Chapter 4. Impedance Matching Networks 4.1 L Matching Networks 4.2 The π Matching Network 4.3 The T Matching Network 4.4 Wideband Matching Networks 4.5 Matching Network Using Smith Chart 56 Reasons for Impedance Matching 1. Maximum power transfer To ensure minimum losses due to impedance mismatch. 2. Impedance transformation Usually for Rx LNA, we need a source resistance to obtain the best NF. (This source resistance is different from that to give maximum power transfer). 3. As an aid in filtering off unwanted signals Due to the frequency response of the matching networks. 4.1 L Matching Networks • A 2 elements L network may be configured as LP or HP depending on design requirements: If a DC path is required between source and load, the LP version must be used. If a DC block is required, HP version must be used. Fig.4-1: The L network • The shunt arm or reactance must be connected to the higher resistance of Rs or RL. The series element and the shunt element must be of opposite polarity. 57 • The Q of the matching network is: Q= RL −1 Rs Once the source and load resistance is fixed, Q is fixed. This means that the bandwidth of the network can’t be controlled by design. (This limitation can be overcome by using 3 elements matching networks). • L matching networks are always wider band compared to 3 elements matching networks like T or π. Note: 1. LP matching networks can be a source of help in harmonic filtering. 2. So far the matching is done only for resistive source and load. For complex source and load, the absorption and resonant methods are used. Derivation of Q formula for L matching network Fig.4-2: L matching network 1000 − j 333 Z1 = = 315∠ − 71.58° = 100 − j 300 Ω 1000 − j 300 58 Fig.4-3: Equivalent circuit Rp Qp = Xp ' X X Qs = s' = s = Q s Rs Rs ' i.e. Qp = Qs = Q Since Rp =Rs (Q2 + 1), we obtain Q= Rp −1 Rs Note: matching using LC is true for one frequency only. 4.2 The π Matching Network • It is two “back to back” L networks that are both configured to match load and source to an “invisible” or “virtual” resistance. The virtual resistance R must be smaller than either Rs or RL as it is “connected” to the series arm of each section i.e. R ≤ Rs, R ≤ RL Fig.4-4: π matching network 59 • In practice, the rule of thumb for loaded Q for the π network is: Q= RH −1 R Where RH = largest terminating impedance of Rs or RL R = virtual resistance The Q of the network is therefore controllable by design. The lowest Q is equal to that of the L network. • Suitable for cases where Rs and RL are large values. 4.3 The T Matching Network • Similar to the π network except that the virtual resistance is larger than either the load or source resistance. Fig.4-5: The T matching network • As a rule of thumb, the loaded Q is: Q= R Rsmall −1 Where R = virtual resistance Rsmall = smallest terminating resistance of Rs or RL The Q of the network is therefore controllable by design. The lowest Q is equal to that of the L network. 60 • Suitable for cases where Rs and RL are small values. 4.4 Wideband Matching Networks • Cascade L sections to reduce the effective Q of the matching networks. Fig.4-6: Cascaded L networks for wideband matching • As in the previous cases, R must be larger than the smallest terminating resistance and smaller than the largest terminating resistance. • Maximum bandwidth (minimum Q) is obtained when: Loaded Q then becomes: Q= R Rsmall −1 = Rl arg e −1 R Where R = virtual resistance Rsmall = smallest terminating resistance Rlarge = largest terminating resistance • If even wider bandwidth were required, more L networks can be cascaded with virtual resistances between each networks. For optimum BW Rl arg e R R1 R = 2 = 3 == Rsmall R1 R2 Rn 61 Where = smallest terminating resistance Rsmall = largest terminating resistance Rlarge R1, R2, …, Rn = virtual resistance of each L network Example 4-1: Design a LP L network to match a source resistance of 100Ω to a load resistance of 1000Ω. Assume that frequency of operation is 100MHz. Solution: Since RL > Rs, shunt arm is at RL. We have Q= RL − 1 = 1000 − 1 = 3 100 Rs Qs = Xs / Rs, Xs = Qs x Rs Therefore Xs = 3 x 100 =300 Ω L = Xs / ω = 300 / (2π x 100 x 106) = 477 nH Qp = RL / Xp , Xp = RL / Qp Xp = 1000 / 3 = 333 Ω C = 1 / ωXp = 1 / [(2π x 100 x 106) x 333] = 4.8 pF 62 Example 4-2: Design a π network to match a source resistance of 100 Ω to a load resistance of 1000 Ω. Assume a loaded Q of 15. Solution: We have Q= RH −1 R Therefore R= RH = 1000 = 4.42 Ω 2 Q + 1 152 + 1 Load Side: X p2 = R p RL 1000 = = = 66.6 Ω 15 Q Q Xs2 = Q x R = 15 x 4.42 = 66.3 Ω Source side: Q= X p1 = Rs − 1 = 100 − 1 = 4.6 R 4.42 R p 100 = = 21.7 Ω Qs 4.6 63 Xs1 =Qs x R = 4.6 x 4.42 = 20.51 Ω Example 4-3: Use the absorption method to match the complex source and load at 100MHz. Solution: 1. Ignore the source and load reactances and design a L network to match Rs and RL. 2. Refer to earlier worked example where we have done the L network matching from the source resistance of 100 Ω to load resistance of 1000 Ω. Series L = 477 nH Shunt C = 4.8 pF j126 Ω @100 MHZ = 200 nH Therefore the elements of this L network are series L = 477 – 200 = 277 nH 64 shunt C = 4.8 –2 = 2.8 pF Example 4-4: Design an impedance matching network that will block the flow of DC from the source to the load. The frequency of operation is 75 MHz. Use the resonant approach. Solution: 1. The need to block DC dictates a series C and shunt L matching network. 2. Use the resonant method to remove the stray capacitance of 40 pF L= 1 1 = = 112.6 nH 6 ω Cstray 2π × 75 × 10 2 × 40 × 10−12 2 3. Proceed to match 50 Ω source to 600 Ω load. Qs = Q p = Rp − 1 = 600 − 1 = 3.32 50 Rs Xs = Qs x Rs = 3.32 x 50 = 166 Ω Xp = Rp / Qp = 600 / 3.32 = 181 Ω 65 Cs = 1 / (ωXs) = 12.78 pF Lp = Xp / ω = 384 nH Leqv = L1L2 /(L1 + L2) = 87 nH 4.5 Matching Network Using Smith Chart • The Smith Chart is the most convenient method for design of all matching networks. The procedure for matching network design is as follows: 1. Plot the normalized load impedance and complex conjugate of the source. 2. Determine if the L network is sufficient or if Q need to be controlled. L network Join the load and conjugate source points together by 2 arcs. 3 elements network a) b) Plot the constant Q circle for the specified Q. Ensure that one out of the 3 network arcs touches the constant Q circle when the graphical manipulation is completed. 66 Fig.4-7: Summary of component addition on a Smith Chart 67 Fig.4-8: VSWR circles and constant Q circles Example 4-5: Given that the source impedance is Zsn = 1.5 + j1 (normalized) and the load impedance ZLn = 0.2 – j0.2 (normalized). Design a 2 elements matching network for maximum power transfer. 68 Solution: Z1 = 0.2 – j0.2 - j0.43 = 0.2 – j0.63 = 0.661∠ -72.38° 69 Z2 = Z1 // j0.88 = (0.661∠-72.38°) x (0.88∠ 90°) / (0.2 – j0.63 + j0.88) = (0.582∠ 17.62°) / (0.32∠ 51.34°) = 1.818∠ -33.72° = 1.5 –j 1 = Zs* There we have maximum power transfer. Example 4-6: Investigate the performance of various matching networks in matching a load of 388 + j245.5Ω at 500MHz to a 50Ω source. Consider the following matching networks: a. 2 elements b. 3 elements c. multi elements wideband 70 Solution: 2 elements matching network 71 3 elements matching network (a) 72 3 elements matching network (b) 73 Multi elements wideband matching network (a) 74 Multi elements wideband matching network (b) 75 Multi elements wideband matching network (c) 76