Chapter 7

Bandgap Reference Circuit

Reference voltages or currents that exhibit little dependence on temperature prove essential in many analog circuits. It is interesting to note that, since most process parameters vary with temperature, if a reference is temperatureindependent, then it is usually process-independent as well.

In order to generate a quantity that remains constant with temperature, we postulate that if two quantities having opposite temperature coefficients (TCs) are added with proper weighting, the result displays a zero TC. For example, for two voltages V1 and V2 that vary in opposite direction with temperature, we choose 1 and 2 such that 1 VREF = 1V1 + 2V2 , with zero TC.
V1 V + 2 2 = 0 , obtaining a reference voltage, T T

7.1 Negative-TC Voltage

The base-emitter voltage of bipolar transistors or, more generally, the forward voltage of a pn-junction diode exhibits a negative TC. For a bipolar device, we can write I C = I S exp(VBE / VT ) , where VT = KT / q . The saturation current I S is proportional to kTni2 , where denotes the mobility of minority carriers and ni is the intrinsic minority carrier concentration of silicon. The temperature dependence of these quantities is represented as 0T m , where m 3 / 2 , and ni2 T 3 exp[ Eg /(kT )] , where E g 1.12eV is the Bandgap energy of silicon. Thus,

E I S = bT 4+ m exp g kT

(7.1)

- 45 -

where b is a proportionality factor. Writing VBE = VT ln ( I C / I S ) , we can now compute the TC of the base-emitter voltage. In taking the derivative of VBE with respect to T, we must know the behaviour of I C as a function of temperature. To simplify the analysis, we assume that I C is held constant. Thus,
VBE VT I C VT I S ln = T T I S I S T

(7.2)

From (7.1), we have

E I S = b(4 + m)T 3+ m exp g T kT

Therefore,

Eg Eg 4+ m + bT exp kT kT 2

(7.3)

Eg VT I S V = (4 + m) T + 2 VT I S T T kT From equations (7.2) and (7.4), we can write E VBE VT I C V = ln (4 + m) T g2 VT T T IS T kT = VBE (4 + m)VT Eg / q T

(7.4)

(7.5)

Equation (7.5) gives the temperature coefficient of the base-emitter voltage at a given temperature T, revealing dependence on the magnitude of VBE itself. With VBE 750mV and T = 300 K ,
VBE 1.5mV / K . T

We note that the temperature coefficient of VBE itself depends on the temperature, creating error in constant reference generation if the positive-TC quantity exhibits a constant temperature coefficient.

- 46 -

7.2 Positive-TC Voltage

If two bipolar transistors operate at unequal current densities, then the difference between their base-emitter voltages is directly proportional to the absolute temperature. As shown in Fig. 7.2.1, if two identical transistors ( I S 1 = I S 2 ) are biased at collector currents of nI0 and I0 and their base currents are negligible, then VBE = VBE1 VBE 2

nI I = VT ln 0 VT ln 0 I S1 IS 2 = VT ln n
Thus, the VBE difference exhibits a positive temperature coefficient, given by:

VBE k = ln n T q
Interestingly, this TC is independent of the temperature or behaviour of the collector currents.

Fig. 7.2.1 Generation of PTAT voltage

- 47 -

7.3 Bandgap Reference

With the negative- and positive- TC voltages obtained above, we can now develop a reference having nominally zero temperature coefficient. To generate a Bandgap reference voltage, we use the circuit shown in fig 7.3.1.

Fig. 7.3.1 Bandgap Reference Voltage Circuit

In our circuit, the transistor pairs M1-M2 and M3-M4 are identical. This makes ID1=ID2 and hence Vx=VY. Using suitable transistor sizes, we obtain a current of approximately 10.5 A in each branch. To be more accurate, the current values obtained after simulation are: ID1 = 10.5 A ID2 = 10.68 A Size of transistor M5 is 3 times that of M3 and M4. Therefore, On simulation, we obtain ID5 3ID2 ID5 = 31.78 A

Since, voltages at X and Y are approximately equal to VX = 749.5 mV and VY = 742.6 mV

- 48 -

We assume that

VX VY = VBE 0

Therefore, Voltage across R1, VR1 = VBE 0 VBE1

= VBE = VT ln n
Here n=10, since m=10 for Q1 and m=1 for Q0. Therefore, current across R1 can be indicated as:
I R1 = VT ln10 R1 3VT ln10 R1

Current through M5 is: Therefore, Voltage output,

ID5 3I R1 =

VREF = VBE 2 + I D 5 R2

V ln10 = VBE 2 + 3 T R2 R1
Differentiating with respect to temperature, we get:
I VREF VBE 2 = + R2 D 5 T T T

(7.6)

Here, we assume that the temperature coefficient of R2 is zero. For zero TC of the Bandgap reference voltage, we need
VREF =0 T

Therefore,

I VBE 2 + R2 D 5 = 0 T T

(7.7)
VBE 2 = 1.27763mV / K T

From simulation plots shown in fig.7.4.2, we obtain Now,

R2

I D 5 3k ln10 R2 = T q R1 R = ( 5.9505 104 ) 2 V/K R1

From equation (7.7), we get

R2 1.27763 103 = = 2.147 R1 5.9505 104 - 49 -

From equation (7.6), we get: or, or, From equation (7.7), we get:

VREF = VBE 2 + I D 5 R2 = 1.2V

0.7179 + 31.78 10 6 R2 = 1.2

R2 15k R1 7k

On selecting these values of R1 and R2, the simulation results showed the following discrepancies: 1. Voltages VX and VY were very different. 2. VREF had a slightly negative temperature coefficient. To solve the above two problems, the value of R1 is reduced and following several iterations we obtain zero TC for VREF at 300K for R1 = 5 K & R2 = 12 K.

The plot in Fig.7.4.3 shows that though the voltage reference obtained is around 1.1112 V, it has nearly zero TC at T = 300K. In the curve shown in Fig. 7.4.3, we notice a finite curvature which may be due to many reasons, some of which are listed below: 1. Temperature variations of base-emitter voltages. 2. Temperature variations of collector currents. 3. Temperature variations of offset voltages.

- 50 -

7.4 Observations and Results

Fig.7.4.1 Schematic for 1.1 V Band-gap Voltage Reference Circuit

- 51 -

Fig.7.4.2 Variation of Vbe with Temperature

Fig.7.4.3 Variation of Vbe with Temperature

- 52 -

Fig.7.4.4 Variation of Output Reference voltage with Temperature

Fig.7.4.5 Magnified look into the variation of output reference voltage with Temperature

- 53 -