|
Index |
Iron-powder materials in HF applications (published in Electron #1, 2006) Introduction This chapter on iron-powder materials is an extension to
the article-series on Ferrites
in HF applications. Next to ferrites iron-powder cores are regularly
applied in comparable conditions and often regarded as just an other HF core material. Qualities of iron powder-cores
however are different enough to dedicate an additional chapter, especially
since these materials are sometimes applied when ferrite-cores where
intended, which may lead to undesired and often dangerous situations, either
to the equipment or to the operator. Two main groups of iron-powder materials may be
distinguished for application 'around' HF circuits, identified by the
manufacturing process: 1. Electrolytic iron powder Flake-like iron particles are formed by an electrolytic
process and cut to microscopic size. This 'flake powder' is mixed with an
isolating binder material and pressed / cured to a high-density material of
the desired shape. Permeability up to 100 may be obtained by this type of
processing. Electrical losses usually will be high since the iron flakes are
sharing mutual contact points that allow relatively large Eddy-currents.
Magnetic domains are somewhat larger than in ferrite materials allowing for
higher flux density before saturation will cut in. Application of
electrolytic iron-powder materials therefore may be found at high(er) DC / low(er) AC current applications, e.g. chokes in switched-mode power supplies
(SMPS). Ferroxcube is manufacturing electrolytic iron-powder toroides, intended for application below one megahertz.
Material 2P40 for instance is characterized by permeability mi = 40 and Q = 16.7, always measured and
specified at 10 kHz. and intended for low-frequency applications. These
materials may also be found in ignition coils for fluorescent lamps. MicoMetals and Amidon also are well known
sources for this material e.g. 'mix-26' type, color-coded yellow-white, and
again intended for SMPS applications. 2. Carbonyl-iron powder Iron particles are formed in a chemical vapor deposition
process from iron-tetra-carbonyl and are much smaller that
in the above process. Again iron particles are mixed and processed with an
isolating binder material but to a lower density. This process yields a lower
permeability material but with very low loss (high Q) since iron particles
are not in mutual contact. Saturation qualities are lower than in
electrolytic iron powder and are comparable or lower than those for ferrites.
Ferroxcube used to apply these materials in tuning
rods for filters in telephony applications, especially because of the low
temperature coefficients as compared to ferrite materials. Other Ferroxcube applications may be found in HF inductors,
usually in rod-type shapes. For this type of application permeability does
not need to be too high and saturation may not quickly be a problem in these
rod-shapes. Materials Q-factor usually will be high to over 10 MHz., allowing for inductor applications for a large
portion of the HF range. Low permeability values in general are related to the small grain size in the carbonyl process. At higher field strength the magnetic domain borders (Bloch walls) will be displaced. When these displacements are crossing the magnetic domain borders, additional energy will be lost. Therefore low permeability iron powder will saturate earlier than higher permeability materials and certainly earlier than ferrites. General carbonyl
properties At the MicroMetals and Amidon
website most carbonyl materials may be found. Higher permeability material
(hydrogen reduced iron powder says Amidon) in the
range of 35 - 100 is recommended for low-frequency applications which is
typical for electrolytic iron powders. Carbonyl material permeability is
ranging from 3 - 35, with highest application frequencies for lowest
permeability types. In general carbonyl type materials will be priced in the
ferrite range or somewhat lower since the first do not require the expensive
sintering process. Iron powder materials in general are mold-pressed after
which a mild-temperature oven process removes / hardens the binder material.
This process may differ somewhat for each manufacturer as Ferroxcube
guarantees its materials up to MicroMetals is currently one of the important iron-powder
manufacturers with a wide range of materials including a 'high performance'
range, which is their indication of carbonyl types. This is not unrealistic
since e.g. 'mix-2' type of material is specified at an initial permeability
tolerance of +/- 5 % and temperature stability of -95 ppm/C, which translates
to high precision and high stability when compared to equivalent ferrite
parameters. It should be noted though Eddy currents to increase at
raised temperatures which means higher materials loss at high(er) power applications. This may start an avalanche of
increasing loss that eventually may destroy the component; 'thermal runaway'
according to MicroMetals. When comparing above 'high performance' materials to widely
applied ferrite types like 4C65 (61), permeability tolerance at the latter is
+/- 20% with a temperature coefficient between 0 and As a first conclusion it may be decided carbonyl iron
powder is the better material for HF resonant applications with the lower permeability ferrites (4C65, 61) to excel in
wide-band / high power circuits. This is in line with MicroMetals,
suggesting: " Broadband transformers with iron powder cores will
not have the wide bandwidth attainable with high permeability ferrite
cores". Also in the professional world, iron-powder materials
will not be regarded as competitors to (NiZn)
ferrites as permeability of the first is rather low and materials loss at
higher flux densities is higher than with ferrites. It should be noted that
all permeability measurements are performed at a very low measuring flux of
0,1 mT, to avoid hysteresis effects. This is a
generally accepted technique which also applies to ferrite materials.
Q-factor will be defined as the 'bare-materials' properties and defined as
μ' / μ", disregarding other loss mechanisms e.g. copper loss.
A practical inductor therefore will always exhibit a somewhat lower quality factor.
When discussing these Q-factors it should be noted that
some manufacturers may be using different definitions. Especially when using
the inverse Q as the materials 'Loss-factor', these manufacturers are
defining this to the initial permeability in stead
of the permeability at frequency! Most
manufacturers are presenting materials overviews as a selection mechanism. MicroMetals is presenting a materials overview that gives
a fair impression of a large portion of the carbonyl iron powder field.
Unfortunately US manufacturers do not often specify permeability (μ' ) and loss (μ")
curves over frequency as in In impression of the materials application area may be obtained from the resonant range figures. This range represents the higher Q-values, usually over 75. MicroMetals Carbonyl iron powder
materials
Table 1: Carbonyl iron-powders by MicroMetals A general trend to much lower permeabilities as compared to the ferrites table is evident as also the higher frequency range for resonant applications. Consequence of the low permeability is the non direct relationship between the (square of the) number of turns and final inductance. Because of this low permeability, some flux leakage may be expected so winding style will influence final induction as may be appreciated from figure 1, also by MicroMetals.
This winding effect is not
present at the (usually much) higher permeability ferrites. Since magnetic 'resistance'
is much lower in the latter, no flux will leak outside the core. Manufacturers
usually present a winding factor for a particular core shape and material
type, usually expressed in nano-Henry per turn
squared (nH/n2). Some popular -
T200-2 (2", mix-2 toroide) is
specified at -
TN36/23/15-4C65 ( For
a 10 turn inductor, the first will show an inductance of around 1,2 μH, depending on winding style (figure 1), while the
second will exhibit an inductance of 17 μH,
independent of the winding style. With the To
arrive at the more technical definition of nH/n2
, the μH/100 figure should be divided by a
factor of In general a number of specific
measures should be taken care of when constructing a high-Q inductor on
low-permeability materials: - Make turns stay close and next to each
other (avoid leakage inductance) - Use the core efficiently (use full winding
space). An inductor that is fully filling the winding space will have a higher Q at equal inductance as an
inductor this will not fully occupied winding space (on a differently shaped coil former) -
Mind the For all inductor types one
should note - Apply - It is not useful (and
should be avoided) to apply litze (multiple isolated strands) type of wire on
HF frequencies above 2 MHz. The gain of the higher surface area is more than
lost by the increasing parasitic parallel capacitance of this type of wire,
which also will have a diminishing effect on system Q. - Do not apply more than one layer of wire.
Parasitic parallel capacitance will increase excessively with each new layer. Core
resistivity Resistivity for iron powder
materials is in the order of 0,5 Ohm.m., as
compared to at least 50 kOhm.m. at NiZn ferrite materials, all measured at 1 MHz. This basic resistivity difference is showing in more
than one way at practical inductors. When making a coil on an iron
powder coil-former, one should always start to apply an isolating layer
before putting on turns. Depending on the coil former finishing, edges could
be sharp and may cut through the wire coating, shortening the inductor. This
additional isolating layer may will add to the leakage flux of these low
permeability materials. The highly conductive,
iron-powder coil former will have an increasing effect on the parasitic
capacitance of an inductor on this material. Increased capacitance will lower
maximum operational frequency. Some
practical measurements Various inductors have been constructed and measured on
popular iron-powder coil formers in the carbonyl range, especially on T200-2,
T68-2 and T50-2 toroides. In table 2 some of these
measurements may be found as made on a 6 turns inductor on a T200-2 toroide.
Table 2: Measurements to mix –2 carbonyl toroide inductor Surprisingly the materials permeability (μ') is
higher than specified by the manufacturer. This permeability is nicely constant
over a wide frequency range which also shows at the inductance column. The
second observation is the low value of the equivalent series resistor (second
column), that is related to the materials loss factor of μ". Above
20 MHz. loss is increasing and the effects of loss
and series resistance are showing in the last column for the Q-values. For
this inductor, highest quality factors will be obtained around 10 MHz. to drop off rather sharply at higher frequencies.
These measurements show this material may be applied in resonating circuits
up to 20 MHz. A series of comparable measurements have been made at a
Table 3: Measurements to a 4C65 (61) ferrite toroide Again permeability is nicely constant across a wide range
of frequencies. Up to 10 MHz. this material also is
very suitable for high-Q (low loss) applications. Above this frequency the
quality factor is dropping-off sharply and this effect is starting at a lower
frequency when compared to mix-2 material. Because of the much higher permeability, total impedance
is seven times higher at all HF frequencies and will stay that way for a long
time thereafter in spite of the lowering Q-values. This effect allows a choke
or a transformer at 4C65 material to consist of less turns for a required
impedance and therefore to also show a lower parasitic capacitance. Ferrite
materials therefore in general offer wider band-width for this type of
application. In the article on Ferrites for HF
applications we derived a formula for the maximum voltage across an
inductor on ferrite core materials, based on core loss mechanisms. With
ferrite materials the maximum induction is in the order of 300 mT, where carbonyl materials may only be saturated at 1
T. The formula for maximum voltage in the latter material may therefore be
adapted tot this new situation, by changing the
loss-correction factor. ______________________ Umax = Ö(Pmax.core . (Q/2 + 1/Q) . XL)
(1) with: Umax = maximum
voltage across the inductor-on-core Pmax = maximum loss power in the core Q = quality
factor (XL / r , also m’ / m”) XL =
reactance of the inductor After measuring a For the following calculation it has been assumed these
thermal considerations also to apply to iron-powder materials. It should be
noted though, these materials to have been constructed differently
(press-molded and dried at low(er) temperatures)
and so the thermal resistance may be different (lower). In the chapters on ferrite materials we found that
roughly above 1 MHz. maximum allowed voltage across
an inductor is determined by the material loss; at lower frequencies the
maximum allowed voltage for linear application will determine the limits. While taking thermal conditions for mix 2 and 4C65
materials to be equal we have calculated maximum allowed system power in a 50
Ohm system to be applied to a
Graph 1 is
showing 4C65 ferrite material to be applied at (much) higher system power
between 1 and 30 MHz. compared to the iron-powder
core. This is mainly due to the (much) higher impedance of the inductor with
the same number of turns. Only at around 30 MHz.
the two cores may handle system power equally, mainly because of the lowering
Q of 4C65 ferrite, meaning loss factors are becoming more important. This
also applies to the mix-2 material. The graph is also showing that at a maximum system power of 100 Watt, the 4C65 component to be fit for a frequency range below 1 MHz. to 30 MHz., with mix 2 to handle this amount of power only at around 30 MHz. The calculation does not take into account the higher parasitic capacitance at mix-2 material because of low core material resistivity. This will influence the high-frequency cut-off.
What if we
were to apply the bigger T200 toroide ( To start off, the impedance would go down by around 7 % because of the less favorable relation between core-area versus magnetic path length. The T220 core however has a bigger volume and since thermal resistance is scaling with the root of the volume ratio, core dissipation will go up by 29 %. Taking both effects into account, maximum allowable system power would go up by 20 %, which is low pay-off for this 40 % size increase. Increasing the
number of turns would be much more effective. The All in all graph 1 illustrates the effectiveness of high permeability materials (ferrites) in power applications. Even at low quality factors, impedances are easily much higher than the same number of turns on a iron powder core and will allow higher power / bandwidth ratio's. In general one should be very much aware when designing baluns for non-resonating antenna systems and / or in high impedance environments like symmetrical (300 - 600 Ohm) feed-lines. To still be effective, equivalent parallel inductances of the particular balun should be very high indeed. To still be efficient over a wide band width, the number of turns of this balun should be low enough to not have parasitic effects spoil the upper frequency limit, while still guarantee a high enough impedance at the lowest operating frequency. Low permeability materials therefore are not very much suited for this kind of applications with damaged components and / or transceiver as a result should this advice be neglected. At the end of our discussion on iron power materials it
may be useful to generate a global overview on the various core materials,
the specific application aria and the frequency range of choice. This may be
found in table 4.
Figure 4: Application area's for inductor core materials The table in figure 4 presents a global overview of
application area's that should be regarded in an un-dogmatical
way. The columns for m, Q, Tco and Bsat are indicating which parameter is the more important
in that particular application area, with '+' for 'important' and '~' for 'no
dominant factor'. The composition column is showing MnZn
for manganese / zinc ferrite materials, the high permeability ferrites
(μ' > 1000), NiZn for nickel / zinc ferrite
materials, the lower permeability ferrites (100 < m' < 1000), electrolytic iron for high
permeability iron-powder materials (35 <
m' < 100) and Carbonyl for low permeability iron-powder
materials (2 < m' < 25). It should be noted that each materials group represents a
large 'community' with indicated frequency limits for the best materials in
that community. In this overview the general tendency is clear for
electrolytic iron-powder materials to be found in LF applications like
switch-mode power supplies, MnZn to be applied up
to the lower HF area (and higher for choking purposes), NiZn
for the greater part of the HF area (and up to 200 MHz in choking
applications) and carbonyl for the higher HF area and above. Although most important parameter have been showed in the
table, for each application all parameters should be regarded to prevent
unpleasant surprises in the final system.
Bob J. van Donselaar, on9cvd@veron.nl |
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
