`TRANSFORMERS
`
`by Jerry Sevick, WZFMI
`
`I Znd Edition I
`
`•
`
`Published b y - - - - - - - (cid:173)
`the American Radio Relay League
`
`225 Main Street
`Newington, CT 06111
`
`Page 1 of 23
`
`Samsung Exhibit 1029
`
`
`
`..
`
`Copyright© 1990 by
`
`The American Radio Relay League, Inc
`
`Copyright secured under the Pan-American
`Convention
`
`This work is publication No. 75 of the Radio
`Amateur's Library, published by the League. All
`rights reserved. No part of this work may be
`reproduced in any form except by written
`permission of the publisher. All rights of
`translation are reserved.
`
`Printed in USA
`
`Quedan reservados todos los derechos
`
`ISBN: 0-87259-296-0
`
`$20.00 in USA
`
`Second Edition
`Second Printing, 1991
`
`Page 2 of 23
`
`
`
`Contents
`
`Preface
`About the Author
`
`Chapter 1 Analysis
`1.1
`Introduction
`1.2 The Basic Building Block
`1.3 The Guanella Analysis
`1.4 The Ruthroff Analysis
`
`Chapter 2 Low-Frequency Characterization
`2.1
`Introduction
`2.2 Low-Frequency Analyses of Ruthroff 's
`1 :4 Transformers
`2.3 Low-Frequency Analyses of Guanella's
`1 :4 Transformers
`2.4 The Rod v the Toroid
`2.5 Rod Parameters
`
`Chapter 3 High-Frequency Characterization
`3.1
`Introduction
`3.2 Experiment v Theory
`3.3 To Twist or Not to Twist
`3.4 The Autotransformer v the
`Transmission Line Transformer
`Ferrites and Frequency Response
`
`3.5
`
`Chapter 4 Transformer Parameters for Low-Impedance
`Applications
`4.1
`Introduction
`4.2
`Stripline Transformers
`4.3 Low-Impedance Coaxial Cable Transformers
`4.4 The Third Winding
`
`Page 3 of 23
`
`
`
`Chapter 5 Transformer Parameters for High-Impedance
`Applications
`5.1
`Introduction
`5.2 High-Impedance Limitations
`5.3 Long Transmission Lines
`5.4 Variable Characteristic Impedance Lines
`5.5
`Series Transformers
`
`Chapter 6 1:4 Unbalanced-to-Unbalanced Transformer Designs
`6.1
`Introduction
`6.2
`Schematics and Pictorials
`6.3
`12.5:50-0 Ununs
`6.4
`25:100-0, 50:200-Q and 75:300-0 Ununs
`
`Chapter 7 Unbalanced-to-Unbalanced Transformer Designs
`with Impedance Ratios Less Than 1:4
`7 .1
`Introduction
`1:1.5Ununs
`7.2
`7.2.1 Tapped-Bifilar Transformers
`7 .2.2 Quintufilar Transformers
`1:2 Ununs
`1:3 Ununs
`
`7.3
`7.4
`
`Chapter 8 Unbalanced-to-Unbalanced Transformer Designs
`with Impedance Ratios Greater Than 1:4
`8.1
`Introduction
`8.2 Guanella Transformers
`8.2.1 5.56:50-0 Ununs
`8.2.2 50:300-0 Ununs
`8.2.3 50:450-0 Ununs
`8.2.4 50:600-0 Ununs
`8.2.5 50:800-Q Ununs
`8.3 Ruthroff-Type Transformers
`8.3.1 5.56:50-0 Ununs
`8.3.2 50:450-0 Ununs
`8.3.3 3.125:50-0 Ununs
`8.4 Ruthroff-Guanella Transformers
`8.5 Coaxial Cable Transformers: Ruthroff-Type
`
`Page 4 of 23
`
`
`
`Chapter 9 Baluns
`9 .1
`Introduction
`9.2 The 1: 1 Balun
`9.2.1 Rod v Toroidal Baluns
`9.2.2 Bifilar v Trifilar Baluns
`9.2.3 Air-Core v Ferrite-Core Baluns
`9.3 The 1 :4 Balun
`9.3.1 50:200-Q Baluns
`9.3.2 75:300-Q Baluns
`9.3.3 25:1 00-Q Baluns
`9.3.4 12.5:50-Q Baluns
`9.4 The 1:9Balun
`9.5 Baluns for Yagi, Quad and Rhombic Antennas
`9.5.1 Yagi Beams
`9.5.2 Quad Antennas
`9.5.3 Rhombic Antennas
`
`Chapter 10 Multimatch Transformers
`10.1
`Introduction
`10.2 Dual-Output Transformers
`10.2.1 1:1.5 and 1:3 Ratios
`10.2.2 1 :2 and 1 :4 Ratios
`10.3 Parallel Transformers
`10.4 Dual-Output, Parallel Transformers
`10.5 8-Ratio Transformer
`
`Chapter 11 Materials and Power Ratings
`11 .1
`Introduction
`11. 2 History of Ferrites
`11.3 Experimental Results
`11.4 Power Ratings
`11.5 Suppliers of Materials
`
`Chapter 12 Simple Test Equipment
`12. 1 Introduction
`12.2 The Wheatstone Bridge
`12.3 A High-Frequency Resistive Bridge
`12.4 Signal Generators
`12.5 Efficiency Measurements- The Soak Test
`12.6 Characteristic Impedance Measurements
`
`Page 5 of 23
`
`
`
`Chapter 13 Hints and Kinks
`13 .1
`Introduction
`13.2 Selecting Ferrites- Substitutions
`13.3 Winding Rod Transformers
`13.4 Winding Toroidal Transformers
`13.5 Constructing Low-Impedance Coaxial Cable
`13.6 The Care and Handling of Ferrite Transformers
`Chapter 14 Summary Statements
`Chapter 15 References
`
`Page 6 of 23
`
`
`
`Chapter 1
`
`Analysis
`
`Sec 1.1 Introduction
`
`T here are two basic methods for constructing broadband, imped(cid:173)
`
`ance-matching transformers. One employs the conventional
`transformer that transmits the energy to the output circuit by
`flux linkages; the other uses the transmission line transformer to trans(cid:173)
`mit the energy by a transverse transmission line mode. With techniques
`exploiting high magnetic efficiency, conventional transformers have
`been constructed to perform over wide bandwidths. Losses on the order
`of one decibel can exist over a range from a few kilohertz to over
`200 MHz. Throughout a considerable portion of this band, the losses
`are only 0.2 dB.
`On the other hand, transmission line transformers exhibit far
`wider bandwidths and much greater efficiencies. The stray inductances
`and interwinding capacitances are generally absorbed into the charac(cid:173)
`teristic impedance of the transmission line. As such, they form no
`resonances that could seriously limit the high-frequency response. Here
`the response is limited by the deviation of the characteristic impedance
`from the optimum value; the parasitics not absorbed into the character(cid:173)
`istic impedance of the transmission line; and, in some transformer
`configurations, the length of the transmission line.
`With transmission lines, the flux is effectively canceled out in the
`core and extremely high efficiencies are possible over large portions of
`the passband-losses of only 0.02 to 0.04 dB with certain core materials.
`Therefore, the power ratings of transmission line transformers are
`determined more by the ability of the transmission lines to handle the
`voltages and currents than by the size and conventional properties of
`the core.
`The earliest presentation on transmission line transformers was
`by Guanella in 1944 (ref 1). 1 He proposed the concept of coiling
`
`1 Each reference in this chapter can be found in Chapter 15.
`
`Analysis
`
`1-1
`
`Page 7 of 23
`
`
`
`transmission lines to form a choke that would reduce the undesired
`mode in balanced-to-unbalanced matching applications. Before this
`time, this type of device, known as a balun, was constructed from
`quarter- or half-wavelength transmission lines and, as such, had very
`narrow bandwidths. By combining coiled transmission lines in parallel(cid:173)
`series arrangements, he was able to demonstrate broadband baluns with
`ratios of 1 :n2, where n is the number of transmission lines. Other writers
`followed with further analyses and applications of the balun transformer
`introduced by Guanella (refs 2-8). In 1959, Ruthroff published another
`significant work on this subject (ref 9). By connecting a single trans(cid:173)
`mission line such that a negative or a positive potential gradient existed
`along the length of the line, he was able to demonstrate a broadband 1 :4
`balun or unun (unbalanced-to-unbalanced) transformer. He also intro(cid:173)
`duced in his paper the hybrid transformer. Many extensions and appli(cid:173)
`cations of his work were published and are included in the reference list
`(refs 10-28).
`In a general comparison, it can be said that the transmission line
`transformer enjoys the advantage of higher efficiency, greater
`bandwidth and simpler construction. The conventional transformer,
`however, remains capable of de isolation. The purpose of this chapter
`is twofold: to review Guanella's and Ruthroff's approaches to the
`analysis and understanding of these new wide-band transformers, and
`to present additional material to form a basis for the chapters that follow.
`
`Sec 1.2 The Basic Building Block
`The single bifilar winding, shown in Fig 1- 1, is the basic building
`block for the understanding and design of all transmission line trans(cid:173)
`formers. Higher orders of windings (trifilar, quadrifilar and so on) also
`perform in a similar transmission-line fashion and will be discussed in
`Chapter 7, which treats the subject of fractional-ratio transformers.
`The circuit of Fig 1- 1 can perform four different functions de(cid:173)
`pending upon how the output load, RL, is grounded. The functions are:
`(A) a phase-inverter when a ground is connected to terminal 4, (B) a
`balun when the ground is at terminal 5 or left off entirely (a floating
`load), (C) a simple delay line when a ground is at terminal 2, and (D) a
`"boot-strap" when + V 1 is connected to terminal 2. The operation of
`these four functions can be explained by simple transmission line theory
`and the choking reactance of the transmission lines. This choking
`reactance, which isolates the input from the output, is usually obtained
`
`1-2 Chapter 1
`
`Page 8 of 23
`
`
`
`Chapter 11
`
`Materials and Power
`Ratings
`
`Sec 11.1 Introduction
`
`I n Chapter 2 it was shown that the role of a transmission line
`
`transformer core is to enhance the choking action of the coiled
`windings to better isolate the input from the output circuit. Even
`though calculations were made for only the low-frequency performance
`of these transformers, experimental results demonstrated that the core
`influenced most of the operating region (see Sec 3.4).
`The analysis for toroidal cores confirms that the low-frequency
`response is directly related to the permeability of the core material. With
`higher permeabilities, fewer turns are required, allowing for higher-fre(cid:173)
`quency operation. It was also shown that rods, although not yielding the
`high reactance of the toroids, exhibit good low-frequency response and
`can be used in many applications. At low frequencies, the rod was found
`to be relatively insensitive to the permeability of the rod material
`because of the high reluctance of the air path around the core.
`Accurate loss measurements have shown that only a limited
`number of ferrite materials are useful in power applications, where high
`efficiency is an important consideration. This chapter describes trans(cid:173)
`mission line transformers that use nickel-zinc ferrite cores with per(cid:173)
`meabilities in the moderately low range of approximately 50 to 300, to
`yield efficiencies in excess of 98%. No conventional transformer can
`approach this performance. The losses are not a function of current as
`in the conventional transformer, but are, in most cases, related to the
`impedance levels at which the transformers are operated. This suggests
`a dielectric loss, rather than the conventional magnetic loss caused by
`core flux.
`Included in this chapter are: (A) a history of ferrites, (B) experi(cid:173)
`mental results on many of the typical ferrites when used with Ruthroff
`
`Materials and Power Ratings
`
`11·1
`
`Page 9 of 23
`
`
`
`l :4 unbalanced-to-unbalanced transformers, (C) new concepts in power
`ratings, and (D) suppliers of materials used in transmission line trans(cid:173)
`formers. Although no efficiency measurements have been obtained on
`Guanella transformers, it can be safely assumed that the results would
`be as good as those found on Ruthroff transformers.
`
`Sec 11.2 History of Ferrites
`With the discovery of magnetic ferrites by T. Takei in Japan, and
`the excitement brought about by the 1947 publication of work done at
`the Philips Research Laboratories in the Netherlands during World War
`II, new and improved devices emerged in the field of magnetics (refs
`30, 31). 1 The chemical formula for ferrites is MFe204, where M stands
`for any of the divalent ions: magnesium, zinc, copper, nickel, iron,
`cobalt or manganese, or a mixture of these ions. Except for compounds
`containing divalent iron ions, ferrites can be made with bulk resistivities
`in the range of 102 to 109 n-cm, compared to 10-5 n-cm for the
`ferromagnetic metals (such as powdered iron). This increase in bulk
`resistivity represents a major step forward for applications in frequency
`ranges heretofore unobtainable.
`Ferrite compositions are made by ceramic technology. This in(cid:173)
`volves intimate mixing of fine powders of appropriate oxides, com(cid:173)
`pressing the mixture and firing it in carefully controlled atmospheres at
`temperatures of about 1100 °C to 1200 °C. Single crystals have been
`made by several techniques. By using different combinations of oxides
`previously listed and variations in ceramic processing, the mixtures can
`be tailored to fit a wide variety of technical requirements. In fact, ferrites
`with similar specifications by various manufacturers have been found
`to exhibit different efficiencies in transmission line transformers. This
`is because the ferrite is not completely defined by its chemistry and
`crystal structure; it is also defined by its processing, that is, powder
`preparation, compact formation, sintering and machining the ferrite to
`its final shape.
`By the early 1950s, it was generally recognized that inductor
`cores of Permalloy dust had reached the point of diminishing returns in
`their application to higher and higher frequencies. A major contribution
`to the solution of the problem was made in 1952 by F. J. Schnettler and
`A. G. Ganz of Bell Labs, who developed a high-permeability manga-
`
`1 Each reference in this chapter can be found in Chapter 15.
`
`11 ·2 Chapter 11
`
`Page 10 of 23
`
`
`
`nese-zinc ferrite for use at telephone carrier frequencies of 100 kHz and
`higher (ref 32). This material has also found widespread use at lower
`frequencies in power transformers, fly back transformers and deflection
`yokes.
`In the 1960s and 1970s, Bell Labs scientists also made several
`important advances in linear ferrite properties, in response to the rising
`need for high-quality linear devices in the transmission area. This
`resulted in the development of a process for making suitable nickel-zinc
`ferrites capable of operating up to 500 MHz (ref 33). These advances
`were made by using cobalt additives and carefully controlled cooling.
`This form of ferrite is the best one for high-power transmission line
`transformers, and it is commercially available.
`While the use of ferrites in inductors and transformers for carrier
`frequencies had a major impact on communications, its impact on
`microwave and computer technology is of equal importance. The avail(cid:173)
`ability of magnetic oxides eventually led to the large family of non(cid:173)
`reciprocal magnetic devices that play a key role in microwave
`technology. The materials effort is credited largely to L. G. Van Uitert
`of Bell Labs. He proposed the substitution of nonmagnetic ions for
`magnetic ions in the ferrite structure to reduce internal fields and
`thereby lower the ferromagnetic resonance frequency.
`In the computer field, A. Schonberg of Steatit-Magnesia AG in
`Germany, and workers at MIT's Lincoln Laboratories, found a family
`of magnesium-manganese ferrites with remarkably square hysteresis
`loops for use in memory and other computer and switching applications.
`These devices subsequently gave way to the semiconductor logic and
`memory circuits of the mid-l 970s.
`Although considerable information is available on the theory and
`application of transmission line transformers, dating back to the classic
`papers of Guanella in 1944 and Ruthroff in 1959, virtually no investi(cid:173)
`gations have been made on the use of ferrites in power applications (refs
`1, 9). Discussions by the author with scientists and engineers from major
`laboratories working in the ferrite field confirm this lack of develop(cid:173)
`ment. The following sections will show the results obtained by the
`author on readily available ferrites and their use in high-power trans(cid:173)
`mission line transformers. It is of some interest to note the differences
`in the properties of ferrites resulting from variations in processing
`techniques used by the different manufacturers.
`
`Materials and Power Ratings
`
`11-3
`
`Page 11 of 23
`
`
`
`Sec 11.3 Experimental Results
`
`Early experiments by the author indicated that the bulk resistivity
`of ferrite material could be related to high-efficiency operation. There(cid:173)
`fore, many of the major suppliers were asked to supply samples of their
`highest-resistivity material. Table 11 - 1 is a list of suppliers who pro(cid:173)
`vided samples, the code symbol for their materials, the low-frequency
`(initial) permeabilities and the bulk resistivities. Powdered iron was
`included because it has been used for some applications, but as will be
`shown, it suffers by comparison because of its very low permeability.
`All of the data presented in this book on loss as a function of
`frequency were obtained at Bell Labs on a computer-operated transmis(cid:173)
`sion measuring set with an accuracy of one to two millidecibels over a
`frequency range of 50 Hz to 1000 MHz (refs 34, 35, 36). As a reference,
`it should be noted that a loss of 44 millidecibels represents a loss of I%,
`or an efficiency of 99%. Actual data shows that many of the transform(cid:173)
`ers, made using the best ferrite materials, exhibit losses over a consid-
`
`Table 11-1
`
`Cores, Suppliers and Specifications
`
`Material
`
`Supplier
`
`01 (NiZn)
`
`Allen-Bradley
`(formerly Indiana General)
`Allen-Bradley
`G (NiZn)
`Allen-Bradley
`02 (NiZn)
`H (NiZn)
`Allen-Bradley
`Ferroxcube
`4C4 (NiZn)
`3C8 (MnZn)
`Ferroxcube
`K5 (NiZn)
`MH&W Intl (TDK)
`KR6 (NiZn)
`MH&W Intl (TDK)
`CMD5005 (NiZn) Ceramic Magnetics
`C2025 (NiZn)
`Ceramic Magnetics
`CN20 (NiZn)
`Ceramic Magnetics
`C2050 (NiZn)
`Ceramic Magnetics
`E (Powdered
`Arnold Engineering,
`Iron)
`Amidon Associates
`
`11-4 Chapter 11
`
`Permeability Bulk Resistivity
`(Q-cm)
`108
`
`125
`
`300
`40
`850
`125
`2700
`290
`2000
`1400
`175
`800
`100
`10
`
`106
`
`109
`104 -105
`
`107 - 108
`102 - 103
`2 x 108
`105 -106
`7 x 109
`5 x 106
`106
`3 x 107
`10-2
`
`Page 12 of 23
`
`
`
`erable portion of their passbands of only 20 millidecibels, equivalent to
`efficiencies of 99.S%. Since short windings (wire lengths of 10 to lS
`inches) are generally used, very little loss is attributed to the windings.
`In fact, wires as thin as no. 18 can easily handle a kW of power. Although
`the previous chapters stress theory and design, many of the experimental
`results presented do display the high efficiency of the transmission line
`transformer. This section reveals the differences between the various
`ferrites and, in particular, stresses the effect of operation at impedance
`levels greater than 100 n.
`Figs 11-1, 11 -2 and 11-3 show the loss results of three 4:1
`transformers operating at three different impedance levels. Two are
`transmission line transformers and one is an autotransformer. The cores
`are similar in size, varying from 2.4 inches in OD for the Q 1 cores and
`2.62S inches for the KS core. The two transmission line transformers
`had S tightly wound bifilar turns of no. 14 wire, and the autotransformer
`had 10 tightly wound turns of no. 14 wire. Fig 11-4 depicts this type of
`winding. Fig 11 - 1 shows that the two transmission line transformers had
`about the same loss above 2 MHz. Below 2 MHz, the KS material was
`
`(/)
`
`0
`0.1
`0.2
`m 0.3
`'O 0.4
`(/) 0.5
`9 0.6
`a:::
`0.7
`LL.I u
`:::> 0.8
`0
`(/)
`z
`0.9
`<(
`a:::
`1.0
`.....
`1 .1
`1.2
`1.3
`1.40.1
`
`4:1 (50:12.5fi)
`TRANSFORMERS
`e TRANSMISSION LINE,
`K5
`6 TRANSMISSION LINE,
`Q1
`o AUlO, Q1
`
`I
`I
`
`1.0
`
`10
`FREQUENCY (MHz)
`
`100
`
`Fig 11 -1- Performance of 4:1 transformers operating at the 50:12.5-n
`level.
`
`Materials and Power Ratings
`
`11 -5
`
`Page 13 of 23
`
`
`
`J::r
`
`/a-0.00 ..
`
`t~f
`.
`
`.
`
`Q ••
`·q
`
`...
`
`tl.
`\
`q
`
`t
`I
`t
`1
`I
`~
`. i • TRANSMISSION LINE,
`K5
`~
`TRANSMISSION LINE,
`6 Q1
`~
`0 Aum, Q1
`
`4:1 (75:18.75il)
`TRANSFORMERS
`
`¢
`
`~
`
`1.0
`
`10
`FREQUENCY (MHz)
`
`100
`
`Fig 11-2- Performance of 4:1 transformers operating at the 75:18.75-Q
`level.
`
`0
`
`0 .1
`
`0 .2
`0.3
`m 0.4
`~
`(/) 0.5
`(/)
`0 0.6
`...J
`a::: 0.7
`w
`B 0.8
`(/) 0.9
`z
`<{ 1.0
`a:::
`I-
`1.1
`
`1.2
`
`1 .3
`
`1.4
`
`0.1
`
`0
`0 .1
`0.2
`0.
`ID 0.4
`~
`(/) 0 .5
`(/) g 0 .6
`a:::
`w 0 .7
`u 25 0 .8
`~ 0.9
`<{ 1.0
`a:::
`I-
`1.1
`1.2
`1.3
`1.4
`0.1
`
`I
`
`~
`'lt:r.0-0·0-o
`f/f
`..
`•
`l
`I
`
`G..
`
`~
`
`'\
`\
`~
`
`\
`
`4:1 ( 100: 25n)
`TRANSFORMERS
`•
`TRANSMISSION LINE,
`K5
`TRANSMISSION LI~,
`6 Q1
`0 AUTO,Q1
`
`~~
`f
`fl
`!/
`
`1.0
`
`10
`FREQUENCY (MHz)
`
`100
`
`Fig 11-3-Performance of 4:1 transformers operating at the 100:25-n
`level.
`
`11 -6 Chapter 11
`
`Page 14 of 23
`
`
`
`superior because of its higher permeability (290 compared to 12S for
`Ql material). The autotransformer showed not only a much narrower
`bandwidth, but also the greater loss of the conventional transformer
`(0.2 dB), which transmits the energy by coupling through flux linkages.
`Figs 11-2 and 11-3 show an interesting phenomenon of the KS material.
`At about 7 MHz, another loss mechanism comes into play. At the higher
`impedance levels of 100:2S Q of Fig 11-3, the loss beyond 10 MHz is
`more than double that of the Q 1 material. In fact, at 10 MHz, the loss is
`only 3S millidecibels for the Ql material and 80 millidecibels for the
`KS material. Fig 11-3 also shows that the 100:2S-Q impedance level
`yields the better high-frequency response for the type of winding used
`in the test. Although it is difficult to see from the idealized drawings in
`these figures, actual data shows that the midband loss of the autotrans(cid:173)
`former decreases from 0.22 dB to 0.19 dB in going to the higher
`impedance level of 100:25 n. This confirms the expected decrease in
`core loss at the higher impedance levels, where the currents are lower.
`To examine further the increase in the loss of transmission line
`transformers as a function of impedance levels, four 4: 1 transformers,
`using different core materials with windings similar to Fig 11-4, were
`
`Fig 11·4-The type of winding used In the 4:1 transformers of Figs 11-1,
`11-2 and 11 ·3.
`
`Materials and Power Ratings
`
`11·7
`
`L
`
`Page 15 of 23
`
`
`
`constructed and tested at the 200:50-Q level. The transformers used KS
`'
`Q 1, Q2 and Carbonyl E materials. The Q2-core transformer has 9 bifilar
`turns instead of 6 because of its lower permeability of 40. The results
`of the experiment are shown in Fig 11-5. The K5 and Ql cores show
`more loss at the higher impedance level. Also, there is a considerable
`slope in the passband for these two materials, which shows that the loss
`is frequency dependent. But the Q2 material exhibited the same constant
`low-loss characteristic of the Ql material at the lower impedance
`levels-about 40 millidecibels. The Carbonyl E core again exhibited a
`much poorer low-frequency response because of its low permeability
`of 10. Note that although the core's percentage bandwidth was small,
`the powdered iron exhibited a midband loss of only 0.07 dB at 50 MHz.
`This is still less than most of the ferrites listed in Table 11-1 operating
`at the 200:50-Q level. And finally, the Carbonyl E core exhibited the
`highest frequency response because its core had a smaller cross sec(cid:173)
`tional area, allowing a shorter transmission line with 6 turns. Con(cid:173)
`versely, the high-frequency response of the Q2 core was the poorest
`because of its much longer transmission line (9 turns).
`
`(,)
`
`ffi 0.7
`i5 0.8
`~ 0.9
`ct a: 1.0
`.... 1.1
`1.2
`1.3
`1.4L-~~~~~L..-...&.L..~~~~~~_.__~~~~~~.L--J
`0.1
`1.0
`10
`100
`FREQUENCY ( MHz)
`
`Fig 11-5-Loss v frequency of four 4:1 transformers at the 200:50-Q
`level.
`
`11-8 Chapter 11
`
`Page 16 of 23
`
`
`
`An interesting observation can be made on an earlier experiment
`which was described in Chapter 3. Fig 3-7 shows the performance of
`two transformers with optimized windings operating at the 40: 10-Q and
`200:50-Q level, each using 4C4 material. Although the higher-imped(cid:173)
`ance transformer exhibited the better high-frequency response, which
`was the point of the experiment, the transformer also maintained about
`the same low-loss characteristic of the lower-impedance transformer. In
`other words, 4C4 material appears to be less sensitive to impedance
`levels. This difference between Ql and 4C4 material may be caused by
`differences in processing. The 4C4 material is known to exhibit the
`normal B-H loop, while the Ql material exhibits the perminvar loop,
`which has steeper sides and a smaller enclosed area.
`Chapter 3 also shows data on G material (Fig 3-6). Losses of
`about 0.1 dB were noted at impedance levels of 75:18.75 n. This is
`about three times the loss of the Q 1 and 4C4 materials. The permeability
`of the G material is also about three times larger than the other two.
`Four ferrite toroids from Ceramic Magnetics were also investi(cid:173)
`gated. The results are shown in Fig 11 -6. These transformers used 6
`
`0
`
`0.2
`
`0.4
`
`Ill
`-0
`
`en 0.6
`en
`0
`....J
`a:: 0.8 -
`w
`u
`::::::>
`0
`en
`z
`<(
`a::
`I-
`
`1.0
`
`1.2
`
`4: 1 TRANSFORMERS I 50: 12.5 .0.
`N =6 TURNS STRIPLINE,
`w = 7/64 11 ,t=0.0028 11
`CERAMIC MAGNETICS
`CORES,OD =1.5"
`• c 2050-66-165-5,µ. = 100
`0 c 2025-18-172-4,µ. =175
`A CN 20-127-16-6, µ.=800
`o CMD 5005-R-107-8,µ.=1400
`
`0.1
`
`10
`1.0
`FREQUENCY (MHZ)
`
`100
`
`Fig 11-6-Loss v frequency for four core materials from Ceramic
`Magnetics with optimized windings for the 50:12.5-n level.
`
`Materials and Power Ratings 11-9
`
`Page 17 of 23
`
`
`
`CD
`~
`Cf)
`
`Cf) g
`a:
`w
`u
`::::>
`0
`Cf)· z
`<! a:
`I-
`
`MAGNETIC CERAMICS-5005
`N=G, L= 10.511
`N0.18 WIRE INSULATED
`
`•-100,25S1
`0-150, 37.5S1
`t;- 200'50S1
`o- 250,62.5S1
`
`1.0
`
`10
`
`100
`
`FREQUENCY (MHZ)
`
`Fig 11-7-Loss v frequency for a CMD5005 toroid from Ceramic
`Magnetics with optimized winding for the 200:50-n level.
`
`bifilar turns of 7/64-inch stripline, with Scotch no. 92 insulation, which
`is optimum for operation at the 50: 12.5-Q level. The cores had an OD
`of 1 V2 inches. Although their permeabilities ranged from 100 to 1400,
`the losses were about the same for the four different cores-less than
`0.08 dB over their passbands. This is particularly noteworthy for the
`CMD5005 material, which has a permeability of 1400. It is the lowest
`loss measured by the author on any of the higher-permeability materials.
`Coincidentally, its bulk resistivity of 7 x 109 0-cm is the highest of any
`of the ferrites. Fig 11-7 shows the results of the CMD5005 material at
`higher impedance levels. This transformer used 6 bifilar turns of
`Teflon-covered no. 18 wire, optimized for the 200:50-Q level. Note that
`the CMD5005 material shows an increased loss similar to that which
`most ferrites exhibit at higher impedance levels.
`These measurements on Ceramic Magnetics materials also bring
`out another important fact about transmission line transformers . Con(cid:173)
`ventional wisdom has stated that increased coupling between transmis(cid:173)
`sion lines is necessary for better low-frequency response. The results in
`Fig 11-6 show that the differences in low-frequency responses between
`these four materials are due to the differences in their permeabilities.
`
`11-10 Chapter 11
`
`Page 18 of 23
`
`
`
`0 ..----------------------~
`0.1
`0.2
`0.3
`
`'° o.4
`
`A~£:'0.
`IY"°~.~~·~·fYfY~
`·r. .. ~."°'
`... 'i
`~ ,..£!.·l':s·
`I i'
`,{:!
`"·
`\
`~ \
`"O
`Cf) 0.5
`~ /;
`~-b
`l
`I
`:'
`~ 0.6
`': \
`:p
`),
`_J
`a:: 0.7
`'(
`:. I
`I
`/
`': I
`~ 0.8
`f'
`" \
`:::::>
`I
`9 /
`~ 0.9
`~
`~ 1.0
`I
`:
`~ 1.1
`f
`I
`1.2
`I
`,:
`I
`1.3
`1.4..___-u-...._ _ _ _ _ _,_ _ _ _ _ _ _ _ ..__ _ _ _ _ _ ----1--J
`0.1
`1.0
`10
`100
`FREQUENCY (MHz)
`
`4:1 (200:50.0.)
`TRANSMISSION LINE
`TRANSFORMERS
`NO.
`MATERIAL TURNS
`5
`3C8
`5
`KR6
`5
`H
`
`•
`
`0
`6
`
`. .
`
`Fig 11-8-Measurements of 3C8, KR6 and H materials at the 200:50-n
`level.
`
`The transmission lines, and hence couplings, are the same. What is
`taking place here is that the reactance of the windings (and hence
`choking action) increases with permeability, thus resulting in better
`low-frequency response.
`Finally, Fig 11-8 shows measurements taken on 3C8, KR6 and H
`materials at the 200:50-n level. Even at lower impedance levels, these
`materials exhibit considerably more loss than the other materials al(cid:173)
`ready described. Incidentally, the 3C8 material is a manganese-zinc
`ferrite and has a high permeability of 2700. It is used extensively at low
`frequencies in power transformers, TV fly back transformers and deflec(cid:173)
`tion yokes. Ruthroff used this type of ferrite in his original work on
`transmission line transformers, but it is unsuitable for applications at
`power levels, where efficiency becomes a factor (ref 9).
`
`Sec 11.4 Power Ratings
`Power ratings are generally determined by two conditions: ( 1) the
`temperature rise due to losses, and (2) the exceeding of some maximum
`value of operating parameter by accident, which can create a cata(cid:173)
`strophic failure. A failure caused by an increase in temperature is usually
`
`Materials and Power Ratings
`
`11 -11
`
`Page 19 of 23
`
`
`
`time dependent, while the breakdown of a device operated over its
`preset value is instantaneous.
`With conventional transformers, losses can be separated into
`three categories: eddy-current losses, hysteresis losses and residual
`losses. These losses, together with the permeability of ferrite material,
`generally increase with temperature to create a possible runaway con(cid:173)
`dition. Catastrophic failures can occur with some ferrite materials at
`flux densities greater than 500 gauss. This is particularly true of ferrites
`exhibiting the perminvar loop; they become excessively lossy. Each of
`these conditions are related to flux densities and hence ampere-turns.
`Increasing the size of the core can, therefore, increase the power rating.
`Different conditions exist when a transmission line transformer's
`power rating is determined. Because of the canceling effect of the
`transmission line currents, little flux is generated in the core. This holds
`true even when tapped multiwindings are involved. Since losses with
`certain ferrites are only on the order of 20 to 40 millidecibels, very small
`transmission line transformers can handle surprisingly high power
`levels. Further, as was shown in the previous section, losses in most
`ferrite materials increase with impedance levels , and these levels must
`be considered when designing a transmission line transformer. Most
`failures in transmission line transformers are of the catastrophic type,
`and are usually caused by poorly terminated (or unterminated) trans(cid:173)
`formers. Such conditions create high voltages and a breakdown in the
`insulation between the windings. This is particularly true of close(cid:173)
`wound enameled or Formvar-type wires.
`Standards for setting power ratings for transmission line trans(cid:173)
`formers have not appeared in the literature, nor are they available from
`any of the suppliers of ferrite material. To the author's knowledge, the
`data presented in this book is the only quantitative information available
`on the losses of these transformers. Because limited reliability informa(cid:173)
`tion is available on transmission line transformers, and losses generally
`increase with impedance levels, an exact formulation of power ratings
`is difficult to make as of this writing. But as a result of the findings by
`the author, some general guidelines can be offered when considering
`ratings. They are:
`1) The power capability of these devices (when energy is trans(cid:173)
`mitted from input to output by a transmission line mode) is determined
`more by the size of the conductors, and not by the cores. Very small
`structures can handle amazingly high power levels. Thus, larger wires
`
`11-12 Chapter 11
`
`Page 20 of 23
`
`
`
`or the use of coaxial cable or stripline can more than double the power
`ratings.
`2) The voltage ratings can be increased significantly by the
`use of polyimide-coated wires. Some commercial brands are ML
`and H Imideze wires. In many cases, in order to optimize the charac(cid:173)
`teristic impedance of the windings, extra layers of Scotch no. 92 tape
`(another polyimide insulation) are used. This also increases the break(cid:173)
`down voltage.
`3) Generally, the lowest-permeability nickel-zinc ferrites yield
`the highest efficiencies. These have permeabilities in the range of 40 to
`50. But, these ferrites can limit the low-frequency response. When
`operating at impedance levels below 100 n, permeabilities as high as
`300 should yield very high efficiencies (98% to 99%). When operating
`at impedances above 100 n, the trade-off is low-frequency response for
`efficiency. Actually, most of the ferrites with permeabilities of 200 to
`300 can still yield acceptable efficiencies (at least 97%) at the 200- to
`300-Q impedance level.
`4) Very few differences in efficiency were observed from the
`ferrites supplied by the manufacturers listed in Sec 11.5. Limited mea(cid:173)
`surements on 4C4 material (µ = 125) from Ferroxcube showed the bes