Application Report SNLA027B – May 2004 – Revised May 2004 AN-807 Reflections: Computations and Waveforms ..................................................................................................................................................... ABSTRACT In this application note, the logical progression from the ideal transmission line to the real world of the long transmission line with its attendant losses and problems is made; specifically, the methods to determine the practicality of a certain length of line at a given data rate is discussed. Transmission line effects on various data formats are examined as well as the effects of several types of sources (drivers) on signal quality. A practical means is given to measure signal quality for a given transmission line using readily available test equipment. This, in turn, leads to a chart that provides the designer a way to predict the feasibility of a proposed data-transmission circuit when twisted-pair cable is used. This application note is a revised reprint of section three of the Fairchild Line Driver and Receiver Handbook. This application note, the second of a three-part series (see AN-806 and AN-808), covers the following topics: 1 2 3 4 5 6 7 8 9 10 11 12 13 Contents Overview ..................................................................................................................... 3 Introduction .................................................................................................................. 3 The Initial Wave ............................................................................................................. 3 Cut Lines and a Matched Load ........................................................................................... 4 Kirchoff's Laws and Line-Load Boundary Conditions .................................................................. 5 Fundamental Principles .................................................................................................... 9 Tabular Method for Reflections— the Lattice Diagram ............................................................... 10 Limitations of the Lattice Diagram Method ............................................................................. 13 Reflection Effects for Voltage Source Drivers ......................................................................... 14 Reflection Effects for Matched-Source Drivers ........................................................................ 16 Reflection Effects for Current-Source Drivers ......................................................................... 17 Summary—Which are the Advantageous Combinations? ........................................................... 21 Effect of Source Rise Time on Waveforms ............................................................................ 23 List of Figures 1 Generator Driving an Infinite Transmission Line ........................................................................ 4 2 Thevenin Equivalent for Initial Wave ..................................................................................... 4 3 Voltage/Current Steps for Three Source Resistances ................................................................. 4 4 Voltages and Current on an Infinite Length Line........................................................................ 5 5 Boundary Conditions at the Line/Load Interface ........................................................................ 5 6 Waveforms for RL = R0 ..................................................................................................... 7 7 ................................................................................................................................ 9 ................................................................................................................................ 9 Superposition of Simple Waveforms to Form More Complex Excitations ......................................... 10 Sign Conventions for Waves ............................................................................................. 11 (a) Line Circuit to be Analyzed........................................................................................... 11 Reflection Bookkeeping with the Lattice Diagram ..................................................................... 11 Model Used for Lattice Diagram Method ............................................................................... 13 RS < R0 ...................................................................................................................... 15 RS = R0 ...................................................................................................................... 15 8 9 10 11 12 13 14 15 All trademarks are the property of their respective owners. SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 1 www.ti.com 16 RS > R0 ...................................................................................................................... 15 17 Source and Load Voltage Waveforms for Various R S and RL ....................................................... 16 18 Approximation of Midline Voltage with RS > R0 and RL > R0 ......................................................... 21 19 Transmission Line Model and Its Lattice Diagram .................................................................... 24 20 Waveforms for tr = 2 ≪ τ ................................................................................................. 24 21 Waveforms for tr = 2τ...................................................................................................... 25 22 Waveforms for tr = 2τ...................................................................................................... 26 23 Waveforms for tr = 2τ...................................................................................................... 26 24 Waveforms for tr = 3τ...................................................................................................... 26 25 Waveforms for tr = 3τ...................................................................................................... 27 26 Waveforms for tr = 4τ...................................................................................................... 27 27 Waveforms for tr = 4τ...................................................................................................... 28 28 Waveforms for tr = 6τ...................................................................................................... 28 29 Waveforms for tr = 6τ...................................................................................................... 28 30 Waveforms for tr = 8τ...................................................................................................... 29 31 Waveforms for tr = 8τ...................................................................................................... 29 List of Tables 2 1 (RS = 2000Ω, R0 = 100Ω, RL = 4000Ω) .................................................................................. 18 2 (RS = 500Ω, R0 = 75Ω, RL = 10 kΩ) ..................................................................................... 19 3 BASIC Program Listing ................................................................................................... 20 4 Summary of Effects ....................................................................................................... 22 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Overview www.ti.com 1 Overview In this application note, the logical progression from the ideal transmission line to the real world of the long transmission line with its attendant losses and problems is made; specifically, the methods to determine the practicality of a certain length of line at a given data rate is discussed. Transmission line effects on various data formats are examined as well as the effects of several types of sources (drivers) on signal quality. A practical means is given to measure signal quality for a given transmission line using readily available test equipment. This, in turn, leads to a chart that provides the designer a way to predict the feasibility of a proposed data-transmission circuit when twisted-pair cable is used. This application note is a revised reprint of section three of the Fairchild Line Driver and Receiver Handbook. This application note, the second of a three-part series (see AN-806 and AN-808), covers the following topics: • The Initial Wave • Cut Lines and a Matched Load • Kirchoff's Laws and Line-Load Boundary Conditions • Fundamental Principles • Tabular Method for Reflections—The Lattice Diagram • Limitations of the Lattice Diagram Method • Reflection Effects for Voltage-Source Drivers • Reflection Effects for Matched-Source Drivers • Reflection Effects for Current-Source Drivers • Summary—Which are the Advantageous Combinations? • Effect of Source Rise Time on Waveforms 2 Introduction In AN-806 it was determined that transmission lines have two important properties: one, a characteristic impedance relating instantaneous voltages and currents of waves traveling along the line and, two, a wave propagation velocity or time delay per unit length. In this chapter, both Z0and δ are used to compute the line voltages and currents at any point along the line and at any time after the line signal is applied. Also, concepts of reflections and reflection coefficients are explored along with calculating methods for voltages and currents. 3 The Initial Wave Application Note AN-806 also showed that for most practical purposes, where fast rise and fall time signals are concerned, the characteristic impedance of the line actually behaves as a pure resistance . Figure 1 shows a generator comprised of a voltage source (magnitude V), a source resistance of RS ohms, and a switch closing at time t = 0 connected to a lossless, infinite length transmission line having a characteristic resistance, R0. Because the relationship of VIN to IIN is known as VIN = R0 IIN, the lossless transmission line can be replaced with a resistor as shown in Figure 2. The loop equation is. IIN (RS + R0) = V (1) Substituting VIN/R0 for IINand collecting terms shows (2) This shows that both source and characteristic resistances act as voltage dividers for the source voltage V. Figure 3 shows voltage and current steps for the various source resistances. Source resistances of less than R0 produce initial voltage steps on the line which are greater than half the compliance of the source voltage, V. A matched source (RS = R0) produces voltage steps exactly half of V and source resistances greater than R0 produce an initial voltage step less than one half V in magnitude. Generators can be classified into three categories: • Voltage source types where RS < R0 • Matched source types where RS = R0 • Current source types where RS > R0 SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 3 Cut Lines and a Matched Load www.ti.com Waveforms of these types will be discussed more fully in AN-808 on long line effects. Suffice to say that initial voltage wave amplitude depends greatly on source resistance. Voltage source type drivers produce higher amplitude initial voltage waves in the line than either matched source or current source type drivers. Figure 1. Generator Driving an Infinite Transmission Line Figure 2. Thevenin Equivalent for Initial Wave Figure 3. Voltage/Current Steps for Three Source Resistances 4 Cut Lines and a Matched Load In examining an infinite, lossless line (Figure 4), it is already known that the ratio of line voltage to current is equal to the characteristic resistance of that line. The line is lossless, and the same voltages and currents should appear at point x down the line after a time delay of xδ. If the line at point x is cut, and a resistor of value R0 is inserted, there would not be a difference between the cut, terminated finite line and the infinite line. The vx and ix waves see the same impedance (R0) they were launched into at time t = 0, and indeed, the waves are absorbed into RL (= R0) after experiencing a time delay of τ = xδ. So, from an external viewpoint, an infinite-length lossless line behaves as a finite-length lossless line terminated in its characteristic resistance. 4 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Kirchoff's Laws and Line-Load Boundary Conditions www.ti.com Figure 4. Voltages and Current on an Infinite Length Line 5 Kirchoff's Laws and Line-Load Boundary Conditions The principle of energy conservation, widely known and accepted in the sciences, applies as well to transmission line theory; therefore, energy (as power) must be conserved at boundaries between line and load. This is expressed in an English language equation as follows. (3) Figure 5 shows power available at the line end is derived by the following formula. (This is assuming inphase current and voltage.) (4) The power absorbed by the load will be (5) while power not absorbed by the load is represented by (6) Here, the r subscript stands for reflected (not absorbed) power, voltage or current, respectively. Applying Kirchoff's laws to point x in Figure 5, the current to the load is iL = ix − ir (7) and voltage across the load is vL = iL RL = vx + vr (8) Figure 5. Boundary Conditions at the Line/Load Interface SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 5 Kirchoff's Laws and Line-Load Boundary Conditions www.ti.com To find the ratio of vr to vx so that it can be ascertained how much power is absorbed by the load, and how much is not absorbed (therefore, reflected), substitute vx/R0for ix and vr/R0 for ir into Equation 7. (9) Rearranging Equation 8 and substituting for iL in Equation 9 yields (10) The minus sign associated with vr/R0means, in this case, that the reflected voltage wave vr travels in the −x direction toward the generator. Collecting like terms of Equation 10 yields (11) So, (12) and the desired relation for vr/vx is (13) This ratio is defined as the voltage reflection coefficient of the load ρVL (14) A similar derivation for currents shows (15) For the remainder of this application note and AN-808, the v or i subscript on the reflection coefficient is dropped, and ρL is assumed to be the voltage reflection coefficient of the load. Similarly, applying Kirchoff's laws to the source-line interface, the voltage reflection coefficient of the source is (16) The current reflection coefficient of the source has the same magnitude as ρS, but is opposite in algebraic sign. When a traveling wave vx, ix meets a boundary such as the line load interface, a reflected wave is instantaneously generated so that Kirchoff's laws are satisfied at the boundary conditions. This is the direct result of the conservation of energy principle. Referring again to Figure 5, the effects of three different termination resistance RL values are shown. Case 1, RL = R0 In this case, RL is equal to the characteristic resistance of the line. Using Equation 14, the voltage reflection coefficient of the load ρL is (17) Since vr/vx = ρL, then vr = ρL vx = 0 and no reflection is generated. This agrees with the discussion of cut lines and matched load where a line terminated in its characteristic impedance behaves the same as an infinite line. All power delivered by the line is absorbed into the load. The waveforms appear as shown in Figure 6. The wave starting at the source at time t = 0 is reproduced at point x down the line after a time delay of t = xδ = τ. 6 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback www.ti.com Kirchoff's Laws and Line-Load Boundary Conditions Figure 6. Waveforms for RL = R0 Case 2, RL > R0 To simplify this case, assume that RS = R0. This means that the initial voltage is (18) Also assume RL = 3 R0, then the load voltage reflection coefficient is (19) The voltage wave arriving at point x at time t = xδ generates a reflected voltage wave of magnitude (20) and the load voltage is (21) The reflected voltage wave vr generated at t = xδ = τ travels back down the line toward the source arriving at the source at time t = 2xδ = 2τ. This wave will be absorbed without generating another reflection because RS was picked to equal R0, making ρS equal to zero. The source voltage is now (22) and equilibrium is achieved. If the circuit in Figure 5 is analyzed using simple circuit theory and neglecting the transmission line effects, it is easily seen that (23) This agrees exactly with Equation 22 and will always be the case. After all reflections cease and the circuit reaches equilibrium, the steady state voltages and currents on the line are the same as those produced using simple dc circuit analysis. Waveforms for RL > R0 (specifically RL = 3 R0) appear in Figure 7. In general, the case where RL > R0is viewed in the following manner. Because the line is capable of delivering more power than can be instantaneously absorbed by the load, the excess power is returned to the source and absorbed in the source resistor (assuming RS = R0). An upper limit on the voltage reflection coefficient is found by allowing RL to go to infinity. In this case, Equation 14 goes to +1. Case 3, RL < R0 In this case, again set RS = R0and allow RL to equal R0/3. The initial wave, as before, is SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 7 Kirchoff's Laws and Line-Load Boundary Conditions www.ti.com (24) and the load voltage reflection coefficient is (25) Therefore, the reflected voltage wave vr is (26) which starts propagating back toward the source at time t = τ. The load voltage at time t = τ is (27) The (−V/4) reflected wave arrives back at the source at time t = 2τ. Because RS is set equal to R0, ρS is, then, equal to zero and no reflected wave will be generated. The voltage at the source is now (28) From a dc circuit analysis, the steady state voltage is (29) This agrees with the result of Equation 28. The waveforms for Case 3 (RL < R0) appear in Figure 8. An interpretation of the actions occurring when load resistance is less than the characteristic line resistance is as follows: when power available at the line end is less than the power the load can absorb, a signal is sent back to the source saying, in essence, “send more power”. It has been shown that a ratio of line and load resistance (ρ) can be used to calculate the voltages and currents in terms of a wave arriving at the boundary, possibly generating a reflected, reverse-traveling wave to satisfy the conservation of energy principle at the line-to-load boundary. This ratio is (30) where RB represents the resistance into the boundary, RB is RS when considering the source-to-line interface and RB would be RL when considering the line-to-load interface. It is obvious that if discussing impedances, then ZSwould be substituted for RS in Equation 30, and there may be some phase angle between the voltage and current waves. The forward traveling wave, vx, plus the reflected wave, vr, is equal to the load voltage (VL). Since vr is ρL vx, this can be expressed as vx(1 + ρL) = vL (31) This quantity (1 + ρ) can be defined as the voltage transmission coefficient of the load and it is known that (32) 8 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Fundamental Principles www.ti.com RS = R0, RL = 3R0 Figure 7. Figure 8. The cases with various load resistances can be summarized. Circuit at time t = τ (one line delay time) Condition 6 1. RL = R0 ρL = 0 No reflection is produced—circuit reaches steady state immediately. 2. RL > R0 ρL > 0 Positive voltage reflection—wave is sent back toward source. Voltage at load is higher than steady stage voltage (overshoot). 3. RL < R0 ρL < 0 Negative voltage reflection—wave is sent back toward source. Voltage at load is lower than steady state voltage (undershoot). Fundamental Principles Before examining the algorithm for keeping track of reflections, there are two principles to keep in mind. • Energy (as power) is conserved at boundary conditions (as explored previously) • The principle of linear superposition applies. This means any arbitrary excitation function can be broken down into step functions, or ramps. The reaction of the circuit to each part can be analyzed, and the results can be added together when finished. This means that a positive pulse of duration t is SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 9 Tabular Method for Reflections— the Lattice Diagram www.ti.com examined by superimposing two step functions, one positive and one negative, starting after a delay of t (Figure 9). It also means the voltage at any point on the line is the sum of initial voltage plus the sum of all voltage waves that have arrived at or passed through the point up to and including the time of examination. Also, the current on the line is, at any point, the sum of initial current plus any forward or reverse traveling currents passing the point up to and including the time the current is examined. It has also been established that the steady state solution for voltages and currents on the line can be found by simple dc circuit analysis. In examining reflection effects for the remainder of this application note, the following conventions are used. A voltage or current wave traveling toward the point of interest will have the subscript “i” for incident wave, A voltage or current wave traveling away from the point of interest will have the subscript “r” for reflected wave, The subscript “S” means the parameter applied to the source (vSfor the voltage at the source, etc.), and The subscript “L” means the parameter applied to the load (vL for the voltage at the load, etc.) Sign conventions for voltage waves and their associated currents are shown in Figure 10. Figure 9. Superposition of Simple Waveforms to Form More Complex Excitations 7 Tabular Method for Reflections— the Lattice Diagram The waves going up and down the line can be monitored by drawing a time scale, as a vertical line with time increasing in the down direction, to represent the location on the line under examination. Because voltages at the source and load ends of the transmission line are normally of primary interest, two time scales are necessary. Drawing arrows from one time scale to the other as in Figure 10 shows the direction of travel of the waves during a specific time interval. Since the main concern is only with the waveforms at the line ends, time scales are ruled off in multiples of the time delay of the line τ. If a unit-step type wave is launched from the source at time t = 0+, it is known that the magnitude of the wave will persist unchanged until a wave arrives back from the load after a round trip delay time of two line delays. The source time scale then is incremented in multiples of 2 mτ where m = 0, 1, 2, 3,… Likewise, the first wave arrives at the load after a single time delay, so the first increment ruling on the load time scale is τ, or one time delay of the line. Because the subsequent waves arrive back at the load in increments of 2τ, the load time scale is ruled off in multiples of (2m + 1)τ where m = 0, 1, 2, 3,… The operation of the lattice diagram is discussed using the example in Figure 12 which is the lattice diagram for the associated circuit. time t = 0− (just before the switch closes) The voltages at the source and load are equal with a magnitude of vinitial. Assume that no initial voltage is present. So, in this case, the voltage at the source and load equals zero. Vinitial = vS(0−) = VL(0−) = 0 (33) time t = 0+ (just after the switch has closed) 10 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Tabular Method for Reflections— the Lattice Diagram www.ti.com Figure 10. Sign Conventions for Waves Figure 11. (a) Line Circuit to be Analyzed (b) Lattice Diagram Figure 12. Reflection Bookkeeping with the Lattice Diagram The first wave vi(1) is launched at the source and begins to travel toward the load end of the line. As previously mentioned, a voltage divider action between RS and R0 is used to derive the magnitude of the initial voltage wave. (34) At this time, the voltage at the source is the sum of the initial voltage plus the voltage wave vi(1) just generated. (35) Because the switch closure represents a step function, the source voltage remains at this level until a wave returns after reflecting from the load at time t = 2τ. time t = τ The incident voltage wave vi(1) now arrives at the load and generates a reflected voltage wave SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 11 Tabular Method for Reflections— the Lattice Diagram www.ti.com (36) where ρL is the voltage reflection coefficient of the load. The reflected voltage wave vr(1) immediately starts traveling back toward the source becoming the incident voltage wave v i(2) which arrives back at the source at t = 2τ. The voltage at the load is now the sum of the initial voltage plus the incident voltage wave vi(1) that just arrived plus the reflected voltage wave that is just departing. (37) Again, because of the step function excitation, the load voltage remains unchanged until the new wave arrives at time t = 3τ. time t = 2τ vi(2) now arrives at the source and generates a reflected voltage wave vr(2) of magnitude (38) where ρS is the source voltage reflection coefficient. The reflected voltage wave vr(2) starts back toward the load end of the line and becomes the incident voltage wave vi(3) arriving at the load at time t = 3τ. The voltage at the source is now the sum of the voltage that was there plus the incident voltage wave just arrived plus the reflected voltage wave just departed for the load. (39) time t = 3τ vi(3) arrives at the load generating vr(3) vr(3) = ρLv i(3) (40) vr(3) departs back toward the source becoming v i(4) to the source. The load voltage is now vL(3) = vL(1) + vi(3) (1 + ρL) (41) time t = 4τ When vi(4) arrives at the source and generates v r(4), then vr(4) = ρSv i(4) (42) starts back toward the load to become vi(5) to the load. The load voltage is now vL(4) = vL(2) + vi(4) (1 + ρL) (43) This process can continue ad infinitum or until no measurable changes are detected. The reflection process at that time is considered complete and the line assumes a steady state condition. Steady state conditions can be found by applying simple dc circuit theory to source load circuits. Summarizing this lattice diagram method, any time t = mτ and m > 1, the following relationships exist: If m is odd, the vi(m) wave is arriving at the load and generates a reflected wave vr(m) = ρLv i(m) (44) This becomes vi(m + 1) as it starts toward the source. The voltage at the load at time t = mτ will be vL(m) = vL(m − 2) + vi(m) (1 + ρL) (45) This is the sum of the voltage that was there before the wave arrived, i.e., vL(m − 2), plus the wave arriving vi (m) and the reflected wave vr(m) departing. If m is even, the vi(m) wave is arriving at the source and generates a reflected wave vr(m) = ρSv i(m) (46) This becomes vi(m + 1) as it starts toward the load. The voltage at the source is now vS(m) = vS(m − 2) + vi(m) (1 + ρS) 12 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated (47) SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Limitations of the Lattice Diagram Method www.ti.com This is the sum of the voltage that was present vS(m − 2) plus the incident wave arriving vi(m) plus the reflected wave departing vr(m). The voltage and current at the source end of the line for a lossless line can be expressed as a summation. (48) (49) where e(t) is the generator voltage as a function of time, and u(t) is the unit step function. Likewise, the load voltage and load current for the lossless line can be expressed as a summation. (50) (51) A similar expression of summation can be developed for the voltage (or current) at any point along the line at any time. Because the lattice diagram is tabular in method, a computer program can be written relieving the designer of bookkeeping and repetitive calculations. A BASIC computer program for lattice diagrams appears in >Figure 13. 8 Limitations of the Lattice Diagram Method Before using the lattice diagram to explore reflection effects with various source and load characteristics, it is necessary to pause at this point and examine the models used by the lattice diagram. First, both the line driver and receiver are simulated either by a constant input or output resistance. The source has two voltage sources and a switch representing the internal source voltage at a time less than zero and equal to (or greater than) zero. The receiver is represented by a single resistor shunting the line end opposite the driver site. The line itself is represented by its characteristic resistance R0 and its total one-way time delay (τ). This is equal to length times propagation delay per unit length. This model is shown in Figure 13. Figure 13. Model Used for Lattice Diagram Method SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 13 Reflection Effects for Voltage Source Drivers www.ti.com Because most data communication circuits are voltage types, that is, the receiver senses the line voltage to decide if a logic One or logic Zero is present, the primary interest is in voltages at the source and load as a function of time. Major exceptions include the current loops used in teletypewriters, telegraphs, and burglar alarm systems. The majority of data communications circuits used in computers, peripherals, and general controllers are voltage types. The lattice diagram method cannot easily use source or receiver current/voltage relationships that are non-linear; i.e., not purely resistive. For non-linear current/voltage characteristics such as found in diodes, a graphic method can be used called the reflection diagram or the Bergeron method. NOTE: A French hydraulic engineer, L.J.B. Bergeron developed the method to study the propagation of water hammer effects in hydraulics. See references, AN-806. Signals exchanged using lattice diagrams are of the unit step variety. When ramps or more complex waves are exchanged, the complexity of the bookkeeping increases dramatically. Additionally, the lines are presumed to be lossless, although a constant line attenuation factor could be accommodated without excessive bookkeeping. These limitations should be kept in mind when examining various source and load resistance combinations and their reflection characteristics. There are three classes of source resistance, RS < R 0, RS = R0 and RS > R0. There are also three classes of load resistance, RL< R0, RL = R0 and RL > R0. This gives nine types of single driver, single receiver line circuits. Each circuit will be examined in turn to determine reflection effects for these combinations with evaluations of each combination for voltage type communications. 9 Reflection Effects for Voltage Source Drivers Initial waves launched by a voltage source type driver (R S < R0) are greater than one-half the magnitude of the internal voltage source. Referring to Figure 13, the initial voltage wave is derived as follows. (52) while the voltage at the source at t = 0+ is (53) If the receiver switching point is at the mean of the driver voltage swing, the initial wave always has sufficient magnitude to indicate the correct logic state as it passes the receiver site. This maximizes the noise margins of the receiver. Since RS < R0, the source voltage reflection coefficient ρS is less than zero. Any voltage waves, then, arriving back at the source are changed in sign, reduced in amplitude (assuming RS > 0Ω), and sent back toward the load. If the load resistance equals the characteristic line resistance (RL = R0), the voltage reflection coefficient of the load is (54) No reflections, therefore, are generated at the load. The voltage wave produced at the source is reproduced at the load after a time delay of τ = ℓδ, and the line assumes a steady state condition. Figure 17 illustrates the source and load voltage waveforms for this case. If RL is greater than R0, ρ L is positive. Waves arriving at the load generate the same polarity reflections as the arriving waves. ρS and ρLare of opposite signs, so a dampened oscillatory behavior of the load voltage is expected. The oscillation period or ringing is 4τ. The overshoot of vL from t = τ to 3τ may cause breakdown of the input circuitry of a receiver, depending on the receiver voltage rating. The undershoot at t = 3τ or 5τ can reduce the noise immunity of a receiver or even cause a logic level misinterpretation—an error in the data. These waveforms are shown in Figure 17. If RL is less than R0, then ρ L is negative and a wave arriving at the load generates a reflection opposite in polarity to the incident wave. This causes the voltage at the source to overshoot steady state voltage at t = 0. Each reflection returning from the load causes the source voltage to continually step down toward the steady state voltage VSS. These steps last for 2τ, or one round trip delay. Load voltage starts an increasing step-up waveform towards VSS at time t = τ, with steps again taking one round trip delay, 2τ. A line receiver placed in the middle of the line sees an entirely different waveform—dampened oscillations much 14 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Reflection Effects for Voltage Source Drivers www.ti.com like the load voltage in Figure 17. This is caused by the negative signs of both source and load voltage reflection coefficients. Each time an incident wave arrives at either source or load, the reflected wave generated at that time has a sign opposite to the sign of the incident voltage wave. The voltage at a distance half way down the line is composed of these forward and reverse traveling waves arriving at that point commencing at time t = 0.5τ, and with each new wave passing that point after one line delay (τ). These waveforms are shown in Figure 17. The optimum load resistance for voltage signal communications on transmission lines driven by a low impedance source (RS < R0) is equal to the characteristic line resistance. Large signal line voltages are produced and there are no reflection effects complicating the waveforms Figure 15. However, a matched load (RL = R0) is a dc load on the driver, thus it increases system power dissipation. But, it does preserve signal fidelity and amplitude allowing use of multiple bridging receivers (Rin ≫ R0) along the line. The unterminated case (RL > R0) reduces dc driver loading and also reduces system power dissipation over the matched load case. The unterminated case does, however, allow the load signal to exhibit pronounced overshoot and undershoot around the steady state voltage. If the load signal undershoot places the receiver in its threshold uncertainty region, data errors result. There is a way to “civilize” the voltage waveform of the unterminated line load by trading off signal rise time versus line time delay. This is discussed later. The final case of RS < R0 and R L < R0 is not generally useful in terms of voltage signals produced (Figure 17). Systems using this case consume more power than the previous two cases and have no particular advantage for voltage mode communications. RL > R0 RL = R0 RL < R0 Figure 14. RS < R0 RL > R0 RL = R0 RL < R0 Figure 15. RS = R0 RL > R0 RL = R0 RL < R0 Figure 16. RS > R0 SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 15 Reflection Effects for Matched-Source Drivers www.ti.com Figure 17. Source and Load Voltage Waveforms for Various R S and RL 10 Reflection Effects for Matched-Source Drivers In all three cases under discussion here, the initial voltage produced by the driver onto the line is (55) since RS = R0. The voltage at the source at time t = 0+ is (56) Assume, for clarity, that initial voltage (V0−) is zero, thus Equation 56 simplifies to (57) Since RS = R0, ρSis equal to zero. This means that load-generated reflections due to load mismatch are absorbed at the source when, at time t = 2τ, the reflected wave arrives back at the source. The line then assumes a steady state throughout. This back match or series termination effect of a matched source allows a wide latitude in choice of load resistance without sacrificing the signal fidelity of the load voltage waveform. If the load resistance equals the characteristic line resistance RL = R0, then ρL equals zero and no load site reflections are generated. The initial voltage wave arrives at the load at time t = τ (one line delay) and voltages (and currents) on the line immediately assume steady state conditions (see Figure 17). The optimum receiver threshold here is one-half the steady state voltage or V0+/4. The main advantage over the voltage source type driver with matched load case (RS < R0, RL = R0) is that RSand RL resistance tolerances may be relaxed without incurring much signal ringing. This effect is due primarily to the termination provided by both line ends, rather than just one line end. Any reflected voltage wave on either system is attenuated by the product of ρS and ρL for each round trip line delay time. Since the ρSρL product for the fully matched case is smaller than the ρSρLproduct for the single matched case, the reflections are attenuated and die out in fewer round trips. For example, if 20% tolerance resistors are used in both cases, ρS and ρL values for the fully matched case become 0.0 ±0.0909, which is a ρSρ L product of ±0.0033. This means that after one round trip (2τ), the reflection amplitude starting back toward the load would be less than 0.33% of the initial wave. Using RS = 10Ω, RL = 100Ω, and R0 = 100Ω as for Figure 17, shows the same 20% tolerances applied to the single matched case —0.8519 ≤ ρS ≤ —0.7857 —0.0909 ≤ ρL ≤ +0.0909 (58) (59) |ρS ≤ ρL |≤ 0.0909 (60) and The voltage reflection amplitude after one round trip is a maximum of 7.7% of the initial wave. The choice between using the single and fully matched system should be carefully considered because the fully matched system does sacrifice signal voltage magnitude to get a decreased dependence on absolute resistor values. If the load resistance for a matched driver circuit is made much greater than the line resistance, the initial wave arriving at the load at time t = τ will be almost double since ρL will be close to +1.0. Because source resistance is set equal to line resistance, ρ S becomes zero, the reflected voltage wave from the load is absorbed by the source at time t = 2τ, and steady state conditions prevail. Waveforms for this case are shown in Figure 17. This is called back matching or series termination. 16 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Reflection Effects for Current-Source Drivers www.ti.com The main advantage of series termination is a great reduction in steady state power consumption when compared with the parallel terminated case (RS ≪ R0, RL = R0). At the same time, series termination provides the same signal fidelity to a receiver placed at the line end. Compare the load voltage waveforms for the two cases in and Figure 17. The main disadvantage to series termination is that receivers placed along the line see a waveform similar to that shown for the source in Figure 17. That is, receivers along the line see the V 0+/2 initial wave as it passes that point on the line, and do not see a full signal swing until the load end reflection passes that point. Consequently, receivers along the line do not see a signal sufficient to produce the valid logic state output until the load reflection returns. Depending on actual line length and receiver characteristics, the receiver may even oscillate, having been placed in its linear operation region. With the benefit, then, of reducing system power, the series termination method has a constraint of allowing only one line receiver located at the line load end. The parallel termination method should be used if other receivers along the line are required. The final case of matched source drivers is with the use of a load resistance less than the characteristic line resistance. The waveforms for this case are shown in Figure 17. A line receiver with a threshold of V0/4 placed at the source responds like a positive, edge triggered one-shot and produces a pulse in response to a +V/2 initial wave of 2τ duration. Aside from its use as a one-shot, this circuit doesn't seem to offer any advantages for voltage mode communications. 11 Reflection Effects for Current-Source Drivers The name current source drivers is somewhat of a misnomer, and might be more properly called currentlimited voltage source drivers. True current source drivers such as the DS75110A are normally used in conjunction with parallel termination resistors to create a matched source. The current source drivers (RS > R0) discussed resemble true current sources in the respect that their output resistance is usually much greater than the characteristic line resistance. The initial voltage step produced on the line is thus usually small vi(1) = (iS(1)R 0). This is due to the voltage divider action of the driver source resistance and the characteristic line resistance. Voltage waveforms for a current source type driver either step up to VSS, reach steady state after 2τ, or execute a dampened oscillation around VSS, depending on whether the load resistance RL is greater, equal, or less than R0, respectively. The second case RL = R0 provides signals much the same as the other two cases where RL = R0, that is, the source voltage steps immediately to VSS, with the load voltage following after one line time delay. Here the amplitude of the signal is much smaller than previous matched load cases. Since the current source type drivers (DS75110A) have high off-state impedances, they allow multiple drivers on the line to produce data bus or party line. Waveforms for the matched load case are shown in Figure 17. The case RL < R0 really provides no useful advantage for voltage mode communications. The negative sign for ρ L and the positive sign for ρS lead to dampened oscillatory behavior, or ringing. The maximum perturbation takes place at the source end of the line. Waveforms for this case are similar to those shown in Figure 17, and are shown to scale in Figure 17. With the given values used to produce the figure, the maximum amplitude ringing appears at the source line end. The RL > R0 case is of interest because it is representative of DTL driving a transmission line with the output going from LOW to HIGH. DTL has a high value RS, (2 kΩ or 6 kΩ) in the HIGH logic state. Since both RS and RL are greater than R0, both ρS and ρL are positive. A small voltage step starts from the source at t = 0+; its magnitude is (61) NOTE: Since the input diode is not represented, the representation of DTL input as a single resistor to ground is not strictly correct. For purposes of approximation, this simple representation is used. Treatment of non-linear current/voltage sources and loads is covered by Metzger & Vabre. (op. cit.) Upon arrival at the load at time t = τ, this initial wave generates a positive voltage reflection since ρL > 0. The voltage reflection arrives back at the source site at time t = 2τ. Since ρS is also positive, another positive voltage reflection is launched back toward the load. The process repeats, and the source and load voltages both execute a step-up approach toward steady state voltage VSS. These waveforms are shown in Figure 17. SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 17 Reflection Effects for Current-Source Drivers www.ti.com In examining voltage at the line midpoint (x = ℓ/2), a step-up type waveform is seen which is the sum of all the incident voltage waves passing the line midpoint up to the time of examination. The midpoint voltage is expressed as follows. vm(t) = VSS(1 − exp[−(t + 0.5τ)/T]) (62) NOTE: This equation is presented without derivation, but a procedure similar to that used by Matick (Ref. 2, AN-806) can be used. for t = n + 0.5τ with n = 0, 1, 2, 3, etc. VSS in Equation 62 is the steady state line voltage (63) and T is a time constant given by (64) with τ being one line delay (τ = ℓδ). NOTE: This equation is presented without derivation, but a procedure similar to that used by Matick (Ref. 2, AN-806) can be used. Equation 62 provides an exact solution for odd multiples of n (n = 1, 3, 5…, so t = 1.5τ, 3.5τ, 5.5τ…), while it approximates vm(t) for even multiples of n (n = 0, 2, 4…, so t = 0.5τ, 2.5τ, 4.5τ…). The closer the ρSρL product is to 1, the better Equation 62 predicts vm(t), particularly for even multiples of n. To illustrate the fitting, Table 1 and Table 2 are generated by the BASIC language computer program Table 3 and their data is plotted in Figure 18. Designers familiar with DTL circuits should quickly recognize that the waveforms shown in Figure 18 are very similar to the rising edge waveforms found when a DTL gate output goes from the LOW to HIGH state. This characteristic waveform has usually been attributed to the series RC circuit (a gate output resistance driving a lumped transmission line capacitance). The time constant for this approach, based on the C(dv/dt) = i rule from simple circuit theory, provides only an approximation. The actual cause of the waveform shape, however, is due to reflection effects. Unfortunately, the only way to speed up the rising edge is to reduce source resistance, (providing an initial step greater than the receiving threshold) and terminate the line to eliminate the load reflections. DTL inability to drive transmission lines at high repetition rates is the direct result of the signal rise time limitation caused by positive reflection coefficients for both the source and load. A transmitted positive pulse may be missed if its duration is less than the time required for the load signal to reach the receiver threshold. The RS > R0 and RL > R0 case provides no definite advantages as voltage mode communication is concerned. This case, in fact, poses a definite hazard to high speed data communications because the reflections cause, in effect, a slow, exponential signal transition. Because line delay is a factor, longer lines will only increase the effect. Table 1. (RS = 2000Ω, R0 = 100Ω, RL = 4000Ω) (1) (1) 18 TIME VM(T) VAPPX %DIFF 0.5 0.04762 0.04820 +1.220% 1.5 0.09292 0.09292 +0.000% 2.5 0.13390 0.13440 +0.373% 3.5 0.17288 0.17288 +0.000% 4.5 0.20815 0.20858 +0.207% 5.5 0.24170 0.24170 +0.000% RHOS = 0.904762 RHOL = 0.951220 TAU = −13.3250 V1(1) = 4.76190H-C2 VSS = 0.666667 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Reflection Effects for Current-Source Drivers www.ti.com Table 1. (RS = 2000Ω, R0 = 100Ω, RL = 4000Ω) (1) (continued) TIME VM(T) VAPPX %DIFF 6.5 0.27206 0.27243 +0.136% 7.5 0.30093 0.30093 +0.000% 8.5 0.32705 0.32737 +0.097% 9.5 0.35190 0.35190 +0.000% 10.5 0.37439 0.37466 +0.073% 11.5 0.39577 0.39577 +0.000% 12.5 0.41512 0.41536 +0.057% 13.5 0.43353 0.43353 +0.000% 14.5 0.45018 0.45038 +0.045% 15.5 0.46602 0.46602 +0.000% 16.5 0.48035 0.48053 +0.036% 17.5 0.49399 0.49399 +0.000% 18.5 0.50632 0.50647 +0.030% 19.5 0.51805 0.51805 +0.000% 20.5 0.52867 0.52880 +0.024% 21.5 0.53877 0.53877 +0.000% Table 2. (RS = 500Ω, R0 = 75Ω, RL = 10 kΩ) (1) (1) TIME VM(T) VAPPX %DIFF 0.5 0.13043 0.13971 +7.112% 1.5 0.25893 0.25893 +0.000% 2.5 0.35390 0.36066 +1.909% 3.5 0.44746 0.44746 +0.000% 4.5 0.51661 0.52153 +0.952% 5.5 0.58473 0.58473 +0.000% 6.5 0.63509 0.63867 +0.564% 7.5 0.68469 0.68469 +0.000% 8.5 0.72135 0.72396 +0.361% 9.5 0.75747 0.75747 +0.000% 10.5 0.78416 0.78606 +0.242% 11.5 0.81046 0.81046 +0.000% 12.5 0.82990 0.83128 +0.167% 13.5 0.84904 0.84904 +0.000% 14.5 0.86320 0.86420 +0.117% 15.5 0.87714 0.87714 +0.000% 16.5 0.88744 0.88818 +0.083% 17.5 0.89759 0.89759 +0.000% 18.5 0.90510 0.90563 +0.059% 19.5 0.91249 0.91249 +0.000% 20.5 0.91795 0.91834 +0.042% 21.5 0.92334 0.92334 +0.000% RHOS = 0.739130 RHOL = 0.985112 TAU = −6.30356 V1(1) = 1.30435 VSS = 0.952381 SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 19 Reflection Effects for Current-Source Drivers www.ti.com Table 3. BASIC Program Listing PRINT'ENTER RS, R0, RL'1 100 INPUT R1, R0, R2 110 P1=(R1-R0)/R1+R0) 120 P2=(R2-R0)/R2+R0) 130 V1=R0/R1+R0) 140 K1=2./LOG(P1*P2) 150 V9=R2/(R1+R2) 160 PRINT'RHOS='; P1;'RHOL=';PS;'TAU=';K1 170 PRINT 'V1(1)=';V1;'VSS=';V9 180 V=V1 190 PRINT'TIME VM(T) VAPPX %DIFF' 200 FOR T=0.5 TO 20.5 STEP 2. 210 V2=V9*(1.-EXP((T+.5)/KL)) 220 P=100.*(V2-V)/V 230 PRINT USING 250,T,V,V2,P 240 :##.# -#.##### -#.##### +###.###% 250 V1=V1*P2 260 V=V+V1 270 REM SOURCE END 280 V2=V9*(1.-EXP( (T+1.5)/k1 ) ) 290 P=100.*(V2-V)/V 300 PRINT USING 250,T+1.,V,V2,P 310 V1=P1*V1 320 V=V+V1 330 NEXT T 340 PRINT 350 PRINT 360 PRINT 370 GOTO 100 380 20 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Summary—Which are the Advantageous Combinations? www.ti.com Table 3. BASIC Program Listing (continued) END 390 Figure 18. Approximation of Midline Voltage with RS > R0 and RL > R0 (65) 12 Summary—Which are the Advantageous Combinations? In examining the basic combinations of source, line and load resistances, and typical waveforms characteristic of each case, advantageous combinations can be determined. The primary results are tabulated in Table 4. those combinations generally used in voltage mode communications circuits are as follows. 1. Unterminated case (RS ≪ R0, R L ≫ R0). This situation provides low steady state power dissipation and large signal levels, but also shows pronounced “ringing” effects. The “ringing” can be reduced by SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 21 Summary—Which are the Advantageous Combinations? www.ti.com controlling signal rise/fall time versus τ, or by clamping diodes to limit load signal excursions. This case is representative of TTL circuits and is thus widely employed. 2. The parallel terminated case (RS ≪ R0, RL = R0) provides large signal levels, and excellent signal fidelity. However, it is power consuming with most of that power dissipated in the load resistor. This case is useful for cleaning up the reflection effects of Case 1 but, at the same time, does require a driver circuit to have its internal current limits set at greater values than those required to produce the desired signal level into the minimum line resistance used. Thus, this case requires specific line driver devices such as the DS75114/DS9614. Ordinary TTL, except for the above mentioned circuits, has too low a current limit point to adequately drive 50Ω lines. 3. The series terminated or backmatched driver case RS = R 0, RL ≫ R0 provides a low steady state power dissipation system for use with one receiver located at the load end of the line. The positive reflection coefficient of the load is used to approximately double the initial wave arriving at the load. Setting R S = R0 terminates the reflected wave when it arrives back at the source site after two line delays, and the line then assumes steady state conditions. The use of other receivers located along the line is not recommended, because they will not see the full driver signal swing until the reflection from the load passes their particular bridging points Such receivers could malfunction, as they would see a voltage very close to their threshold, and perhaps even place the line receiver in its linear operating region. This could make the line receiver sensitive to oscillatory, parasitic feedback. If these constraints are acceptable, the series termination method can be used to good advantage in providing the same signal fidelity and signal amplitude as with the parallel termination method, while at the same time, contributing a significant savings in steady state power consumption. 4. The fully matched case RS = R0, RL = R0 not only provides excellent signal fidelity all along the line, but also has reduced signal amplitude over that of the parallel terminated case. Additionally, the power consumption is somewhat less than the parallel termination case and the power is divided equally by the source and load. The primary advantage of the fully matched system is that termination resistor tolerances can be relaxed somewhat without incurring large amounts of ringing. This is because both the source and load act as line terminations. Table 4. Summary of Effects Signal Characteristics Optimum Receiver Threshold Line Receivers Allowed at Other Than Load End of Line? ≫ R0 Ringing Pronounced 0.5 VSS Yes Undershoot May Cause Errors ≪ R0 = R0 Excellent Fidelity 0.5 VSS Yes Load Resistor Consumes Power ≪ R0 ≪ R0 Awful—Different Signals at Each Point on the Line NA No Not Generally Useful Series Terminated or Backmatched Driver = R0 ≫ R0 Load Signal Excellent 0.5 VSS No Reduced Power Consumption Over Parallel Termination Fully Matched = R0 = R0 Excellent Fidelity 0.25 VSS Yes Greater Tolerances on Resistors Allowed for Same Fidelity as Parallel Termination = R0 < R0 Load Signal Like a One-Shot NA NA Not Generally Useful for Data, is Useful as Pulse Generator ≫ R0 ≫ R0 Exponential Like Signal Waveforms 0.5 VSS Yes Low Power Consumption. Increased Delay due to Signal “Rise” Times. (Driver) Source Resistance (Receiver) Load Resistance Unterminated ≪ R0 Parallel Terminated Configuration Name (if any) Comments (66) 22 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Effect of Source Rise Time on Waveforms www.ti.com Table 4. Summary of Effects (continued) Configuration Name (if any) 13 (Driver) Source Resistance (Receiver) Load Resistance ≫ R0 = R0 ≫ R0 < R0 Optimum Receiver Threshold Line Receivers Allowed at Other Than Load End of Line? Small Signal Amplitude and Excellent Fidelity 0.5 VSS Yes Produces Only Small Signal Voltages Compared with Other Methods. Uses Current Sinking Drivers such as the 75110A. Very Small Signal Amplitudes, also Ringing NA NA Not Generally Useful Signal Characteristics Comments Effect of Source Rise Time on Waveforms Previously, it was assumed that the source-produced signal rise time was always much less than the line time delay (τ). Because the waveforms for the source and load voltage were the superposition of incident and reflected waves occurring at their proper times, and because the shape of each wave was a square edged step function, the resultant source and load waveforms were thus also square edged, or ideal in nature. In many practical cases, particularly when line length is short, the source excitation possesses a finite, and non-negligible, rise time. Therefore, depending on the ratio of rise time to line delay, it is possible to have a new wave start arriving at the point of interest before the previous wave can reach its final value. The net waveform for voltage or current at that point, then, would consist of the superposition of two or more waves during their time of overlap. To study the superposition effect on signal waveforms, the source excitation is represented as a simple linear ramp rise to its final value of V0+, so e(t)— = 0 for t < 0 e(t)— = V0+•t/tr for 0 ≤ t ≤ tr e(t)— = V0+ for t > tr and where tr represents the 0-to-100% source rise time. The circuit model and its lattice diagram are shown in Figure 19. The values of RS, R0 and RL were chosen to equal those of an actual circuit on hand, allowing the theoretical waveforms, obtained by graphical superposition, to be compared with the measured response of an actual circuit. Figure 20 shows the load voltage vL, source voltage vS and source current iS waveforms versus time for a circuit with a source rise time very much less than τ. The actual waveforms for vL, v S and iS are composed of the superposition of both incident and reflected waves in their proper time sequence. In the figures, these waves are shown as dotted lines. Each wave represents the sum of the incident wave plus its reflection. The resultant vL, vS and i S waveforms (shown as solid lines) are the superposition of the waves represented by the dotted lines. With the exception of a slight rounding of the edges, the actual waveforms for the circuit, shown in the oscilloscope photograph in Figure 20, closely approximate the waveforms predicted by theory. Source Load t vi + vr ii + ir vS iS t vi + vr ii + ir vL iL in (τ) (V) (mA) (V) (mA) in (τ) (V) (mA) (V) (mA) 0 0.9400 12.53 0.9400 12.53 1 1.8500 0.40 1.8500 0.40 2 0.1224 −22.64 1.0624 −10.10 3 −1.5500 −0.34 0.3000 0.06 4 −0.1026 18.97 0.9599 8.87 5 1.2986 0.28 1.5986 0.35 6 0.0859 −15.90 1.0458 −7.03 7 − 1.0881 −0.24 0.5106 0.11 8 −0.0720 13.32 0.9738 6.29 9 0.9116 0.20 1.4222 0.31 10 0.0603 −11.17 1.0341 −4.87 11 −0.7638 −0.17 0.6584 0.14 12 −0.0505 9.36 0.9836 4.48 13 0.6399 0.14 1.2983 0.28 SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 23 Effect of Source Rise Time on Waveforms www.ti.com Source Load t vi + vr ii + ir vS iS t vi + vr ii + ir vL iL in (τ) (V) (mA) (V) (mA) in (τ) (V) (mA) (V) (mA) 14 0.0424 −7.84 1.0259 −3.36 15 −0.5362 −0.12 0.7622 0.16 16 0.0355 6.57 0.9904 3.21 17 0.4492 0.10 1.2114 0.26 Figure 19. Transmission Line Model and Its Lattice Diagram Figure 20. Waveforms for tr = 2 ≪ τ 24 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Effect of Source Rise Time on Waveforms www.ti.com If the source excitation is adjusted so that its 0-to-100% rise time tr is equal to 2τ, each of the vi + vr and ii + ir waveforms must be modified to include this rise time. The waves will have the same final value as predicted by the lattice diagram, but they now require two line time delays to reach this final value. The vL, vS and iS waveforms consist of the superposition of these linear ramps. Because each wave reaches its final value just as a new wave arrives, their superposition converts the square edged vL, vS and iS waveforms into triangular waveforms. This is shown in Figure 21. The accompanying oscilloscope plot shows the close correspondence between the actual and theoretical waveforms whereas an additional oscilloscope photograph in Figure 21 shows the actual waveforms for the case where tr = τ. Not surprisingly, the tr = τ case changes the vL, vS and iS waveforms of the tr ≪ τ case into trapeziodal forms because each arriving wave reaches its final value well before a new wave arrives. If the source excitation is adjusted such that its rise time equals three line delays tr = 3τ, the vi + vr and ii + ir waves overlap for a period of time equal to τ. That is, each wave reaches only ⅔ of its final value when a new wave starts arriving. Considering the waveform, the load voltage from time τ to 3τ is vi(1) (1 + ρL) e (t − τ) (67) Starting at t = 3τ, the wave vi(3) = vi(1)ρ SρL e(t − 3τ). (68) begins arriving from the source, and the load voltage then is the superposition of these two waves. Because vi(3) is a negative wave (ρS < 0), the algebraic sum of the last third of the first wave and the first third of the second wave vi(3) arriving at the load causes the load voltage to reduce in amplitude from the (t r ≪ τ) case. Likewise, the source voltage and source current show reduced amplitudes over the ideal case, due to the overlap period of the waves arriving at the source. Theoretical and actual waveforms for the tr = 3τ case are shown in Figure 24. Notice that load voltage perturbations and source current iS requirements are reduced from those of the tr ≪ τ case. Similarly, the ratio of tr to τ can be successively increased. This results in reduced ringing on the load voltage and reduced source current due to the overlapping of more and more vi + vr (or i i + ir) waves. Actual and theoretical waveforms for tr equal to 4τ, 6τ, and 8τ are shown in Figure 26, Figure 28 and Figure 30, respectively. In each case, as the tr to τ ratio is increased, the instantaneous source and load voltages become more equal. The source current is also reduced so that the circuit exhibits fewer reflection effects and the transmission line itself can be considered as a simple interconnection from dc circuit theory. Figure 21. Waveforms for tr = 2τ SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 25 Effect of Source Rise Time on Waveforms www.ti.com Figure 22. Waveforms for tr = 2τ Figure 23. Waveforms for tr = 2τ Figure 24. Waveforms for tr = 3τ 26 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Effect of Source Rise Time on Waveforms www.ti.com Figure 25. Waveforms for tr = 3τ Figure 26. Waveforms for tr = 4τ SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 27 Effect of Source Rise Time on Waveforms www.ti.com Figure 27. Waveforms for tr = 4τ Figure 28. Waveforms for tr = 6τ Figure 29. Waveforms for tr = 6τ 28 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback Effect of Source Rise Time on Waveforms www.ti.com Figure 30. Waveforms for tr = 8τ Figure 31. Waveforms for tr = 8τ Using the tr to τ ratio to reduce reflection effects has many practical advantages in digital design. The low source and high input resistance of TTL or ECL circuits allows one gate to drive many receiving gates. The reflection effects of this unterminated combination, however, can cause data errors or at least lead to reduced noise immunity due to the pronounced load voltage undershoot. Since the rise and fall times of these devices are easily measured, a maximum line length can be set such that the resulting tr to τ ratio provides the desired reduction in ringing. This is the primary basis for the wiring rules of each logic family and, usually, the tr to τ ratio is chosen somewhere between 3:1 and 4:1. As an example, the rise and fall time for normal TTL is t10%–90% = 6 ns. When this is converted to an equivalent linear 0% to 100% time, t r = 8 ns. A common propagation delay of 1.7 ns/ft, in combination with the requirement that tr = 3τ, gives the maximum line length of approximately 18 inches. This corresponds with the published recommendation of the various manufacturers for the 74 series TTL circuits. A similar computation of the rise and fall times for other logic families yields their respective line length recommendations. The faster families require shorter SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated 29 Effect of Source Rise Time on Waveforms www.ti.com line lengths for the same trto τ ratio, and slower logic families allow relatively longer line length. This ratio can also be used to make stubs or taps on lines “disappear”. In other words, if the stub's time delay is made very short when compared to the tr of the signal at the stub line location, the stub reflections will have a minimal effect on the line signals. A stub length to generate a tr to τ ratio of greater than 8:1 is usually considered adequate to negate the stub reflections. The third primary application of the tr to τ ratio for controlling reflection effects is that used in some standard data communications interfaces such as EIA/TIA-232-E (RS-232). Here, driver slew rate is explicitly controlled. This, along with the implied maximum interconnect cable length serves to produce a tr to τ ratio of 3:1 or greater. This, in turn, reduces the reflection effects inherent in a voltage source driver, unterminated line system. The main disadvantage of using the tr to τ ratio to control reflection effects is in the overall time for the signal representing the data to rise above the receiver threshold level. With the parallel terminated method, the minimum time delay was τ or one line delay. When the tr to τ ratio is used, an additional delay time of approximately 0.5 tris added to the line delay yielding, therefore, a greater effective signal propagation delay. This increased delay may or may not be acceptable in the desired system so the trade-off between ease of usage of the unterminated case must be weighed against the increased effective signal delay over that delay obtainable with the terminated case. 30 AN-807 Reflections: Computations and Waveforms Copyright © 2004, Texas Instruments Incorporated SNLA027B – May 2004 – Revised May 2004 Submit Documentation Feedback IMPORTANT NOTICE Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, enhancements, improvements and other changes to its semiconductor products and services per JESD46, latest issue, and to discontinue any product or service per JESD48, latest issue. Buyers should obtain the latest relevant information before placing orders and should verify that such information is current and complete. All semiconductor products (also referred to herein as “components”) are sold subject to TI’s terms and conditions of sale supplied at the time of order acknowledgment. TI warrants performance of its components to the specifications applicable at the time of sale, in accordance with the warranty in TI’s terms and conditions of sale of semiconductor products. Testing and other quality control techniques are used to the extent TI deems necessary to support this warranty. Except where mandated by applicable law, testing of all parameters of each component is not necessarily performed. TI assumes no liability for applications assistance or the design of Buyers’ products. Buyers are responsible for their products and applications using TI components. To minimize the risks associated with Buyers’ products and applications, Buyers should provide adequate design and operating safeguards. TI does not warrant or represent that any license, either express or implied, is granted under any patent right, copyright, mask work right, or other intellectual property right relating to any combination, machine, or process in which TI components or services are used. Information published by TI regarding third-party products or services does not constitute a license to use such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property of the third party, or a license from TI under the patents or other intellectual property of TI. Reproduction of significant portions of TI information in TI data books or data sheets is permissible only if reproduction is without alteration and is accompanied by all associated warranties, conditions, limitations, and notices. TI is not responsible or liable for such altered documentation. Information of third parties may be subject to additional restrictions. Resale of TI components or services with statements different from or beyond the parameters stated by TI for that component or service voids all express and any implied warranties for the associated TI component or service and is an unfair and deceptive business practice. TI is not responsible or liable for any such statements. Buyer acknowledges and agrees that it is solely responsible for compliance with all legal, regulatory and safety-related requirements concerning its products, and any use of TI components in its applications, notwithstanding any applications-related information or support that may be provided by TI. Buyer represents and agrees that it has all the necessary expertise to create and implement safeguards which anticipate dangerous consequences of failures, monitor failures and their consequences, lessen the likelihood of failures that might cause harm and take appropriate remedial actions. Buyer will fully indemnify TI and its representatives against any damages arising out of the use of any TI components in safety-critical applications. In some cases, TI components may be promoted specifically to facilitate safety-related applications. With such components, TI’s goal is to help enable customers to design and create their own end-product solutions that meet applicable functional safety standards and requirements. Nonetheless, such components are subject to these terms. No TI components are authorized for use in FDA Class III (or similar life-critical medical equipment) unless authorized officers of the parties have executed a special agreement specifically governing such use. Only those TI components which TI has specifically designated as military grade or “enhanced plastic” are designed and intended for use in military/aerospace applications or environments. Buyer acknowledges and agrees that any military or aerospace use of TI components which have not been so designated is solely at the Buyer's risk, and that Buyer is solely responsible for compliance with all legal and regulatory requirements in connection with such use. TI has specifically designated certain components as meeting ISO/TS16949 requirements, mainly for automotive use. In any case of use of non-designated products, TI will not be responsible for any failure to meet ISO/TS16949. Products Applications Audio www.ti.com/audio Automotive and Transportation www.ti.com/automotive Amplifiers amplifier.ti.com Communications and Telecom www.ti.com/communications Data Converters dataconverter.ti.com Computers and Peripherals www.ti.com/computers DLP® Products www.dlp.com Consumer Electronics www.ti.com/consumer-apps DSP dsp.ti.com Energy and Lighting www.ti.com/energy Clocks and Timers www.ti.com/clocks Industrial www.ti.com/industrial Interface interface.ti.com Medical www.ti.com/medical Logic logic.ti.com Security www.ti.com/security Power Mgmt power.ti.com Space, Avionics and Defense www.ti.com/space-avionics-defense Microcontrollers microcontroller.ti.com Video and Imaging www.ti.com/video RFID www.ti-rfid.com OMAP Applications Processors www.ti.com/omap TI E2E Community e2e.ti.com Wireless Connectivity www.ti.com/wirelessconnectivity Mailing Address: Texas Instruments, Post Office Box 655303, Dallas, Texas 75265 Copyright © 2012, Texas Instruments Incorporated

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising