  I’m monitoring myself and have noticed less than 0.1Hz drift over the last 24 hours!

G6NHU monitored on 40m

G6NHU monitored on 40m

At time of posting, conditions haven’t been great on 40m and the best WSPR spot I’ve had has only been to the east coast of the USA.  I’m hoping to be received by ZL2IK’s QRSS grabber once things improve a bit.

The original output power of the U3S was around 150mW but I wanted a little more.  I’ve added a second PSU to the box with a slightly higher voltage for the output stage (three BS170 FETs) and fitted some heatsinks and I’m now running about 500mW output.

Posted in Amateur radio, Construction, HF, QRSS, WSPR | Tagged , , | Leave a comment



It’s been a while since I made a semi-serious contest entry and I happened to notice a couple of weeks ago that the CQ WPX RTTY contest was coming up.  I currently hold the England record for the 15m high power section and so I planned to try and beat that this year.

I was on the air prior to sunrise on Saturday and operated until after sunset the same day, making a total of 436 QSOs which I thought was well on the way to being able to beat my previous record.

Sunday was a different matter though.  Conditions were poor and I struggled to get up to 630 QSOs by mid afternoon.  I went out for a walk to get some fresh air and clear my head and then worked another twenty when I got back for a total of 650 contacts.  I’m told that the band did still stay open for another couple of hours to the west but I’d really had enough by that time.

My final tally was:

  • 650 QSOs including six duplicates
  • 71 countries
  • 6 continents
  • 427 prefixes

My claimed score is 693,875 which is only a few thousand below my current record but there are duplicates to remove and I’m sure there will be a few logging errors.  Interestingly, I managed around 35 more multipliers this year than I’ve ever worked before.

One thing I really need to sort out is my logging software.  I use fldigi because there’s a native Mac version but although it’s good enough for normal data comms, it really doesn’t cut it for contesting.  I’ve done OK using it in the past but the big thing I really need to make the next step is DXCluster integration along with the ability to flag new stations seen on the cluster, mark whether they’re new multipliers and just generally help boost my score.  I’m sure that with decent tools, I could have worked a lot more stations and picked up a lot of new multipliers.  Out of my 650 QSOs, 645 of them were worked by running a frequency and just five by search and pounce.

To wrap, here’s my QSO map for the contest.  You can click for a full size version.

Stations worked during CQ WPX RTTY 2015

Stations worked during CQ WPX RTTY 2015

Posted in Amateur radio, Contesting, Data, HF, RTTY | Tagged , , | Leave a comment