What is Current Mirror? – Circuit Diagram and its Workings

The two-transistor current source, also called a current mirror, is the basic building block in the design of integrated circuit current sources. Figure 20.25 (a) depicts the basic current-source circuit, which consists of two matched or identical transistors Q₁ and Q₂, operating at the same temperature, with their base terminals and emitter terminals connected together. The base-emitter voltage, V_BE is, therefore, the same in the two transistors. Transistor Q₁ is connected as a diode; consequently, when the supply voltage is applied, the base-emitter junction of transistor Q₁ is forward biased and a reference current I_REF is established.

Although there is a specific relationship between I_REF and V_BE1, V_BE1 can be thought of as the result of I_REF. Once V_BE1 is established, it is applied to the base-emitter junction of transistor Q₂. The applied V_BE2 turns Q₂ on and generates the output current, which is employed to bias a transistor or transistor circuit.

The reference current in the two-transistor current source can be established by connecting a resistor to the supply source +V_CC as shown in Fig. 20.25 (b). The reference current is then

where V_BE is the base-emitter voltage corresponding to collector current (I_REF).

Connecting the base and collector terminals of a BJT effectively provides a two-terminal device with ampere-volt characteristics that are identical to the i_C versus v_BE characteristic of the BJT.

Current Relationships: Figure 20.25 (a) depicts the current in the two-transistor current source. Since V_BE is the same in both devices, and the transistors are identical, then I_B1 = I_B2 and I_C1 = I_C2. Transistor Q₂ is assumed to be biased in the forward-active region. Applying Kirchhoff’s current law to the collector node of transistor Q₁, we have

Replacing I_C1 by I_C2 (i.e., I_C1 = I_C2) and noting that I_B2 = I_C2/β, above equation becomes

i.e., the constant current provided at the collector of transistor Q₂ mirrors that of transistor Q₁

Since

The current I_REF set by V_CC and R₁ is mirrored in the current into the collector of transistor Q₂.

Transistor Q₁ is referred to as a diode-connected transistor because its base and collector are shorted together.

The operation of the current mirror may be described as follows:

1. Q₁ acts like a diode and a current is established through it.
2. Q₁ will develop a voltage drop V_BE, in response to the current.
3. The base-emitter of Q₂ is in parallel with that of Q₁ (V_BE1 = V_BE2 = V_BE).
4. Q₂‘s collector current will be established in response to its V_BE.
5. Since the two transistors are matched, Q₂‘s collector current will be approximately equal to Q₁‘s collector current. Q₂‘s collector current is said to “mirror” the current through Q₁.

Q₁ is often called a compensating diode because it automatically compensates for variations in temperature. Matching of diode curve means matching at all temperatures as well as voltages. When the temperature increases, the voltage across the emitter diode falls approximately 2 mV per degree. Since the voltage across the compensating diode also decreases by 2 mV per degree, the collector current is little affected with temperature increase.

Modified Current Mirror:

Equation (20.48) reveals that the collector current in the transistor Q₂ is related to the reference current I_REF by a factor (β+2/β). For β >> 2, I_C2 = I_REF1, but for low-current transistors I_C2 can differ significantly from I_REF. However, this error can be reduced by introducing another transistor as an emitter follower. The resulting circuit, as shown in Fig. 20.26, is a modified current mirror circuit or three-transistor current source. We again assume that all transistors are identical; therefore, since the V_BE is the same for transistors Q₁ and Q₂, I_B1 = I_B2 and I_C1 = I_C2. Transistor Q₃ supplies the base currents to Q₁ and Q₂, so these base currents should be less dependent on the reference current. Also, since the current in transistor Q₃ is substantially smaller than that in either Q₁ or Q₂, the current gain of Q₃ is expected to be less than those of transistors Q₁ and Q₂. We may define the current gains of transistors Q₁ and Q₂ as β₁ = β₂ = β, and current gain of the transistor Q₃ as β₃.

Applying Kirchhoff’s current law to the collector node of transistor Q₁, we have

Combining above Eqs. (20.49), (20.50) and (20.51), we have

Replacing I_C1 by I_C2 and noting that I_B2 = I_C2/β, the above Eq. (20.52) may be rewritten as

The output or bias current is then

The reference current is given by

As a first approximation base-emitter voltages of transistors Q₁ and Q₃ are assumed to be equal (i.e. V_BE3 = V_BE1) as indicated in Eq. (20.55).

A comparison of Eq. (20.54) for the three-transistor current source and Eq. (20.48) for the two-transistor current source reveals that the approximation of I_out ≡ I_REF is better for three-transistor circuit.

Note : For a two-transistor current source (Fig. 20.25) and for a three-transistor current source (Fig. 20.26), a large valued resistor R₁ is required for a small biasing current, which is not feasible because it becomes costly in terms of chip area. Usually, these current sources are employed for producing a current of about 0.3 mA-0.5 mA.

Multiple Current Source:

In the previous current sources, we established a reference current I_REF and one load current I_out. In the two-transistor current source in Fig. 20.25 (a), the base-emitter junction of the diode-connected transistor Q₁ is forward biased, when the bias voltage +V_CC is applied. Once V_BE is established, the voltage is applied to the base-emitter junction of transistor Q₂, which turns Q₂ on and generates the output current I_out. The base-emitter voltage of transistor Q₁ can also be applied to additional transistors, to generate multiple load currents. Consider the circuit shown in Fig. 20.27. Transistor Q_R, which is the reference transistor, is connected as a diode. The resulting base-emitter voltage of reference transistor Q_R, established by reference current I_REF, is applied to N output transistors, producing N load currents.

The relationship between each load current and the reference current, assuming all transistors are matched, is

The collectors of multiple output transistors can be connected together, changing the load current versus reference current relationship. For an example, the circuit depicted in Fig. 20.28 has three output transistors with common collectors and a load current I_out. Assuming that transistors Q_R, Q₁, Q₂ and Q₃ are all matched if the current gain β is very large, the base currents can be neglected and I₁ = I₂ = I₃ = I_REF, and the output current l_out = 3I_REF. This process is not recommended for discrete devices, since a mismatch between devices will usually cause one device to carry more current than the other devices. Connecting transistors in parallel increases the effective base-emitter area of the device. In actual IC fabrication, the base-emitter area would be doubled or tripled to provide an output current twice or three times the value of I_REF.

Wilson Current Source:

This is another configuration of a three-transistor current source, known as a Wilson current source. The circuit is shown in Fig. 20.29. This circuit has a large output resistance. We again assume that all the transistors are identical, with I_B1 = I_B2 = I_B; I_C1 = I_C2 and V_BE1 = V_BE2.

Applying KCL at the emitter of transistor Q₃, we have

Comparing Eqs. (20.57) and (20.58), we have

Applying KCL at the collector of transistor Q₁

The difference

is extremely small error for modest values of β.

In the Wilson current source, the output resistance looking into the collector of Q₃ is R_out ≡ βr₀₃/2, which is approximately a factor β/2 larger than that of either the two-transistor source or the basic three-transistor source. This means that, in the Wilson current source the change in bias current I_out with a change in output collector voltage is much smaller.

Widlar Current Source:

The basic current mirror shown in Fig. 20.25 (b) has a limitation that for low value current source the resistance R₁ required is sufficiently large and cannot be fabricated economically in IC circuit. For a two-transistor current shown in Fig. 20.25 (b), if a load current of I_out = 10μA is required, then for V_CC = 10 V, the required resistance value is

The solution to above problem is a circuit shown in Fig. 20.30, called a Widlar current source. This circuit is particularly suitable for low value of current. The circuit differs from the basic current mirror [Fig. 20.25 (b)] only in the resistance R_E that is included in the emitter lead of transistor Q₂. A voltage difference is caused across emitter resistor R_E1 so that the V_BE2 is lesser than V_BE1. A smaller V_BE produces a smaller collector current. It means that the load current I_out is lesser than reference current I_REF.

For β >> 1 for transistors Q₁ and Q₂ and for transistors to be identical

Subtracting Eq. (20.67) from Eq. (20.66), we have

From the circuit, we see that

Comparing Eqs. (20.68) and (20.69), we have

Equation (20.70) provides the relationship between the reference and bias currents. Sometimes, emitter resistances are used in both the transistors Q₁ and Q₂ as depicted in Fig. 20.31.

From the circuit shown in Fig. 20.31.

Applying KVL in the base-emitter loop, we have

Comparing Eqs. (20.72) and (20.73), we have

For the range 0.1 < I_C2/I_C1 < 10 it can be assumed that I_C2/I_C1 ≡ R_E1/R_E2. Thus even large ratio is obtained by the modified circuit (Fig. 20.31).