Cobwebb antenna impedance matching scheme

From time to time, correspondents have asked how the Cobwebb antenna works, and particularly how the impedance matching scheme works.

Firstly, what is the Cobwebb?

It is an innovative antenna for small spaces, quite compact and as I recall originally intended to cover five amateur bands from 20-10m.

Essentially, each element is an approximately half wave dipole folded into a square, the ends not quite meeting in the middle of one side, and fed in the centre of the opposite side. Five such elements tuned for each of the bands are supported on a frame parallel to each other and their feed points all drawn together to a common connection point in the style of the well-known Fan Dipole.

Though the inventor makes the claim full size half wave dipole on each band, without nulls, it clearly isn’t full size, but is folded and the folding affects its radiation pattern and feed point impedance. The feed point impedance of a plain Cobwebb shaped dipole is around 12Ω at resonance, a consequence of the fact that current flowing in one leg is offset to some extent by current flowing in the opposite leg, so higher current needs to flow into the feed point to obtain radiated power comparable to that of a half wave dipole… no magic there!

The first versions of the Cobwebb that I recall seeing used Fig-8 twin line for the antenna conductors, though there seem to have been variants since, this article discusses that original twin line version only.

Analysing a single element, in the twin line version, one conductor only is broken in the middle for the feed point (much like a folded dipole), and the two wires of the twin are bridged about half way along each leg. This type of structure contains key elements of the ordinary folded dipole with equal conductors:

  1. at any point along the line, the current in adjacent conductors is approximately equal; and
  2. the short circuit transmission line stubs formed by the pair of conductors and the bridge for each side of the feed point appear in series, then in parallel across the feed point.

Lets look at the consequence of 1. The current flowing into the feed point is accompanied by an equal current flowing in the unbroken wire, so the effect is that the feed point current is one half of the total current flowing in the conductors near the feed point, so achieving a 4:1 impedance transformation, the feed point terminals being four times the 12Ω of a plain Cobwebb dipole.

The consequence of 2 is that there are two inductive stubs in series across the feed point. If we assume that Zo of the twin line is around 120Ω, then the reactance of a sub of about 70° stub is about 330Ω, two in series are about 660Ω. The dipole will need to be slightly short to offset that shunt reactance and so deliver min VSWR at the desired frequency. This is quite like an ordinary equal conductors folded dipole, except that the stubs are about half as long, but they have such high reactance that it is of little real consequence. It is possible to fine tune the impedance transformation ratio by careful tuning of the element overall length and the bridging point, but the author does not suggest that he has done that. Rather, the stub lengths are probably chosen to not interfere with each other impedance wise harmonically (the reason that the Fan Dipole connection is incompatible with simple folded dipoles on even harmonics).

Now the parallel feed points places the other four radiators and stub pairs in shunt with the main radiator and its stub pair, the stub pairs complicating the effects seen in the conventional Fan Dipole arrangement. The extra dipoles are not likely to affect radiation pattern much, but the stub losses may degrade antenna performance significantly.

So, is it a good antenna?

Well, it is compact, and that is innovative.

The author, like so many antenna designers, embellishes his functional description with unsubstantiated and sometimes outrageous claims… such as no lossy traps, stubs or loading coils. As I have pointed out, there are in fact two stubs in each dipole in the design, ten in all. Everything in an antenna system contains loss, it is the overall efficiency that is relevant… and assessing that comes down to some quantitative design and measurement. Using emotive language to a dumbed down ham audience as the inventor has done just plays to prejudice and he promotes his own product by talking down others, falsely perhaps.

To my mind, in the absence of accessable credible measurement of antenna performance, or credible models, there are two issues that deserve attention:

  1. if the twin line is flexible PVC insulated, my experience has been that such cables when measured as a transmission line show that the PVC is a relatively poor dielectric and that will increase the loss of the stubs (PTFE or polyethylene dielectric would be better); and
  2. since higher current flows, albeit in shorter conductors, the effective RF resistance of conductors is important and needs to be sufficient that conductor loss is reasonable.

Issue 1 is easily answered by simply measuring the input impedance of the actual stubs isolated from the antenna at all frequencies of operation, the R term of both stubs in series appears as a loss element along with the reactance in parallel with the feed point as already mentioned. Effectively, all five stub pairs appear in shunt with the feed point, and the current that flows into each stub pair creates I^2R loss is the equivalent resistance of the stub pair. The inventor’s pretence that there aren’t any lossy stubs glosses over the issue rather than rationally and quantitatively addressing it.

Issue 2 is easily answered by modelling the dipoles and determining the conductor loss from the model for given wire sizes.

All antenna systems are compromises of some sort, it is part of the design challenge, and users of multi band antennas need to accommodate higher system losses than otherwise, but pretending that an antenna system is lossless when it can’t be is delusional.