Sunday, May 19, 2013

Phased 40M Wire Verticals

Phased 40M Wire Verticals

About the 40M twin-vertical antenna array

Since early 2007, I have had two 40M wire verticals suspended from a rope between two pines in the forest on the slope below our back yard.

The wire verticals are 1/4-wavelength (about 33' tall at 7.050 mhz).

Each vertical has four 33' raised radials each, feedpoints 10' to 15' off the ground. The verticals are spaced about 30' apart, and in switchable end-fire mode they aim at about 30 degrees (northeast to Europe), or 210 degrees (southwest to the deep South Pacific and California).
Basic schematic for end-fire in two directions
Christman phasing calculator
2009 interaction with half-square array
2010 Update -- phasing lines more precise
Audio samples of antenna doing its thing

Early experiments to reach Europe

Initially, I fed the array at the south vertical, and had a 3/4-wave phasing line of 75-ohm RG6 running to the north vertical. This produced a single end-fire pattern to the north (Europe), but this effect was only slight. I suspect the actual velocity factor of the RG6 required an adjustment to the length of the 3/4-wave phasing line (removing 2.5 feet).

I was also able to feed both verticals with 1/4-wave lines for a bi-directional broadside (US/JA) pattern.

In May 2007, I ran an exact 3/4-wave phasing line of RG-6 between the verticals, feeding at the south end, supposedly to end-fire at EU just east of north. Exhibited some gain, but never worked much to EU.

In 2008, I changed the feed of these verticals to use the Christman method -- each vertical fed by 84-degrees, and a 71-degree phasing loop inserted between the antenna feedpoints at the common control box. All coax is 50-ohm RG-58 for now. Relays gave me switchable end-fire patterns to the north or south (for either EU or W6 coverage).

Basic schematic for end-fire in two directions

Note: This diagram does not show a second relay used to create a broadside pattern (simply shorts across the A and B terminals, so each antenna is fed by the same length of feedline). If you want to keep each antenna's radial system isolated from the other antenna, use a DPDT relay instead -- which will let you switch both the shield and the center conductors of each feedline.

See notes on my two-element 80M verticals for details of the phasing and relay setup.

Christman Phasing Calculator

Operating frequency: Mhz
Coax velocity factor: Vf

The 71-degree phasing line should be: ft or m

The 71-degree phasing line is 90 degrees at Mhz

Cut the coax to the suggested length plus a few inches, in case your velocity factor is not quite right.

Leaving one end of the coax open, set your RF analyzer to Mhz and trim the coax until you see minimum Z impedance. You now have a length of 71 degrees at your desired operating frequency.Each 84-degree phasing line should be: ft or m

The 84-degree feedlines are 90 degrees long at Mhz

Cut the coax to the suggested length plus a few inches, in case your velocity factor is not quite right.

Leaving one end of the coax open, set your RF analyzer to Mhz and trim the coax until you see minimum Z impedance. You now have a length of 84 degrees at your desired operating frequency.

January 2009 -- Interaction problems with half-square array

I had to remove the 40M vertical array in January 2009, however, as these verticals were too close to my 40M half-square array. They severely skewed the pattern of the half-squares when beaming East to the U.S. and Canada.

Here's how skewed the half-square antenna pattern was, when the 40M twin verticals were in place:

And here's the half-square antenna pattern to the East with the 40M twin verticals removed:

May 2009

With the installation of a SteppIR yagi with 30M/40M dipole, I have decommissioned the two-element 40M verticals for now. Plan is to eventually resurrect the array as a three-vertical triangle array for 40M. I am currently working on a three-vertical triangle for 80M to test the concept.

February 2010

I have two major and seemingly perpetual deficiencies for contesting: 40M access to Europe, and 160M to anywhere. I'm working on addressing both. The first order of business in this pursuit was to resurrect the two-element 40M vertical array and get it working properly, firing NE and SW.

I pulled the verticals back into the air in January 2010, but they didn't work very well. Same old story. Seeing as the array has never achieved the performance I expected, my next suspect was the Christman phasing lines. Perhaps I had cut them to the wrong lengths using the old analog noise bridge.

I had to wait until a warm spell in February to get outdoors and bring the two 84-degree feedlines and 71-degree phasing line inside. There, I checked all three coax lines with my new Autek VA1 RF analyzer. Turns out that they were all too long -- as much as 5 inches too long on the phasing line, and about 3 inches on the feedlines. Enough to make a significant difference in the antenna's gain and F/B.

Using the noise bridge had brought me "close" to the required lengths, but not close enough. The VA1 allowed me to cut them precisely... well, I probably should have cut the cables an inch or so too short to compensate for the extra length of the relay leads inside the switch box, but for now I think I'm "close enough."

Here's how I cut them to the correct length:

Connect the antenna analyzer (VA1 in my case) to one end of a cable,other end left open. I manually held the coax centre and shield against the analyzer's connector, so there was no extra length presented by a PL259 -- even a 1/4-inch of extra length makes a difference.

Trim the coax until the analyzer indicates minimum Z impedance -- this is where the line is precisely 1/4 wavelength (90 degrees) long, as follows:
For 7.050 mhz, the 84-degree feedline is 90 deg. at 7.554 mhz.
For 7.050 mhz, the 71-degree phasing line is 90 deg. at 8.937 mhz.

I recommend using the simple calculator to get the correct frequency.

In my set up, I'm using ring terminals on each end of the coax lines, and these only add a tiny bit of extra length to the finished cables. Using PL-259 connectors would add more length, so take that into account.

Propagation to Europe hasn't been very good lately, but tonight I was hearing more European countries than I've heard ever before. There is pronounced front-to-back when switching from NE to SW. The corrected phasing arrangement has dramatically improved how the array works.

For now, I will try out the 40M vertical array in the upcoming 2010 XE RTTY and CQWW WPX RTTY contests.

This spring, I may actually take one more attempt at trimming the coax lines a tiny bit more to allow for the extra length of the connecting wires inside the phasing box.

August 2010

The vertical wires have been remeasured for resonance at 7.050 Mhz with the VA1 RF analyzer. They are both exactly 10.127m (33.23 feet) tall and SWR minimum for each measured at 1.2:1 at the target frequency -- pretty close to a 50-ohm feedpoint impedance, due to the fact that the radials slope down from the feedpoint.

When both verticals are in place, the system SWR rises to about 1.8:1 to the north, and nearly 2:1 to the south. That's caused by the expected mutual coupling -- and no doubt some unwanted radial interaction -- when they're both in the air. The ON4UN's Low Band DXing (Chapter 11-8, part 3.4.2, Fig. 11-7) suggests the common feedpoint impedance (at the relay) should produce an SWR of 2.3:1 for this array, so I think I am pretty close, taking into account the sloping radial effect).

The book says any suitable matching network could be employed to provide a good 50-ohm feed. I'm going to just match it at the shack end with a tuner, and run only low power into this system (a 2:1 SWR at the relay could burn things with sustained high power, particulary on RTTY).

Care was taken to arrange the radial wires so none of them overlap. Radials of one antenna within a few feet of the other antenna's radials detuned the combined system quite a bit. All radials in this array slope down from the feedpoints at various angles, depending on where I've been able to tie off their ends.

I need to confirm the length of all the radials, as some broke and were spliced back together over the past year or two. If a radial on one of the antennas is significantly off-frequency, that could explain why SWR is worse in one of the directions. Another explanation: some radials slope down more than others, which definitely affects the feedpoint impedance of these elevated antennas (though sloping radials generally offer the benefit of raising the impedance closer to 50 ohms).

Audio Recordings of 40M vertical array in action

Aiming Northeast -- Here's ES3AX (Estonia) on 7.010 mhz at around 0140z on Feb. 17, 2010. You'll hear him clearly with the 40M array switched to the Northeast (about 30 degrees) then fade away when I switch the array to aim Southwest.
Audio recording 40M_1 (258kb, MP3)
FT-2000 filter = 300hz

Aiming Southwest -- This is HK1KRY (Colombia) on 7.010 mhz at around 0150z. The difference between directions isn't as noticeable in this one, but it's still prominent.

It starts pointed at EU, then I switch SW (not directly at HK1) and he comes up quite a bit. He really disappears into the noise when I switch the array back totoward Europe. You'll hear some fading on the signal, and I think his bearing is actually on one of the rear-facing lobes in the cardiod pattern, so he's still fairly strong at times even off the back quarter.
Audio recording 40M_2 (436kb, MP3)
FT-2000 filter = 300hz

Hear how the 80M version of this vertical array sounds

No comments: