Delta Loop Examined

Regular readers (are there any?) will know I am a big fan of delta loops.  These wonderful triangles of wire, easily propped up by just one pole for a vertically polarised version (ideal for my QTH) took me from noisy dipoles and less-than-ideal long wires to working global DX with relative ease.

Sure, it's not a beam, but then, you have to look deeper than that.  You have to look, especially, at the angles of arrival and departure into and from the antenna.  These are important, no matter what someone else has written.

This weekend past, the pole broke on the 20m delta, so I took the opportunity to change from corner fed to 0.08 lambda up from the bottom corner (0.25 lambda down from the apex.)  Before I did that, I ran the corner fed delta against a dipole held aloft at 10m on a tower, using WSPR to eliminate any human bias.

Here, firstly, is the pattern of the dipole at 10m above ground.  Remember that these are my approximate ground conditions on top of an old copper mine; your ground will almost certainly be worse as a result.  Also, my delta base ('radial') wire is at 2m.  A lot of US-based hams comment this is much too low.  But that's wrong, because the pattern develops higher angle lobes much beyond about 4 or 5m (for a 20m loop.)   It's also almost impossible to mount a 20m delta loop with its base at 5m, especially in windy locations and keeping to the non-conductive mast requirement.

Plenty of gain at higher angles, which falls off rapidly as you near the horizon - where good DX is found.

 
Here is the rather odd pattern of a corner-fed delta loop as modelled by MMANA-GAL (others yield the same pattern):

A corner-fed delta worked the world in a very short time for me.  But it can be improved upon.  The 'loaf of bread' elevation pattern is a problem for rejecting higher angle signals.

The result?  The delta loop came in very slightly ahead, by about 0.5dB, of the dipole.  In essence, and because the statistical significance needs to be built up a bit, it's pretty clear the two antennas are neck-and-neck, except that the delta has an almost omnidirectional field against the dipole's distinctly directional two lobes.  This fact was of course taken into account in making comparisons, using only stations within the peak beam of the dipole and delta.

So what?  Well, some people have tall trees, poles, or other supports.  But an awful lot of people don't.  As I've noted previously, getting anything 10m in the air is not the simple operation antenna books and magazine articles endlessly try to make out.  It is an awful lot easier to stick a single fishing pole with a triangle of wire attached to it than even putting up a wire dipole, which needs at least two, strong supports.

For those looking at claims of ground gain from certain aluminium dipoles up on towers or long poles and wondering whether it's a good idea to part with £300, you may want to take stock of the above, equal result.  That, and the roughly £50 it costs to make and install a delta loop.

A test 20m dipole going up to 10m - the expensive way!

So, if you have a fairly clear site and like working interesting DX, it seems to make much more sense to put up a delta than a dipole.  Argue amongst yourselves, or leave a comment!

Next, it was time to take advantage of a calm, starry night and change the delta's feed point.  Off came the cumbersome RG213 and 4:1 balun, on went 300Ohm twin to a dipole centre at 0.08lambda (about 1.72m) up from the corner at which it used to be fed.  Greasing of the conections helps avoid SWR changes during rain.  The old corner was soldered easily together, strung up, and the twin secured.  The twin runs to a 4:1   current balun just outside the shack, from where a short section of RG213 connects it to the ATU or rig 'tuner'.

The new, twin-fed, 0.08lambda-up-from-bottom-corner delta.

The result?  Immediately obvious is the much quieter reception.  That's because the pattern is much more directive in elevation.  Italian stations, notorious for high-power, higher-angle QRM in the UK, have become several 'S' points weaker, entirely in keeping with the modelled pattern.  VK stations, coming in on extremely low chordal long path hops are now about an 'S' point stronger.  An Italian Antarctic station was easy to hear at 55 (58 with preamp), and I got the same 55 report back.  I should also note that, if you can only corner feed, that version managed three previous Antarctic QSOs for me, so it's by no means essential to use this dog-legged feed.

Here is how the pattern looks with the new feed:

The upper elevation angles are now much attenuated, and the low angle gain improved significantly.

What's more, with the twin and 4:1 combination, the delta is matchable, if not the most efficient, on all bands to 6m (15m could not be matched with the previous, corner feed).  But, primarily, the delta is good only on the band for which it is cut, and its second harmonic at 10m, where it has a perfect match.

When I get a chance, I'll run the dipole vs. delta again, see what difference, if any, becomes evident.

Further reading:

Large Loops by ON4UN.