场效应管，半导体二极管和放大器（英文版）

field effect transistors ,semiconductor and amplifier
one,field effect transistors
     the field effect transistor of fet,  ①  find its greatest use in integrated circuit, especially in the digital area. the circuitry of a single chip often contains several thousand fets, which are used not only as active devices but also as resistors and capacitors.
    in linear electronics the fet has special properties of particular significance. one of these is the extrermely high input resistance of the insulated-gace field-effect transistor(igfet), more comonly referred to as a mosfet. the junction field-effect transistor(jfet) also has high input resistance, and low noise as well. the difference in the cost is usually not a significant factor in the selecction of the type of a discrete device to be used in a circuit. for most purposes the bjt has superior performance. the transconductance of a device is the incremental change in its output current divided by the corresponding change in the input voltage. the transistor capacitances have a dominant effect on the bandwidth. for the same bandwidth the bjt has a larger transconductance and hence more voltage gain than the fet, which is a major consideration in amplifiers. in addition, the low saturation voltage of the bjt combined with its higher gain makes it generally more suitable for use in power amplifiers. however, power fet are not as susceptible to thermal runaway and can provide less distortion, both of which are desirable in audio amplifiers.
    a simplified structure of a p-channel mosfet is illustrated in fig. a-1.  the metallic gate is electrically isolated from the silicon by a very thin insulator. this insulator is usually an oxide, and the most common choice is silicon dioxide sio2. the letters mos represet metal-oxide-semicoductor,  and the device is often referred to as a mos transistor, ② or simply most.
the sketch of fig. a-2 is not  to scale. the dimension are given in micrometers or microns, for a typical mosfet in a digital integrat dcircuit. as indicated, the channel length is the distance between the doped source and drain regions, and may be substantially greater than or less then 4μm. the channel width w is the width of each p region along the surface in the direction normal to the length 1. it may be more or less than the 50μm shown on the figure. in a power device designed to handle a large current the width would be much greater. the gate-oxide thickness tox is generally between 0. 08 and0. 20μm. both p regions are heavily doped, and the n-substrate is lightly doped, having a resistivity of from 0. 01 to 0.20 ohm-m. consequently, the space-charge layers of the pn junctions wholly within the substrate.
                                                                                               　　　　　　　　　　　
figure a-1　　　　　　　　　　　　　　　　　　　　 figure a-2

    let us suppose that the substrate and gate terminals are connected to the source, along with a battery between d and s that makes vgs =- 6v.
    now suppose vgs is changed from zero to - 10v with  ③ vds remaining at - 6v. this is illustated in fig. a-3. the negative gate attracts positive holes that are present in the n substrate and in the source and drain regions and repels free electrons. in a thin surface region the equilibrium hole density in the n substrate becomes greater than the equilibrium free-electron density, with the result that this surface region has changed from n-type to p-type. as shown in the figure, there is a continuous p region from the source to the drain, and vds causes holes to flow from the source through the channel to the drain. this is a majority-carrier current which flows by the drift process. it crosses no space-charge layers, and there are no diffusion currents. because their operation depends on only a single type of charge carrier, fets are unipolar transistors. in contrast, the bipolar junction transistor requires both hole and electron currents.
  

figure a-3
    the metal gate, the oxide insulator, and the channel form a capacitor. within the insulator of this capacitor is an electric field associated with the gate voltage. the field controls the channel conductance and consequetly the current. with zero gate voltage and vgs negative, there is no conduction but when vgs is sufficiently negative, a channel is formed and current flows. the more negative we make vgs the greater the current. thus the gate voltage enhances the conductance. an fet that conducts appreciably only when a nonzebro voltage is applied to the gate is called an enhancement-mode field-effect transistor.
    it is informative to compare the basic operation of a p-channel device with that of a pnp junction transistor.   ④ in the p-channel mosfet the holes flow from the source through the channel to the drain, with the flow controlled by the gate voltage. in the pnp transistor holes flow from the emitter through the base to the collector, with the flow controlled by the base current. accordingly, there is a functional correspondence between the source and the emitter, the gate and the base, and the drain and the collector.
two,semiconductor
a semiconductor diode is a two-terminal device containing a single p-n junction.the  general circuit symbole for a semiconductor diode is shown in fig. b-1 along with its relationship to the p-n structure.
    generally, diode applications can be classified according to which of the three regions of diode operation (see fig. b-2) are used. switching and rectifying applications involve transitions between the reverse-bias and forward-bias region. in such applications care must be taken to choose a diode with a reverse breakdown voltage sufficiently large to prevent underesired reverse breakdown.  the reverse breakdown region is employed primarily in voltage reference appliciations. there, the diode is choosen for the specific value of reverse voltage at with reverse breakdown occurs.
                  
fig. b-1  circuit symbol                  fig. b-2  typical diode charcteristic
for a semiconductor diode                illustrating three regious of operation

the v-i characteristic of a typical semiconductor is shown in fig.b-2.
    the diode is the first network element we encounter that is strikingly nonlinear in the middle of its normal operating range. both kvl and kcl can be used, since their validity dose not depend on the linearity or nonlinearity of the network elements. however, we must exercise caution in the use of superposition, thevenin equivalents, and norton equivalents, because these methods are explicitly restricted to linear networks.
    one generally useful approach is to separate the linear network from the nonlinear elements and carry out a graphical solution for the voltage current in the nonlinear elements.  to illustrate the method of graphical solution, let us consider the network shown in fig. b-3(a). a single nonlinear element, a diode, is connected to a network of arbitrary complexity, but containing only linear resistive elements and sources. the liner portion of the network may be its thevenin equivalent network as shown in fig. b-3(b).
     the first step in the solution is to separate the linear network from the nonlinear element, as shown in fig. b-4(a), next, we determine the relationship between vd and
                                
fig. b-3(a)                                fig. b-3 (b)  linear network
general nonlinear network                  replaced by thevenin equivalent circuit

id, the v-i characteristic of the nonlinear element, by experimental measurement or from another source such as a manufacturer's data sheet, and plot this relationship as shown in fig. b-4(b). the third step in the solution is to find the relationship between vl and il. the v-i characteristic for the linear network. from fig. b-4(a), we have
                        vl=voc－ilrt
    where voc and rt are the thevenin equivalent voltage and resistance of the linear network. equation (1-1) is then plotted on the graph containing the v-i characteristic of the nonlinear element. since eq.  (1-1) is a linear relationship between vl and il., the equation plots as a straight line, and only two points on the line need be calculated to determine the entire line. two convenient points are the intercepts at vl= 0 and il=0 corresponding respectively to a short circuit and an open circuit at the linear network terminals. for vl=0,eq. (1-1) yields.
il =voc/rt     vl=0
whereas for il=0 we obtain
      vl=voc il=0
     these points, which are the x-axis and y-axis intercepts of the line in question, are indicated in fig. a-4(b) along with the resulting line. the slope of the line is seen to be -1/rt, so that for small value of rt the line approaches the vertical and for large valuse of rt the line becomes horizontal.
     if the nonlinear element is connected to the linear network, as shown in fig. b-3(b), then we have the following circuit constraints imposed by the connection.
vl=vd=v1             il=id=i1
     from fig.b-4(b) there is only one point where vl equals vd and il. equals id: the intersection of the two v-i characteristic curves. thus the required values of vl and i1 must be the valuse of voltage and current at this intersection point, and can he read directly off the graph.
     the linear v-i characteristic plotted in fig. b-4(b) is known as a load line, since it represents the locus of all possible loads the linear network can present to the nonlinear
      
(a)                                  (b)
                               figure b-4
element. also, the intersection v1-i1 is often called the operating point or q-point of the nonlinear element.
three,amplifier
a single transistor amplifier stage can be arranged in any of the three configurations in fig c-1. these are known as common emitter, common collector and common base,respectively.
                                  
(a) common emitter         (b) common collector            (c) common base
figure c-1  basic transistor amplifier configurations
     the most popular form of aplifier circuit is the common emitter. in its simplest form it is arrangedas in fig c-2(a). to keep a transistor conduction with current flow from collector to emitter, a much smaller current has to flow from base to emitter and under these conditions a voltage will exist between emitter and base of approximately 0. 6v.
                        
(a) simple circit      (b) temperature stabilization     (e)practical amplifier
fig c-2
     the ratio of collector current to base current is called the common emiter current gain, and has the symbol hfe (or sometimes the greek letter β). values of hfe vary greatly, even within the same type of transistor, for example, the common audio transistor bg108 can have an hfe in the range 100 to 800.
in fig c-2(a), base current is provided by rb and the base current is given by
ib=(vcc-0.6)/rb
this causes a collector current to flow;
              ic=hfeib
in turn this causes a voltage drop across rl:
                     vl=icrl
     ideally rl and rb are chosen such that vl = 0. 5vcc, allowing an equal positive or negative swing of voltage at the collector.
     if a small a. c. signal is now applied to the base, the base current will change in sympathy, causing a large change in collector current. this, in turn, produces a large voltage change across rl.
     note that a positive increse at the base causes more base current to flow, causing more collector current to flow and the collector voltage to fall. the amplified output is the inverse of the input signal.
     this simple amplifier circuit has many shortcomings. the values of rb and rl. have to be adapted very precisely to the characteristics of the particular transistor. as mentioned before, hef varies widely from transistor to transistor, even of the same nominal type of more importance,  however,  is the sad fact that transistor characteristics are very temperature-dependent. a simple circuit sueh as in fig c-2 (a) would not, in fact, work reliably over a temperature range of more than a few degrees centigrade.
     there are several transistor parameter that are temperature-dependent, but the most important are the current gain hfe and the collector to emitter leakage current.
     the leakage current (denoted iceo) is the current flowing from collector to emitter with the base disconnected. this current is highly temperature-dependent and doubles for each 8℃ rise.
     an improvement can be made with the circuit shown in fig c-2 (b). the base resistor rb is now returned to the collector. suppose we have chosen rb and rl such that the collector is sitting correctly at 0. 5vcc, and changes in temperature cause the leakage current to rise. the change in leakage current will cause more collector current to flow, causing the collector volts to fail. the fall in collector volts will reduce the base current flowing through ru, reducing the collector current and compensating, to some extent, for the change in leakage current.
     the circuit shown in fig c-2 (c) gives almost perfect compensation for changes in transistor characteristics,  resistors rbl and rb2 are a voltage divider defining the base voltage. the emitter voltage is thus defined, since the base emitter voltage is effectively constant at 0. 6v. the emitter curent is given by
                            ie=ve/re=(vb-0.6)/re
in modern transistors, with high values of hfe it is a reasonal approximation to say that
                                   ic=ie
hence the voltage drop across rl is defined.
     in this particular circuit arrangement, variations in hfe only affect the base current being drawn from the voltage divider rbl, rb2. the variation in base current will cause a negligible change in operating conditions if the standing bleed current through, rbl, rb2 is significantly larger than the base current, rbl and rb2 must not be made too small, however, or the input impedance of the stage will be unacceptable low.
     calculation of the gain of a single-stage amplifier can be made very mathematical, with complex models,  for most purposes, however, simple approximations will give adequate accuracy. to define the gain of single stage amplifier we need two parameters.
     the first is hfe. this is similar to the d.c. gain hfe above, with the exception that it is the small signal a.c. gain (or a. c. β), i.e.
hfe=δic/δib
     where δ denotes " small change".  the parameter hfe, is a ratio (and hence is dimensionless) and has a typical range of 50 to 800.
     the second parameter is hfe. this relates the variation of base current to small signal changes in base emitter voltage. it is defined as
hie=δvbe/δib
the parameter hie has the dimensions of resistance, and has a typical value of several hundred ohms.