Lecture 13 Date:

Transcription

1 ecture 13 Date: Quarter-wave Impedance Transformer The Theory of Small Reflections

2 The Quarter Wave Transformer (contd.) The quarter-wave transformer is simply a transmission line with characteristic impedance Z 1 and length l = λ/4 (i.e., a quarter-wave line). R l = λ/4 This λ/4 line is the matching network! the quarter wavelength case is one of the special cases that we studied. We know that the input impedance of the quarter wavelength line is: Solving for Z 1, we find its required value to be: Z1 Z0R Z in Z Z Z R

3 Multiple Reflection Viewpoint The quarter-wave transformer brings up an interesting question in μ-wave engineering. z = l z = 0 R Q: Why is there no reflection at z = l? It appears that the line is mismatched at both z = 0 and z = l. A: In fact there are reflections at these mismatched interfaces an infinite number of them! We can use signal flow graph to determine the propagation series, once we determine all the propagation paths through the quarterwave transformer. l = λ 4

4 Multiple Reflection Viewpoint (contd.) T T R a b T Γ -Γ T j j l = λ 4 z = l z = 0 Γ b a p n1 n ow, let s try to interpret what physically happens when the incident voltage wave reaches the interface at z = l. We find that there are two forward paths through the quarter-wave transformer signal flow graph.

5 Multiple Reflection Viewpoint (contd.) Path 1. At z = l, the characteristic impedance of the transmission line changes from Z 0 to Z 1. This mismatch creates a reflected wave, with complex amplitude p 1 a : z = l z = 0 R p1 Γ a b T j Γ -Γ Γ T j

6 Multiple Reflection Viewpoint (contd.) Path 2. However, a portion of the incident wave is transmitted (Τ) across the interface at z = l, this wave travels a distance of βl = 90 to the load at z = 0, where a portion of it is reflected (Γ ). This wave travels back βl = 90 to the interface at z = l, where a portion is again transmitted (Τ) across into the Z 0 transmission line another reflected wave! R So the second direct path is: p Te e T T 2 j90 j90 2 note that traveling 2βl = 180 has produced a minus sign in the result.

7 Multiple Reflection Viewpoint (contd.) a T j b p a 2 b Γ -Γ Γ T j Path 3. However, a portion of this second wave is also reflected (Γ) back into the Z 1 transmission line at z = l, where it again travels to βl = 90 the load, is partially reflected (Γ ), travels βl = 90 back to z = l, and is partially transmitted into Z 0 (Τ) our third reflected wave! R

8 Multiple Reflection Viewpoint (contd.) SFG a T Γ -Γ j Γ ote that path 3 is not a direct path! b T j 3 2 p Te e e e T T j90 j90 j90 j90 2 Path n. We can see that this bouncing back and forth can go on forever, with each trip launching a new reflected wave into the Z 0 transmission line. ote however, that the power associated with each successive reflected wave is smaller than the previous, and so eventually, the power associated with the reflected waves will diminish to insignificance!

9 Multiple Reflection Viewpoint (contd.) Q: But, why then is Γ = 0? A: Each reflected wave is a coherent wave. That is, they all oscillate at same frequency ω; the reflected waves differ only in terms of their magnitude and phase. Therefore, to determine the total reflected wave, we must perform a coherent summation of each reflected wave this summation results in our propagation series, a series that must converge for passive devices. It can be shown that the infinite propagation series for this quarter-wavelength structure converges to the closed-form expression: b a b a p n1 T 2 2 pn 2 n1 1 n Thus, the input reflection coefficient is: Using our definitions, it can be shown that the numerator of this expression is: b in a T 2 2 T Z Z R Z 2 2 Z1 Z0R 1 0 1

10 Multiple Reflection Viewpoint (contd.) It is evident that the numerator (and therefore Γ in ) will be zero if: Z Z R Z Just as we 1 Z0R expected! Physically, this result ensures that all the reflected waves add coherently together to produce a zero value! ote all of our transmission line analysis has been steady-state analysis. We assume our signals are sinusoidal, of the form exp(jωt). This signal exists for all time t the signal is assumed to have been on forever, and assumed to continue on forever. In other words, in steady-state analysis, all the multiple reflections have long since occurred, and thus have reached a steady state the reflected wave is zero!

11 The Theory of Small Reflections Recall that we analysed a quarter-wave transformer using the multiple reflection view point. T T R l = λ 4 We found that the solution could be written as an infinite summation of terms (the propagation series): b a p n where each term had a specific physical n1 interpretation, in terms of reflections, transmissions, and propagations.

12 The Theory of Small Reflections (contd.) For example, the third term was path: 3 2 p T e 2 j2l R SFG a 1 b 1 T b 2 e jβl Γ -Γ Γ T e jβl a 2 a 5 b 5 ow let s consider the magnitude of this path: j2l p T e p 2 2 T 3 Recall that Γ = Γ for a properly designed quarter-wave transformer: R R Z Z 1 1 p T For the case where values R and Z 1 are numerically close, R Z R Z 1 1, the magnitude of the reflection coefficient will be very small: R R Z Z

13 The Theory of Small Reflections (contd.) As a result, the value Γ 3 will be very, very, very small. Moreover, we know (since the connector is lossless) that: We can thus conclude that the magnitude of path p 3 is likewise very, very, very small: T T 1 p T 1 This is a classic case where we can approximate the propagation series using only the forward paths!! Recall there are two forward paths: a T e jβl R Γ -Γ Γ p 1 = Γ T e jβl p 2 = T 2 Γe j2βl

14 The Theory of Small Reflections (contd.) Therefore if Z 0 and R are very close in value, the approximate reflected wave using only the direct paths of the infinite series can be found from the SFG: ow, if we likewise apply the approximation that T 1.0, we conclude for this quarter wave transformer (at the design frequency): 2 j2l 1 2 b p p a T e a j2l 1 2 b p p a e a This approximation, where we: 1. use only the direct paths to calculate the propagation series, 2. approximate the transmission coefficients as one (i.e., T = 1.0). is known as the Theory of Small Reflections, and allows us to use the propagation series as an analysis tool (we don t have to consider an infinite number of terms!).

15 The Theory of Small Reflections (contd.) Consider again the quarter-wave matching network SFG. ote there is one branch ( Γ = S 22 of the connector), that is not included in either direct path. a T e jβl Γ -Γ Γ With respect to the theory of small reflections (where only direct paths are considered), this branch can be removed from the SFG without affect. p 1 = Γ T p 2 = T 2 Γe j2βl e jβl a Γ T e jβl Γ a 1.0 e jβl p 1 = Γ T p 2 = T 2 Γe j2βl e jβl Γ p 1 = Γ 1.0 p 2 = T 2 Γe j2βl e jβl Γ Moreover, the theory of small reflections implements the approximation, T = 1.0, so that the SFG becomes:

16 The Theory of Small Reflections (contd.) Reducing this SFG by combining the 1.0 branch and the e jβl branch via the series rule, we get the following approximate SFG: The approximate SFG when applying the theory of small reflections! Γ e jβl e jβl Γ b in a e j2l ote this approximate SFG provides precisely the results of the theory of small reflections! Q: But wait! The quarter-wave transformer is a matching network, therefore Γ in = 0. The theory of small reflections, however, provides the approximate result: j2 e l in Is this approximation very accurate? How close is this approximate value to the correct answer of Γ in = 0?

17 The Theory of Small Reflections (contd.) A: et s find out! Recall that Γ=Γ for a properly designed quarter-wave matching network, and so: e 1 e j2l j2l in ikewise, l = λ/4 (but only at the design frequency!) so that: 2 2l 2 4 Thus: j 2 l e e j 1 (1 1) 0 in where you of course recall that β = 2π λ! Q: Wow! The theory of small reflections appears to be a perfect approximation no error at all!?! A: ot so fast.

18 The Theory of Small Reflections (contd.) The theory of small reflections most definitely provides an approximate solution (e.g., it ignores most of the terms of the propagation series, and it approximates connector transmission as Τ = 1, when in fact Τ 1). As a result, the solutions derived using the theory of small reflections will generally speaking exhibit some (hopefully small) error. We just got a bit lucky for the quarter-wave matching network; the approximate result Γ in = 0 was exact for this one case! The theory of small reflections is an approximate analysis tool!

19 Example 1 Indraprastha Institute of Use the theory of small reflections to determine a numeric value for the input reflection coefficient Γ in, at the design frequency ω 0. l 1 = 3λ 0 8 l 2 = λ 0 8 Z 0 Γ in Z 1 Z 2 Z Γ 0 = 0.1 Γ 0 = 0.05 Γ 0 = 0.15 ote that the transmission line sections have different lengths!

20 Frequency Response of a λ/4 Matching etwork Q: You have once again provided us with confusing and perhaps useless information. The quarter-wave matching network has an exact SFG of: a b T Γ -Γ T e jβl e jβl You could have left this simple and precise analysis alone BUT OOO!! You had to foist upon us a long, rambling discussion of the propagation series and direct paths and the theory of small reflections, culminating with the approximate (i.e., less accurate!) SFG: Γ Using our reduction rules, we can quickly conclude that: in b T e a 1 Γ e jβl e jβl 2 j2l Γ

21 Freq. Response of a λ/4 Matching etwork (contd.) From the approximate SFG we were able to conclude the approximate (i.e., less accurate!) result: in b a e j2l The exact result was simple and exact! Why did you make us determine this approximate result? A: In a word: frequency response*. * OK, two words. the mathematical form of the result is much simpler to analyze and/or evaluate (e.g., no fractional terms!). Q: What exactly would we be analysing and/or evaluating? A: The frequency response of the matching network, for one thing. Remember, all matching networks must be lossless, and so must be made of reactive elements (e.g., lossless transmission lines). The impedance of every reactive element is a function of frequency, and so too then is Γ in.

22 Freq. Response of a λ/4 Matching etwork (contd.) Say we wish to determine function Γ in (ω). Q: Isn t Γ in ω = 0 for a quarter wave matching network? A: Oh my gosh no! A properly designed matching network will typically result in a perfect match (i.e., Γ in ω = 0) at one frequency (i.e., the design frequency). However, if the signal frequency is different from this design frequency, then no match will occur (i.e., Γ in (ω) 0). Recall we discussed this behavior before:

23 Freq. Response of a λ/4 Matching etwork (contd.) 2 j2l T e Q: But why is the result: in or its approx form: 1 dependent on frequency? I don t see frequency variable ω anywhere in these results! A: ook closer! Remember that the value of spatial frequency β (in radians/meter) is dependent on the frequency ω of our eigen function (aka the signal ): e j2l in where you will recall that v p is the propagation velocity of a wave moving along a transmission line. This velocity is a constant (i.e., v p = 1 ), and so the spatial frequency β is C directly proportional to the temporal frequency ω. Thus, we can rewrite: l l v p T 1 v p Where T = l v p is the time required for the wave to propagate a distance l down a transmission line.

24 Freq. Response of a λ/4 Matching etwork (contd.) As a result, we can write the input reflection coefficient as a function of spatial frequency β: ( ) e j2l in Or equivalently as a function of temporal frequency ω: j2 ( ) e T in Frequently, the reflection coefficient is simply written in terms of the electrical length θ of the transmission line, which is simply the difference in relative phase between the wave at the beginning and end of the length l of the T. l T So that: j2 ( ) in e ote we can simply insert the value θ = βl into this expression to get Γ in (β), or insert θ = ωt into the expression to get Γ in (ω). ow, we know that Γ = Γ for a properly designed quarter-wave matching network, so the reflection coefficient function can be written as: j2 in( ) 1 e

25 Freq. Response of a λ/4 Matching etwork (contd.) ote that: j0 j( ) j j 1 e e e e 2 ( ) And that: j2 And so: in( ) 1 e e j e j e j e j e e e e j j j j e j e j e j e j 2cos ow, magnitude of our result is: ( ) e j 2 cos 2 cos ote: Γ in (θ) is zero-valued only when cosθ = 0. This of course occurs when θ = A: Remember, θ = βl. Thus if θ = in ( ) 2 cos 0 2 π in 2. /2 Q: What the heck does this mean? π 2: /2 l 2 / 4 As we (should have) suspected, the match occurs at the frequency whose wavelength is equal to four times the matching (Z 1 ) transmission line length, i.e. λ = 4l.

26 Freq. Response of a λ/4 Matching etwork (contd.) In other words, a perfect match occurs at the frequency where l = λ 4. ote the physical length l of the transmission line does not change with frequency, but the signal wavelength does: Q: So, at precisely what frequency does a quarter-wave transformer with length l provide a perfect match? A: Recall that θ = ωt, where T = This frequency is called the design frequency of the matching network it s the frequency where a perfect match occurs. We denote this as frequency ω 0, which has wavelength λ 0, i.e.: v p T 2 l v p. Thus, for θ = π 2: 1 2 T T 2l 2T 2l v p v p l v p f v p 2T 2l

27 Freq. Response of a λ/4 Matching etwork (contd.) Given this, yet another way of expressing θ = βl is: vp f l vp f Thus, we conclude: f in(f) 2 cos 2 f0 This expression helps in the determination (approximately) of the bandwidth of the quarter-wave transformer! First, we must define what we mean by bandwidth. Say the maximum acceptable level of the reflection coefficient is value Γ m. This is an arbitrary value, set by you the microwave engineer (typical values of Γ m range from 0.05 to 0.2). et us denote the frequencies where f m this maximum value Γ in(f f m) m 2 cos m occurs f m. In 2 f0 other words:

28 Freq. Response of a λ/4 Matching etwork (contd.) There are two solutions to this equation, the first is: And the second: m2 2 f0 cos 1 You will find that f m1 < f 0 < f m2. So the values f m1 and f m2 define the lower and upper limits on matching network bandwidth. f 2 m f 2 f0 1 m m1 cos 2 Important note! Make sure cos 1 x is expressed in radians! All this analysis was brought to you by the simple mathematical form of Γ in (f) that resulted from the theory of small reflections! f m1 f m2

29 The Multi-section Transformer Consider a sequence of transmission line sections; each section has equal length l, but dissimilar characteristic impedances: Γ 0 Γ 1 Γ 2 Γ 1 Γ Z Where the marginal 0 reflection coefficients are: If the load resistance R is less than Z 0, then we should design the transformer such that: Conversely, if R is greater than Z 0, then we will design the transformer such that: l l l Z Z Z Z n Z Z n1 n1 Z Z n n Z Z Z Z Z 0 > Z 1 > Z 2 > Z 3 > > Z > R Z 0 < Z 1 < Z 2 < Z 3 < < Z < R

30 The Multi-section Transformer (contd.) In other words, we gradually transition from Z 0 to R! ote that since R is real, and since we assume lossless transmission lines, all Γ n will be real (this is important!). ikewise, since we gradually transition from one section to another, each value: Z Z will be small. n1 n As a result, each marginal reflection coefficient Γ n will be real and have a small magnitude. This is also important, as it means that we can apply the theory of small reflections to analyse this multi-section transformer! The theory of small reflections allows us to approximate the input reflection coefficient of the transformer as: Γ 0 Γ 1 Γ 2 Γ 1 Γ l l l R

31 The Multi-section Transformer (contd.) The approximate SFG when applying the theory of small reflections! a 0 b 0 a 1 e jβl e jβl e jβl Γ 0 Γ 1 Γ 2 Γ 1 Γ = Γ e jβl e jβl e jβl b 1 a 2 b 2 a 1 b 1 a b b a 0 0 ( ) in j 2 l j 4 l 2 0 1e 2 e... j l e Γe n n0 j2nl We can alternatively express the input reflection coefficient as a function of frequency (βl = ωt): ( ) e e... e in j2t j4t j2t Γe n n0 j(2 nt ) where: T = l v p propagation time through 1 section

32 The Multi-section Transformer (contd.) We see that the function Γ in (ω) is expressed as a weighted set of basis functions! i.e., c n = Γ n Γ in (ω) = n=0 We find, therefore, that by selecting the proper values of basis weights c n (i.e., the proper values of reflection coefficients Γ n ), we can synthesize any function Γ in (ω) of frequency ω, provided that: 1. Γ in (ω) is periodic in ω = 1 2T. 2. we have sufficient number of sections. c n Ψ(ω) Ψ(ω) = e j(2nt)ω Q: What function should we synthesize? A: Ideally, we would want to make Γ in ω = 0 (i.e., the reflection coefficient is zero for all frequencies). Bad ews: this ideal function Γ in ω = 0 would require an infinite number of sections (i.e., = )!

33 The Multi-section Transformer (contd.) Therefore, we seek to find an optimal function for Γ in ω, given a finite number of elements. Once we determine these optimal functions, we can find the values of coefficients Γ n (or equivalently, Z n ) that will result in a matching transformer that exhibits this optimal frequency response. To simplify this process, we can make the transformer symmetrical, such that: Γ 0 = Γ, Γ 1 = Γ 1, Γ 2 = Γ 2,. ote: this does OT mean that: Z 0 = Z, Z 1 = Z 1, Z 2 = Z 2,. We then find that: Γ ω = e jωt Γ 0 e jωt + e jωt + Γ 1 e j( 2)ωT + e j( 2)ωT + Γ 2 e j( 4)ωT + e j( 4)ωT +

34 The Multi-section Transformer (contd.) and since: e jx + e jx = 2cos(x) we can write for even: Γ ω = 2e jωt Γ 0 cosωt + Γ 1 cos( 2)ωT + + Γ n cos( whereas for odd: Γ ω = 2e jωt Γ 0 cosωt + Γ 1 cos( 2)ωT + + Γ n cos( The remaining question then is this: given an optimal and realizable function Γ in ω, how do we determine the necessary number of sections, and how do we determine the values of all reflection coefficients Γ n?? Multi-section transformer is often used to maximize the bandwidth of transformer.

35 The Multi-section Transformer (contd.) Alternatively, we can say that one way to maximize bandwidth is to construct a multi-section matching network with a function Γ(f) that is either maximally flat or can be considered flat albeit with pass-band ripple. Binomial Function satisfies the condition of maximum flatness Chebyshev Polynomial can be considered flat with pass-band ripple

36 The Binomial Multi-Section Transformer Recall that a multi-section matching network can be described using the theory of small reflections as: ( ) e e... e in j2t j4t j2t Γe n n0 j(2 nt ) where: T = l v p propagation time through 1 section ote that for a multi-section transformer, we have degrees of design freedom, corresponding to the characteristic impedance values Z n. Q: What should the values of Γ n (i.e., Z n ) be? A: We need to define independent design equations, which we can then use to solve for the values of characteristic impedance Z n. First, we start with a single design frequency ω 0, where we wish to achieve a perfect match: in 0 0 That s just one design equation: we need -1 more! These addition equations can be selected using many criteria one such is to make the function Γ in (ω) maximally flat at the point ω = ω 0.

37 The Binomial Multi-Section Transformer (contd.) To accomplish this, we first consider the Binomial Function: This function has the desirable j2 A1 e j properties that: A e A and that: d n Γ (θ) dθ n = 0 for n = 1,2,3,, 1 θ= π 2 In other words, this Binomial Function is j2 A1 e maximally flat at the point θ = π 2, where it has a value of Γ θ = π 2 = 0. Q: So? What does this have to do with our multi-section matching network? A: et s expand (multiply out the identical product terms) the Function: j2 A1 e j2 j4 j6 j 2 A C 0 C1 e C2 e C3 e... C e where: C n! n! n!

38 The Binomial Multi-Section Transformer (contd.) it is obvious the two functions have identical forms, provided that: j2 A1 e n AC n T Moreover, we find that this function is very desirable from the standpoint of the a matching network. Recall that Γ θ = 0 at θ = π 2 a perfect match! Additionally, function is maximally flat at θ = π 2, therefore Γ θ 0 over a wide range around θ = π 2 a wide bandwidth! Q: But how does θ = v p 1 0 T 2 l 2 π 2 relate to frequency ω? A: Remember that ωt = θ, so θ = π 2 corresponds to the frequency: This frequency (ω 0 ) is therefore our design frequency the frequency where we have a perfect match. ote that the length l has an interesting relationship with this frequency: l v p

39 The Binomial Multi-Section Transformer (contd.) Binomial Multi-section matching network will have a perfect match at the frequency where the section lengths l are a quarter wavelength! Thus, we have our first design rule: Set section lengths l so that they are a quarter-wavelength the design frequency ω 0. λ 0 4 Q: I see! And then we select all the values Z n such that Γ n = AC n. But wait! What is the value of A?? A: We can determine this value by evaluating a boundary condition! at Specifically, we can easily determine the value of Γ(ω) at ω = 0. R l l l As ω approaches zero, the electrical length βl of each section will likewise approach zero. Thus, the input impedance Z in will simply be equal to R as ω 0.

40 The Binomial Multi-Section Transformer (contd.) As a result, the input reflection coefficient Γ(ω = 0) must be: However, we likewise know that: j2(0) 0 A 1 e A 11 A2 Zin 0 Z R Z 0 Z 0 Z R Z in Equating the two expressions: A R R 2 Z Z 0 0 therefore: A R R 2 Z Z 0 0 (A can be negative!) We now have a formulation to calculate the required marginal reflection coefficients Γ n : A! R Z0! n ACn 2 ( n)! n! R Z ( n)! n! 0 we also know that these marginal reflection coefficients are physically related to the characteristic impedances of each section as: n Z Z n1 n1 Z Z n n

41 The Binomial Multi-Section Transformer (contd.) Equating the two and solving, we find that that the section characteristic impedances must satisfy: 1 1 AC Z Z Z n n n1 n n 1 n 1 ACn ote this is an iterative procedure we determine Z 1 from Z 0, Z 2 from Z 1, and so forth. Q: This result appears to be our second design equation. A: Alas, there is a big problem with this result. ote that there are +1 coefficients Γ n (i.e., n {0,1,,}) in the Binomial series, yet there are only design degrees of freedom (i.e., there are only transmission line sections!). Thus, our design is a bit over constrained, a result that manifests itself the finally marginal reflection coefficient Γ. ote from this iterative solution, the last transmission line impedance Z is selected to satisfy the mathematical requirement of the penultimate reflection coefficient Γ -1. Z Z 1 AC 1 1 Z Z 1

42 The Binomial Multi-Section Transformer (contd.) 1 AC 1 Therefore the last impedance must be: Z Z 1 1 AC 1 But there is one more mathematical requirement! AC The last marginal reflection coefficient must likewise satisfy: where we use the fact that C = 1. 2 R R But, we selected Z to satisfy the requirement for Γ -1, we have no physical design parameter to satisfy this last mathematical requirement for Γ! As a result, we find to our great R Z AC consternation that the last requirement R Z is not satisfied: Q: Yikes! Does this mean that the resulting matching network will not have the desired Binomial frequency response? A: That s exactly what it means! Z Z 0 0

43 The Binomial Multi-Section Transformer (contd.) Q: You big Why did you waste all my time discussing an overconstrained design problem that can t be built? A: Relax; there is a solution to our dilemma albeit an approximate one. You undoubtedly have previously used the approximation: y x 1 y ln y x 2 x This approximation is especially accurate when y x is small (i.e., when y x 1).

44 The Binomial Multi-Section Transformer (contd.) ow, we know that the values of Z n+1 and Z n in a multi-section matching network are typically very close, such that Z n+1 Z n is small. Thus, we use the approximation: ikewise, we can also apply this approximation (although not as accurately) to the value of A: let s start over, this time we ll use these approximations. ( 1) A 2 ln First, determine A: Z Z 1 Z ln n1 n n1 n Zn 1 Zn 2 Zn R Z0 ow use this result to calculate the mathematically required marginal reflection coefficients Γ n : A R Z R 0 ( 1) 2 2 ln R Z0 Z0 (A can be negative!) n AC n A! ( n)! n!

45 The Binomial Multi-Section Transformer (contd.) Of course, we also know that these marginal reflection coefficients are physically related to the characteristic impedances of each section as: Equating the two and solving, we find that that the 1 Z ln n1 n 2 Zn Z section characteristic impedances must satisfy: n 1 Zn exp2 n This is our second design rule. ote it is an iterative rule we determine Z 1 from Z 0, Z 2 from Z 1, and so forth. Q: Huh? How is this any better? How does applying approximate math lead to a better design result?? A: Applying these approximations help resolve our over constrained problem. Recall that the over-constraint resulted in: R Z R Z AC

46 The Binomial Multi-Section Transformer (contd.) But, as it turns out, the approximations leads to the happy situation where: 1 R ln 2 Z AC provided that the value A is the approximation as well. Effectively, these approximations couple the results, such that each value of characteristic impedance Z n approximately satisfies both Γ n and Γ n+1. Summarizing: a. If you use the exact design equations to determine the characteristic impedances Z n, the last value Γ n will exhibit a significant numeric error, and your design will not appear to be maximally flat. b. If you instead use the approximate design equations to determine the characteristic impedances Z n, all values Γ n will exhibit a slight error, but the resulting design will appear to be maximally flat, Binomial reflection coefficient function Γ(ω)!

47 The Binomial Multi-Section Transformer (contd.) ote that as we increase the number of sections, the matching bandwidth increases. Q: Can we determine the value of this bandwidth? A: Sure! But we first must define what we mean by bandwidth.

48 The Binomial Multi-Section Transformer (contd.) As we move from the design (perfect match) frequency f 0 the value Γ(f) will increase. At some frequency (say, f m ) the magnitude of the reflection coefficient will increase to some unacceptably high value (say, Γ m ). At that point, we no longer consider the device to be matched. Γ(f) Γ m f m1 f f 0 ote there are two values of frequency f m one value less than design frequency f 0, and one value greater than design frequency f 0. These two values define the bandwidth f of the matching network: f f f 2 f f 2 f f m2 m1 0 m1 m2 0 f m2 f

49 The Binomial Multi-Section Transformer (contd.) Q: So what is the numerical value of Γ m? A: I don t know it s up to you to decide! Every engineer must determine what they consider to be an acceptable match (i.e., decide Γ m ). This decision depends on the application involved, and the specifications of the overall microwave system being designed. However, we typically set Γ m to be 0.2 or less. Q: OK, after we have selected Γ m, can we determine the two frequencies f m? A: Sure! We just have to do a little algebra. We start by rewriting the Binomial function: j2 j j j A1 e Ae e e ow, we take the magnitude of this function: j Ae 2cos j 2 Ae cos 2 A cos

50 The Binomial Multi-Section Transformer (contd.) ow, we define the values θ where Γ(θ) = Γ m as θ m. i.e., : 2 A cos m m m We can now solve for θ m (in radians!) in terms of Γ m : m1 1 2 A 1 m cos 1/ m2 1 2 A 1 m cos 1/ ote that there are two solutions (one less than π 2 and one greater than π 2 )! ow, we can convert the values of θ m into specific frequencies. Recall that ωt =θ, therefore: v T l 1 p m m m

51 The Binomial Multi-Section Transformer (contd.) But recall also that l = λ 0 4, where λ 0 is the wavelength at the design v frequency f 0 (not f m!), and where λ 0 = p f0. Thus we can conclude: vp 4vp 2 f0 where θ m is m m m m m 4 f0 m fm expressed in l 0 2 radians. Therefore: f m1 2f 1 2 A 0 1 m cos 1/ f m2 2f 1 2 A 0 1 m cos 1/ Thus, the bandwidth of the binomial matching network can be determined as: 4f 1 2 A 0 1 m f 2 f0 fm 1 2 f0 cos 1/ ote that this equation can be used to determine the bandwidth of a binomial matching network, given Γ m and number of sections.

52 The Binomial Multi-Section Transformer (contd.) 4f 1 2 A 0 1 m f 2 f0 fm 1 2 f0 cos 1/ It can also be used to determine the number of sections required to meet a specific bandwidth requirement! Finally, we can list the design steps for a binomial matching network: 1. Determine the value required to meet the bandwidth ( f and Γ m ) requirements. 2. Determine the approximate value A from Z 0, R and. 3. Determine the marginal reflection coefficients Γ n = AC n required by the binomial function. 4. Determine the characteristic impedance of each section using the iterative approximation: Z n+1 = Z n exp 2Γ n. 5. Perform the sanity check: Γ 1 ln R = AC 2 Z n. n 6. Determine section length l = λ 0 4 for design frequency f 0.

53 Chebyshev Multi-section Matching Transformer Self Study

54 Tapered ines We can also build matching networks where the characteristic impedance of a transmission line changes continuously with position z. We call these matching networks tapered lines. ote all our multi-section transformer designs have involved a monotonic change in characteristic impedance, from Z 0 to R (e.g., Z 0 < Z 1 < Z 2 < < R ). ow, instead of having a stepped change in characteristic impedance as a function of position z (i.e., a multi-section transformer), we can also design matching networks with continuous tapers. R Z 3 Z 2 Z 1 Z 0 Z 3l Continuous Impedance Taper 2l l Multi section Impedance Taper 0 z

55 Tapered ines (contd.) A tapered impedance matching network is defined by two characteristics its length and its taper function Z 1 (z). R z = 0 z = There are of course an infinite number of possible functions Z 1 (z). Your book discusses three: the exponential taper, the triangular taper, and the Klopfenstein taper.

56 Tapered ines (contd.) For example, the exponential taper has the form: Z z Z e 1 0 az 0 < z < where: 1 Z ln a Z0 ote for the exponential taper, we get the expected result that Z 1 z = 0 = Z 0 and Z 1 z = = R. Recall the bandwidth of a multi-section matching transformer increases with the number of sections. Similarly, the bandwidth of a tapered line will typically increase as the length is increased.

57 Tapered ines (contd.) Impedance variations for the triangular, exponential, and Klopfenstein tapers. Resulting reflection coefficient magnitude versus frequency for the tapers

58 Tapered ines (contd.) Q: But how can we physically taper the characteristic impedance of a transmission line? A: Most tapered lines are implemented in stripline or microstrip. As a result, we can modify the characteristic impedance of the transmission line by simply tapering the width W of the conductor (i.e., W(z)). In other words, we can continuously increase or decrease the width of the microstrip or stripline to create the desired impedance taper Z 1 (z).