Predicting space weather – and how to get it wrong!

Space weather prediction is an inexact science. We still don’t fully understand the Sun or how it works.

This means that predicting what HF radio conditions are going to be like over the next seven days can be challenging – especially when you have to produce the forecast a couple of days before it is published.

This week was a perfect example of how to get it wrong!

I prepare the HF part of the RSGB’s GB2RS Propagation Prediction forecast on Thursday afternoon. Jim G3YLA looks at VHF and Sporadic E, before it is sent to John G4BAO who adds some more info on any microwave propagation such as rain scatter and also adds the EME report.

I then edit it, add my HF predictions and send it to RSGB HQ so it can be collated, published and sent to GB2RS readers on Friday.

Sunspots on Thursday 23rd August 2018.

At that point this week the Sun looked fairly innocuous with one tiny sunspot group, a solar flux index of 70 and a sunspot number of 15, representing one sunspot group with five tiny spots.

There were no obvious coronal holes and NOAA predicted the K index would likely be two all week due to a lack of geomagnetic disturbances. The US Air Force agreed.

So far so good!

At 04:45UTC on Monday 20th NOAA had reported:

“Region 2719 (S06, L=133, class/area Bxo/010 on 19 Aug)
developed in the SE quadrant on 19 Aug. No significant flare events
occurred from either region. Other activity included a filament
eruption centered near S11W04 observed lifting off the solar disk at
approximately 19/0538 UTC. An associated coronal mass ejection was
observed off the SW limb in SOHO/LASCO C2 imagery at 19/0812 UTC.
WSA/Enlil modelling of the event suggested the ejecta was primarily
directed westward of the Sun-Earth line and is not expected to cause
any significant effects.”

So the CME wasn’t earth-directed so shouldn’t be a problem – bear with me!

The NOAA report for Monday 20th August said there were no events.
The NOAA report for Tuesday 21st August said there was an RSP (a sweep-frequency radio burst) at 18:46UTC observed at Palahua, Hawaii, USA
The NOAA report for Wednesday 22nd August said there were no events.
The NOAA report for Thursday 23rd August said that at 20:23UTC on Thursday evening the older region 2719 emitted an an “A class” X-ray solar flare. “A class” flares are the weakest and it was after sunset in the UK so didn’t have any major impact for us according to the ionosonde data.

But there were further A class flares on Friday 24th August, which were at 12:47-12:50 UTC, 21:54-21:58 UTC and 21:57-22:01 UTC.

The Chilton and Fairford Ionosondes in the UK show Friday’s flare’s effect at 12:47 as there is a gap in the data.

You can see this at Jim G3YLA’s Propquest site at: http://www.propquest.co.uk/graphs.php?type=archive

But by Saturday 25th August a new large sunspot
group had appeared.

By Friday another new sunspot group (2720) had emerged, which grew rapidly and by Saturday it was quite large. Sunspot AR2720 is not only large, but also strange. Its magnetic polarity is reversed. The North and South ends of its magnetic field are backwards compared to the norm for sunspots in decaying Solar Cycle 24.

So could AR2720 be from the next solar cycle, Solar Cycle 25? We’ve already had one reverse polarity sunspot which has been attributed to Cycle 25, but that was last year. This is going to be discussed over the coming weeks.

Action on Friday 24th August continued.

From 02:00UTC a series of solar flares were recorded, mainly from the new region 2720. These were B-class solar flares.

By the early hours of Sunday morning, 26th August the K index had started to climb, peaking at Kp7 at 03:00 and 06:0UTC. The bands on Sunday were lousy with a lack of signals across the board in the UK as the ionosphere collapsed. The 14.100MHz IARU beacons were all inaudible.

The Kp index hit 7 in the early hours of Sunday
26th August 2018.

A sweep with my SDRPlay receiver and SDRUno software saw very few signals on 20m at all and virtually nothing higher in frequency. There were a few weak and watery CW signals from Germany on 40m.

So what was the cause?

NOAA now says there had been an earth-directed coronal mass ejection on Monday 20th August that no-one seemed to comment on at the time. NOAA’s own records don’t show any solar flares on Monday August 20th – see (ftp://ftp.swpc.noaa.gov/pub/indices/events/20180820events.txt)

So was it this CME or the one on the 19th that is now reported as causing the high-speed solar wind stream that brought about the Kp7 event and auroral conditions? The experts seem to think so – now!

Or was it a CME as a result of the solar flares on Friday 24th August? That to me would seem more likely and fit in with the roughly two-day solar wind transit time to Earth.

SpaceWeather.com described it as a “surprise geomagnetic storm” and I think that’s a pretty apt description as no-one predicted it.

If that is the case there is no way I could have known about it when I filed the copy to RSGB on Friday morning. This report predicted quiet geomagnetic conditions for the week commending Sunday 26th August – I don’t think so somehow!

Oh, the joys of space weather and HF radio propagation forecasting!

Fixing the 20m QCX radio

I've had my 20m QCX kit working now since just after Christmas. While the build was quite straightforward I did have a few problems.

Click to enlarge the photograph by the way.

The first was that I managed to blow the 7805 regulator while leaving the radio running on WSPR one morning. Having replaced that I found that the rig was quieter and there was no TX output any more. Thinking I had also blown the PA transistors I replaced all three of the BS170s, but that didn't help either.

Anyway, after a lot of help on the QCX reflector I decided that the fault had also taken out the 74ACT00N chip. This had been soldered onto the board directly so had to be cut out with a Dremel and a new IC socket installed.

And after finding a short somewhere near the PA transistors all was well - success! Output using a 12.6V Lithium battery pack was about 1.8W

I have been using it on and off ever since as it has a really nice receiver - even took it to North Norfolk as I detailed in another post.

But one thing always bugged me - why only 1.8W? When I built it originally I was getting more like 3W, so perhaps it was the replacement BS170 transistors, which came off Ebay on a slow boat from China.

So having a few hours free today I ignored my own "if it ain't broke don't fix it" mantra and pulled it apart. I replaced all three BS170s with new ones bought from the GQRP club and while I was at it I replaced C22 with a 0.1uF capacitor to try and soften the sidetone a little.

I managed to break one of the wires to one of the control buttons on the front in the process, but it is now all back together.

And the power output is now just under 3W with 12.6V - success! I have no doubt I could get more if I messed around with the band pass filter, but I don't think it is worth it. I tend not to run it on 13.8V as the regulator gets VERY hot indeed.

Anyway, it is a great little radio and will be going on some more journeys no doubt.

I can thoroughly recommend it as a kit project - see QRPLabs site. My case came from Banggood in China and only cost  a few pounds. It did take about 10 hours to machine (posh word for drilling out with a Black and Decker and attacking it with a file).

The label was produced in Photoshop and printed off by Photobox.co.uk. I ruined the first one by spraying it with varnish, which made the dye run. Luckily I have a few more spares and may eventually cover one in sticky-back plastic. But for now it looks great.

The red buttons were bought just before Maplin went bust so are a lasting tribute to what was one of my favourite shops and the result of much ribbing by the family every time I went there - RIP Maplin.

"Bumps on the Air" 2018 - QRP in North Norfolk, UK

Today was 2018 ”Bumps on the Air” day for me. Because we don’t have any summits in Norfolk I have to make do with tiny hills, or bumps!

Two years ago Jim G3YLA and I climbed up Beeston Bump (63m) and did some amateur radio operating. This year I wanted to beat that so drove to Norfolk’s mightiest peak – Beacon Hill near West Runton at 103m. Ta dah!

Actually, this wasn’t a very exciting spot as there are just two rotting benches, a flag pole and trees all around.

Nevertheless, I set up my end fed half weave vertical on a 10m fishing pole and connected up my Yaesu FT817. There wasn’t too much about but I worked John F5VKU (also G8MM) near Cannes with 5W SSB. He said he was struggling to hear me.

My 20m QCX radio - click to enlarge any image.

After playing with SSB for a bit and failing to break some pile ups I connected up my 20m QCX radio and tried some CW. I was really surprised as Karl IV3RJH came back to me, 559/549 both ways. I was running about 2W, he was running 3W.

I do have to fix the sidetone on the QCX as calling CQ sounds more like “thump-de-thump-thump-thump-thump-de-thump”. I have the new capacitors, but am waiting for some new BS170 transistors so I can hopefully up the power level to more like 4W too.

But the QCX is an amazing radio for the money and hey, I built it myself, complete with the case and custom label, so any QSO makes me smile.

Another 20m SSB session with the FT817 and I bagged Jack OH3GZ and Juha OH6QAZ.

At this point I packed away and after lunch headed for Beeston Bump on the coast near Sheringham. I’ve written about this before as it has a fantastic view over the North Sea.

Anyway, I put up the EFHW vertical for 20m and set to with 5W CW from my Elecraft K1. This raised Gert OH/DL7UG and Dima RW4C. I then bagged Fabi IK5IiS near Florence.

Also heard were stations in Cypus and Canada, but the Cyprus station on SSB couldn’t hear me and the Canadian seemed to be sending his life story in CW to another station and I got fed up waiting to call him!

I also had a play with my 3W MTR3, but didn’t work anyone as I could see rain clouds coming and decided to pack up.

So not a bad day’s radio. The Elecraft K1 and Yaesu FT-817 worked flawlessly. I think I was getting some RF into the keyer on the QCX as I had trouble sending “/P” – the “/” turned into a right mess sometimes, although I had no trouble sending it on the K1 with the same Palm Paddle.

The moral is that life is NOT too short for QRP!

Update: I checked the reverse beacon network last might and saw that I had been picked up across Europe while calling CQ, but often with no response. Some of these SNR figures are quite large too. In other words I was loud enough.

That's a shame as it would have been nice to have worked some more stations.

Is this because of a lack of activity? Or is it because, as I often wonder, that a lot of people have moved to FT8?

IBP propagation predictions for the UK for May 2018

Propagation predictions for the International Beacon Chain
for the UK, May 2018.

In an effort to try to make HF propagation predictions more accurate I have been playing with a different method of producing them using the ITURHFPROP software.

Previously I had been calculating the field strength of each beacon and then converting that to an S point level. It was pointed out that this isn't the best way and that calculating the signal strength in dBm and then converting it would be better.

They are based on 100W to a dipole at 10m - we know that the IBP chain use a Cushcraft R5 antenna.

I then hit the problem as to what antenna gain do you use. This may sound simplistic, but in fact it isn't as the actual gain off a dipole will depend upon its height, its orientation and the required take-off angle.

Now, the take-off angle will depend upon what ionospheric mode is dominant at that point in time. That is, is it one F-layer hop, two F-layer hops or a combination of E and F-layer hops.

After a lot of thought and discussion I settled on setting the gain at both ends at 0dBi, which appears to give reasonable results - setting it at 2.15dBi gave over-optimistic results

Anyway, attached is the plot of predictions for the beacon chain for May 2018. I'll check to see how accurate it is and adjust the input files accordingly. The ultimate goal is to the transfer these settings to other HF predictions.

If anyone wants to report on whether they find these accurate or not please let me know. They are median predictions, which means some days may be better, some days worse.

Just click on the chart to enlarge it.

International Marconi Day 2018 from Caister Lifeboat

Chris G0DWV on the mic. with Roger G3LDI looking on.

Norfolk Amateur Radio Club managed to contact 116 other radio amateurs in 23 different countries on Saturday 22nd April 2018 when we took part in the annual International Marconi Day (IMD) event to mark the inventor's birthday.

This was the ninth year we've taken part.

Using the call GB0CMS and a mixture of Morse code, telephony (speech) and data (FT8), contacts were made with other radio amateurs across the UK, Europe, and the USA.

Notable contacts were made with other IMD stations in Italy, Ireland, Weston-Super-Mare and Poldhu, Cornwall – home of the Marconi Centre from where the inventor made the first transatlantic transmission in 1901.

We ran the all-day special event station at Caister Lifeboat to commemorate the village's original Marconi Wireless Station, which was established at Caister in 1900. The station was in a house in the High Street known as Pretoria Villa and its original purpose was to communicate with ships in the North Sea and the Cross Sands lightship.

Conditions weren’t brilliant due to a major solar disturbance (effects of a coronal hole), but we were still able to cross the Atlantic on three occasions.

We were picked up in Pennsylvania, US via FT8 at -20dB on the reverse beacon network, but didn't make any transatlantic FT8 QSOs.

The equipment used was 100W maximum from an Icom IC-7400 (20m) and an Icom IC-7300 (80/40/30m). Antennas were a W5GI dipole on 40m and my design of a monoband end-fed half-wave (EFHW) vertical for 20m. The FT8 digital mode was used for the first time to show what can be worked with just 20-25W.

We also used a Flex Maestro to remotely link into Malcolm G3PDH's home station via the internet and worked back to ourselves using his special First Class Operators' Club call for April, G4FOC. The Maestro is very impressive!

Above all, it was great fun and a good social event

Commonwealth Contest 2018 and QRP

RSGB certificate for 5th place QRP section.

Update: It turns out I came 5th in the QRP section of the 2018 Commonwealth Contest. I lost two QSOs for being "not in log",  and lost 4th place by 10pts to Dave G3YMC (which is how it should be!). Great fun and I'll do it next year, probably with my Yaesu FT817 and the addition of an EFHW vertical for 20m too.

I took part in the 2018 Commonwealth Contest this weekend on HF CW. This was always going to be a casual entry as I had other things to get on with over the weekend, not least was helping tidy the house for Mothers’ Day on Sunday!

I decided to enter the QRP category with a maximum of 5W. One reason was I knew that my good friend Dave G3YMC had won the category last year with 30 QSOs. I mean, how hard can it be to beat 30 Qs with 5W? Answer: very hard apparently!

I got going with 20m and quickly worked four stations in Malta and Cyprus. I also cracked off some of the UK HQ stations on 80m for 25pts apiece.

The afternoon saw openings to Canada, which brought in a few VEs. I was also happy to work the Cayman Islands and G3TXF on Mauritius.

My multiband fan dipoles in the loft.
Only used with QRP for EMC reasons.

Ionospheric propagation conditions were lousy with no sunspots and a geomagnetic disturbance the night before.  Fifteen metres was closed so it was a 80, 40 and 20m only event.

What you quickly realise is that many stations are using the full legal limit, so might be running 400-1,000W. If they aren’t very strong they probably won’t hear your 5W QRP station - fact of life!

So there were lots of “NR?” being sent for repeats, if they came back at all.

I went to bed about 23:30hrs and set to again at 06:45hrs. At that time 20m was dead, but there were some 40m and 80m contacts to be had and I managed to get VE in the log again. I also swept up the remaining G, GD, GI, GJ, GM and GW HQ stations as these were quite easy with 5W.

The antennas were my loft-mounted fan dipoles for 40-10m, which beat the outdoor multiband 80-10m end-fed half wave and W5GI dipole on 20m - go figure. I used the EFHW or W5GI for 40m and 80m, depending on which gave the best signal.

The “nearly” category included QSOs with Singapore, Australia, Gibraltar, New Zealand and Zambia - I just wasn’t loud enough. But hey, Cayman Islands, Canada, and Mauritius on 5W will do me!

Next year if I do it again with 5W I might put up a monoband 20m EFHW and perhaps use my Yaesu FT-817 (I don’t currently have a CAT cable for it), although the IC-756 Pro3 band scope is handy and it is an excellent rig, even when backed off to 5W. Logging/keying was with N1MM+, which worked flawlessly.

All good fun and 30 Qs won’t win it ( the highest QRP entry so far is C6AKT with 92 QSOs), but my unconfirmed overall score was higher than the 2017 winner’s, so not too shabby.

The Commonwealth Contest is a great one-weekend opportunity for UK stations to work DX with no big pileups.

IBP beacon predictions for UK using ITURHFPROP

Over Christmas and the New Year I got better acquainted with the ITURHFPROP propagation prediction software.

This is the software that Gwyn G4FKH uses to produce the monthly HF predictions for RSGB’s RadCom magazine and which was produced by the ITU.

Gwyn adapted it to produce a graphical output, which can be found at www.predtest.uk

But I was intrigued to see if I could use it to produce some predictions for the worldwide International Beacon Project (IBP) chain that runs on 14, 18, 21, 24 and 28MHz. I won’t go into too many details here about the IBP network as you can find all you would want to know at http://www.ncdxf.org/beacon/

The good news is that as they all use the same power output (100W max.) and the same antenna (pretty much a unity gain or +2.15dBi antenna) it should be fairly easy to model what the expected signal strength should be here in the UK.

ITURHFPROP runs in Windows and uses an input file with all the parameters needed, such as your QTH lat. and long., the transmitter’s lat. and long., required SNR, bandwidth, smoothed sunspot number, power etc etc.

You then run it from the command line and it creates an output file.

I ended up creating input files for each of the beacons, and a spreadsheet that can take each of the output Field Strength (dB(1uV/m)) figures and convert them into S units.

I then created a batch file that runs all of the predictions for each of the beacons in one go. With an I5 processor this takes less than about 15 seconds.

The spreadsheet then formats these so that I can produce the chart you see on the page. This may sound long-winded, but it really doesn’t take long. I’m sure someone with Python programming skills could automate the whole thing, but I wanted to make sure it worked first before going any further.

The end result has proved to be quite accurate and shows that at this point in the sunspot cycle we can’t expect miracles in terms of hearing the distant beacons. While some are audible on 14 and 18MHz, generally we are not hearing much above that.

In the UK, what have been audible quite regularly this month on 14.100MHz are RR90 (Novosibirsk, Siberia), CS3B (Madeira), 4U1UN (New York) and occasionally 4X6TU (Tel Aviv).

It also shows what the effect of an elevated K index can be as this generally means the beacons are less audible (if there at all) as the MUF declines. We have been suffering from the adverse effects of coronal holes recently and that isn't going to end any time soon.

You can see the predicted MUF over different path lengths in near real-time using the graphed ionosonde data for the UK at http://www.convectiveweather.co.uk/ionosphere/graphs.php?type=live

This was produced by fellow RSGB Propagation Studies Committee member Jim G3YLA and his colleagues at Weatherquest in Norwich.

I’ve posted the predictions here to see if they are of interest to anyone. If they are I’ll update them monthly.

To view the January IBP prediction chart full size just click here.