ArticlePDF Available

The Effect of Dielectric Inside an Inductance Coil (Issue 3)

Authors:
  • Alan Payne Associates

Abstract and Figures

Inductance coils are often wound around an insulating former which then provides mechanical support. The magnetic field is not affected by this former and so the low frequency inductance is not changed. However the inductance of all coils increases with frequency due to self-resonance, and the presence of the insulating former will increase this change of inductance because it lowers the self-resonant frequency (SRF). Losses in the former will reduce the Q of the coil, but interestingly it is found that the biggest reduction in Q is due to the reduction in the SRF. So a material with a low dielectric constant is preferable to one with a high dielectric constant, even if the loss factor of the low dielectric material is higher.
Content may be subject to copyright.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
1
THE EFFECT OF DIELECTRIC INSIDE AN INDUCTANCE COIL
Inductance coils are often wound around an insulating former which then provides
mechanical support. The magnetic field is not affected by this former and so the low
frequency inductance is not changed. However the inductance of all coils increases with
frequency due to self-resonance, and the presence of the insulating former will increase
this change of inductance because it lowers the self-resonant frequency (SRF).
Losses in the former will reduce the Q of the coil, but interestingly it is found that the
biggest reduction in Q is due to the reduction in the SRF. So a material with a low
dielectric constant is preferable to one with a high dielectric constant, even if the loss
factor of the low dielectric material is higher.
1. INTRODUCTION
Inductance coils are often wound around an insulating ‘former’ which then provides mechanical support.
The magnetic field around the inductor is not affected by this and so the low frequency inductance is not
changed. However the inductance of all coils increases with frequency due to self-resonance, and the
presence of the insulating former will increase this change of inductance because it lowers the self-resonant
frequency (SRF).
It is shown here that the effective dielectric constant of the former material is much less than its actual
dielectric constant because the electric field inside the coil is very small, and is dependent upon the factor
(πdc)2 /(p λ) where dc is the coil diameter, p the pitch of the winding, and λ the wavelength. Also the electric
field is strongest close to the wire and so it is here that the dielectric has most effect.
Losses in the former will reduce the Q of the coil, but interestingly it is found that the biggest reduction in
Q is due to the reduction in the SRF, so that in practice a lossy material with a low dielectric constant can
be preferable to a low loss material with a high dielectric constant.
Key equations are coloured brown, and a summary of these is given in Section 6, with a practical example
in Section 7.
2. THE EFFECTS OF DIELECTRIC INSIDE A COIL
2.1. Introduction
The permittivity of a material is given by the product εo εr, where εo is the permittivity of free space (=8.854
pf/metre) and εr is the relative permittivity of the material, also called the dielectric constant of the material.
The dielectric constant has real and imaginary components as follows :
εr = ε’ – j ε’’ 2.1.1
ε’ is the normally quoted dielectric constant and ε’’ is the loss component. Often ε’’ is not given in tables of
constants, and instead the ‘dissipation factor’ tan δ is given and this is equal to ε’’/ ε’. In RF work it is more
convenient to express loss in terms of Q, and this is equal to :
Qdi = 1/ tan δ = ε’ / ε’’ 2.1.2
The dielectric constant and Qdi for a number of materials are given in the table below at 1 MHz and 20oC
(ref 12):
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
2
0
50
100
150
200
250
300
0 2 4 6 8 10
Inducatnce µH
Frequency MHz
Inductance v Frequency
Air Inductance µH
Inductance with water Tube
ε’ Qdi
Wood (Scots Pine) 8.2 17
PVC-U 2.7-3.1 59 - 167
Acrylic (Perspex) 3-3.5 33 50
Polystyrene 2.56 1400
PTFE 2 - 2.1 ≥ 5000
The effective dielectric constant is much lower than these values, and its calculation is considered in the
Sections 3 onwards.
In the following paragraphs it is assumed that the effective dielectric constant ε’eff is known.
2.2. Increased Inductance
The inductance of all coils increases with frequency due to self-resonance (Payne ref 1), and the presence
of the insulating former will increase this change of inductance. An example of the frequency change with
frequency is shown below :
Figure 2.2.1 Measured Inductance of coil with air core and with tube of water
The coil with an air core (pink curve) has reasonably constant inductance up to about 2.5 MHz, where it is
about 5% above its low frequency value. At higher frequencies the inductance rises rapidly as the
frequency approaches 10.5 MHz, the Self-Resonant Frequency (SRF).
A tube of water was added to this coil (described in Appendix A3.3) and the SRF dropped to 6.86 MHz, as
shown by the blue curve. The inductance at 2.5 MHz was now 15% above its low frequency value. Water
was used for this example because it has a high dielectric constant of 78 and so the effect is easily seen.
The rising inductance is due to the fact that the coil is actually a transmission-line and resonance occurs
when the wave reflected from the far end returns in phase with that at the sending end. Adding the water
reduces the phase velocity so that resonance now occurs at a lower frequency.
In principle the effect is predictable from transmission-line theory if the Zo of the line is known along with
the phase velocity, but equations for these are not very well developed. Fortunately there is an alternative
approach based upon the assumption that the inductance is constant with frequency and that it has self-
capacitance across its terminals, and the effect of this capacitance is an apparent change of inductance with
frequency.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
3
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0 2 4 6 8 10
Inducatnce µH
Frequency MHz
Normalised Inductance v Frequency
Normalised Air Inductance
Welsby Equation
Normalised Inductance with water
Welsby Equation
Figure 2.2.2 Self-Capacitance Model
In this representation the coil has a fixed inductance at all frequencies, and this inductance is equal to its
low frequency value (ie low compared to its SRF). The capacitor also has a fixed value, chosen to produce
a change in reactance at the terminals which matches that of the real inductor.
Welsby (ref 2) has analysed this equivalent circuit and derives the following equation for the measured
inductance (for Q > 3):
Lm = Lo / [ 1- (f / fr )2] 2.2.1
where Lm is the inductance measured at frequency f
Lo is the low-frequency inductance
fr is the self-resonant frequency (SRF).
This Equation is plotted below for the two values of SRF, with the inductance normalised to its low
frequency value :
Figure 2.2.2 Normalised Inductance compared with Welsby equation
Correlation between the above equation and the measurements is within 5% up to frequencies of about 75%
of the SRF. Since a coil would generally be used at frequencies far below the SRF this correlation is more
than acceptable. The exception is antenna loading coils at the lower HF frequencies where operation close
to the SRF is common.
The self-resonant frequency fr in the above equation is given by the following for a coil with one end
grounded (ref 1) :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
4
fr ≈ [300*0.25 / ℓw] 0.8 / [dc 2/ (73 p)]0.2/ (ε’eff)0.5 MHz 2.2.2
where w = w (1+ 0.225 dc / c)
w is the length of wire in the winding
dc is the coil diameter to the centre of the wire
p is the pitch of the winding
ε’eff is the effective dielectric constant of the former
2.3. Resistive Loss due to Dielectric
It is shown in Appendix 9 that the dielectric loss can be represented by a resistance in parallel with the coil
of value :
Rdp = [2π fr2 L0 ] [ε’ / ε’’] ]/ [f (1 - 1/ε’eff)] 2.3.1
Where ε’eff is the effective dielectric constant
[ε’ / ε’’] is the Q of the dielectric
Lo is the low-frequency inductance
The Q of the dielectric is usually independent of frequency in the RF range and so Rp reduces with
frequency.
Often it is more convenient to express this loss as a series resistance, and the conversion is given by :
Rds = (Rdp X2) / (Rdp 2 +X2) 2.3.2
where Rdp is given by Equation 2.3.1
X = (2πf Lm )
Lm = Lo / [ 1 - (f / fr )2]
Lo is the low-frequency inductance
2.4. Combined Wire Loss and Dielectric Loss : Parallel Resistance
The overall coil loss can be modelled as a parallel resistance Rp, comprising the loss due to the dielectric
Rpd, and that due to the conductor Rpw . The overall parallel resistance is then given by :
Rp = (Rdp Rwp ) / (Rdp + Rwp ) 2.4.1
where Rdp is given by Equation 2.3.1
Rwp is given by Equation A10.2
An example is shown below, along with separate curves for Rpd and Rpw. The SRF in this case was 5.35
MHz, and the dielectric Q was 29.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
5
0
20000
40000
60000
80000
100000
120000
0 1 2 3 4 5
Parallel Resistance R p
Frequency MHz
Rp v Frequency
Dielectric Parallel loss Rdp
Conductor Parallel Loss Rpw
Predicted Combined R
Parallel Measurements (at TOP of coil)
1
10
100
1000
0 1 2 3 4 5
Series Resistance R s
Frequency MHz
Rs v Frequency
Predicted Dielectric Rs
Predicted Conductor Rs
Predicted Combined Rs
Measurements (at TOP of coil)
Figure 2.4.1 Example of overall Parallel Loss Resistance
There is an increasing error at high frequencies between the prediction and the measurement, and this is
largely due to measurement error since the VNA was not accurate at these very high resistance values
(80kΩ).
Notice that the parallel resistance due to the dielectric loss is constant with frequency in this case (blue
line), because the Q of the water was found to be proportional to frequency (see Appendix 12).
2.5. Combined Wire Loss and Dielectric Loss : Series Resistance
The overall coil loss can also be modelled as a series resistance Rs, comprising the series loss due to the
dielectric Rsd, and that due to the conductor Rsw . The overall resistance is then given by :
Rs = (Rds + Rws ) 2.5.1
where Rds is given by Equation 2.3.2
Rw is given by Equation A10.1
An example is shown below:
Figure 2.5.1 Example of overall Series Loss Resistance
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
6
The predicted values are shown in red and the measurements in purple (see Appendix 11 for details). The
correlation is very close and generally within ± 5% over the 3 octave range of resistance. Also shown is the
series resistance due to the wire (brown) and that due to the dielectric (blue). It should be noted that the
effect of the dielectric is unusually large in this example (by a factor of up to 100) because the dielectric
constant was much higher at 78 compared with a normal former of say less than 2, and the dielectric Q was
very low at 45.
3. PUBLISHED INFORMATION ON ELECTRIC FIELD INSIDE A COIL
3.1. Introduction
The above equations require the effective dielectric constant of the winding former ε’eff, and this is
dependent upon the intensity of the electric fields within the coil. There are three possible electric fields : a
radial field which goes across the diameter of the coil, a longitudinal field which follows the axis of the
coil, and a circumferential field which is circular, following the wire.
The author has been able to find only a few references to field intensity and these are summarized below.
3.2. Cutler
Cutler (ref 3) has carried out some detailed measurements, and the results of these are shown below along
with his theoretical predictions :
Figure 3.2.1 Cutler’s measurements
This shows the longitudinal field to be the largest, decreasing approximately exponentially from a
maximum at the centre of the wire to a minimum, non-zero, value at the axis. The radial field is smaller at
the wire and then drops to zero at the axis. As for the circumferential field Cutler had difficulty in
measuring this and warns that his measurements are not accurate.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
7
In evaluating Cutler’s results it is important to realise that he was not measuring a normal inductance coil,
but one intended for accelerating electrons in a Travelling Wave Tube. As such his measurement frequency
was very high and the length of each turn corresponded to about λ/4, whereas in inductance coils the whole
length of the wire (ie N turns) would be somewhat less than λ/4. The voltage across each turn would
therefore be smaller by at least 1/Nth of that in his coil, so if there are 10 turns we could expect the radial
field to be 1/10th the value he measured, making it negligible compared with the longitudinal field
(sometimes called the axial field).
Cutler’s coil had gaps between turns and so he was able to measure the fields both between turns and
opposite turns (see Figure 3.2.1). Taking the average of these and normalising to unity gives the following
field intensity (blue) :
Figure 3.2.2 Cutler’s Longitudinal Measurement
Also shown is an empirical equation which matches his data (red), and the equation for this is :
e = E[(r/r0)n +A] v/m 3.2.1
where n = 3.6 and A= 0.38
When normalised for e = 1 at r/r0 =1 this becomes :
e/E = [(r/r0)n +A] /(1+A) = [(r/r0)3.6 +0.38]/1.38 v/m 3.2.2
3.3. Lee et al
The assumption that the longitudinal field (or axial field) is the strongest is confirmed by the experiments
of Lee et al (ref 4), on an inductance coil and their results are illustrated below:
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
8
Figure 3.3.1 Measurements by Lee et al
They say ‘The field strength is generally weaker inside the solenoid than outside, because the fields ……
are largely cancelled inside. Inside the solenoid, the field directions are mostly axial except near the ends.
The radial component decreases rapidly approaching the axis of the solenoid, while the axial component
remains relatively constant. On the axis of the solenoid, the field direction is purely axial except near the
hot end, as it should be considering the symmetry of the coil’
They also say that no azimuthal component ( ie circumferential) is observed either inside or outside within
experimental error.
One unexpected result is that they show the electric field going to zero towards the top end of the coil,
before rising again at the top and beyond, as shown below
Figure 3.3.2 Axial field from Lee et al
‘The experimental result shows the axial field has two peaks, one near the centre of the solenoid and the
other above the hot end. The axial component of the electric field becomes zero near the hot end…….and
the sign of the total axial field changes crossing this position’. They explain this effect by a linear charge
distribution along the length of the coil, however it is difficult to understand how this would lead to a field
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
9
in the opposite direction. Such a null might be explained by resonance but in their experiment the coil had
15 turns, of diameter 50 mm, and length 72 mm long, and this will have a first SRF of 38.1 MHz ( Equation
2.2.2 ). Their measurement frequency was 3.92 MHz, so much below the SRF.
In summary they say ‘The major field component inside a solenoid is confirmed to be the axial field’.
3.4. Knight
One of the difficulties in measuring the electric field is that most detectors have metallic parts and these
affect the electric field being measured. Lee et al (above) used a piezoelectric detector inserted into the coil
via a quartz rod which transmitted the data mechanically. Knight (ref 5) used gas discharge tubes to display
the field intensity, although these absorb energy and reduce the Q considerably. His website contains many
photographs of these illuminated tubes. His experiments are not able to quantify the field intensity but the
following comments by him are useful : This photograph shows that the field is strongest near the helical
conductor, and weak along the solenoid axis’ and It appears to confirm that the e-field is tilted, ie, the low-
field region at the end of the coil is funnel-shaped’.
Significantly in Knight’s photographs there is no evidence of Lee’s null within the coil (Knight’s photos do
show nulls but only when the coil is excited at frequencies above its first SRF).
3.5. Choy
Choy (ref 6) gives the following equation for the effective dielectric constant at low frequencies for a
former with a fractional volume filling η and relative dielectric constant εr :
εeff = 1+ η / (a+b) 3.5.1
where a = 1/ (εr -1)
b = (1- η)/2
He also gives a related equation for high frequencies nearer the SRF. Neither of these equations includes
the wavelength but it is shown by Sichak (below) that the effective dielectric is very dependent upon the
wavelength, and this agrees with the author’s experiments.
3.6. Sichak
No closed solution is known for the intensity of the electric fields within a coil. However, Sichak has
analysed a similar problem, that of a coaxial transmission-line with a helical inner conductor (ref 7). He
considered the effect of a dielectric material inside the helix and a different dielectric between the helix and
the outer conductor. In relation to this he says ‘The significant parameter is (2πa N)( 2πa/λ), where N is the
number of turns per unit length, a is the mean radius of the helix and λ the wavelength. When this
parameter is considerably less than 1 …….the dielectric inside the helix has only a second order effect,
while the dielectric outside the helix has first order effect’. When the parameter above is greater than 1, the
effective dielectric constant ε’ is :
ε’eff = (εx + εi)/2 3.6.1
where εx is the external dielectric constant, and εi the internal.
So Sichak defines the two limits which the effective dielectric constant can have, the lower limit being the
external εx, and the upper limit given by the above equation. Between these limits the actual value is
dependent upon his factor Fs = (2πa N)( 2πa/λ) and this is illustrated below :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
10
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with Dielectric Former
Empirical Equation
Upper Limit
Lower Limit
Figure 3.6.1 Upper and Lower Limits of Effective Dielectric Constant
In the above curve εx is assumed to be unity (ie air) and the effective dielectric constant of the core is
assumed to be 4, so that Equation 3.6.1 gives the upper limit as (1+4)/2 = 2.5. The pink curve has no
theoretical basis and is given for illustration only.
Note that Sichak’s factor Fs = (2πa N)( 2πa/λ) can be more conveniently expressed as Fs = (πdc)2 /(p λ)
where dc is the coil diameter and p the pitch of the winding.
Unfortunately Sichak does not give an equation for the effective dielectric constant over the transition
region from ‘considerably less than 1’, to greater than 1, and this is the region of most interest here.
However from his statement we know that it changes from ε’= x + εi)/2 when FS >1, to εx when FS is
‘considerably less than 1’. An empirical equation which fits this criteria is for the effective dielectric
constant to be :
ε’eff = (εx + ψ )/2 3.6.2
where ψ ≈ [εdi (FS ) β + α εx]/[ (FS ) β+ α]
εx is the dielectric constant outside the coil
εdi is the dielectric constant inside the coil (Section 4)
α = 4.2 (determined from experiments)
β = 1.1 (determined from experiments)
FS = (πdc)2 /(p λo)
dc is the mean radius of the coil to the centre of the
conductor
p is the pitch of the winding
λ0 is the free space wavelength = 300/f for f in MHz
Generally for an inductance coil there is air outside the helix and so εx =1.
The wavelength λ in Sichak’s equation is assumed to be that on the wire. However it is more convenient if
Fs is expressed in terms of the free-space wavelength λ0, and it is shown in Appendix 7 that FS then
becomes equal to 0.74 [( πdc)2 /(p λo)] 1.33. However, given that α and β in Equation 3.6.2 are determined
empirically from measurements we can define FS in terms of λo as FS = (πdc)2 /(p λo) and chose α and β
accordingly.
The change in wavelength when ferrite is introduced is discussed in Section 5.
3.7. Coil Diameter dc
There is potentially an uncertainty in the coil diameter as there is for the determination of inductance,
especially when the wire diameter is large. However clarification comes from Sichak’s statement that ‘
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
11
……. dielectric inside the helix has only a second order effect, while the dielectric outside the helix has
first order effect’. The dividing line between the inside and the outside of the helix is thus along the centre
of the conductor, and this is confirmed by Figure 3.2.1. So for all the equations the coil diameter dc is
defined as that diameter which runs through the centre of the conductor.
Support for this is given by the Alumina experiment described in Appendix 3 where the wire diameter was
very large at 17% of the coil diameter, and agreement between the prediction and experiment was 2%.
3.8. Summary of Published Information on Electric Field
In the above references Sichak gives the dependence of the effective dielectric constant on the coil
diameter, its winding pitch and the wavelength, leading to Equation 3.6.2. This equation alone would be
sufficient if the dielectric former totally filled the winding, but often the former is tubular and so
consideration must be given to the change of field intensity across the coil diameter, and this is supplied by
Cutler (Equation 3.2.2). These equations are used in the following Section to derive the effective dielectric
constant of a former, be this solid, tubular or grooved to locate the wire.
4. DIELECTRIC FORMERS
4.1. Introduction
The dielectric constant of the material making the former is diluted by an inevitable gap between the centre
of the conductor and the outer surface of the former (material C in Figure 4.2.1 below). Also the former is
often tubular and so the overall dielectric constant is further reduced. For convenience here these reductions
are called the ‘dilution’ of the dielectric.
4.2. Dilution of Dielectric Constant
The geometry of a tubular former and its winding is shown below:
Figure 4.2.1 Tubular Former and wire
In this diagram the tubular former is B, of radii r1 and r2, having an air core A. Clearly if the winding
former is solid then r1=0.
The material C consists of a mixture of materials : the conductors, their insulation and air gaps. No electric
field can exist in the conductors and so the dielectric is the combination of the air and the insulation. The
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
12
latter is likely to be polyamide with a thickness of about 7% of the wire diameter, and a dielectric constant
of 2.5. The dielectric constant of the combination will depend upon the ratio of the wire spacing to the wire
diameter, but assuming this ratio is similar to the illustration above it can be seen that the field is probably
70% in air. The overall dielectric constant of material C is therefore assumed to be around 1.1.
Note that the longitudinal electric field in the gap between wires is not normally affected by the winding
former (be it tubular or solid) and so the dielectric constant of material C is always close to unity. The
exception is where the former is grooved to locate the winding and then the dielectric constant of material
C is increased.
A theoretical analysis of this configuration is given in Appendix 2 assuming that the electric field is
longitudinal. That is, there is no radial component of the field so the field in one layer does not pass across
the boundary into the adjacent layer. This seems to be essentially true in the centre of a long coil since the
analysis yields results which agree well with measurements, but this cannot be true at the ends and this is
covered in Paragraph 4.3.
Equation A2.1 gives the diluted dielectric constant, where n and A describe the field intensity across the
diameter. If it is assumed that the intensity of the electric field is given by Equation 3.2.2 (ie Cutler’s
measurements) and εgap = 1.1 then the diluted dielectric constant for εair = 1, becomes :
εdi = 1.1 kg + εformer kf + ka 4.2.1
where εdi is the diluted dielectric constant
kg = 1 0.5 (r’2 5.6 + r’2 2)
kf = 0.5 (r’2 5.6 + r’2 2 - r’1 5.6 - r’1 2)
ka = 0.5 (r’1 5.6 + r’1 2)
The radii r’ are normalised to the coil radius r3, and so r’1 = r1/ r3 and r’2 = r2/ r3.
4.3. End Effect
Equation 4.2.1 assumes that the coil is very long compared with its diameter, so that the field intensity can
be assumed to be constant down the length. In fact the intensity of the electric field reduces towards the
ends of the coil, and this further dilutes the effective dielectric constant of the winding former. A simple
model of this field variation which agrees with experiment is for it to be constant at unity in the centre
section and that it reduces to 0.42 at the ends over a length of 0.22 dc from each end (see Appendix 5).
Notice that the length over which the field is reduced is assumed to be proportional to the coil diameter dc,
and also that it applies to each end whether this is connected to ground or not, and experiments support
these assumptions. Experiment shows that the diluting effect of this field taper is given by :
Longitudinal dilution = e x 4.3.1
This dilution applies to the susceptibility of the coil former (rather than its permittivity), and so Equation
4.2.1 becomes :
εdi = 1.1kg + [(εformer -1) e x + 1] kf + ka 4.3.2
where x = - 0.33 dc / lc
dc is the coil diameter to the centre of the wire
lc is the coil length
4.4. Dielectric Extension
Equation 4.2.1 assumes that the dielectric former has the same length as the coil. In practice the former is
usually slightly longer and experiment shows that for a coil earthed at one end, extending the former at this
end has no measurable effect. At the other end (the hot end) the effect of the extension is shown below (see
Appendix 4 for details) :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
13
1.00
1.02
1.04
1.06
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Permittivity Multiplier
Protrusion normalised to Diameter
Permittivity Multiplier : Ceramic Tube
Permittivity Multiplier
Figure 4.4.1 Increase at ‘hot end’ due to coil former being longer than coil
To use the above curve, the dielectric constant ε’eff is calculated from Equations 3.6.2 and 4.2.1, and then
the calculated value is multiplied by the above factor.
An equation for the above is given in Appendix 4. However for short extensions of less than one dielectric
diameter, this can be simplified and the total effective dielectric constant εtotal , including the end extension,
is then given by :
εtotal [ε’eff 1] [ 1+ 0.045 (l’die /ddie )] + 1 4.4.1
where l’die is the length of the dielectric protrusion at one end
ddie is the diameter of the dielectric protrusion
ε’eff is given by Equation 3.6.2
This equation is for a coil grounded at one end, so having one ‘hot’ end only. If the coil is balanced so that
both ends are ‘hot’ the equation becomes εtotal [ε’eff 1] [ 1+ 0.09 (l’die /ddie )] + 1
Notice that for a normal extension of say 0.5 ddie the correction is only 2% and can be ignored.
4.5. Experimental Support
Experimental support for Equations 3.6.2, and 4.3.2 is given by the results illustrated below for a range of
dielectrics :
a) A coil filled with distilled water dielectric r = 78)
b) A coil filled with a PTFE dielectric (εr = 2.3)
c) A coil with a tube of distilled water
d) A coil with a tube of ceramic (alumina) εr = 9.68) .
For each experiment the length of the dielectric former was the same as that of the coil, so the correction
provided by Equation 4.4.1 was not required.
The theoretical curves given in the following graphs are given for α = 4.2 and β =1.1 (Equation 3.6.2), and
these values were optimized to give the best agreement with all the experiments (see Appendix 3 for details
of these experiments).
Each graph shows three or four experimental points, but these are not for three different coils but for the
same coil/dielectric combination at the first three/four resonant frequencies.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
14
0.1
1.0
10.0
100.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with Water Dielectric
Empirical Equation
Measurements
Upper Limit
Lower Limit
0.9
1.0
1.1
1.2
1.3
1.4
1.5
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with PTFE Dielectric
Coil Filled with distilled water :
Figure 4.5.1 Coil Filled with water of same length as Coil
Notice that both scales are logarithmic. The diluted dielectric constant was 56 (Equation 4.3.2).
Coil Filled with solid PTFE :
Figure 4.5.2 Coil Filled with PTFE of same length as Coil
Notice that the x axis is linear. The diluted dielectric constant was 2.17 (Equation 4.3.2).
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
15
1.0
10.0
100.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity 0.73d Hole in Water Dielectric
Empirical Equation
Measurements
Upper Limit
Lower Limit
1.0
1.1
1.2
1.3
1.4
1.5
1.6
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with Ceramic tube Dielectric
Empirical Equation
Measurements
Upper Limit
Lower Limit
Coil with tube of distilled water:
Figure 4.5.3 Tubular Former of Water
Notice that both scales are logarithmic. The dielectric constant was 36 (Equation 4.3.2).
Coil with tube of Alumina:
Figure 4.5.4 Tubular Alumina Former
Notice that the x axis is linear. The diluted dielectric constant was 3.7 (Equation 4.3.2).
There is a trend in the graphs above for the agreement between experiment and theory to be better at the
first resonance than at the higher resonances but the reason for this not known.
The values chosen for n and A in the above curves were 3.6 and 0.38 respectively (describing the electric
field intensity), and these gave good agreement with experiment and have a sound basis from Cutler’s
measurements (albeit with different values of α and β). However the values are not critical, and a wide
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
16
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity of Water Tube with Ferrite Core
Empirical Equation
Measurements
range of alternatives give equally good correlation with experiment. The reason for this is that, even if the
field intensity was constant across the coil, the displacement current reduces as the enclosed area, so the
current in the centre of the coil is very much reduced over that near the winding. For instance at half the
radius (r/ro =0.5), the area enclosed is only ¼ of the total and so dielectric material at larger radii will have
3 times the effect of dielectric in the centre. This area factor greatly diminishes the effect of changing the
rate at which the field decreases.
5. FERRITE CORES
Ferrites have a dielectric constant with a value of around εr 10. In applying the previous equations account
needs to be made for the effect of the permeability, since this changes the wavelength on the wire. Sichak’s
factor then becomes :
FS = (πdc)2 /(p λo) c/Vw 5.1
where Vw/c is the wire velocity as a proportion of c
When the ferrite is introduced the ratio c/Vw is equal to √µcoil, where µcoil is the effective permeability of the
ferrite, and is equal to the ratio of the inductance with and without the ferrite. For a toroid µcoil = µr , the
relative permeability of the ferrite. For an open core such as a ferrite antenna or the experiment here, the
permeability is diluted because the flux has to flow partially in air. This is discussed in the authors
reference 8. So Equation 5.1 becomes :
FS = (πdc)2 /(p λo) (µcoil)0.5 5.2
where µcoil is the ratio of low frequency inductance with and without the ferrite (ref 8)
dc is the coil diameter to the centre of the wire
p is the winding pitch
λo is the free-space wavelength = 300/f where f is in MHz.
Appendix 8 describes an experiment to test this equation with µcoil = 4 with the following results :
Figure 5.1 Tubular Alumina Former
The pink curve is that of Equation 3.6.2, with α = 4.2 and β =1.1 (ie the same as all other experiments). FS
for the measurement point was calculated from Equation 5.2. Agreement is better than 2%.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
17
6. SUMMARY OF EQUATIONS
The Q of a coil is given by (Equations 2.2.1 and 2.5.1) :
Q = ωLo /[1 (f/fr )2] / (Rds + Rws ) 6.1
where
Lo is the low frequency inductance
fr ≈ [300*0.25 / ℓw] 0.8 / [dc 2/ (73 p)]0.2/ (ε’eff)0.5 MHz (Equation 2.2.2)
where w = w (1+ 0.225 dc / c)
w is the length of wire in the winding
dc is the coil diameter to the centre of the wire
p is the pitch of the winding
ε’eff is the effective dielectric constant of the former
ε’eff = (εx + ψ )/2 (Equation 3.6.2)
where ψ ≈ [εdi (FS ) β + α εx]/[ (FS ) β+ α]
εx is the relative dielectric constant outside the coil (normally =1)
εdi is the relative dielectric constant inside the coil (see below)
α = 4.2 (determined from experiments)
β = 1.1 (determined from experiments)
FS = (πdc)2 /(p λo)
dc is the mean radius of the coil to the centre of the conductor
p is the pitch of the winding
λ0 is the free space wavelength = 300/f for f in MHz
εdi = 1.1kg + [(εformer -1) e x + 1] kf + ka (Equation 4.3.2)
where x= - 0.33 dc / lc
kg = 1 0.5 (r’2 5.6 + r’2 2)
kf = 0.5 (r’2 5.6 + r’2 2 - r’1 5.6 - r’1 2)
ka = 0.5 (r’1 5.6 + r’1 2)
εformer is the relative dielectric constant of the coil former
dc is the coil diameter to the centre of the wire
lc is the coil length
The radii r’ are normalised to the coil radius r3, and so r’1 = r1 / r3 and r’2 = r2 / r3.
Rds = (Rdp X2) / (Rdp 2 +X2) (Equation 2.3.2)
where X = (2πf Lm )
Lm = Lo / [ 1 - (f / fr )2]
Lo is the low-frequency inductance
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
18
0
20
40
60
80
100
120
140
160
180
0 1 2 3 4 5
Quality Factor Q
Frequency MHz
Predicted Q v Frequency
Predicted Q, wire loss only
Rdp = [2π f12 L0 ] [ε’ / ε’’] ]/ [f (1 - 1/ε’eff )] (Equation 2.3.1)
where [ε’ / ε’’] is the Q of the dielectric
Rws = K √f / [ 1- (f / fr )2]2 (Equation A10.1)
where K= Rdc φ 0.25 dw/0.067
Rdc = 4 ρ ℓw /(π d2w)
dw is dia of the wire in metres
ρ = resistivity (1.77 10-8 for copper)
w is the length of the conductor (straight length) in metres
φ is the proximity effect (see ref 1 Appendix 1, and note below).
NB if the ratio of the wire diameter to the pitch is 0.5 (which often gives the highest Q) then φ 1.78 for
coils with a length equal to twice the diameter or longer.
7. PRACTICAL APPLICATIONS
Antenna loading coils are often operated at frequencies close to the SRF, and then the effect of the
dielectric is greatest. The following example of a loading coil has been extracted from an ARRL handbook
for resonating an 8ft whip at 3.8 MHz with the coil at the centre of the whip:
Coil length : 254 mm
Coil dia : 65 mm
Number of turns : 100
Wire diameter : 1.3 mm (16 AWG) (giving dw/p =0.5)
Low frequency inductance Lo = 140 µH (calculated from Wheeler’s equation)
SRF : 5.8 MHz (calculated from Equation 2.2.2)
If there is no dielectric former, there will be only conductor loss and the coil Q is then calculated to be
(from Equation 6.1 with Rds = 0) :
Figure 7.1.1 Coil Q with only conductor loss
Notice that the Q peaks at around half the SRF, and at the operating frequency of 3.8MHz the Q has
dropped to about 86% of this peak value. If a lossless tubular former is introduced having a dielectric
constant of 3 and a wall thickness 10% of the coil diameter, the Q becomes :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
19
0
20
40
60
80
100
120
140
160
180
0 1 2 3 4 5
Quality Factor Q
Frequency MHz
Predicted Q v Frequency
Predicted Q wire loss only
0
20
40
60
80
100
120
140
160
180
0 1 2 3 4 5
Quality Factor Q
Frequency MHz
Predicted Q v Frequency
Predicted Q wire loss only
Predicted Q including dielectric loss
Figure 7.1.2 Coil Q with lossless dielectric former
Notice that although the coil former is lossless the Q has dropped significantly at 3.8 MHz, because of the
decrease in the SRF.
If this former has loss with a dielectric Q of 100, the overall coil Q becomes (red curve) :
Figure 7.1.3 Coil Q with lossy former
Notice that when the dielectric former is introduced the greatest reduction in Q is due to the reduction in the
SRF, rather than the loss in the dielectric, even though the dielectric was assumed to be of relatively low Q.
So formers with a high dielectric constant should be avoided even if they have high Q (e.g Ceramic), and
indeed a dielectric with a poor loss tangent can perform better if its dielectric constant is very low.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
20
8. TEST EQUIPMENT AND CALIBRATION
All measurements were made with an Array Solutions UHF Vector Network Analyser. Calibration of this
analyser required an open circuit, a short circuit and known resistive load, and these are shown below.
Figure 8.1 Calibration loads
To ensure that the resistive load had minimal stray reactance a thick-film resistor was used, and this had the
added advantage that it could be located in the same plane as the short circuit. Its value was 47 Ω ± 1%.
SMA connectors were used because they are small and therefore have a small stray capacitance, and so any
error in calibrating this out would also be small.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
21
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0 0.2 0.4 0.6 0.8 1.0
Electric Intensity
r/ro
Electric Intensity across Coil radius
Cutler LongL normalised
APPENDIX 1 : DILUTED DIELECTRIC CONSTANT OF A SOLID FORMER
Consider a coil with dielectric cylinder inside having a radius r, thickness δr and a length equal to that of the
coil. Given that the main electric field is longitudinal this will be along the length of the cylinder, so the
displacement current will also be in this direction. For this cylinder the displacement current will be
proportional to the cross-sectional area 2π r δr , and to the electric intensity e:
i = e εo εr 2π r dr A1.1
where e is the electric intensity in v/m
From Cutler the intensity will have an intensity similar to that below :
Figure A1.1 Intensity of longitudinal field at radius r
Assuming the curve is proportional to rn and reduces to A at the centre, then in general the field will have
the form :
e = E ( r n +A) v/m A1.2
where E, n and A are constants
Combining the above two equations gives :
ide = E εo εr 2π ( r n +A) r dr
= εo εr 2π E ( r n+1 +A r) dr A1.3
The total longitudinal current between a radius of r1 and r2 will be :
r2
Ide = 2π E εo εr ( r n+1 +A r) dr
r1
= 2π E εo εr [{ r2 (n+2) / (n+2) + A r22/2} - { r1 (n+2) / (n+2) + A r12/2}] A1.4
To find the dielectric constant this displacement current can be compared to that with no dielectric Id1, and
this is given by the above equation when εr =1, r2 = ro (the radius of the coil) and r1 =0 :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
22
Id1 = 2π E εo [ ro (n+2) / (n+2) + A r02/2] A1.5
So the effective dielectric is the ratio of the currents with and without the dielectric :
eeff = εr [{ r2 (n+2) / (n+2) + A r22/2} - { r1 (n+2) / (n+2) + A r12/2}] / [ ro (n+2) / (n+2) + A r02/2] A1.6
It is convenient to normalise all radii to that of the coil r0 so Equation A1.6 becomes :
eeff = εr [{ r’2 (n+2) / (n+2) + A r’22/2} - { r’1 (n+2) / (n+2) + A r12/2}] / [ 1/ (n+2) + A/2] A1.7
where r’2 = r2/ r0
r’1 = r1/ r0
APPENDIX 2 : EFFECTIVE DIELECTRIC CONSTANT OF A TUBULAR FORMER
With a tubular former there are three dielectrics rather than the two above, as shown in the diagram below
(see also Section 4):
Figure A2.1 Dielectrics with Tubular Former
Assuming the electric field is longitudinal, and therefore that the displacement current is longitudinal and
does not go from one dielectric to another, then the combined displacement current is the sum through each
one and so the combined dielectric constant is (from Equation A1.7:
εrt = εgap kg + εformer kf + εair ka A2.1
where kg = [{ r’3 (n+2) / (n+2) + A r’32/2} - { r’2 (n+2) / (n+2) + A r’22/2}] / [ 1/ (n+2) + A/2]
kf = [{ r’2 (n+2) / (n+2) + A r’22/2} - { r’1 (n+2) / (n+2) + A r12/2}] / [ 1/ (n+2) +A/2]
ka = [{ r’1 (n+2) / (n+2) + A r’12/2} / [ 1/ (n+2) +A/2]
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
23
APPENDIX 3 : MEASUREMENTS OF EFFECTIVE DIELECTRIC CONSTANT IN COILS
A3.1. Introduction
It is shown in the main text that the effective dielectric constant can be described by Equation 3.6.2, using
the diluted dielectric constant determined by Equation 4.3.2. However, there are then four unknown
parameters, α and β in the first of these equations and n and A in the second equation. It was intended to
determine these by experiment but the parameters are slightly interactive and so as a starting point it was
assumed that the electric field was as measured by Cutler (Section 2.2), and this gave n=3.6 and A=0.38.
With these values the best agreement with all the following experiments was with α = 4.2 and β = 1.1. So
the modeled dielectric constant in the following experiments uses :
n = 3.6 and A = 0.38 in Equation 3.2.1 A 3.1
with α = 4.2 and β = 1.1 in Equation 3.6.2.
It is shown that these values give good agreement with experiment, but equally it is found that these values
were not critical and a wide range of alternatives give equally good correlation with experiment.
The experiments measured the SRF of coils with both an air core and a dielectric core, both solid and
tubular. The change in SRF was assumed to be due to the dielectric reducing the phase velocity by √ε’
(Equation 2.2.1), so that the measured effective dielectric constant was given by:
Measured ε’ = [ (SRF air)/(SRF dielectric) ]0.5 A3.2
In the initial experiments only the first SRF was measured but it was then realised that the higher SRF’s
would also give useful data since these higher frequencies would correspond to larger values of the Sichak
factor Fs.
All coils were operated with one end grounded, and this excited the n λ/4 modes where n is an integer, 1,2,3
4….. The measured SRF’s are not exact integers as this would imply because of end effect, and also
because the effective dielectric constant increases with frequency, as predicted by Figure 3.6.1.
The length of the dielectric was the same as the coil in each case. In practice the former is likely to be
slightly longer and the effect of this is discussed in Section 4.4.
A ground plane soldered to the lower end of the coil was needed for reproduceable measurements except
for the large coil figure A3.2.1.
Measurements were made with an Array Solutions UHF VNA, coupled into the coil by either a tap on the
coil, or alternatively a coupling loop. The SRF’s were defined as those frequencies where the phase of the
input impedance went through zero.
A3.2. Total fill of water
Water has a very high dielectric constant of 78, and so the reduction of the SRF will be large, thereby
reducing some of the experimental error. Distilled water was used because it has a much lower loss than tap
water, but nevertheless the Q at resonance was only around 15 at 12 MHz, and this tended to smear one
resonance into another (see below) (Note : the distilled water used here was from a de-humidifier, and it
was found that the loss of this water doubled if it was left for several weeks in an open beaker-see
Appendix 12).
A coil was wound with 59 turns of 1 mm dia copper wire, onto a thin plastic bottle of outer diameter 49.2
mm, giving a mean wire diameter of 50.2 mm. The length of the winding was 125 mm and inside diameter
of the tube was 47.2 mm. To minimize loading of the coil by the VNA and connection leads, connection to
the VNA was at a tap on the coil 16 turns from the earthed bottom end. The top end had no lead and no
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
24
connection. The bottle was filled with water to a height of 125 mm, ie the top of the winding. The water
extended by 10 mm beyond the bottom (earthed) end of the coil, but it is shown in Appendix A4 that this
will have had no significant effect.
Figure A3.2.1 Test Coil ready for filling with Water
NB the size of the coil can be judged by comparison with the SMA connector which has on outer metal
length of 10mm.
Measurements gave the following (frequencies are in MHz):
λ/4
λ/2
3λ/4
4λ/4
SRF air
12.38
28.7
41.7
56.3
SRF water
6.37
9.78
14.03
19.93
Calculated ε'
3.78
8.63
8.83
7.99
Sichak Fs
0.484
1.124
1.632
2.204
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
25
0.1
1.0
10.0
100.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with Water Dielectric
Empirical Equation
Measurements
Upper Limit
Lower Limit
Figure A3.2.2 Measurements of water dielectric cf empirical model.
The diluted dielectric constant of the water, εdill (Equations 4.2.1 and 4.3.1) was calculated to be 62.5
assuming the effective dielectric constant of the gap between water and the centre of the conductor was 1.1.
The measured impedance with water is shown below, for the first four resonances. It is noticeable that the
Q at resonance is very low, and this is largely due the loss in the distilled water because the coil without
water showed a Q of 64.
Figure A3.2.3 Measured Impedance with water dielectric
A3.3. 36.4 mm Tube of water
A tube of water was made by inserting a 36.4 mm plastic tube into the coil shown in Figure A 3.2.1 and
filling around it with distilled water. The inside diameter of the tube of water was 0.73 that of the coil.
Measurements gave the following :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
26
1.0
10.0
100.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity 0.73d Hole in Water Dielectric
Empirical Equation
Measurements
Upper Limit
Lower Limit
λ/4
λ/2
3λ/4
4λ/4
SRF air
12.21
28.2
40.6
55.2
SRFwater
7.82
12.13
16.15
22.09
ε'
2.44
5.41
6.32
6.24
Sichak Fs
0.48
1.10
1.59
2.16
Figure A3.3.1 Measured Impedance with water tube dielectric
The diluted dielectric constant of the water, εdill was calculated to be 40.3 assuming the effective dielectric
constant of the gap between water and the centre of the conductor was 1.1.
A3.4. Solid PTFE
A coil was close wound onto a solid PTFE bar of 14 mm diameter using copper wire of diameter 0.44 mm
over its enamel insulation, over a length of 29.3 mm, thus giving approximately 67 turns :
Figure A3.4.1 Coil with PTFE Dielectric (ground-plane not shown)
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
27
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with PTFE Dielectric
NB the size of the coil can be judged by comparison with the SMA connector which has on outer metal
length of 10mm.
It was found that a well-defined ground-plane was necessary (not shown above) to give repeatable results
and this was formed from a thin sheet of copper 150 mm x 110 mm. One end of the coil was connected to
this ground plane and also to the earth of the test jig
The SRF of this coil was to be measured with and without the PTFE and so it was important that the PTFE
could be easily removed, while minimising any air gap between it and the winding. This was achieved by
unwinding the coil by a fraction of a turn while it was on the rod and then securing it with cellotape (see
picture). The wire was cut-off close to the former at the top end and the bottom was connected to ground
terminal of the test connector via a short lead of about 5 mm length. Coupling to the coil was via a 5 turn
coupling loop at the earthed end of the coil.
Measurements gave the following:
λ/4
λ/2
3λ/4
SRF air
34.07
73.89
103.68
SRF water
33.00
68.66
95.83
ε'
1.07
1.16
1.17
Sichak Fs
0.61
1.33
1.87
Figure A3.4.2 Measured Impedance with solid PTFE dielectric
The dielectric constant of pure PTFE is 2.1 but often a small proportion of glass is introduced to improve its
mechanical properties. This increases the dielectric constant slightly and measurements of the author’s
material gave 2.3 (see Appendix 12). The diluted value was calculated as 2.18
A3.5. Tubular Alumina
A coil of 30 turns was wound onto an Alumina tube over a length of 30 mm so the pitch was 1 mm. The
outer diameter of the tube was 6.25 mm and the inner diameter 4.85 mm. The wire was 20 swg enamelled
copper wire having a quoted copper diameter of 0.914 mm and a measured diameter over insulation of 0.98
mm. The thickness of the insulation was therefore (0.98 -0.914)/2 = 0.033 mm.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
28
1.0
1.1
1.2
1.3
1.4
1.5
1.6
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with Ceramic tube Dielectric
Empirical Equation
Measurements
Upper Limit
Lower Limit
Figure A3.5.1 Test Coil and Alumina Core (ground-plane not shown)
NB the size of the coil can be judged by comparison with the SMA connector which has on outer metal
length of 10mm.
Resonance was in the VHF range and it was found that a well-defined ground-plane was necessary (not
shown above) and this was formed from a thin sheet of copper 50 mm x 65 mm. One end of the coil was
connected to this ground plane and also to the earth of the test jig. The VNA was connected to a tap on the
coil 10 turns from the earthed end. The other end of the coil was left free with no lead.
Measurements gave the following :
Figure A3.5.2 Measured Impedance with Ceramic tube dielectric
One potential problem with ceramic materials is that they can be porous and absorb moisture. To test this
the ceramic tube was heated to soldering iron temperature for one hour, cooled to room temperature and the
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
29
0.96
0.98
1.00
1.02
1.04
-2.0 -1.0 0.0 1.0 2.0 3.0
Normalised Permittivity
Protrusion normalised to Diameter
Permittivity v Protrusion Ceramic Tube
Permittivity Multiplier : Cold end
Permittivity Multiplier : Hot end
Flux equation
SRF re-measured. The SRF had changed less than 0.1 MHz in 180 MHz, and this is within the
experimental repeatability.
APPENDIX 4 : DIELECTRIC EXTENSION
A4.1. CERAMIC TUBE
To test the end effect of the coil shown in Figure 3.5.1 the SRF was measured for various protrusions of the
ceramic tube from one end of the coil. The tube was very much longer than the coil so that it always
protruded a great distance beyond the other end, so that in effect the amount of protrusion at this end was
constant. This was repeated for each end, the earthed end and the free end. For completeness the effect on
the SRF of the end of the dielectric tube being within the coil was measured. Measurements gave the
following :
Figure A4.1.1 Extension of Ceramic Tube Dielectric
In the above the protrusion has been normalized to the coil diameter. The effective dielectric constant has
been normalized to the value it has when the former is the same length as the coil. So the values on the y
axis are the multiplier for protrusion.
The measurements show that extending the dielectric beyond the cold end has no effect on the SRF (the
pink curve) and so there can be no significant electric field beyond the coil at this end. In contrast there is a
larger electric field beyond the free end, with the effective dielectric constant increasing rapidly for short
protrusions (brown curve).
Also shown in yellow is a semi-empirical equation described in paragraph A4.4.
A4.2. Solid PTFE
A similar experiment was carried-out wuth the coil shown in Figure A3.4.1, again with a dielectric much
longer than the coil.
Measurements gave the following :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
30
0.96
0.98
1.00
1.02
1.04
-2.0 -1.0 0.0 1.0 2.0 3.0
Normalised Permittivity
Protrusion normalised to Diameter
Permittivity v Protrusion Solid PTFE
Permittivity Multiplier : Cold end
Permittivity Multiplier : Hot end
Flux equation
0.84
0.88
0.92
0.96
1.00
1.04
1.08
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2
Normalised Permittivity
Protrusion normalised to Diameter
Permittivity v Protrusion Water fill
Permittivity Multiplier : Hot end
Flux equation
Figure A4.2.1 Extension of Solid PTFE Dielectric
A4.3. Distilled Water
A similar experiment was carried-out with the coil shown in Figure A3.2.1. Distilled water was poured into
the bottle to various levels including beyond the end of the coil, with the following results :
Figure A4.3.1 Extension of Water Dielectric
Notice that the normalised effective dielectric constant is very similar to the measurements with the
ceramic tube and the PTFE, as indicated by the empirical equation plotted for all three. However at large
protrusions the water is having a larger effect and this is probably due to its lower electric reluctance (ie
higher dielectric constant).
A4.4. Equation for Extension
The effect of extending the dielectric beyond the end of the coil is likely to be similar to that of extending
magnetic material beyond the end of the coil. This has been analysed by the author (ref 8), assuming that
magnetic material inside the coil affects only the internal field, and material outside the coil affects only the
external field. The sudden change in slope as the dielectric protrudes indicates that this may also be true for
the electric fields.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
31
In the case of the magnetic field it was shown that as the ferrite is extended beyond the end of the coil the
effective permeability increased logarithmically. However with dielectric the extension will be relatively
small and so a linear approximation can be assumed. The total effective dielectric constant εtotal , including
the extension, can then be expressed by the following empirical equation:
εtotal [ε’eff -1] [ 1+ k1 (l’die /ddie )] +1 A4.4.1
where l’die is the length of the dielectric protrusion at one end
ddie is the diameter of the dielectric materia
Where ε’eff is the effective dielectric constant when the dielectric is the same length as the coil ie l’die = 0,
(and given by Equation 4.2.1) and the extension is given in terms of an increase in this effective dielectric
constant. Best agreement with the water experiments was with k1= 0.055 but this gave an error with
ceramic and PTFE. The reason is that the water has a much higher effective dielectric constant of 78
compared with 3.75 for the ceramic (diluted from 9.6 because it is a tube) and 2.3 for the solid PTFE. This
is a similar to the effect of low permeability with magnetic materials, and the reduction in this case was
found to be given by the following semi-empirical equation :
Φ/ Φmax ≈ 1 / [1+ (l’f /df ) 1.4/ (k2 μr )] A 4.4.2
Where l’f is the protruding length ( lf - lcoil)
Φ is the electric flux
So in terms of protrusion at one end only this becomes (changing the symbols for dielectric rather than
ferrite) :
Φ/ Φmax ≈ 1 / [1+ (2 l’die /ddie ) 1.4/ (k2 εe )] A 4.4.3
where l’die is the length of the dielectric protrusion at one end
ddie is the diameter of the dielectric material
εe is the permeability of the extension (diluted if a tube)
Combining Equations A4.4.1 and A4.4.3 gives :
εtotal {[ε’eff -1] [ 1+ k1 (l’die /ddie )]+1} / {1+ (2 l’die /ddie ) 1.4/ (k2 εe)} ] A4.4.4
The best match with the above experiments was with k2 equal to 1.55, and this equation is plotted in yellow
in each of the curves above, and shows good agreement with the ceramic and water experiments.
Agreement with the PTFE experiment is less good but the experimental error is higher here because the coil
winding was not mechanically stable and could have changed during insertion of the dielectric.
The above equation can be simplified for a practical situation where the protrusion will be not more than
0.5 ddie and also the effective material dielectric constant will be close to unity, because a low dielectric
constant material will be used and as a tube. The diluted dielectric constant εe can be assumed to be 1.5.
Equation A 4.4.4 can then be simplified to :
εtotal ≈ [ε’eff 1] [ 1+ 0.045 (l’die /ddie )] + 1 A4.4.5
where l’die is the length of the dielectric protrusion at one end
ddie is the diameter of the dielectric material
ε’eff is given by Equation 3.6.2
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
32
0.1
1.0
10.0
100.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with Water Dielectric
Empirical Equation
Measurements 18 turns
Upper Limit
Lower Limit
0.00
0.20
0.40
0.60
0.80
1.00
1.20
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Normalised Permittivity
Protrusion normalised to Diameter
Permittivity v Protrusion Water fill
Permittivity Multiplier : Hot
end
APPENDIX 5 : MEASUREMENTS OF LONGITUDINAL FIELD TAPER
The longitudinal taper of the electric field was measured with coils of five different ratios of length to
diameter of 0.38, 0.5, 1.16, 1.99, and 2.79. Each coil had 18 turns of 1 mm dia wire wound onto a thin
plastic bottle of outer diameter 49.2 mm, giving a mean wire diameter of 50.2 mm (ie the plastic former
shown in Figure A3.2.1).
The lower end of the winding was connected to earth and the VNA was connected to a tap on the coil. The
top of the winding had no lead and no connection. At each length the SRF was measured, water was poured
into the bottle, and the SRF re-measured with the following results :
Figure A5.1 The measured dependence of ε’eff on coil length
The empirical curve in pink is arbitrary, and is included merely to illustrate the trend. The two shorter coils
clearly have a much lower effective dielectric constant than the trend of the longer coils, and this is due to
the field reducing towards the ends. This can be seen from an expansion of Figure A4.3.1, shown below :
Figure A5.2 Change of Slope close to end of Coil
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
33
It can be seen that close to the end of the coil (ie l’die /ddie =0) the slope reduces. Since the curve amounts to
the integral of the flux v length, this implies a reduction of flux density towards the end, to around 40% of
the central flux over a distance of around 0.2 ddie. Clearly a very short coil with a length say of only 0.4 ddie
will have only 40 % of the flux of a longer coil, with a commensurate reduction in the effective dielectric.
To allow for this effect there are three possibilities: a) the Sichak factor Fs = (2πa N)( 2πa/λ) could be
modified, b) Equation 3.6.2 could be modified, or c) the diluted dielectric constant of the dielectric could be
modified. The latter seems the most appropriate in that the diluted value is already a function of the radial
field intensity, and this would make it now also dependent on the longitudinal intensity. It is shown in
Appendix 6 that the diluted dielectric constant is now multiplied by the following :
Longitudinal dilution = 1 x [dc / lc] [ 1- φmin / φmax] A5.1
The factor x is the relative distance from end at which the electric flux intensity reduces, and φmin / φmax is
the reduced flux intensity at the end compared with that in the middle of the coil. We have seen that these
are approximately equal to x= 0.4 and φmin / φmax = 0.4, but the experiment above can give these values
more accurately, and it is shown that these are 0.42 and 0.43 respectively. Using these values and Equation
A6.4 for each of the lengths gives the dilution as :
Coil Length/ Diameter Longitudinal Dilution Overall diluted Dielectric Constant
0.38 0.37 23.6
0.5 0.52 33.4
1.16 0.8 51
1.99 0.88 56.5
2.79 0.91 58.7
(NB in the above the plastic former dilutes the water from εr = 78 to εr = 64.2)
Using the new diluted dielectric constant for each length gives :
Figure A5.3 Results with Longitudinal Dilution
NB A true representation of the results would have a different light-blue line (and thus pink curve) for each
measured point, and equal in value to the overall diluted dielectric constant in the table above. However
this would be confusing and would not allow direct comparison with Figure A5.1. So in the above figure,
all measured values have been increased by the longitudinal dilution as calculated from Equation A5.1, so
that the blue points represent the values that would obtain if there was no longitudinal dilution.
The pink curve is Equation 3.6.2 with α = 4.2 and n=1.1.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
34
1.0
10.0
100.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity with Water Dielectric
Empirical Equation
Measurements 18 turns
Upper Limit
Lower Limit
Balanced Coil
The above experiments used coils with one end grounded, and the results indicate that the flux tapering was
at both ends (in contrast to the effect of extending the dielectric : Appendix 4). To confirm this an 18 turn
coil was wound over a length of 22 mm (length/dia = 0.44) and used in a balanced configuration, whereby
the centre was grounded and both ends left free. Water was used as the dielectric. The VNA was connected
to a tap on the coil 4 turns from the centre. The coil was also measured with one end grounded, with results
two shown below :
Figure A5.4 Balanced and unbalanced Coil
The same longitudinal correction was applied to both configurations (Equation and it can be seen to apply
to both. So the flux tapering along the length is the same whether one end is grounded or not.
The pink curve is Equation 3.6.2 with α = 4.2 and n=1.1 (ie the same as Figure A5.3).
APPENDIX 6 : LONGITUDINAL FIELD TAPER
In the following analysis it is assumed that the longitudinal flux intensity is φmax down the whole length of
the coil except close to each end where it decreases to φmin. It is further assumed that the length of this flux
reduction is related to the diameter of the coil, so that its length is x dc . So assuming the flux intensity is
unity over the length (lc x dc ) and φmin / φmax over a length x dc (ie the length at each end of the coil is 0.5
x/ dc), then the combined flux φt is equal to :
φt = 1 ((lc x dc ) + (φmin / φmax) (x dc ) A6.1
Normalising to the coil length lc gives :
Φ’t = [1 x (dc / lc ) ] + (φmin / φmax ) (x (dc / lc )) A6.2
Where Φ’t is the longitudinal dilution of flux. Collecting terms gives :
Longitudinal dilution = 1 x [dc / lc] [ 1- φmin / φmax] A6.3
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
35
The experiments in Appendix 5 show that best match with experiment was obtained with x = 0.42 and
φmin / φmax = 0.43, so that the above equation becomes :
Longitudinal dilution = 1 0.25 [dc / lc] A6.4
This equation losses accuracy for dc / lc >2.5 and it fails when dc / lc =4 and a better approximation is given
by :
Longitudinal dilution = e x A6.5
where x= - 0.33 [dc / lc]
APPENDIX 7 : THE WAVELENGTH IN SICHAK’S FACTOR
Sichak’s factor FS is equal to :
FS= (π dc /p)( π dc /λ) A7.1
The wavelength λ is assumed to be that down the wire. However it is more convenient to express FS in
terms of the free-space wavelength λ0, and so the above equation becomes :
FS = (π dc /p)( π dc 0) c/Vw A7.2
where Vw/c is the wire velocity as a proportion of c
For an air-cored coil the phase velocity down the wire, Vw/c is less than unity and is given by (see Payne
ref 1):
Vw/c 1/ k
Where k = √20/π [dc2/(λ p)]0.25 A7.3
Therefore k = 0.8 Fs0.25, and thus Vw/c 1.25/ Fs0.25. Inserting this into Equation A7.2 gives :
FS = 0.74 [2 dc2 ) /( p λ0)] 1.33 A7.4
Where p is the pitch of the winding
dc is the diameter of the winding to the centre of the wire
λ0 is the free-space wavelength = 300/f for f in MHz
APPENDIX 8 : MEASUREMENT WITH FERRITE
Appendix 3 describes an experiment with a tube of water formed by inserting a closed plastic tube into a
coil and filling around it with water. To test the effect of ferrite, into this inner tube was inserted 7 ferrite
rods each having a diameter of 10 mm and a length of 120 mm, equal to that of the coil. The permeability
of the rods was not known but inserting them increased the inductance by exactly 4, when measured at a
frequency of 0.1MHz (ie at a frequency much lower than the SRF ).
The effect of the ferrite was determined by measuring the SRF with the ferrite present and then again with
water surrounding it (strictly surrounding the plastic tube into which the rods had been placed). This gave
5.82 and 3.47 MHz respectively, so that the effective dielectric constant was (5.82/3.47)2 = 2.8. From
Equation 5.2 Sichak’s factor was calculated as Fs = 0.532, for √µcoil = 2. This is plotted below along with
Equation 3.6.2 for α = 4.2 and β =1.1 (ie the same as all other experiments) :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
36
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
0.01 0.1 1 10 100
Effective Permittivity
Sichak Factor (πD/p)(πD/λ)
Effective Permittivity of Water Tube with Ferrite Core
Empirical Equation
Measurements
Figure A8.1 Effect of Ferrite permeability
The measured effective dielectric constant was 2.3% below that predicted from the curve.
One aspect not considered above was the dielectric constant of the ferrite itself. This is likely to have a
small effect of the experiment because the two resonant frequencies were measured with this present and it
can be assumed that its effect cancelled in the ratio.
APPENDIX 9 : CALCULATING DIELECTRIC LOSS
The dielectric constant can be written as :
ε = ε’ - jε’’ A9.1
ε’ is the normally quoted dielectric constant and ε’’ is the loss component. Often ε’’ is not given in tables of
constants, and instead they quote the ‘dissipation factor’ tan δ and this is equal to ε’’/ ε’. In RF work it is
more convenient to express loss in terms of Q, and this is equal to :
Qdi = 1/ tan δ = ε’ / ε’’ A9.2
The Q of most materials which might be used as coil formers is 100 or more. For instance Alumina has a Q
of 5000 and glass around 250, depending on the particular glass. Generally the Q is constant over the RF
range of frequencies and this will be assumed to be so in the following analysis.
To determine the extra loss produced by the dielectric, we can use the self-capacitance model and calculate
the loss from the increase in capacitance due to the dielectric former. The self-capacitance is given by :
C = 1/ [(2πfr)2 L0] A9.3
where fr is given by Equation 2.2.2
L0 is the low frequency inductance of the coil
When the former is introduced the capacitance will increase by ε’eff . If the capacitance with no former is C1
and with a former is C2 then the increase in capacitance is given by
C’ = C2 - C1 = [1 1/ε’eff] / [(2πfr)2 L0] A9.4
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
37
where fr is the SRF with former
L0 is the low frequency inductance of the coil
ε’eff is the effective dielectric constant
The reactance of this parallel capacitance is :
Xc = 1/ (2π fr C’) A9.5
Where fr is the SRF with the former present
The value of the parallel resistance Rdp, due to the loss in this capacitance is therefore :
Rdp = Xc Q = [1/ (2π f C’)] [ε’ / ε’’] A9.6
Inserting the value for C’ from Equation A9.4 gives
Rdp = [2π fr2 L0 ] [ε’ / ε’’] ]/ [f (1 - 1/ε’eff)] A9.7
Where ε’eff is the effective dielectric constant
[ε’ / ε’’] is the Q of the dielectric
fr is the SRF with the former present
Notice that if the dielectric Q is constant (normally the case) Rp reduces with frequency.
If this dielectric loss was the only loss then the Q of the inductor would be :
Qf = Rp / (2π f Lm) A9.8
where Lm is given by Equation 2.2.1
If the Q due to the conductor losses is Qc, then the overall Q will be :
Qt = (Qf Qc ) / (Qf + Qc ) A9.9
Often it is more convenient to express the loss as a series resistance, and the conversion is given by :
Rds = (Rdp X2) / (Rdp 2 +X2) A9.10
where Rdp is given by Equation A9.7
X = (2πf Lm )
Lm = Lo / [ 1- (f / fr )2]
Lo is the low-frequency inductance
fr is the SRF with the former present.
APPENDIX 10 : CONDUCTOR LOSS
The series resistance due to the conductor, including the effect of resonance, is (ref 1) :
Rws = K √f / [ 1- (f / fr )2]2 A10.1
where K= Rdc φ 0.25 dw/0.067
Rdc = 4 ρ ℓw /(π d2w)
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
38
1
10
100
1000
0 1 2 3 4 5
Series Resistance R s
Frequency MHz
Rs v Frequency
Predicted Dielectric Rs
Predicted Conductor Rs
Predicted Combined Rs
Measurements (at TOP of coil)
dw is dia of the wire in metres
ρ is the conductor resistivity (1.77 10-8 for copper)
w is the length of the conductor (straight length) in metres
φ is the proximity effect (see ref 1 Appendix 1, and note below).
NB if the ratio of the wire diameter to the pitch is 0.5 (which often gives the highest Q) then φ ≈1.78 for
coils with a length equal to twice the diameter or longer.
Sometimes it is more convenient to express this loss as a parallel resistance Rcp, and the transformation is
given by :
Rcp = (Rcs 2 +X2)/ Rcs A10.2
where Rcs is given by Equation A10.1
X = (2πf Lm )
Lm = Lo / [ 1- (f / fr )2]
Lo is the low-frequency inductance
fr is the SRF with the former present.
APPENDIX 11 : EXPERIMENT TO PROVE LOSS EQUATIONS
A coil was wound similar to Figure A3.2.1, with 59 turns of 1 mm dia wire, onto a thin plastic bottle of
outer diameter 49.2 mm, giving a mean wire diameter of 50.2 mm. The length of the winding was 117 mm
and the inside diameter of the tube was 47.2 mm. Connection to the VNA was between the ends of the coil.
The bottle was filled with distilled water the top of the winding. The water extended by 10 mm beyond the
bottom (earthed) end of the coil, but it is shown in Appendix 4 that this will have had an insignificant
effect.
Measurements of the series resistance were made with the following results (purple curve):
Figure A11.1 Series Resistance including Dielectric Loss
Also shown is the total predicted series resistance being the sum of Equation A10.1 for the conductor loss
(brown) and Equation A9.10 for the series dielectric loss (these are also shown). The error was generally
within ± 6% over the whole resistance range of more than 3 decades.
In the predictions the dielectric Q was assumed to reduce as 1/f (see Appendix 12.2 ) with a value of 30 at
5.35 MHz, the SRF. This value of Q was measured from the bandwidth of the self-resonance, and assumed
to be all due to the dielectric loss since the conductor loss was negligible in comparison at this frequency.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
39
The effective dielectric constant was determined by measuring the SRF’s with and without water and
taking the square of the ratio, so the above results do not include any errors in the theoretical prediction of
the effective dielectric constant.
APPENDIX 12 : DIELECTRICS USED IN EXPERIMENTS
A12.1. Introduction
Three dielectric materials were used in the experiments, water, PTFE and Alumina, and the dielectric
properties of these are discussed below.
A12.2. Water
Water is a very useful dielectric for proving the equations because its dielectric constant is very high.
However it was found that tap water was very lossy, having a Q of only 1.3 at 5.5 MHz. Distilled water has
a much lower loss and so this was used.
The dielectric constant of distilled water is often given as 78.2 in the HF range (ref 9). However, it is
known that the purity of distilled water can affect its dielectric properties and the water used here was taken
from a simple domestic de-humidifier, so its purity was unknown. Consequently the following test was
done to determine ε’ and ε’’.
The UHF capacitor shown below was immersed in the water and the capacitance measured along with the
Q.
Figure A12.2.1 Capacitor used to determine the dielectric properties of water used.
The air capacitance with the vanes fully closed was 12 pf, but not all of this would be affected by the water
since some of this capacitance was due to the ceramic insulation. The vanes at minimum capacitance gave 4
pf, and experience had shown that around half of this is due to the ceramic supports. The capacitance which
would be affected by the water was therefore estimated as 10 ± 0.5 pf. With water the capacitance increased
to 783 pf (averaged over 0.6 to 6 MHz), giving ε’ = (783-2)/10 = 78.1 ± 5%, which is close to the published
value.
As for ε’’ it was expected that this would follow the Debye characteristic shown below :
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
40
0
5
10
15
20
25
0 1 2 3 4 5
Quality Factor Q
Frequency MHz
Q v Frequency
Q
ε'p
Figure A12.2.2 The Debye characteristic
The Debye relaxation time for water is about 10ps at 25oC, giving the peak value of ε’’ at around 16 GHz.
So at the much lower frequencies used here it was expected that ε’’ would be small and decreasing only
slightly as the frequency was reduced. However ε’’ increased as the frequency was reduced, giving a lower
Q at low frequencies as shown below :
Figure A12.2.3 Measured Q of distilled water
This lower Q at low frequencies is due to ionic conduction, as explained in reference 10 :At low
frequencies, the overall conductivity can be made up of many different conduction mechanisms, but ionic
conductivity is the most prevalent in moist materials. er’’ is dominated by the influence of electrolytic
conduction caused by free ions which exist in the presence of a solvent (usually water). Ionic conductivity
only introduces losses into a material. At low frequencies the effect of ionic conductivity is inversely
proportional to frequency and appears as a 1/f slope of the er’’ curve’.
The measured curve has the equation Q = K f, where f is in MHz, and this is consistent with the above
mechanism. K is equal to 3.85 in this case, giving a Q at 5 MHz of 19.4. However it was found that the Q
degraded with time, due probably to the absorption of CO2, and the absorption of sodium ions from the
glass bottle. As an example water straight from the de-humidifier could have Q as high as 80 at 5 MHz. It
was therefore important that the K was determined for each experiment and this was determined from the
bandwidth of the SRF resonance. So for instance the Q of the SRF resonance in Appendix 11 was 30 at
5.35 MHz, and so the equation for the dielectric Q used in the equations was Q = 30 f / 5.35, where f is in
MHz.(the proportionality with f did not change with time).
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
41
A12.3. PTFE
The dielectric constant of pure PTFE is widely reported as 2.1, and constant over a wide frequency range.
However it is common to add glass to PTFE to improve its mechanical properties, and this will increase its
permeability to 2.4 (for 25% glass). The material used here had been bought as ‘virgin’ and so presumably
without a glass filling but to be sure it needed to be measured.
A parallel plate capacitor could have been made by cutting a thin slice of the PTFE from the rod (diameter
15 mm). However the author did not have the tools to cut this accurately, especially the very thin slice
which would be necessary to minimize the fringing field around the edges.
The alternative was to measure the resonant frequency of a twin wire transmission line when it was
immersed in the dielectric. For this a section of the PTFE rod, 186 mm long, was slotted along its length of
two opposing sides to a depth of 4 mm and a copper wire was inserted into these slots and the two wires
connected together at one end. The wire needed to have a diameter large enough to be a tight fit in the slots
so as to minimize any air gap and the wire used was 1.69 mm over its enamel insulation. The slots were
then packed with PTFE tape (plumbers tape). This is shown below :
Figure A12.3.1 Transmission line within PTFE rod
The wires were connected together at one end and the VNA connected to a tap on each wire, about 10 mm
from this end. To contain the field within the PTFE rod it was inserted into a conducting tube, made from
copper foil.
The objective now was to measure the λ/4 resonant frequency, calculate the velocity and compare this with
c. In practice it was easier to compare the resonant frequency with that which would be obtained if the
velocity was c, and this was 300/(4*(0.186) = 403 MHz, for the length of 0.186 meters. Resonance was
measured at 281.2 MHz so the effective dielectric constant was (403/281.2)2 = 2.06. However this assumes
that there is no end effect (whereby the fields extend beyond the end making the wires appear longer). To
test this a copper sheet was placed very close to the end of the dielectric (about 0.5 mm away) and the
resonant frequency reduced by less than 1%. So it would appear that the interface with the air at the end
provided a very good reflection to the waves.
However the resonant frequency was sensitive to hand proximity and so some field was leaving the sides.
To contain this a copper tube of length 0.16 metres was wrapped around the PTFE and its joint soldered. It
was positioned so that one end was flush with the end of the PTFE away from the VNA connections. The
resonant frequency dropped to 271.9 MHz, giving εr =2.2.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
42
The actual dielectric constant will be greater than this because there is air space around the wires (they are
circular in rectangular slots), and the PTFE tape filling the outer part of the slots had air gaps within it.
So it would be reasonable to assume that the dielectric constant was 2.3 ±0.1.
A12.4. Alumina
A tube of Alumina was bought having the following specification :
99% Purity Aluminium Oxide (Alumina). Balance is Silicon Dioxide
Fired to full density so not porous
OD : 6.35 ±0.33 mm (measured by author at 6.25 ±0.1 mm )
ID : 4.75 ±0.33 mm (measured by author at 4.8 ±0.1 mm )
Wall thickness : 0.75 mm average
The supplier did not give the dielectric constant but Reference 11 gives 9.8 for 99.5% at 1MHz, 9 for 96%,
and 9.1 for 94% purity, and assuming a linear increase in dielectric constant versus purity, for 99% purity εr
= 9+0.8*3/3.5 = 9.68.
Payne : The Effect of Dielectric Inside an Inductance Coil (Issue 3)
43
REFERENCES
1. PAYNE A N : ‘Self-Resonance in Coils’,
https://www.researchgate.net/publication/353572229_SELF-RESONANCE_IN_AIR_COILS
2. WELSBY V G : The Theory and Design of Inductance Coils’, Second edition, 1960, Macdonald,
London
3. CUTLER C C : ‘Experimental Determination of Helical-wave Properties’, Proc IRE, Feb 1948,
Vol 36, Pt 2, pp 230-233.
4. LEE S, LEE Y & YU I : ‘Electric Field in Solenoids’, Japanese Journal of Applied Physics, Vol
44, No 7A, 2005, pp 5244-5248 (available on the web).
5. KNIGHT D : ‘Inductor Resonance and Self Resonance Experiments’ .
http://www.g3ynh.info/zdocs/magnetics/appendix/self_res/gallery.html
6. CHOY T : ‘On the effects of dielectric or permeable formers on the inductance and self-
capacitance of solenoids’ .
http://www.g3ynh.info/zdocs/magnetics/appendix/self_res/Choy2015_Effect_of_Dielectric_forme
r.pdf
7. SICHAK W : ‘Coaxial Line with Helical Inner Conductor’ Proc IRE, 1954, pp 1315-1319.
8. PAYNE A N : ‘The Inductance of Ferrite Rod Antennas’,
https://www.researchgate.net/publication/351552681_The_Inductance_of_Ferrite_Rod_Antennas
9. ZAHN M, OHKI Y, et al : ‘ Dielectric Properties of Water and Water/Ethylene’ .
http://www.rle.mit.edu/cehv/documents/35-Proc.IEEE.pdf ).
10. Agilent Application Note : Basics of Measuring Dielectric Materials .
.http://cp.literature.agilent.com/litweb/pdf/5989-2589EN.pdf
11. Accuratus data sheet : http://accuratus.com/alumox.html
12. KNIGHT D : ‘Components and Materials Pt6’ http://www.g3ynh.info/zdocs/comps/part_6.html
Issue 1 : November 2015
Issue 2 : February 2020 : Changes to Equations 4.3.1 and 4.4.1 (and consequently 6.1) to improve accuracy.
Also minor changes and corrections.
Issue 3 : July 2021 : Minor changes
© Alan Payne 2021
Alan Payne asserts the right to be recognized as the author of this work.
Enquiries to paynealpayne@aol.com
... The resonance frequency and the quality factor significantly change with increasing temperature. Variations in the temperature and humidity change the dielectric permittivity and the electrical conductivity of the clay former [28], resulting in changes in the parasitic capacitance C Coil [29]. Losses in the former changes [29], leading to changes in the quality factor. ...
... Variations in the temperature and humidity change the dielectric permittivity and the electrical conductivity of the clay former [28], resulting in changes in the parasitic capacitance C Coil [29]. Losses in the former changes [29], leading to changes in the quality factor. For these experiments, the plastic and clay formers were exposed to the same humidity for the same period of time. ...
Article
Full-text available
Eddy current displacement sensors (ECDSs) are widely used for the noncontact position measurement of small displacements (lift-offs). Challenges arise with larger displacements as the sensitivity of the ECDSs decreases. This leads to a more pronounced impact of temperature variations on the inductance and, consequently, an increased position error. Design solutions often rely on multiple coils, suitable coil carrier materials, and compensation measures to address the challenges. This study presents a single-coil ECDS for large displacement ranges in environments with high temperatures and temperature variations. The analysis is based on a sensor model derived from an equivalent circuit model (ECM). We propose design measures for both the sensing coil and the target, focusing on material selection to handle the impact of temperature variations. A key part of improving performance under varying temperatures includes model-based temperature compensation for the inductance of the sensing coil. We introduce a method to calibrate the sensor for large displacements, using a modified coupling coefficient based on field simulation data. Our analysis shows that this single-coil ECDS design maintains a position error of less than 0.2% full-scale for a temperature variation of 100K for the sensing coil and 110K for the target.
Article
Full-text available
and the self-capacitance myth All coils show a self-resonant frequency (SRF), and as this frequency is approached the inductance and resistance increase while the Q decreases until a frequency is reached where the coil resonates in a similar way to a parallel tuned circuit. Because of this similarity the effect has been attributed to self-capacitance in the coil, and many researchers have tried to reduce this capacitance in order to raise the Q. However nowhere in the coil can this capacitance be measured or deduced, and in fact the rising inductance and loss are explained if the coil is seen as a helical transmission line. This article discusses these issues, and gives accurate equations for the changes with frequency.
Article
Full-text available
and the self-capacitance myth All coils show a self-resonant frequency (SRF), and as this frequency is approached the inductance and resistance increase while the Q decreases until a frequency is reached where the coil resonates in a similar way to a parallel tuned circuit. Because of this similarity the effect has been attributed to self-capacitance in the coil, and many researchers have tried to reduce this capacitance in order to raise the Q. However nowhere in the coil can this capacitance be measured or deduced, and in fact the rising inductance and loss are explained if the coil is seen as a helical transmission line. This article discusses these issues, and gives accurate equations for the changes with frequency.
Article
Full-text available
When a ferrite rod is inserted into an air coil its inductance increases by a large factor, but the widely quoted equations for predicting the new inductance are shown to be flawed. A new theory is presented, based upon the magnetic reluctance, and this gives accurate predictions compared to experiment. Interestingly this shows that the increase in inductance when the ferrite is introduced is independent of the number of turns or their spacing or the inductance of the original air coil. Also if the ferrite permeability is high the increase is dependent only on the overall physical dimensions of the coil and ferrite.
Article
Incluye bibliografía e índice
Article
To obtain an approximate solution for the fields in a coaxial line with a helical inner conductor, the helix is replaced by a fictitious surface that is conducting only in the helix direction, an approximation used in the early work on traveling-wave tubes. Maxwell's equations are solved for the lowest "mode" (all fields independent of angle) when the medium inside the helix has permittivity and permeability different from that of the medium surrounding the helix. Equations for the velocity along the axis, characteristic impedance, attenuation constant, and Q are given. The significant parameter is (2¿Na) (2¿a/¿). N=number of turns per unit length, a = helix radius, and ¿ = wavelength. When this parameter is considerably less than 1, the velocity and characteristic impedance depend only on the dimensions. The dielectric inside the helix has only a second-order effect, while the dielectric outside the helix has a first-order effect. The wave appears to propagate along the helix wire with the velocity of light only when the outer conductor is very close to the helix; as the outer-conductor diameter is increased, the apparent velocity along the wire gradually increases and reaches a limiting value when the outer conductor is infinitely large. For the shapes generally used, the apparent velocity along the wire is rarely more than 30 per cent greater than the velocity of light, but with an infinitely large outer conductor this velocity can be 2 or 3 times the velocity of light.
Article
The properties of the wave propagated along a helix used in the traveling-wave amplifier are discussed. A description is given of measurements of field strength on the axis, field distribution around the helix, and the velocity of propagation. It is concluded that the actual field in the helix described is slightly weaker than would be predicted from the relations presented by J. R. Pierce for a hypothetical helical surface.
Inductor Resonance and Self Resonance Experiments
  • Knight D
KNIGHT D : 'Inductor Resonance and Self Resonance Experiments' .
On the effects of dielectric or permeable formers on the inductance and selfcapacitance of solenoids
  • Choy T
CHOY T : 'On the effects of dielectric or permeable formers on the inductance and selfcapacitance of solenoids' .
Dielectric Properties of Water and Water/Ethylene
  • Zahn M
  • Ohki Y
ZAHN M, OHKI Y, et al : ' Dielectric Properties of Water and Water/Ethylene' .
Components and Materials Pt6
  • Knight D
KNIGHT D : 'Components and Materials Pt6' http://www.g3ynh.info/zdocs/comps/part_6.html Issue 1 : November 2015