A Mastodon Oddity

A couple of days ago, I thought the good people of Mastodon - a calm, peaceful place - might like to see the latest iteration that I've drawn of the Marconi 'Carnarvon' VLF station as it appeared between 1925 and the station's closure in 1939.

The latest version isn't some fundamental new interpretation. It's just a more realistic positioning, rather than a merely indicative one, of the final VLF antenna (initially intended for shortwave use), operating from the 'GLJ' valve transmitter (in pink).

Here's what this now looks like (click on the image for a full-sized view):

Since the first iteration of this schematic, I've learned that the No. 2 antenna was fed not by buried twin feed, but by a caged twin feed that ran along timber telegraph poles. This has, by a very good stroke of luck, been confirmed by a 1925 photograph that had narrowly escaped being put in the bin after its keeper had died.

Caged twin feed to No. 2 antenna on poles at left. Poldhu pots were installed at regular intervals along the route and equally regularly needed replacement. The antenna in view is No. 1.

 

Photos from the same collection also show distinct arms or standoffs emerging from near the top of the northernmost line of masts for No. 1 antenna. They are absent on the southern line of masts and it has long been the case that no concrete anchors or bases for GLJ/No. 3 have ever been identified.  This almost certainly means they were for suspending the GLJ, No. 3 antenna from No. 1's masts. 

Images in the Oxford Bodleian collections confirm the alignment and extent of No. 3 antenna, though it was experimentally both shortened and lengthened somewhat early on in its life, owing to considerable interaction problems across the No.1 and No.3 circuits - something that took nearly two years to fully resolve.

So that's all good and, hopefully, useful to the wider world; the IEEE History Center certainly think so, in a plaudit they issued at the end of last year.

Having posted the schematic on Mastodon, I very quickly received a question from VK6VCC, simply asking "is this your source?", providing a link to a Welsh Country magazine article that contained very similar diagrams.

 

I have to say I immediately didn't like the tone of the post which, in the absence of any further question or comment, could be interpreted as suggestive of some skulduggery on my part.  It was pretty likely - confirmed by doing it myself - that VK6VCC had conducted a 'reverse image' search on my schematic and had been presented with the similar image linking back to the magazine.

Output (right panels) of a reverse image search based on my latest schematic (left panel)

 

I'm not sure about VK6VCC's motivation in conducting a reverse image search and it's possible he didn't mean to imply some odd 'goings on' at my end at all. But he certainly seems to have entirely missed, in asking his question, that the author of the Welsh Country magazine article was, er, me!  

Hardly surprising, then, to find a reverse image search match. It even gives my name at the end of the article, and that I'd written a full, 230-odd page history of the 'Carnarvon' station, which is available, now for free (but fully copyrighted, all the same), from here (about 70Mb).

VK6VCC didn't respond in the time I gave him as to his intention in asking his original question. Looking him up (13/01/2025), he says he's a new operator. He may wish to spend more time reading and less time rushing to punch a keyboard. He is now one of very few people over the past year I've blocked on Mastodon. He didn't, after all, ask anything useful, because you don't ask the author of a work whether their source is their own - and same - work!

K4FMH - Analysis, not Assertion.

 

The article below is well worth a read, mindful that the general attitudes behind such assertion-making is something that doesn't help the hobby in the long term, because they are simply wrong.

We've also seen something similar with the RSGB in its claims of membership benefits; for a long time, it claimed, and may still claim (I don't know) that members were "more likely" than non-members to gain planning consent for antennas. 

When asked, no evidence was available to support this claim and it's not surprising; they would have had to run a double-blind assessment with the exact-same officers and exact-same applications for hundreds if not thousands of cases to reach statistically-valid conclusions. Designing such an experiment is, itself, highly non-trivial and clearly beyond the abilities of the RSGB board.  

There is also the point that the claim related to membership per se - advice about planning matters is, of course, available from many other sources not related to RSGB membership, including now from admittedly questionable but improving AI. 

So the real test would be more along the lines of those members who gained consent after using the RSGB planning service and those who were not members but used an alternative advice service of some sort. You could also assess the success of members who, despite that fact, didn't use the RSGB planning service.

https://k4fmh.com/2025/01/12/are-amateur-radio-national-society-members-really-more-active-than-non-members/

VLF Receive Loop - More Verticality?

I had a great time listening to SAQ's centenary VLF transmission over the weekend, despite my battery connections for the amplified (WellGood) loop breaking. I had a spare but forgot about it, instead using my teeth to strip wire and make bare connections to the battery!

If you missed the transmission, you can hear the whole thing here.

My standard receive set-up for VLF is an 18m-long loop of RG-213, slung as neatly as I can manage in tree branches, with vertical sections of the loop about 4m tall and two horizontal sections of about 5m each. The lower horizontal part simply runs along the ground to the amplifier. 

The standard configuration, pole or tree branch supports.

 

This set-up gives a pretty consistent S6-7 signal strength for SAQ and there is no need for any improvement.

But I did come to wonder whether reducing the ground-run of the loop might improve things somewhat. I thought I'd try to lift the centre up to around 6 metres, thus giving, allowing for sag, roughly 14m vertical sections and the rest as a now clear-of-ground, roughly 4m horizontal section running to the amplifier. 

 

I aligned the whole thing on the German time signal DCF-77, at 77.5kHz CW. 

The signal from this 'pseudo-delta loop' arrangement was very slightly under S9+20.

The tall vertical version under test.

 

I then lowered the whole thing to around 4m at the apex,  so that it resembled what I would normally use (see first image), but slung from trees. The signal was indistinguishably different from the previous arrangement, despite having significantly shorter vertical runs.

When the loop was allowed to collapse such that it had no vertical runs to speak of, the DCF-77 signal dropped to around S9, or some 20dB below where it was with reasonable vertical runs.

It seems that the lesson from this quick experiment is that so long as there are a couple of metres of vertical run and that there is approximate alignment with the direction of the transmitter, then that's all that is needed for successful VLF operation. The length of the loop, however, remains important; with a standard 3m-circumference loop, SAQ is only weakly detectable whilst it is strongly detectable at S6-7 on the 18m loop.

The benefit of using a pole rather than trees, though it's something else to carry, is that it's somewhat easier to stick the whole thing in the air - especially if you have no handy trees available or, as is often the case here, those that are available have rotten lower branches!  The 'pseudo-delta' shape can be formed simply with a couple of ropes and tent pegs tugging at the lower corners.

Vertical vs. 2-Ele Quad. Nuts, right?

Sometimes, the heavens do align such that everything becomes possible.

First stroke of luck this week was finding someone willing to tell me about their WSPR system. G4HSB told me about his 2-ele quad and that it was firing along 300 degrees. He was using 0.2W output, as I was.

The second piece of luck was that we've had a long run of calm and relatively warm weather - perfect for mounting verticals in wet sand on the beach and standing around there, in the moonlight, for two hours!

Beach WSPRing last autumn, when there was light to see!

 

I had hoped to run a comparison of the two antennas as received by VE6JY, but unfortunately, I hit on a bit of bad luck as well, inasmuch as geomagnetic conditions were slightly rough on the evening and that station just wasn't hearing anyone from Europe at the time.

Oh, and there was the inevitable 'how did I forget that?' moment, when I realised I had left the USB battery keep-alive board at home!  I overcame that problem by taking a charging cable from the car's radio set-up and then connecting my phone to charge from the battery pack to prevent it shutting down when the WSPRlite was in idle, between transmissions. It worked!

Yellow line shows the 300 degree line from my beach location.

After a beautiful, full moon couple of hours on the beach, it was time for the analysis. I start by taking the inherent 'design' gain of the 2-ele quad to be 11dB. It might be a dB or so higher than this in practice, but that's academic for this purpose. My 3-wire, elevated vertical is assumed to have a design gain of 1dB.

Because of the complexity of terrain modelling (read: it takes a lot of hard work to do), I'm not even going to attempt that for the quad; for my vertical, it's irrelevant (vertical radiation and a flat surface from source to infinity).  

Here's the plot of how KD2OM, a well-known station for having good ears and always available, on a bearing of 290 degrees from G4HSB, heard both stations. Blue is my vertical, orange the quad.

KD2OM hearing the vertical (BLUE) and 2-ele quad (ORANGE)

 

I couldn't find a station that was further north in the US such that the 10 degree error in beam heading was reduced for G4HSB. But it turns out that, looking at a model for the quad, that the drop-off in gain (line and red dot indicate) is very low - only around 0.2dB or so. The model also predicts just over 12dB gain, but I'm still using 11dB for my calculations, because that, being a subtraction from the received signal, is a best-case scenario for the quad and allows for the possibility of a less-optimally spaced antenna. If you want to assume 12dB instead, just drop the orange line for G4HSB by 1dB in the plot above; it's somewhat pointless!

If I compute the mean for both antennas, there is 0dB difference; exactly the same output on average, though my vertical achieves significantly stronger signals, by around 7dB, on numerous occasions. The much stronger dips in the vertical's signal are likely indicative of a changing propagation path that is probably at somewhat higher angles relative to the horizon; the quad accesses a wider elevation range, though with a significant null, than the vertical, the latter putting out most of its radiation much closer to the horizon, including, often very importantly, angles at a small fraction of a degree to the horizon itself.

2-ele quad model in MMANA-GAL (14MHz) at 60 feet boom height, good ground.

Modelled radiation pattern for the sloped-radial, elevated vertical (14MHz), 1.4m feedpoint height, perfect ground. At a beach, where the ground conditions change rapidly from near-perfect to far less so, the azimuth pattern is much modified from that in the model and can be considered to have a beam-like pattern, the rear diminished and the front enhanced in gain (this is proven by experiment as well as logic!)

So that was KD2OM's account. The only station that I did reach from the beach who was further away in the US and on a better-aligned 300 degrees to match the quad's beam direction, was KA7OEI-1, another high-quality 'regular' on the WSPR scene. Here's how that looked:

Vertical (BLUE) and quad (ORANGE) heard by KA7OEI-1.

Now the beach vertical wins out by a significant 4dB or so on average, almost certainly a product of the shallower angle of propagation to that station that may not be available to the quad, but is available to the vertical. I haven't drawn lines between the points in this plot, because there is a gap where my signal wasn't heard for a short while and the software doesn't cope well with gaps for some reason, drawing a misleading line in my favour. During this short period, the quad does get through to the receiver. That could be down to propagation changing to a higher elevation for a while, due to the rapidly-varying geomagnetic field at the time. There may be other explanations!

Do also note that the MMANA-GAL model shows what is certainly true - an antenna-and-ground gain of 6.65dBi, this partly being a product of a near-perfect image in the damp, saltwater-laden sand beneath the vertical. The quad, meanwhile, might benefit from ground reflections, depending on the site elevation and ground layout in the beaming direction. We can only allow for so much in practice!

Sure, if you are nowhere near the sea, this is all rather academic; you'll need a big antenna to achieve the same thing as a simple vertical at the beach. The point being that coastal and other watery environment operating is widely overlooked in the ham press, despite being a truly superb way to get places other sites cannot reach!