The Bipolar (Junction) Transistor

[디바이스마트 바로가기]

Abstract

In the lab, we explore several common transistor circuits; we build a common-emitter amplifier, a pnp current source, an npn emitter follower, and class B and class AB (i.e., biased) push–pull amplifiers. Here, we introduce basic operating guidelines for bipolar transistors and outlines of these circuits.


Contents


1. Bipolar Junction Transistor Model
2. The Ideal Bipolar Junction Transistor
3. Common-Emitter Amplifier
4. pnp Current Sources
5. pnp Current Mirror
6. npn Emitter Follower
7. Simple Class-B Push–Pull Amplifier
8. Biased Class-B Push–Pull Amplifier
A. Bootstrapping for Higher Input Impedance
B. Parts

Note about diode/bandgap conventions: We use the convention that a typical silicon BJT base– emitter diode drop is 0.65V and a standard general purpose silicon diode drop is 0.6V. Other conventions use 0.6V or 0.7V for one or both. Measured laboratory results will most likely be between these two values. Diodes and BJTs implemented on the same integrated circuit (i.e., on the same piece of silicon) may have equivalent characteristics. That is, the diodes and transistors can be matched. Matched components are convenient to use in many circuit designs. We use discrete elements in this laboratory, and so it is not possible to match components unless they’re all implemented within the one part. In the laboratory, a “diode-connected” transistor (i.e., shorted base and collector) may match the base–emitter characteristics of another transistor better than a diode. Diode drops are strongly temperature dependent. Room-temperature transistors have base–emitter drops around 0.65V, but hot transistors have drops near 0.5V because they need less excitation for conduction. So temperature matching is just as important as component matching. Internal temperature compensation in “bandgap voltage references” lets them provide a temperature-independent voltage reference. Their output reference of ∼1.22V is the Silicon diode drop at absolute zero (i.e., 0K or −273.15◦C). It is not a coincidence that the Silicon bandgap (i.e., the energy separating valence and conduction electron bands) is ∼1.22eV. Temperature dependence and manufacturing variations (and the “Early effect”) are always a concern.


1. Bipolar Junction Transistor Model

A bipolar junction transistor (BJT) can be in three modes:
cutoff mode: Transistor acts like an open switch between collector and emitter (i.e., collector–emitter “resistance” is infinite).
active mode: Transistor acts like a dynamic resistor between collector and emitter that adjusts its resistance in order to keep collector current at a set level (i.e., collector–emitter resistance is finite and positive).
saturation mode: Transistor acts like a closed switch between collector and emitter (i.e., collector–emitter “resistance” is very low).
In the active mode, the transistor adjusts the collector current to be a version of the base current amplified by some constant β > 0. If the base current falls to 0, the transistor enters cutoff mode and shuts off. When the base current rises too far, the transistor loses its ability to decrease the collector–emitter resistance to linearly increase the collector current. In this case, the transistor enters saturation mode. To keep the transistor out of saturation mode, the collector and emitter should be separated by at least 0.2V. A simple model for the operation of a transistor in active mode is shown in Figure 1.1 It requires knowing the current gain β in order to design the circuit. In both of these models,

iC = βiB and iE = (β + 1)iB,

and the emitter is separated from the base by a diode. In order for this diode to conduct current, it must be forward biased with 0.65V.

Simple models of active mode bipolar junction transistors (BJTs)
Figure 1.1: Simple models of active mode bipolar junction transistors (BJTs).



2. The Ideal Bipolar Junction Transistor

Because the current gain β is typically unknown or varies greatly with temperature, time, collector–emitter potential, and other factors, good designs should not depend on it. In this laboratory, we assume that β is sufficiently large (i.e., β ≫ 1, where β ≈ 100 in our laboratory) so that

iB ≈ 0 and iC ≈ iE.

These simple rules are similar to the rules we use with operational amplifiers. The analysis approach usually follows these steps:

1. Calculate the transistor base potential vB by assuming that no current enters the base (i.e., iB ≈ 0).
2. Calculate the potential vE at the emitter of the transistor using vB. For an npn transistor,

vE = vB − 0.65V,

and for a pnp transistor,

vE = vB + 0.65V.

3. Calculate the emitter current iE using the emitter voltage vE and the rest of the circuit.
4. Assume that iC ≈ iE and analyze the rest of the circuit.

• Because we know vE, we usually know iE as well. So our iE dictates what iC should be.

However, keep these notes in mind.

• For an npn transistor, active mode requires vC − vE > 0.2V. For a pnp transistor, active mode requires vE − vC > 0.2V. If this condition is violated, the transistor is saturated, and the analysis cannot continue using these simple rules. In design problems, change parameters (e.g., resistors, supply rails, etc.) to prevent saturation.
• Sometimes it’s easier to find vE first and use it to calculate vB.
• How “small” iB must be to neglect its effect depends on the circuit. In particular, iB × RB must be very small, where RB is the the Th´evenin equivalent resistance looking out of the transistor base.

The Base-Emitter Diode: Always keep Figure 1.1 in mind. The Ebers-Moll model of a BJT treats the current-voltage relationship of the base–emitter junction just like a Shockley ideal diode whos current is mirrored by the collector with gain β. When vB and vE are not obvious, remember the base–emitter diode.

Saturation Mode and Compliance

These simple rules are only valid when the transistor is in active mode. In this mode, the transistor is able to dynamically adjust its collector–emitter resistance from near 0Ω (e.g., a closed switch with some small finite resistance) to ∞ (e.g., an open switch). These simple rules fail when they imply that the collector–emitter resistance is negative (i.e., when they imply that the transistor is acting like a battery).

• When the collector–emitter potential in an npn transistor drops to below 0.2V, the transistor saturates and leaves active mode.
• When the emitter–collector potential in a pnp transistor drops to below 0.2V, the transistor saturates and leaves active mode.

In nearly every design in this laboratory, we will keep the transistors in active mode. The only exception to this is when we use the transistor as a switch. For the switch case, the transistor base current is made sufficiently high to drive it into saturation, which brings its collector–emitter potential as close to 0 as possible.

Compliance: A circuit’s compliance refers to the output voltages possible that keep its active components (i.e., its transistors) all in their active modes. If the output voltage is outside of the compliance range, the device will not operate correctly (e.g., a current source will stop providing constant current).

3. Common-Emitter Amplifier

An npn common-emitter amplifier (with emitter degeneration) is shown in Figure 3.1 for some time t ≥ 0.
Single-ended npn common-emitter amplifier (a level-shifter amplifier).
Figure 3.1: Single-ended npn common-emitter amplifier (a level-shifter amplifier).

From the boxed expressions in Figure 3.1, it should be clear that at any time t,

function 3.1

Pay attention to the inequalities; they keep the transistor out of cutoff and saturation modes. So

function 3.2

For maximal bandwidth and minimal current draw, it’s best to make R1 and R2 as large as possible. However, to keep the circuit as insensitive to variations in β as possible, make sure that

function 3.3

That is, keep the parallel combination R1kR2 less than the approximate input impedance of the BJT. Otherwise, even a small base current will cause a significant drop across R1kR2. The capacitor C is needed to AC couple the signal to the base of the transistor. In particular, C, R1, and R2 form a high-pass filter with single time constant (R1kR2)C, and so C ≫ 0. If f is the lowest frequency of interest in the input (i.e., all signals reach the amplifier with at least half power above this frequency), then

function 3.4

At signal frequencies, the circuit’s input impedance may be very low. If the circuit loads the input source (i.e., input signal amplitude drops when connected to your circuit) and you cannot increase R1 and R2 any more, then try the configuration discussed in Appendix A. The output impedance of this circuit can be prohibitively high. Its use often necessitates also using a buffer like the one in section 6. To prevent attenuation or low-pass effects, keep RC low (e.g., RC ≪ 300kΩ).

4. pnp Current Sources

For Figure 4.1, pick vB and RE for current iout. Device in compliance when function 4.1
Resistor-biased pnp current source—output iout, vcomp compliance, load L.
Figure 4.1: Resistor-biased pnp current source—output iout, vcomp compliance, load L.

For Figure 4.2, pick vB for current iout. Device is in compliance when function 4.2
Diode-biased pnp current source—output iout, 10.65V compliance, load L.
Figure 4.2: Diode-biased pnp current source—output iout, 10.65V compliance, load L.

Notice the simplicity of diode bias in Figure 4.2 For greater compliance, use two diodes instead of three.

5. pnp Current Mirror

Instead of using one of the pnp current sources in section 4, you may want to try a pnp current mirror. See below for caveats. As shown in Figure 5.1, a current mirror uses two transistors and a resistor. The transistor on the left sets up a constant current, and the transistor on the right mirrors that current.
pnp current mirror—output iout, 11.8V compliance, load L.
Figure 5.1: pnp current mirror—output iout, 11.8V compliance, load L.

The pnp current mirror is relatively simpler than other current sources and has high compliance.

1. The left pnp transistor is “diode-connected,” which means that its collector and base are shorted together. Because the base is 0.65V below the emitter, the collector is held at 0.65V below the emitter as well. Therefore, the transistor looks like a diode.
2. Potential acrossresistor RC is 11.35V, and so the collector current on the left transistor is (11.35V)/RC.
3. The left and right pnp transistors share the same base and emitter nodes. If we assume that they are perfectly matched transistors, they should have the same collector current. Therefore,

function 5.1

4. In order to keep vCE,right > 0.2V, vout < 11.8V. Therefore, the device has a compliance of 11.8V.

Additional transistors can be added in the same way as the right transistor in order to have multiple outputs all with the same current.

Other current mirrors are (more) immune to these problems, but they usually require more complex configurations and/or tightly matched BJTs. IF YOU USE THIS CURRENT MIRROR IN THE LAB, BE SURE TO MINIMIZE THE TIME THE CIRCUIT IS POWERED ON.

6. npn Emitter Follower

The emitter follower in Figure 6.1(a) produces an output that follows the input, less one diode drop. However, it can source current to the output without loading the input. That is, it provides current/power gain.
npn emitter follower with load L
Figure 6.1: npn emitter follower with load L and function 6.1

Emitter resistance: When output vout is not clipped, the input current required is

function 6.2

where β ≥ 100 for small-signal BJTs. That is, small-signal input impedance is approximately β(REkL). A low RE increases transistor power dissipation and input current required, but it is required for clipping immunity, as shown below. High-β transistors reduce input current demand even for low RE.

Clipping on negative swing: Consider the simplified circuit in Figure 6.1(b), which replaces the transistor with a “dynamic variable resistor” that “pulls up” the output toward 6V. Increasing the size of this resistor moves the output lower. The lowest output occurs at transistor cutoff, as the pull-up resistor is infinite (i.e., an open circuit). Hence, the lowest output is (−6V) × L/(RE + L). A small RE reduces this distortion but also increases power dissipation. Not surprisingly, there is a power–distortion tradeoff.

Unable to “sink” current: The output cannot swing lower than (−6V) × L/(RE + L) because ground current from the load can only travel through the RE resistor. The npn emitter diode cannot sink current, so it all must go through RE. If RE were replaced with another “dynamic resistor,” more negative output swings could be produced. In other words, RE should be replaced with another transistor, as in section 7.

Negative swing caveat: Pulling the input signal down very far will cause the transistor’s base–emitter diode to breakdown. Care should be taken to prevent this from happening. Prevent vin from swinging more than 6V lower than the emitter. For protection, add a reverse-biased diode from the base to the emitter.


7. Simple Class-B Push–Pull Amplifier

In section 6, the npn emitter follower cannot track some negative signals because it lacks the ability to sink current away from its emitter resistor. If the emitter resistor is replaced with a “dynamic resistor” able to alter the current through it, then the output will be able to swing over a wider range of negative values. The circuit in Figure 7.1(a) is a “class-B push–pull amplifier.” It is a combination of an npn emitter follower and a pnp emitter follower. The pnp transistor is the “dynamic resistor” we needed in section 6. Thus, the push–pull amplifier is a kind of “two-tailed follower.” Compare Figure 7.1(b) to Figure 6.1(b).
Class-B Push–Pull Amplifier with load L
Figure 7.1: Class-B Push–Pull Amplifier with load L function 7.1

Pushing, Pulling, Sourcing, and Sinking: It is a “push–pull” amplifier because the upper npn transistor “pushes” (i.e., sources) current for positive input and the lower pnp transistor “pulls” (i.e., sinks) current for negative input. It is a “class-B” amplifier because only one transistor is active at a time. That is, for sufficiently positive signals, the upper transistor will be active and the lower transistor will be in cutoff mode (and vice versa). In Figure 7.1(b), this case corresponds to the PNP resistor being ∞ and the NPN resistor being some finite resistance that puts the desired vout between 0V (from ground) and 6V. For the class-B amplifier, only one “dynamic resistor” has a finite resistance at a time.

Distortion and Power: The class-B amplifier does not “waste” power because transistors only dissipate power (i.e., activate) when absolutely necessary to produce output. This property is attractive in low-power applications, but it comes with a power–distortion tradeoff. For signals between −0.65V and 0.65V, both “dynamic resistors” are infinite (i.e., “off”), and so the output is 0V. Larger signals that swing through this “dead zone” suffer significant distortion. We show in section 8 that some of this crossover distortion can be removed at the cost of additional power dissipation; we use extra biasing power so that the “dynamic resistors” turn “on” more easily.

Practical Applications: Circuits like this one (and the one described in section 8) are often used as output stages of buffering devices, like operational amplifiers. As the combination of two emitter followers, it provides very high power gain and is still be able to handle positive and negative signals (i.e., both sourcing and sinking load current). Because of the distortion it adds, this simple push–pull stage will only be used in low-power applications that sacrifice distortion for reduced power consumption.


8. Biased Class-B Push–Pull Amplifier

The circuit inFigure 8.1 is called a class-B biased push–pull amplifier. It is nearly identical to Figure 7.1, but a new biasing circuit keeps the transistors on for “longer” to reduce crossover distortion.
Class-B Biased Push–Pull Amplifier with load L and |vin| < ∼5.4V.
Figure 8.1: Class-B Biased Push–Pull Amplifier with load L and |vin| < ∼5.4V.>

Biasing diodes: The diodes move some of the “dead zone” from the middle of the input range to the ends, which reduces distortion while decreasing maximum input swing. Ideally, the diodes would have the same potential drop as the base–emitter diode in each of the transistors, which is usually true for a prefabricated transistor/diode array (i.e., inside an integrated circuit). In our case, our discrete components are not matched, so the output still has crossover distortion (e.g., consider signals between −0.05V and 0.05V). Additionally, as vin swings and ibase varies, vD changes slightly, so there is a new source of distortion.

Class-B push–pull amplifier: The amplifier is still a class-B amplifier because only one transistor is active at a time. However, the two diodes make the transistors more sensitive to small signals, and so they activate more readily than in the case in section 7. The “dead zone” that causes output distortion is smaller.

Practical Applications: Circuits like this one (and the one described in section 8) are often used as output stages of buffering devices, like operational amplifiers. As the combination of two emitter followers, it provides very high power gain and is still be able to handle positive and negative signals (i.e., both sourcing and sinking load current). Because of the distortion it adds, this simple push–pull stage will only be used in low-power applications that sacrifice distortion for reduced power consumption.

Class-AB push–pull amplifier: The amplifier bridges on being a “class-AB amplifier.” If additional biasing diodes and emitter resistance are added, both transistors will be active simultaneously for all small signals, which is like a class-A amplifier (i.e., both transistors are active even though only one is “needed”). However, if these signals swing out of a small region (determined by load and how many diodes are added), the amplifier behaves like a class-B amplifier (i.e., only one transistor is active). The class-A operation “wastes” power for less distortion, and the class-B operation introduces distortion to save power.

Power and distortion: The circuit “wastes” power in its biasing diodes in order to reduce output distortion. Increasing the biasing current will further reduce output distortion as the vD drop will depend less on the load. There is always a tradeoff between distortion and power. Clean amplifiers use extra power for low distortion of small signals; the extra power is not simply for production of large (e.g., high volume) signals.


A. Bootstrapping for Higher Input Impedance

Depending on your component choices and signal source, the circuit in Figure 3.1 may load the source so that the input signal is noticeably attenuated when connected to the circuit. That is, at signal frequencies, the input impedance of the circuit may be low compared to the output impedance of the signal source, and so dissipation in the signal source causes attenuation of the signal entering the circuit. To ensure the current into the base of the transistor is negligible, the biasing network must have a relatively low equivalent resistance at DC when looking out of the base. However, there is a clever method we can use to raise the impedance of the network at signal frequencies when looking out of the capacitor. By bootstrapping some of the transistor’s output signal back into the input, we can make the input impedance (at signal frequencies) very large (i.e., approximately βRE, the input impedance of the transistor). Consider the modified circuit in Figure A.1.

Bootstrapped single-ended npn common-emitter amplifier.
Figure A.1: Bootstrapped single-ended npn common-emitter amplifier.>

Instead of Equation (3.3) and Equation (3.4), assume that

function A.1

Otherwise, components can be chosen exactly as before. The bootstrapping capacitor CB must be very large so that it looks like a short circuit to signal frequencies. Theoretically, the resistor RB can be chosen arbitrarily. As long as Equation (A.1) can be met, a high choice of RB (e.g., RB > 1kΩ) is a good idea. The signal at the transistor’s emitter follows the signal at its base. At signal frequencies, CB acts like a short circuit, and so both ends of RB see the same potential. Hence, RB carries no current at signal frequencies. Thus, the R1–R2 divider cannot load the input source because no current from the source makes its way across RB (i.e., RB ≈ ∞ at signal frequencies). The current through R1–R2 that would normally come from the source comes from the output instead. This method is called bootstrapping because we use the circuit’s own output to reduce current required from the input.


B. Parts

Part pin-outs.
Figure B.1: Part pin-outs.>