Now that the blog is back up and running, I thought I'd step back and give the station a proper tour — one piece at a time. When I drew the system diagrams for QRZ it struck me just how much there is here, so this is the first of an occasional series. Where better to start than outside, with the wires and aluminium?

The RF side of G6WPJ (click to enlarge)
Three antennas, three jobs
The Hexbeam. A six-band Hexbeam for 20, 17, 15, 12, 10 and 6m, at 15 m on a crank-up tower. This is the DX and voice antenna — gain, a useful front-to-back, and it turns.
The Batwing. My G4MKT twin end-fed, known round here as the Batwing: two legs of 41.2 m and 29.4 m running east–west in a straight line, fed at the junction through a 36:1 transformer. Each leg is an end-fed half-wave (or a multiple of one) on different bands, which between them gives 80, 60, 40, 30 and 12m. Higher-angle and quieter on the low bands, it's my inter-G antenna — it carries the Sunday GB2RS news on 80m, and through an MFJ-948B manual tuner it gets me onto Top Band too. How well it really does down there turned out to be a story in itself, and it's the subject of the next post.
The vertical. A DX Commander Signature 18: an 18 m fibreglass pole carrying a fan of quarter-wave wires up its length, with a couple of bands thrown in free as harmonics of the longer wires. I built it to the instructions — and then couldn't resist adding Top Band.
Teaching the DX Commander 160m
The trick was to borrow the 60m element, which is already 13.4 m of vertical wire. At its top I added a 60m trap (20.5 µH with 43 pF, resonant at 5.35 MHz), then a 19 µH loading coil, then 14 m of wire running off near-horizontally. On 60m the trap blocks everything beyond it, so 60m carries on as before. On 160m the whole lot — vertical wire, trap, coil and horizontal wire — becomes a centre-loaded inverted-L. The trap and coil sit almost exactly halfway along the 27 m of wire, right at the corner, so from the vertical's point of view everything beyond them acts as top loading.
The part I hadn't fully appreciated until I modelled it: on Top Band the trap does more of the loading than the coil. Below its resonance a parallel trap looks like an inductor, and with plenty of inductance and not much capacitance this one adds about 270 Ω of reactance at 1.85 MHz, against 220 Ω from the coil itself. And because the loading sits at the top of the vertical rather than at its base, the full 13.4 m carries plenty of current — which is what you want for low-angle radiation.
The coil has a second, quieter job. On 10m a 60m quarter-wave is about five quarter-waves long, so it has a natural resonance close to the band — and as there's no 10m wire at all, that resonance is the antenna's 10m match. At 28 MHz the coil is an RF choke of about 3.4 kΩ, so the new wire is effectively disconnected. I modelled it in MMANA-GAL and NEC-2: with the coil, that 10m resonance stays put; without it, it slides down to 25 MHz, its impedance doubles, and new resonances pop up near 20m for good measure. The 10m match isn't perfect, but it's fine for FT8 and WSPR — voice goes to the Hexbeam anyway.
Physics always sends a bill, of course. A 13 m vertical on 160m is short, so the feed impedance is low (the model says about 12 Ω) and the bandwidth is narrow. That makes the L ideal for FT8 and WSPR, where you sit on one frequency, and less so for SSB — which is why the next project is a proper inverted-V dipole for the SSB end of Top Band.
What the VNA says
The whole antenna, swept at the feed with a miniVNA through the Radio Cockpit VNA tool. Amateur bands shaded; dashed line is 2:1.
Models are lovely, but the VNA has the last word. Swept at the feed with my miniVNA, through the VNA tool in Radio Cockpit, every band in use has its dip in the band. Top Band bottoms out just under 2:1 at 1.85 MHz, about 17 kHz wide at 2:1 and nearer 90 kHz at 3:1 — that narrow bandwidth again. The real antenna resonates about 4% lower than the model, which is about what you'd expect once the other DX Commander wires and the pole are in the picture. And 10m shows a 1.8:1 dip at 28.67 MHz with no 10m wire to explain it.
Left: the 160m inverted-L. Right: 10m as built, and with the 60m wire removed.
The proof of where that comes from turned up by accident during some fault-finding. With the 60m wire off the feed plate, the 10m dip simply vanished. Always nice when the model and the measurement agree.
The box that chooses
All three antennas come into the SPE Expert 1.3K-FA, whose built-in antenna switch and ATU do the selection and matching. Change band and the SPE puts the right antenna on and recalls the tune. The one exception is the Batwing on Top Band, where the MFJ-948B sits in the path to do the heavy lifting. Most of the time the FDM-DUO drives the SPE at just 5 W; when I want the full output, the Elad DUO-ART steps in as a driver. The antenna side of the station simply disappears when I'm operating — which, in a shack that's mostly about software, is exactly what I want.
The ones that didn't make it
Every antenna farm has a cupboard. Mine holds a double-size G5RV that I built years ago and never really took to, and before that a Carolina Windom that worked well from 160m upwards but left gaps on 60, 30 and 15m. Neither was a failure exactly — they taught me what I actually wanted, which is what's up there now.
Next time
Next up: the road to that Top Band dipole — and what the Batwing is really doing on 160m, which surprised me. After that, back to the tour with the RF deck itself: the FDM-DUO, the SPE, the GPSDO, and how the SME (my Station Management Entity) takes charge of all of it.
73, Matthew G6WPJ
No comments:
Post a Comment