See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/335571862

The Radiation Efficiency of a Small Loop Antenna

Article · September 2019

CITATIONS READS

0 1,090

1 author:

David Gibson
University of Exeter
33 PUBLICATIONS 62 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

INDIRES – Information Driven Incident Response View project

CREGJ - Cave Radio and Electronics Group Journal View project

All content following this page was uploaded by David Gibson on 03 September 2019.

The user has requested enhancement of the downloaded file.

 Antenna Design

The Radiation Efficiency of a

Small Loop Antenna
For cave radio applications, we are not normally interested in the radiation from a
loop antenna because the distance over which we are working is small (relative to a
wavelength) and so only near-field effects need to be considered. However, it is still
interesting to consider the radiation field, and to express the radiation efficiency in
terms of the specific aperture where, just as for near-field operation, the number of
turns on the antenna does not affect the result. David Gibson explains radiation
resistance, summarises the equations and shows how the skin effect in the wire and
the proximity effect between turns of the antenna winding can be taken into account.
The main purpose of this note was derivation for a cave radio system, (see but you may see this quoted in different
originally to prepare the ground for any [Gibson, 2010] §2.2.4). Radiation remains forms. We can note that…
reader who wished to consult papers on of interest, because it can be expressed in • The radiation resistance is proportional to the
the proximity effect, e.g. [Smith, 1972a; b], the same terminology (specific aperture) square of the number of turns. This immedi-
ately tells us that Rr is not a ‘normal’ resistance,
that discuss it alongside radiation efficien- that we use for near-field systems. which would be proportional to N, not to N2.
cy. One difficulty with those papers is in • Rr is proportional to the inverse fourth power
understanding the different notation in Radiation Resistance of frequency, so at the low frequencies we
use for cave radio, it is practically zero.
use, which this note hoped to clarify. How- If we imagine the antenna as its • For a source of zero size (A = 0) the radiation
ever, I have now decided that, for cave Thévenin equivalent circuit, it is clear that resistance is zero. You may find, on the Inter-
net, that Rr for an ‘infinitesimal’ dipole is fre-
radio purposes, a detailed study of the the radiated power must be represented, quently given as 0.3 Ω. This is not true!
proximity effect is probably not required. from the point of view of the power
source, by a resistance in which the power Radiation Efficiency
Introduction and Background: is dissipated. We call this the radiation Although textbooks give formulas for
Maxwell’s Equations resistance. Clearly, it is not a physical or radiation resistance, they do not, gener-
The equations of electrostatics and ‘ohmic’ resistance but, as far as the power ally, discuss radiation efficiency. This is
magnetostatics allow us to describe the source is concerned, it exists; and it allows because, for normal radio applications, Rr
fields and forces that arise from electric us to easily model the system. Radiation is high (e.g. ≈73 Ω for a half-wave dipole)
charge and current. They tell us, for exam- resistance also plays a part in receiving and it swamps the low ohmic resistance R.
ple, that the magnetic field of an induction antennas, but that is a difficult concept, But in our case, Rr is much lower than R.
loop falls off with an inverse cube law, and which I will avoid discussing here. The radiation efficiency is Rr/(R+Rr).
that the magnetic field from a short wire Given Maxwell’s equations, the radia- But if Rr << R, as it usually is for cave radio,
falls off with a square law. The equations tion resistance of a simple antenna is then a simpler formula, Rr/R – the radia-
also tell us that, for a long wire, the mag- straightforward to calculate. We write tion fraction – will give a similar result.
netic field falls off with an inverse linear down the fields produced by the current It is useful to express R in terms of a
law. Thus, there is clearly much we can do elements and we integrate those over a figure of merit that I call specific aperture
with these laws, but the one thing they do spherical surface that encloses the source Φ [Gibson, 1999], which is defined by
not predict at all is radiation. at a sufficiently large distance that the
md = NIA = Φ P (3)
In this respect, the equations are simi- near-field effects are played out inside it.
lar to Newton’s laws of motion, which do This allows us to calculate the power flux from which the ohmic resistance is just
not, in any way, predict relativity. And, just across the surface of the sphere. Since we
as Einstein built on Newton’s work, show- know the power P and the source current I R = (Φ NA)2 (4)
ing that very rapid motion gave rise to the radiation resistance is simply R = P / I2. With the specific aperture Φ defined as
hitherto unexpected effects, so Maxwell This procedure is covered in many
built on the work of Ampère, Faraday, textbooks. It can be shown that the σ
Φ = 12 b M (5)
Gauss, Ørsted and others, showing that a radiated power of a small (relative to a ρ
rapid time variation of current gave rise to wavelength) circular loop is 1 we can substitute (4) in (2) to get
another hitherto unexpected effect.
Z 0 k0 4 Z k 4
Maxwell’s equations demonstrate that Pr = md 2 (1) Rr Pr
= = Φ2 0 0 (6)
6π R P 6π
the quasi-static near-fields are accomp-
anied by far-fields for E and H that fall off If the number of turns is N and the
The salient point is that the radiation
in an inverse linear fashion, giving rise to a area of the loop is A then, assuming the
fraction (and the radiation efficiency) does
square law for a radiation of power. current is uniform along the total length of
not depend on the number of turns! The
For cave radio applications, we are not the wire (Nbk0 << 1), we can write the
result can be used alongside formulas for
normally interested in the radiation from a radiation resistance directly from (1) as
Q-factor and bandwidth which, likewise,
loop antenna because it is so small, Z 0 k0 4 do not depend on N when expressed in
= (NA)2
Pr
Rr = (2)
although Maxwell’s equations are necess- I2 6π terms of Φ but this note is too short to give
ary to explain the ‘optimum frequency’ examples and develop the concept further.
1 Symbols are listed at the end of this article.

CREG JOURNAL 107, SEPTEMBER 2019 17

 Antenna Design
Skin Effect effect, provided that we space our conduc- Rs
=
1
(14)
In addition to assuming ‘small’ anten- tors by c/a > 2 and preferably c/a > 4,  2πaδσ
nas, (6) also assumes a uniform distribu- where c is Smith’s notation for the half- but ℓ = 2πbN and so we can write
tion of current across the antenna wire. In spacing. That is, the conductors should be
bN
practice, this is not the case, due to the four radiuses apart. This conclusion is also Rs = Z s (15)
skin effect, which causes it to fall off as reached by [Paul, 2009]. This spacing has a
exp(–r/δ) where r is the distance from the other advantageous effects such as redu- There is no point in trying to substi-
surface of the wire and δ is the skin depth, cing the mutual inductance [Gibson, 2019] tute that into (10) because Zs offers no
and the self-capacitance. advantage to us over δ as a way to charac-
2
δ= (7) Why all the fuss about proximity effect terise the antenna – especially when we
ωµσ
in the past? Originally the problem was have already eliminated N .
There is the additional assumption how to wind high-efficiency air-cored coils
that δ << a. With this assumption, it can be for RF applications in situations where the Concluding Remarks
shown that the total current is equivalent space for a winding was limited. One ques- We have characterised the radiation
to a uniform current flowing in a skin of tion was how many turns should be efficiency of a loop antenna in terms of the
width δ. The cross-sectional area of the crammed into the space. Our applications specific aperture, allowing for skin depth
skin is 2πaδ so R is increased by are slightly different, and we usually have in the wire and the proximity effect (10).
the luxury of designing an antenna from However, we have asserted that a detailed
πa 2
= 1 a δ (8) scratch. It is therefore better to design to study of proximity effect is not required –
2πaδ 2
avoid both the skin and proximity effects. we just need to space the conductors by
and Pr/P drops by this factor and is now about two diameters. This principle
Rr Pr Z k 4 1 a Non-Radiating Antennas applies to all antennas – even if non-
= = Φ2 0 0 2  (9) For most cave radio antennas (10) is not radiating – and it should be remembered
R P 6π  δ
helpful, because there is very little radia- that in most cases, small loop antennas do
tion. But the net effect of (10) is to reduce not radiate any appreciable power. In such
Multi-Turn Antennas – a situation (6), (9) and (10) are clearly
Φ, whether there is radiation or not, to
The Proximity Effect useless, but the general principles remain
The proximity effect is similar in con- a  Rp 
Φ→Φ 1  + 1 (12) intact, with (3) and (12) applying.
cept to the skin effect – that is, a redistri- 2 
δ  Rs 
bution of the current in the wires – but the References
proximity effect operates independently of Example: suppose we design a cave
Gibson, David (1995), The Proximity Effect,
the skin effect and has the effect of forcing radio loop antenna. We space the turns to CREGJ 19, p 23
the current to the outermost edges of a eliminate the proximity effect, and we Gibson, David (1999), A Methodical Approach to
bundle of wires. We can consider two calculate δ=0.3 mm with 2 mm dia. wire. Loop Antenna Design, CREGJ 37, pp 17-20
extreme cases – if the wires in the bundle (12) tells us that Φ will be reduced to 77% Gibson, D. (2010), Channel Characterisation and
are very close together then the current of the value we expected. So, using 2 mm System Design for Sub-Surface Communications.
ISBN 978-1-4457-6953-0. Available at
distribution is like that of a single larger wire to boost the mass of the antenna was
lulu.com/content/5870557.
wire and, if they are well spaced, they will possibly a counter-productive idea.
Gibson, David (2019), The Inductance of a Wire
behave like individual wires. In neither (8) seems to suggest that there is a Hoop, CREGJ 107, p6
case is there any specific ‘proximity’ effect skin depth advantage if a < 2δ. This would Paul, C. R. (2009), Inductance: Loop and Partial:
but, between the extremes, there is an be impossible and what it really indicates John Wiley. ISBN 978-0-470-46188-4
additional effect. I described the proximity is that the formulation does not apply Smith, G. S. (1972a), Proximity Effect in Systems
effect in [Gibson, 1995], referring to the when δ << a is not true. Thus, it also indi- of Parallel Conductors, Journal of Applied Physics
43(5), 2196-2203
classic Butterworth paper, to which cates that a < 2δ is a good design rule.
Smith, G. S. (1972b), Radiation Efficiency of
[Smith, 1972a] also refers. In summary we Electrically Small Multiturn Loop Antennas,
can add a term to (6) to get Smith’s Analysis IEEE Transactions on Antennas and Propagation
If you wish to study Smith’s results, 20(5), 656-657
Rr Pr Z k 4  1 a  Rp 
= = Φ2 0 0   + 1  (10) you will hopefully find this introductory
R P 6π  2 δ  Rs
  

note useful. Smith does not use the con- Symbols Used in this Article
λ Wavelength m
where Rp is the resistance due to the prox- cept of a radiation fraction – he uses radia- σ Electrical conductivity of the wire S/m
tion efficiency – so his expressions are a ρ Mass density of wire kg
imity effect and Rs is due to the skin effect ω Angular frequency Hz
alone. There is the assumption that the little more complicated. He also uses a µ Magnetic permeability of free space H/m
different terminology, using “20” for δ Skin depth m
current is uniform over the length of the Φ Specific aperture m2/√Ω
wire; i.e. the length is much less than λ: Z0/6π, and n, β0, Rs, R0 where I use N, k0, Zs, A Area of loop m2
Rs. Also, he does not simplify using Φ. But a Radius of wire m
Nk 0b << 1 (11) b Radius of loop m
my (10) is equivalent to his (9). c Half-spacing between turns m
The million-dollar question is, of course, Smith uses surface resistance (a.k.a. I Current A
k0 Wave number, 2π/λ m–1
“how do we calculate Rp / Rs?”. surface impedance) in his formulations for ℓ Length m
Smith’s papers go into this in detail, radiation efficiency, rather than skin M Mass of wire in an antenna kg
md Magnetic dipole moment, NIA Am2
but the analysis is complicated. Fortun- depth. The surface impedance Zs is N Number of turns of wire –
ately, Figures 5 and 6, and Tables I and II in P Pr Antenna power: dissipated / radiated W
1 ωµ R Ohmic resistance of antenna Ω
[Smith, 1972a] provide some answers, but Zs = = (13) Rp Rr Rs Resistance: proximity, radiation, skin Ω
δσ 2σ Ω
my current thinking is that, in reality, we Z0 Zs Impedance: free space, surface

do not need to worry about the proximity Now the skin resistance per unit length is

18 CREG JOURNAL 107, SEPTEMBER 2019

View publication stats