3 Electronic Switches

Transcription

1 EL 3 Electronic Switches n the digital circuit-technology are used diodes and transistors as electronic switches. Switching elements have a nonlinear characteristic, e.g. a nonlinear resistor (EE-03001). NLW = g() = f() 0 niversity of NONLNEA ESSTO EE The characteristic of a nonlinear element can be approximated by straight lines. n this case by constant resistors. The actual effective resistor is a function of the voltage. So one gets an equivalent circuit, that consists of resistors, a switch with one position for every resistor and voltage sources. n this lecture every electronic switch will be reduced to a simple equivalent circuit, containing only resistors, switches and voltage sources. So a calculation of the behavior will be very simple. For the technological realization the following parameters are of interest: - Switch-resistor in the ON- and OFF-state ( ON and OFF ), - switching time and propagation time, - allowed signal-levels (current and voltage), - control of the switch (powerless or not, potential-free) and - price of the circuit. These parameters are discussed in the following for the different technically realizable switching elements. 3.1 PN- and Schottky-diodes The simplest electronic switch is the semiconductor-diode. For diodes with PN- and Schottkytransitions the behavior is shown. With help of diodes combinatorical logic can be implemented how AND and O. These basic-circuits are used frequently at TTL-circuits

2 EL Static behavior of diodes Starting of the ideal characteristic of a diode the courses of voltage shown in the EE (t) (t) D 2 (t) (t) A 0 1 t (t) A 2 0 t niversity of DEAL DODE AS SWTH EE Through the valve-effect of the diode only the positive parts of the generator-voltage (t) appear at the load-resistor. The negative parts of the generator-voltage with the amplitude of 1 fall off at the closed diode. eal semiconductor-diodes originate through diffusion of p- and n-doped areas (PN-diode) or through a metal/ semiconductor-transition. As semiconductor material today is used Silicium. n the past Germanium because of the low conducting voltage was used also. Germanium diodes are manufactured today no longer, because of their high reverse current as well as low reverse voltage and because of the technology has become obsolete. For superfast circuits today is used doped Galliumarsenid as diode-material. There can be produced PN-diodes and Schottky-diodes. eal diodes show final values for the conducting- and reverse-resistors F and (= ON respectively OFF ) and for the conducting- (forward-) and reverse-voltages F and. ecause of the nonlinear characteristic these values depend on the operating point. The reverse-resistor and the conducting voltage are depending on the temperature too. The conducting voltage has a temperature-coefficient of 2 mv/ K, that leads in the allowed temperature-range (e.g. of 55 until +150 ) to extreme operating-point-displacements. The figure EE shows the equivalent circuit

3 EL niversity of S 1 F F EQVALENT T OF A DODE 2 S EE At the equivalent circuit (EE-03106) the diode for > F is in the conducting state (switchposition S = 1) and for < F in the reverse state (switch-position S = 2). ecause the conducting resistor (forward resistor) of diodes is very small ( F of 1 until 20 Ω), in many cases the circuit is calculated with a constant conducting voltage (forward voltage, threshold voltage) F ( F >> * F ). The typical values F = 0.3 until 0.45 V for Schottky-diodes and F = 0.6 until 0.8 V for Silizium-PN-diodes determine therefore the static switching behavior at the logic-circuits with bipolar transistors. The reverse current of a diode is a function of the temperature. At a temperature of 300 K a silicum-diode has a reverse current of 0,1 until 10 na Dynamic behavior of diodes n the equivalent circuit (EE-03106) the capacitor S is to see. The depletion layer capacitor S is voltage-dependent. t appears in the forward and in the reverse state, also if there is no current. Parallel is also a constant capacitor of the housing. The dynamic behavior of a diode is characterized by the switching times at the changes of the state. From blocking to conducting state only appears a very short time t F. At the transition from conducting to blocking state appears the longer reverse recovery time t rr, that is the sum of the memory time t S and the discharging time t r of the capacitor S. At PN-diodes the minimum of t rr is 4 ns. This value is lower at Schottky-diodes Schottky-diode At the Schottky-diode the diode-effect is effected by a metal-semiconductor-transition. t is a contact between the metal and n-doped silicium. The structure shows figure EE

4 EL metal n- epitaxy n + - substrate SiO 2 space charge region A K niversity of STTE OF A SHOTTKY-DODE EE The Schottky-diodes havy much faster memory times as PN-diodes. The metalsemiconductor-transition effects small foreward voltages, but also small reverse voltages. 3.2 Diode-circuits The valve-effect of diodes makes it very simple to manufacture combinatorical logic. EE shows the both basic circuits for AND and O. D 0 D 1 D 0 D 1 1 Q Q niversity of GATE WTH DODES EE-03180

5 EL The figure shows two different gates: - input level at the anodes and load-resistor grounded - input level at the katodes and load-resistor to supply-voltage The kind of combinatorical logic will be calculated using the equivalent circuit of a diode. For example: Left gate: 1 = H = 5 V, 2 = L = 0V, F = 0,7 V, F = 0 Ω, = : Q = 4,3 V = H 1 = 2 = L = 0V: Q = 0 V = L 1 = 2 = H = 5V: Q = 4,3 V = H O-circuit for positive logic AND-circuit for positive logic nputs Output nputs Output 2 1 Q Q 2 1 Q Q L L L 0 L L L 0 L H H 1 L H L 0 H L H 1 H L L 0 H H H 1 H H H

6 EL 3.3 The transistor as switch With diodes itself only AND and O logic can be implemented. The inverter (NOT-function) can not be realized with diodes. The input-current at diode-circuits are high produced through the level of the input voltages and the resistor. An application of transistors let avoid these disadvantages, because the bipolartransistor is controlled over the base (respectively over the gate of the MOS-transistor). For the digital circuits the grounded emitter is used mainly. The transistors mainly are manufactured out of silicium. There are existing two kinds of bipolar transistors: niversity of TANSSTO E E SYMOLS NPN - TANSSTO PNP-TANSSTO EE The npn-transistor with the dotation layers n p n and the pnp-transistor with the dotation layers p n p. The electrodes or contacts are named collector, base and emitter. ollector and emitter have the same dotation, but a different construction because the collector-base-diode operates in reverse direction and takes the most power consumption. So it often must be well cooled. The base-emitter-diode operates in forward direction. The figure EE shows the voltages and currents at the grounded emitter. t is valid: and + + E = 0 E - E = 0 The most important characteristic is the transistor current gain: = N * or: = -A N * E With the current amplification factor for emitter grounded A N

7 EL E E E + DG GS D DS S D + NPN N-KANAL E E E - DG GS D D DS S - PNP P-KANAL niversity of TANSSTOS EE urrent- and voltage-switches With help of electronic switches voltages or currents are switched at a load-resistor (ONstate). n the OFF-state no more voltage at the load-resistor falls off (figure EE-03210). 0 0 A off on off on voltage switch current switch niversity of PNPLES OF SWTHNG EE At the voltage-switch in the ON-state over the closed switch the voltage 0 falls off at the resistor. At the current-switch the current of the current-source over an alteration switch is

8 EL led to the resistor. n the position ON at a voltage = 0 * falls off. n the position OFF the current is led over a parallel resistor A. Then at consequently no voltage falls off. At electronic switches it is usual, to mark the ON-state with the index "X" or "ON." The OFFstate is marked with the index "Y" or "OFF". The behavior of the ideal switch can be shown by a characteristic (figure EE-03212) with the belonging load-resistor. On the intersections of the axes and the load-resistor one gets SX, SX, SY and SY. A mechanical switch shows approximately the ideal behavior. Electronic switches like diodes and transistors have in the ON-state residual voltages SX. This follows from the ON-resistors and saturation-voltages. n the OFF-state one gets a terminated blocking resistor and consequently a leak-current SY. Therefore one gets noit ideal values for the switch-voltages and -currents. The advantage of electronic switches is the approximately inertia-free switching process. Advantages of electronic switches - maintenance-free - small - bounce-free - high durability - high circuit speed - small requirement of power for the control of the switch S S 0 0 S S S S S ON OFF niversity of S ELETON SWTH S EE As electronic switch can be used the NPN-transistor shown in figure EE The switch is between collector and emitter E. Over the control-electrode base the transistor is supplied by a voltage E or a current. The level of the control signal definites the intersections in the field of the output characteristics with the load-resistor. For example one gets the operating points at = 0 and at Ü respectively X

9 EL / E E E niversity of SAT. TANSSTO SWTH FAMLY OF HAATESTS EE The broken line is the border for the voltage = 0 V. n this case is E equal E. This border-line is named "saturation-border". The intersection of the border-line with the working characteristic of the resistor indicates the current at the saturation-border Ü. For values greater than Ü the transistor is in the saturation ( E E ). One leaves the linear controlarea, the collector current is no longer the product of base-current and the transistor current gain N. An increase of the base-current to X only leads to a small increase of the collector current to X, but one gets a small voltage EX (LOW-level). 3.4 nverter with bipolar transistors n this chapter the inverter with grounded emitter is treated. The saturated voltage-switch is investigated and calculated. The saturation takes care for small residual voltages at the blocked transistor-switch. This and consequently the control determine the switching times of the transistor. To decrease the switching times are used several methods. The dynamic behavior is very influenced by the load at the output. This is especially valid for capacitive and inductive loads

10 EL Saturated transistor inverter The figure EE shows a simple equivalent circuit for the static transistor-switch. niversity of E 0 A * N E E E SMPLE EQVALENT T OF THE STAT TANSSTO- SWTH E E EE From = 0 - A N * E and N = A N / (1 - A N ) follows = N * + 0 * (1 + N ) The collector leakage current 0 of 0.5 until 20 na generally can be neglected at Silicium transistors. The current amplification factor N at grounded emitter is valid for the normal operation with blocked collector diode. N is not constant, it is depending on the collector current. For our calculations we take a constant N. For the control of a transistor-switch are to regard several intersections of the working characteristic of the resistor with characteristic lines in the field of output characteristics. The figure EE shows four areas. The blocking state region, the for digital applications forbidden linear control area, the overload area with the termination of the load-hyperbola and the saturation-area V, that is terminated by the saturation border. The blocking state region is terminated by = 0. The intersection of the characteristic for = 0 and a working characteristic gives the operation point P0 in the blocking state. The points P1 and P2 bound the overload-area. The operation-point P3 lies on the saturation border. Here is valid: N = Ü / Ü

11 EL / E E niversity of TANSSTO-PAAMETES EGON OF SATATON EE The belonging collector-emitter-residual-voltage EÜ is between 0.1 and 1 V. f the collector current will be increased, so the transistor operates in the saturation area (above the operation-point P3). One gets a collector current X = ( - X ) / > Ü The belonging base current is a m-fold of the base current at the saturation-border Ü. This definites the saturation factor m: m = X / Ü = N * X / Ü N * X / X ecause of the saturation the residual-voltage between collector and emitter decreases from EÜ to EX. For a high reliability one chooses m > 1, so that currents > Ü are reached safely also at tolerances of the electronic components. For a transistor-switch in the ON-state follows for the ON-resistor ON = EX / X n the OFF-state flows the collector leakage current 0. One gets an OFF-resistor OFF = Y / Y = / 0 The simplest transistor inverter contains a transistor as switch, a base resistor for impressing of the base current and the load-resistor at the collector. n the figure DST is the additional base shunting resistor A, that improves with the voltage the blocking behaviour of the switch. At a NPN-transistor is zero or negative. So one gets in the blocking state a base-emitter-voltage EY lower than ES = 0,4 Volts

12 EL G Q E Q E G A niversity of SATATED TANSSTO-NVETE DST Transistor inverter in the ON-condition. t is valid: m = N * X / X Or for the base-current: X = m * X / N For the collector current is valid: X = ( EX ) / + QX For the base current is valid: X = ( GX EX ) / ( GX + ) - ( EX ) / A From these equations one gets for example the minimum value of A : A = GX GX + EX m EX N EX + QX Transistor inverter in the OFF-state. The value of EY must fulfill the following condition: EY < ES

13 EL Out of the circuit in figure DST follows: EY = GY GY A + + A GY GY A Then one gets a maximum value for A A = ( ) ( + ) EY GY GY EY Push-pull switch The until now discussed transistor-switches all operate after the principle of single phase, that means, that in the ON-state is made by a switch a connection with the supply-voltage (e.g. GND). n the OFF-state the switch is blocked. This blocking behaviour will not be reached inertia-less. So this will limit the maximum frequency of the system. sing two phases or push-pull switches is possible fast switching to all states. Two complementary driven switches (one is in the ON-state, the other in the OFF-state or viceversa) connect the load either at the positive or at the negative supply-voltage. At digital logic circuits (e.g. TTL, MOS) with a positive supply-voltage (e.g. ) it is switched between this and ground (e.g. GND). This has the consequence, that for each state one gets a low impedance path. This leads with capacitive loads to small time constants. τ 1,2 = ( ON1,2 l ) * L One kind of push-pull switch is the Totem pole circuit. The principle was used first for TTL-circuits. The name "Totem pole" should imply, that this outputs are "singular" is and only have the levels of the belonging logic-circuit. t is not allowed to connect further outputs parallel (at a totem pole is also only a victim!). At connected push-pull outputs it leads to undefined levels or short cuts, if at two circuits at the same time the complementary transistors are conducting. At a "Totem pole" circuit are the two NPN-transistors T 2 and T 3 in the output driven of a driver-transistor T 1 "Split phase". This transistor T 1 has two operating resistors ( 1 and 2 ) at his collector and his emitter. T 1 generates two signals, which have two phases. At the emitter the signal is in phase and at the collector in opposite phase

14 EL DD QH = ON1 QL = ON2 L L Q SS Q 0 t niversity of PSH-PLL SWTH DST T 2 T 1 E2 E1 D F T 3 E3 2 E3 niversity of "TOTEM POLE" - OTPT DST At the calculation of the voltages at the transistors it is noticed, that the transistor T 2 doesn't block safely, if the transistor T 1 is conducting. t is valid E2 = EX1 + EX3 EX3 t follows with EX1 = EX3 E2 = EX3-40 -

15 EL With insertion of a diode D between T 2 and T 3 one gets E2 = EX3 F The transistor T 2 now is surely blocked. All transistors exept transistor T 2 operate in conducting state in the saturation area. The resistor 3 is a protection at a short-cut. At short-cuts against ground by outside connections or by not allowed parallel switching of TTL-outputs the current into the circuit is limited ( 3 30 Ω until 100 Ω) and the output is protected. ecause of the resistor 3 and especially of the operating-point of T 2 in the linear area the H- level is lower than the supply-voltage (typical for TTL-circuits is EY3 = QH 3,2 V until 3,8 V, the manufacturers guarantee in the worst case QHmin = 2,4 V) Switching times of the transistor The calculation of the switching times of the bipolar transistor is very difficult and will not be made in this lecture. The most important part of the switching time is the result of the operating point in the saturation. Avoiding the saturation decreases the switching times Measures to the decrease of switching times To the decrease of the switching times two different circuits are used. At discrete circuits an acceleration-capacitor ( "Speed up" capacitor) parallel to the base resistor. For integrated circuits this method cannot be used, because the implementing of capacitors needs wide areas. n integrated circuits with bipolar transistors the deep saturation is avoided by the use of Schottky-transistors. t is valid EX = EX + FSD or EX = EX - FSD 0,7 V 0,3 V = 0,4 V This is the lowest value for EX. That means, that the transistor is in the ON-state in the saturation area near the saturation-border. The mounting of a Schottky-diode decreases the switching times about the factor of 10 until 20. The realization of a Schottky-transistor shows figure EE

16 EL S E EX E E niversity of TANSSTO-NVETE WTH SHOTTKY-DODE DST E niversity of THE SHOTTKY-TANSSTO T AND DESGN EE