Winter mesospheric echoes drift by...

I was glad I checked the MST radar site in mid-Wales today because, if you look on the extreme left hand side, you can see some weak, but coherent polar non-summer mesospheric echoes, or simply winter echoes, were present early morning to mid-afternoon yesterday.

There are no singificant meteor showers at present, and it's a month since the Quadrantids peaked in early January.  So the explanation is interesting to think about.  I don't think anybody really understands these winter echoes, beyond those that arise in association with meteor showers which is, in itself, not a perfect correlation.

One possibility is that a modest-sized bolide entered somewhere over the planet recently, leaving a trail of charged debris behind it.

I'm not able to check data from the MST archive this morning, but a fascinating visual phenomenon that occurred over Wales four days after the Chelyabinsk bolide entry in February 2013 was this stratospheric layer; it's ablated fine dust from the meteorite's passage through the atmosphere.  It travelled eastwards, the 'long way' around, before first being spotted (only once, with no photo) from Alaska, then once in Ireland (with photos) and Wales:

Overnight WSPR report, 2019 January 29-30

Very quiet geomagnetic conditions, with a Kp of zero.  As a result, there were no anomalous spots from DX distances (and only one or two spots from the near-continent).

Image: Tromso Geophysical Laboratory.

2E0WVY, who seems to be running a two-mobile whip dipole in the attic from his Liverpool QTH, provided something interesting to examine once more:

The 2E signal appears from the noise at 20:58UT, and then trends very slowly upwards until a very stable signal is received after about 01:10UT.  There is a small reduction in signal between 23:40UT and a minute or two after UT midnight.

The appearance, variation and slow increase through the early morning seems to correlate well (but again, may not necessarily be caused by), with the variation in the vertical field component, more particularly, the field variations as recorded at Bjørnøya (Bear Island), half way between the northernmost tip of the Norwegian mainland and Spitsbergen.

This contrasts with the 2E signal as received the previous evening:

Vee beam progress (antenna abandoned)

At the end of last week, I had good results with an experimental V-beam antenna, showing a 4.2dB improvement on my vertical delta loop.

Yesterday afternoon, I set up the V-beam with better wire and, so I thought, a better, slightly wider angle of separation of 75 degrees (accurately measured using a calibrated set of homebrew timber callipers) between the two legs of the antenna.  This is about the angle given in the gospels of various radio 'bibles'.  Wire length is 47m per leg.  It's important to make each leg closely of equal length.

 Hinged timber with ~52 degrees marked.  Simply hold up to the sky and align with the wires for reliable separation.

Well, I gathered plenty of data, which I won't bore you with this time, but that resulted in the V-beam being typically 1dB weaker than my vertical delta loop.  Only on one occasion, to the most distant west coast US, did the V-beam show a slight, 1dB improvement over the delta.

What had happened for a previous 4.2dB improvement to collapse into a 1dB loss over the delta?

This morning, to try and find an answer, I turned my hopeless modelling skills to the question of wire spacing.  Happily, MMANA-GAL already has a 160m dipole input into the software, so I set about modifying that to two 47m long wires, 10m high at the feed point, sloping to 1.5m at the far end of the wires; exactly as my antenna is configured.  If you can, get the ends of the wire as high as possible; just 0.5m higher makes a significant difference to overall gain.

Well, it quickly became apparent that the V-beam, which relies critically on interactions between the radiation patterns of the two legs of the antenna, is quite sensitive to wire separation angle.  So sensitive, that simply guessing the angle just won't do if you want the cleanest pattern and best gain.

When I modelled my 75 degree wire separation, the horizontal pattern looks like this, using real ground characteristics:

The gain is very good, despite the pattern clearly not showing the best interaction between wires. 

With a separation of 53 degrees, the pattern snaps into a very clean beam, as we want:

Now, although the pattern changes are not dramatically different, there is a 2.05dB gain difference between a 75 degree and 53 degree separation.  I suspect that this gain difference, plus the added gain at moderately low elevation angles from the V-beam, relative to what is available from the delta, is what explains last night's poor performance. For the first test with the V-beam, I had used a narrower separation of about 57 degrees (which was a semi-guess of what would work).

The other breakthrough I made last night was perfecting the matching.  Whilst a 2:1 balun gives a good match with 300 Ohm twin wire, I found that matching on bands other than 17m was not possible.

So I went back to the original, 4:1 balun, but this time, using a much better, MFJ 941E match box than the poor L-match unit I was using previously.

Now I found I could match the antenna on all bands from 10Mhz to 50MHz (with the wire separation remaining constant for this test).  At 50Mhz, with the wire separation reduced to 32 degrees, this is nearly 8 wavelengths per leg - and 15.3dBi gain.  Putting 10W out would bring you to 339W EIRP!

It's now decided to start snowing, with a moderate wind here in Wales, so I will wait until Wednesday (30/01) before I venture out to test the V-beam again with the ~50 degree spacing.

UPDATE: I have abandoned the V-beam!  It has too high a radiation pattern for normal deployment heights.  This can yield positive comparisons against my delta on occasion, but in general, the V-beam is some 4dB down on my delta across all stations at which the main lobe is aimed.

Excellent results with a 4:1 balun and the MFJ 941E matching unit.

18MHz results also very good, without wire spacing adjustment.

21MHz continues to look good, wire spacing unadjusted.

24MHz still looking good (poor frequency selection, as I was running late!)

28MHz still producing the goods - now at over four wavelengths long!

50MHz, still without wire spacing changes (which are certainly needed at this frequency)

Final test at 10MHz, no wire spaceing change, still perfect matching.

One Thousand Kilometres

The overnight WSPR report is a little different this morning.

There were no anomalous DX WSPR spots at 14MHz, the last DX station heard being 5W from K6MCS (8130km), at the unusually late time of 22:44UT.  That spot may be explained by a limited patch of ionisation at mid-Atlantic longitudes, as the OVATION model suggests:

The field was fairly quiet, at Kp ~2, though the auroral oval disturbance around magnetic midnight was fairly pronounced.  The Z component underwent a northerly deviation at very high latitudes, but a small southerly deviation at sub-Arctic latitudes:

Image: Tromso Geophysical Laboratory

The peculiarity with this overnight period was the incredible consistency of the distance - just over 1000km - from which almost all WSPR stations were heard, as this plot, first with a logarithmic scale (broadly representative of global DX distances), and then a linear scale (with purple mean line) shows:

Clearly, the spread of distances is very narrow.  The mean is 1046km, the mode is 1087km, and the median also 1087km, with a maximum distance of 1349km, and a minimum of 329km.  As the log plot demonstrates, considering the global possibilities, signals were arriving only from one, short distance, which I guess is broadly a single hop distance.  I suspect this is also explained by the ionisation patches that occurred equatorward of the auroral oval.

I'm quite happy to have gathered all that data, as yesterday was quite windy, with sustained winds of 60km/h, and frequent gusts of 100km/h.  The cable tie that attached the pole to the windward rope stay on my delta had broken away just as I left for the day out.  With the wind load being very heavy on the pole, making it impossible to retract the fibreglass sections, there was nothing to do except use a stay not bearing any load and reposition it into wind.  That kept the antenna going until this morning, when a full repair was completed by 08:00!

Overnight WSPR report, 2019 January 25-26

Raul, LU8DPV, started transmitting 5W 14MHz WSPR before midnight last evening, just in time to reveal an interesting variation in signal strength in the depths of solar minimum, midwinter night:

This is how the Z component of the geomagnetic field varied last night:

The peak at 00:04UT in LU8DPV's signal correlates nicely with a rapid northerly deviation in the vertical component.  The trough in the signal seems to correlate well with a recovering field after a minor south-going rebound at around 01:00UT.  The second peak at 02:00UT does not correlate with any particular event, when the field was slowly undergoing a restoration to quieter levels.

The most interesting aspect of the signal is perhaps not the variations, which are generally explained quite well by a changing field, but the fact that LU8DPV was the only station, apart from the very short-skip G0CCL and one or two other, short skip European stations, that I heard overnight:

G0CCL's signal, though not heard very frequently, reappeared and also tended to show the same peaks at 00UT and 02UT (blue vertical lines).

V-Beam update.

You've got nothing on me, Bear Grylls!

Well, like a very large, ugly frog, I spent a few hours stomping around in the marsh, in the dark, last night.

The aim?  To bring my new, 14MHz V-beam antenna idea under control!

Initial tests with the ~2-wavelength V-beam have been very good; over 4dB improvement on my vertical delta loop, despite being poorly matched at the time.

The problem since then has been in finding difficulty in (a) providing the correct wire lengths (not as easy as you might imagine) and, (b) providing the correct matching arrangements.

As is typically the case, there is not very much reliable, well-considered material on the internet from others who have tried the V-beam to assist us.  A look in the usual radio 'bibles' give happy accounts of how easy it is to provide any multiple of a wavelength for each leg, hook it up to some twin wire and an ATU and, voila! 

Sadly, real antenna life is rarely like this. Hardly anybody, for a start, has an antenna farm on a marsh.  The ground is very different to American deserts or rocky hills.  That last fact also probably explains why so many angry Americans rage against a V-beam, saying it can 'never be better than a dipole' and has 'exaggerated gain'.  All this anger is invariably based on amateur modelling, not real antennas. 

The main problem I had was two resonances falling either side of the 14MHz band.  One was at about 16MHz, the other down somewhere around 13MHz.  At 14MHz, the impedance was only about 8 Ohms.  No ATU I have will match that kind of presentation, nor would it be a good antenna.

I tried first to change the wire length from 50m to 42.3m (two lambda).  Sitting dejectedly on my old ammo can, I watched the SARK-110 draw its plots, finding that this resulted in almost precisely no change.

I dismantled the antenna, removed the wire and cut new wire lengths of 52m for a starting point, with the intention of cutting until I found the best length for about 2 wavelengths at 14MHz.  I was lucky in that, as I got to to the end of the second leg, my wire ran out!

I had calculated this ~52m length from observing the antenna's matching at 30m with the earlier length of 40m, which was fairly easily matched.  So I guessed that 4/3 wavelength leg lengths might be appropriate also for 20m, which would be (4/3)*21.25m per wavelength (allowing a velocity factor of ~0.94, a total length of 53m.  Close enough, and I happen to run into a ditch beyond 52m!)

After failing with a 1:1 balun, I then tried to use 6:1, 4:1 and 2:1 baluns in reverse, to step up the impedance.  This worked reasonably well with a 2:1 in reverse, but I could only get down to 1.4:1.  Certainly more than usable and, with 300 Ohm twin wire as the feed line, of no loss consequence.

SARK sweep with 52m of wire and a 2:1 balun in reverse.  Good, but not good enough!

But, I decided to press on for a better match.  By now it was even more foggy and drizzly.  At least it was warm!

Of course, life could not be simple.  As I grabbed the second leg of my antenna to cut it to a new length, I found it was loose.  The wire had broken at the feed point!  Arrghhhh!  Take down the pole, remove the old wire piece, resolder a new spade connector, start again...

After more cursing and falling about in the water, I came into the shack to see how a new length of 47m would perform. This was really just a case of seeing what difference 5m less wire would make.

After plenty more switching baluns and ATU controls around, the magic happened using a 2:1 balun in the normal, step down sense - a 1.09:1 match!

Finally, this arrangement came in at a perfect match!

Well, that is the situation with my experimental wire, which is very thin enamelled copper wire.  I have now plenty of lightweight PVC-insulated wire on order from SOTABeams (very cheap at about £8.50 per 100m).  I guess the wire length will be a bit shorter for that wire type, but I think 47m will still be fine to start with.

Finally!  47m of wire per leg, and a 2:1 in the normal, step-down sense.

Incidentally, the included angle between legs is roughly 75 degrees, which can only be deployed reasonably accurately by measuring with a compass and noting some obvious tree or gate in the distance to walk towards with your wires.  There is no need to be super-accurate, anyway, but there is a tendency to under-estimate the angle if simply guessed.

Measuring out new wire.

The weekend is busy, contests likely, and a modest storm with gusts up at around 100km/h is due on Sunday.  So I hope I can test the V-beast when things calm down, at the start of next week.  I spent an hour or so in 80km/h winds and heavy rain showers measuring some new SOTABeams wire out this afternoon.  Only directly into wind could I do this successfully!

You know you're a real radio operator when...

...you have spots of solder on your morning mug of tea:

No milk or sugar thanks, just a drop of solder...

Overnight WSPR Report, 2019 January 23-24

A modest sunspot group has appeared on the solar disk over the past day or so, but remains a member of the old cycle.  Kp reached 6 for a while:

Image: SolarHam.com

Image: IRF/Kiruna.

G0CCL, which is spotted from here throughout the night when the disturbance is more modest, was apparently suppressed by the field deviations last night.  Interestingly, a very strong and momentary rebound of the signal from the noise was heard at 19:46UT, just ahead of the first of two sharp southerly deviations in the vertical field component:

An interesting and very anomalous spot came in at 05:58UT, from KK4PP (running a long wire and just 10mW).  Whilst this was well into the morning period here, KK4PP was anomalous in being the only US station heard and, indeed, other than the very close-by G0CCL, the only station heard at that time.  The first US spots do not come in under quiet conditions at the moment until mid- or late morning.  Note also the anomalous spots between 21:44UT and 21:52UT from CX8AT (confirmed as transmitting continuously), which came in during a sharp, north-going rebound in the field, ahead of a larger southerly swing that came about 23:00UT:

The fascinating thing with many of these spots is that they clearly occur due to some smaller or larger areas of ionisation permitting propagation, but that they are stable enough under quite rapidly-changing field conditions to allow a  2 minute-long WSPR signal to remain coherent enough for detection.  The other intriguing question is why singular stations like KK4PP are 'selected' for propagation from the very large population of active WSPR stations in the USA.

Radio and 'The Youth'.

One of the most intractable problems with amateur radio is how to make the hobby appealing to those who are much younger than the typical operator.

A survey conducted in 2015 by the RSGB (fairly recently removed from their web site), found the typical operator in the UK was highly skewed towards the late middle age and retired groups.  There were essentially no members in the teenager to young adult cohort.

The problem is difficult to crack because not only is radio a niche, technical hobby, but also that the people charged with solving the problem - if indeed they really are passionate about cracking the problem at all - are themselves the white, middle class, retired people that typify the membership.

Now, my daughter passed her MW6 Foundation call sign at the age of 11.  She quite enjoyed studying the simple course, helped along by friendly, local club members.  But she has never really taken much of an interest in operating.

SSB is a killer for a youngster, as many older men find talking to a youngster highly difficult and not the kind of slow-paced, three hour long discussion about an RX-210s valve that they managed to find for their ageing RT-999 transceiver.  Sadly, some are rude and don't continue the conversation when they realise the other operator is young.  A very few are friendly and try to at least say words of encouragement.

So, imagine my delight last evening, when I dimmed the lights, tuned in to Radio Uganda with the SDRPlay RSP1a receiver, and placed the earphones on my daughter's head.  Click!  You could see the mind focus its attention and imagination on what was coming through (inevitably, it was 'Praise the Lord Jesus' stuff).  I showed her that peaks meant radio stations, and off she went on her own, exploring the SW world.  This is a lot more interest than the amateur radio licence ever generated.

Encouraging signs...

That's a very big hint on what youngsters find appealing. They want to know about the world, and SW radio brings them, as it did all of us, a different perspective from that which we are fed via domestic TV, radio and the internet.  Arguably, with the internet AIs deciding for us what we should see and hear, SW radio returns the decisions on what we want to know back to us.  Sure, most SW stations are propaganda disseminators, they always have been.  But at least we know that, can see in total what is on offer, and can pick what propaganda we want to listen to on any given day.

I've made the argument on various Facebook groups that RSP1a receivers or similar units should really be sent out to schools under some grant-funded project operated by national societies.  It may appear there is no money, but the likes of the RSGB are linked to various charitable spin-offs, some of which pay their managers quite handsomely (the RSGB itself paid its General Manager,until recently, over £60,000 per annum (and refused to say what the exact salary was)).

So there is plenty of money to buy a bulk order of SDRs and send them out to schools - if we have the real intention of turning back the tide of a membership profile that is getting older with each passing year.

It won't be easy, because schools are very closed communities that, these days, are more concerned with 'security' and 'safety' than engaging with the wider community in a rational, simple way.  They may well welcome very much someone giving them an SDR, but will they allow someone to install a wire antenna for it?  Will they allow someone to demonstrate it?  Will they simply use it for a month and then forget it, because the curriculum is just so busy and there's no time, blah, blah, blah?

That is why we need a recognised organisation behind any such promotion.  I hope the RSGB and others take the hint and actually do more than just run workshops for those who already have a Foundation licence.  That is just easy preaching to the converted. 

Long WSPRnet outage

WSPRnet has been getting progressively slower recently, often failing to load at all for several minutes.

Today, it's been inaccessible for most of the day so far.  This happened a few years ago, and took quite a while for the site owner to address - something he apologised for in due course.

This is a serious weakness in the WSPR system, in that all data ultimately derives from one source: the WSPRnet database.  When that goes, all the alternative databases and plotting systems, such as DXPlorer, the software associated with WSPRlite transmitters, also go down the hole.

WSPRnet is, in fact, quite reliable given the demands placed on it.  But I think it's now time someone looked at how several alternative reporting databases can be set up, to avoid us either suffering too much from long outages like the one currently underway, or from the owner of the site losing interest, dying, or some other of the many things that life tends to result in.

UPDATE: I made much the same comments on a Facebook page dedicated to WSPR.  I was interested to see one person come forward to offer an alternative server, but I am not sure if the interest will last past this outage (fixed after a full day's outage, but collapsed again by the morning of 23/01), or whether it will be possible:

Overnight WSPR report, 2019 January 20-21

Extremely quiet geomagnetic conditions overnight (Kp ~1), with nothing other than the passing auroral oval causing deviations at Arctic latitudes:

Image: Tromso Geophysical Laboratory.

This led to the quite unusual condition where 14MHz propagation ceased completely from mid evening until quite late in the pre-sunset period:

It's interesting to consider that G0CCL, which I postulated yesterday must be geomagnetically-induced very short skip propagation, was also entirely absent overnight.  It was also the case that G0CCL's transmissions ceased being heard during much the same period as I ceased hearing anyone:

Even more interesting is the case of 2E0XLG, now using an OCFD running N/S.  Being at very short distance (158km) from me, I assumed, because of the very fact of the expected ground wave, and through the very constant signal received the previous day, that the geomagnetic field would not have very much effect on this signal.  It's important to note that I later learned Chris had changed his antenna from a mobile whip arrangement to the OCFD, so this may complicate matters.

In fact, 2E0XLG's signal entirely vanished to the rest of the world in the same fashion as other stations last night:

So it seems that, even at this very close range, propagation at 14Mhz during midwinter, solar minimum night is reliant on some degree of geomagnetic influence. 

V-Beam vs. Vertical delta loop

This winter has been very unusual in producing very few windy days.  At the time of writing, we've had only one or two mildly windy days in over a month.

This is a good time to test large wire antennas!

I decided I would make an unterminated (i.e. bi-directional) V beam by taking advantage of strong moonlight, a spare spool of thin copper wire used to make an LF antenna over Christmas, and a big field!  Incidentally, a ~320m spool of thin copper wire is a very cheap way to test long antenna principles, at just £13 each.

There is enough room to fit at least ten wavelengths at 14MHz, but this would probably lead to birds' nest tangles and test my patience.  So I settled for what I could fit between the shack and the nearest boundary, which was 50m, or two-and-a-bit wavelengths at 14MHz.

Red lines show the 50m legs of the V beam, yellow is midline beam direction (~280 degrees)

Theory says that this wire length should yield about 4.5dBd (6.65dBi), which is only 1.7dBi less than a 3-element Yagi for the same band.

Now, remember, all you have to do to deploy this antenna is attach two long wires to a 300Ohm twin wire dipole centre, lift it up as high as you can to make a sloping 'vee', connect to an appropriate balun (a 4:1 is a good starting point), feed through an ATU if needed, and away you go!  Changing beam direction is nothing more difficult than lifting two sticks out of the ground and replacing them in the required beam direction. 

This, folks, is infinitely easier and, mindful of the need for a substantial metal pole, if not a tower, a minimum of £800 cheaper than erecting a Yagi.  And with suitable wire, the V antenna is invisible.  Unlike a Yagi, which would need a portable pole or tower, and a van or trailer to transport it, the V beam can be wrapped up into a shopping bag. The drawback is simply one of the sheer space required for 360 degree coverage.

The thin copper wire I used was ideal for the fibreglass pole.  It is quite strong in tension, but doesn't need a lot of tension to keep it from sagging.  The balance of forces between the pole bending and wire tension is very good.  But it probably wouldn't survive strong winds for more than a season or so in a permanent installation.

Each leg of the V beam ran from about 9m at the pole to 1.5m at its furthest points.  I guessed an angle of about 70 degrees or so, appropriate for this multiple of wavelengths (angle decreases as leg length increases).  I put the centre between legs at a heading of about 280 degrees; ideal for all but the far west of the USA (which is about 305 degrees from here).

So, how did it work?  First off, the 50m leg length was a little tricky to match.  Here's the plot of the impedance and SWR at the end of a 4:1 balun, at both 14MHz and, below, full HF scan:

At 50m, it's expected the antenna is too long for 14Mhz, but that can easily be corrected later.   After cutting it to 40.3m, I found the situation was much the same!  After experimenting with this 40m leg length at 10MHz, not 14MHz, and running it through a 1:1 balun, not a 4:1 version, I found that matching was straightforward.  It seems you should have legs that are about 4/3 wavelengths long.  For update on wire length, see here.  The full HF scan shows the antenna has interesting possibilities at many wavelengths without demanding too much of any tuner that may be needed, if at all.

Despite the mismatch, I ran the V beam against my vertical delta loop, which itself is very often the only one to hear many stations during midwinter night.  It's quite hard work to plough through all the data, so I settled for five stations across the US from east to west.

In the case of AA7FV, the there was one spot where the V beam did not hear him when the vertical delta loop did.  But there were five occasions when the V beam heard AA7FV, where the delta did not.  Those are very important missing spots, of course, because our effort is aimed at hearing people far away.

For AA7FV, the V beam was 2.93dB better than the delta loop, mindful again that the V beam often returned a spot when the delta didn't, which is more important in the end.

For AL7CR, the V beam heard him only once, at -22dB, where the delta didn't hear him at all.

For VE7WAE, there were four occasions out of eight total spots where the delta did not hear the signal, whilst the V beam did.  The V beam was on average 5dB better than the delta.

For KD6RF, there were four out of eleven spots where the delta didn't hear the signal, and only one where the V beam didn't hear it.  On average, the V beam was 2.7dB better than the delta.

For AD0MO, of seven possible spots, the delta captured all of them, whilst the V beam missed one.  On average, the V beam was 5dB better than the delta.

For KE6NFH, of eight possible spots, the delta didn't hear five of them, whereas the V beam missed only one (a spot that was heard by the delta).  On average, the V beam was 5.6dB better than the delta.

So, taking all average differences into account, the V beam was 4.2dB better than the vertical delta loop, which is pretty much what theory predicts.

Interestingly, on later examining a greater number of east-coast US stations, I found that the difference between V beam and delta became gradually greater as late afternoon turned to evening, which is when signals are arriving at shallower angles.  The difference was quite consistent regardless of station selected, so this suggests the V beam has a very good, low angle pattern, no doubt improved by being located in this case on a marsh where fresh water is essentially at the surface.  The consistent difference between antennas on the west coast US stations supports this low pattern hypothesis.

Next step is to increase the length to 53.3m so that the antenna (hopefully!) becomes much better matched.  I've also ordered a handful of non-inductive 500Ohm, 50W resistors from China to try a terminated (unidirectional) version, though this has somewhat less gain and is more difficult to deploy for 360 degree coverage than the bidirectional version, which only needs half the space.

A tale of two short skips.

Not an awful lot happening that was unusual overnight, with the field staying at Kp = 1.  But I was interested to see 2E0XLG, who I think is new to WSPR and only 158km from me, appear as a remarkably constant ground wave signal from 0.2W output, throughout the darkness (blue vertical line = UT midnight):

2E0XLG received by GB0CQD (MW1CFN), 2019 January 18-19.

The small positive deviation in signal strength at 21:36UT is correlated to (but not necessarily caused by) a field restoration from a modest southerly deviation at very high latitudes, as indicated by Svalbard magnetometers (line labelled 'bjn' (Bjørnøya, or Bear Island, 74.5 degrees north)):

Image: Tromso Geophysical Laboratory.

This is very interesting to compare with G0CCL, running 5W, at just over twice the distance from me (353km), which seems to be undergoing very intermittent propagation via non-ground wave paths, as the upward curve of much more consistent signals around sunrise suggests:

USB Isolators.

Some days ago, Bas, PE4BAS, told me about USB isolators.  This was in response to a problem with RFI emitted by the Raspberry Pi 3B+ and associated 7" touchscreen under certain circumstances.

I had never before heard of USB isolators so, at just £6.36, I bought one from Ebay.  If you really want to, you can pay as much as £35 (or, if you are really stupid, a lot more).

A few days later, the isolator arrived direct from China.  First, I tested it as a fairly high transfer-rate isolator on the family Windows 7 PC's external backup hard drive.  Apart from the tiny, yet blindlingly-bright on-board LED, no problems there at all!

USB isolator in action.  To relieve stress on the laptop's USB port, support from underneath the board is advisable.

This morning, I decided to risk sticking it in my main station laptop (the Raspberry Pi operates my remote shack).

Result?  Well, it didn't cause Windows 10 to crash, or the computer to burn up - a good start!

Using it with FT8, the isolator doesn't cause any problems at all in either direction.

BUT!  A problem!  Connecting the RSP1a SDR via the isolator leads to the RSP1a not being detected!  So, I will have to consult others to see if there is a resolution to this issue.

This video is a good with/without comparison using other SDR software:

HC6PE - WSPR during Kp=5 event.

Some pretty good geomagnetic conditions last night, with Kp reaching 5:

Image: NOAA/SWPC.

Image: IRF/Kiruna.

Unfortunately, my friend Raul, LU8DPV and I had agreed to terminate the experiment we ran over the past few days.  Raul seems to have noticed the disturbance, as he transmitted for a short time around 22:30UT.  In fact, whilst the WSPR database does not show I heard Raul, my WSJT-X record shows that I heard him twice, at 22:36 (-15dB) and 22:46 (-10dB).  With no further transmissions, we can reach no conclusions at all.

Put simply, we need much more WSPR activity in Latin America!

I was able to follow HC6PE, who is a very regular WSPR operator.  This is how his spots looked:

The disturbance in the field brings HC6PE to an early close, just before 17:00UT.  But he then reappears at 21:24UT at -28dB, climbing rapidly by 21:32UT to -16dB, dropping again by 22:44UT to -19dB, and then finally dropping off into the noise after 22:54UT, where he was not heard again during the night.

This strong and very sharp peak seems to coincide very nicely with a rapid recovery from a southerly swing that continues into a sharp northerly spike, most closely aligned with the magentometer at Leirvogur, Iceland (the green line ending at 'lrv'):

Image: Tromso Geophysical Laboratory.