c Copyright 2010. W. Marshall Leach, Jr.

, Professor, Georgia Institute of Technology, School of

Electrical and Computer Engineering.
The BJT Dierential Amplier
Basic Circuit
Figure 1 shows the circuit diagram of a dierential amplier. The tail supply is modeled as a current
source I
Q
. The object is to solve for the small-signal output voltages and output resistances. It
will be assumed that the transistors are identical.
Figure 1: Circuit diagram of the dierential amplier.
DC Solution
Zero both base inputs. For identical transistors, the current I
Q
divides equally between the two
emitters.
(a) The dc currents are given by
I
E1
= I
E2
=
I
Q
2
I
B1
= I
B2
=
I
Q
2 (1 +)
I
C1
= I
C2
=
I
Q
2
(b) Verify that V
CB
> 0 for the active mode.
V
CB
= V
C
V
B
=

V
+
I
E
R
C

I
E
1 +
R
B

= V
+
I
E
R
C
+
I
E
1 +
R
B
(e) Calculate the collector-emitter voltage.
V
CE
= V
C
V
E
= V
C
(V
B
V
BE
) = V
CB
+V
BE
1
Small-Signal AC Solution using the Emitter Equivalent Circuit
This solution uses the r
0
approximations and assumes that the base spreading resistance r
x
is not
zero.
(a) Calculate g
m
, r

, r
e
, and r
ie
.
g
m
=
I
C
V
T
r

=
V
T
I
B
r
e
=
V
T
I
E
r
ie
=
R
B
+r
x
+r

1 +
r
0
=
V
A
+V
CE
I
E
(b) Redraw the circuit with V
+
= V

= 0. Replace the two BJTs with the emitter equivalent

circuit. The emitter part of the circuit obtained is shown in Fig. 2(a).
Figure 2: (a) Emitter equivalent circuit for i
e1
and i
e2
. (b) Collector equivalent circuits.
(c) Using Ohms Law, solve for i
e1
and i
e2
.
i
e1
=
v
i1
v
i2
2 (r
ie
+R
E
)
i
e2
= i
e1
(d) The circuit for v
o1
, v
o2
, and r
out
is shown in Fig. 2(b).
v
o1
= i
c1(sc)
r
ic
kR
C
= i
e1
r
ic
kR
C
=
r
ic
kR
C
2 (r
ie
+R
E
)
(v
i1
v
i2
)
v
o2
= i
c2(sc)
r
ic
kR
C
= i
ie2
r
ic
kR
C
=
r
ic
kR
C
2 (r
ie
+R
E
)
(v
i2
v
i1
)
r
out1
= r
out2
= r
ic
kR
C
r
ic
= r
0

1 +
(2R
E
+r
ie
)
R
B
+r

+r
x
+ 2R
E
+r
ie

+ (R
B
+r
x
+r

) k (2R
E
+r
ie
)
(e) The resistance seen looking into either input with the other input zeroed is
r
in
= R
B
+r
x
+r

+ (1 +) (2R
E
+r
ie
)
The dierential input resistance r
ind
is the resistance between the two inputs for dierential input
signals. For an ideal current source tail supply, this is the same as the input resistance r
in
above.
Di Amp with Non-Perfect Tail Supply
Fig. 3 shows the circuit diagram of a dierential amplier. The tail supply is modeled as a current
source I
0
Q
having a parallel resistance R
Q
. In the case of an ideal current source, R
Q
is an open
circuit. Often a di amp is designed with a resistive tail supply. In this case, I
0
Q
= 0. The solutions
below are valid for each of these connections. The object is to solve for the small-signal output
voltages and output resistances.
2
Figure 3: BJT Dierential amplier.
DC Solutions
This solution assumes that I
0
Q
is known. If I
Q
is known, the solutions are the same as above.
(a) Zero both inputs. Divide the tail supply into two equal parallel current sources having a
current I
0
Q
/2 in parallel with a resistor 2R
Q
. The circuit obtained for Q
1
is shown on the left in
Fig. 4. The circuit for Q
2
is identical. Now make a Thvenin equivalent as shown in on the right
in Fig. 4. This is the basic bias circuit.
(b) Make an educated guess for V
BE
. Write the loop equation between the ground node to
the left of R
B
and V

. To solve for I
E
, this equation is
0

I
0
Q
R
Q

=
I
E
1 +
R
B
+V
BE
+I
E
(R
E
+ 2R
Q
)
(c) Solve the loop equation for the currents.
I
E
=
I
C

= (1 +) I
B
=
V

+I
0
Q
R
Q
V
BE
R
B
/ (1 +) +R
E
+ 2R
Q
(d) Verify that V
CB
> 0 for the active mode.
V
CB
= V
C
V
B
=

V
+
I
E
R
C

I
E
1 +
R
B

= V
+
I
E
R
C
+
I
E
1 +
R
B
(e) Calculate the collector-emitter voltage.
V
CE
= V
C
V
E
= V
C
(V
B
V
BE
) = V
CB
+V
BE
(f) If R
Q
= , it follows that I
E1
= I
E2
= I
0
Q
/2. If the current source is replaced with a resistor
R
Q
only, the currents are given by
I
E
=
I
C

= (1 +) I
B
=
V

V
BE
R
B
/ (1 +) +R
E
+ 2R
Q
3
Figure 4: DC bias circuits for Q
1
.
Small-Signal or AC Solutions
This solutions use the r
0
approximations.
(a) Calculate g
m
, r

, and r
ie
.
g
m
=
I
E
V
T
r

=
(1 +) V
T
I
E
r
ie
=
R
B
+r
x
+r

1 +
r
0
=
V
A
+V
CE
I
E
(b) Redraw the circuit with V
+
= V

= 0 and I
0
Q
= 0. Replace the two BJTs with the emitter
equivalent circuit. The emitter part of the circuit obtained is shown in 5(a).
Figure 5: (a) Emitter equivalent circuit. (b) Collector equivalent circuits.
(c) Using superposition, Ohms Law, and current division, solve for i
e1
and i
e2
.
i
e1
=
v
i1
r
ie
+R
E
+R
Q
k (r
ie
+R
E
)

v
i2
r
ie
+R
E
+R
Q
k (r
ie
+R
E
)

R
Q
R
Q
+r
ie
+R
E
4
i
e2
=
v
i2
r
ie
+R
E
+R
Q
k (r
ie
+R
E
)

v
i1
r
ie
+R
E
+R
Q
k (r
ie
+R
E
)

R
Q
R
Q
+r
ie
+R
E
For R
Q
= , these become
i
e1
=
v
i1
v
i2
2 (r
ie
+R
E
)
i
e2
=
v
i2
v
i1
2 (r
ie
+R
E
)
(d) The circuit for v
o1
, v
o2
, r
out1
, and r
out2
is shown in Fig. 6.
Figure 6: Circuits for calculating v
o1
, v
o2
, r
out1
, and r
out2
.
v
o1
= i
0
c1
r
ic
kR
C
= i
e1
r
ic
kR
C
=
r
ic
kR
C
r
ie
+R
E
+R
Q
k (r
ie
+R
E
)

v
i1
v
i2
R
Q
R
Q
+r
ie
+R
E

v
o2
= i
0
c2
r
ic
kR
C
= i
e1
r
ic
kR
C
=
r
ic
kR
C
r
ie
+R
E
+R
Q
k (r
ie
+R
E
)

v
i2
v
i1
R
Q
R
Q
+r
ie
+R
E

r
out1
= r
out2
= r
ic
kR
C
r
ic
= r
0

1 +
R
te
R
B
+r

+R
te

+ (R
B
+r
x
+r

) kR
te
R
te
= R
E
+R
Q
k (r
ie
+R
E
)
(e) The resistance seen looking into the v
i1
(v
i2
) input with v
i2
= 0 (v
i1
= 0) is
r
ib
= R
B
+r
x
+r

+ (1 +) R
te
(f) Special case for R
Q
= .
v
o1
=
r
ic
kR
C
2 (r
ie
+R
E
)
(v
i1
v
i2
) v
o2
=
r
ic
kR
C
2 (r
ie
+R
E
)
(v
i2
v
i1
)
(g) The equivalent circuit seen looking into the two inputs is shown in Fig. 7. The resistors
labeled r
0

are given by
r
0

= r
x
+r

+ (1 +) R
E
The dierential input resistance r
id
is dened the same way that it is dened for Fig. ??. That
is, it is the resistance seen between the two inputs when v
i1
= v
id
/2 and v
i2
= v
id
/2, where v
id
is
the dierential input voltage. In this case, the small-signal voltage at the upper node of the resistor
(1 +) R
Q
is zero so that no current ows it. It follows that r
id
is given by
r
id
= 2

R
B
+r
0

5
Figure 7: Equivalent circuits for calculating i
b1
and i
b2
.
Dierential and Common-Mode Gains
(a) Dene the common-mode and dierential input voltages as follows:
v
id
= v
i1
v
i2
v
icm
=
v
i1
+v
i2
2
With these denitions, v
i1
and v
i2
can be written
v
i1
= v
icm
+
v
id
2
v
i2
= v
icm

v
id
2
By linearity, it follows that superposition of v
icm
and v
id
can be used to solve for the currents and
voltages.
(b) Redraw the emitter equivalent circuit as shown in Fig. 8.
Figure 8: Emitter equivalent circuit.
(c) For v
i1
= v
id
/2 and v
i2
= v
id
/2, it follows by superposition that v
a
= 0 and
i
e1
=
v
id
/2
r
ie
+R
E
i
e2
=
v
id
/2
r
ie
+R
E
v
o1
= i
e1
r
ic
kR
C
=
r
ic
kR
C
r
ie
+R
E
v
id
2
=
r
ic
kR
C
r
ie
+R
E

v
i1
v
i2
2

v
o2
= i
e2
r
ic
kR
C
=
+ r
ic
kR
C
r
ie
+R
E
v
id
2
=
+ r
ic
kR
C
r
ie
+R
E

v
i1
v
i2
2

6
The dierential voltage gain is given by
A
d
=
v
o1
v
id
=
v
o2
v
id
=
1
2
r
ic
kR
C
r
ie
+R
E
(d) For v
i1
= v
i2
= v
icm
, it follows by superposition that i
a
= 0 and
i
e1
=
v
icm
r
ie
+R
E
+ 2R
Q
i
e2
=
v
icm
r
ie
+R
E
+ 2R
Q
v
o1
= i
e1
r
ic
kR
C
=
r
ic
kR
C
r
ie
+R
E
+ 2R
Q
v
icm
=
r
ic
kR
C
r
ie
+R
E
+ 2R
Q

v
i1
+v
i2
2

v
o2
= i
e2
r
ic
kR
C
=
r
ic
kR
C
r
ie
+R
E
+ 2R
Q
v
icm
=
r
ic
kR
C
r
ie
+R
E
+ 2R
Q

v
i1
+v
i2
2

The common-mode voltage gain is given by

A
cm
=
v
o1
v
icm
=
v
o2
v
icm
=
r
ic
kR
C
r
ie
+R
E
+ 2R
Q
(e) If the output is taken from the collector of Q
1
or Q
2
, the common-mode rejection ratio is
given by
CMRR =

v
o1
/v
id
v
o1
/v
icm

v
o2
/v
id
v
o2
/v
icm

=
1
2
r
ie
+R
E
+ 2R
Q
r
ie
+R
E
=
1
2
+
R
Q
r
ie
+R
E
This can be expressed in dB.
CMRR
dB
= 20 log

1
2
+
R
Q
r
ie
+R
E

Example 1 For I
0
Q
= 2 mA, R
Q
= 50 k, R
B
= 1 k, R
E
= 100 , R
C
= 10 k, V
+
= 20 V,
V

= 20 V, V
T
= 0.025 V, r
x
= 20 , = 99, V
BE
= 0.65 V, and V
A
= 50 V, calculate v
o1
, v
o2
,
v
od
, r
out
, and CMRR.
Solution.
I
E
=
0

I
0
Q
R
Q

V
BE
R
B
/ (1 +) +R
E
+ 2R
Q
= 1.192 mA
V
CB
= V
C
V
B
=

V
+
I
E
R
C

I
E
1 +
R
B

= 8.209 V
g
m
=
I
E
V
T
= 0.0472 S r

=
(1 +) V
T
I
E
= 2.097 k
r
e
=
V
T
I
E
= 20.97 r
ie
=
R
B
+r
x
1 +
+r
e
= 31.17
r
0
=
V
A
+V
CE
I
C
= 49.869 k R
te
= R
E
+R
Q
k (r
ie
+R
E
) = 230.83
r
ic
= r
0

1 +
(2R
E
+r
ie
)
R
B
+r

+ 2R
E
+r
ie

+ (R
B
+r

) k (2R
E
+r
ie
) = 390.5 k
v
o1
=
r
ic
kR
C
r
ie
+R
E
+R
Q
k (r
ie
+R
E
)

v
i1
v
i2
R
Q
R
Q
+r
ie
+R
E

= 36.84v
i1
+ 36.75v
i2
v
o2
= 36.84v
i2
+ 36.75v
i1
7
r
out
= r
ic
kR
C
= 9.75 k
A
vd
=
1
2
r
ic
kR
C
r
ie
+R
E
= 36.80
A
vcm
=
r
ic
kR
C
r
ie
+R
E
+ 2R
Q
= 0.0964
CMRR
dB
= 20 log

A
vd
A
vcm

= 51.63 dB
The Di Amp with an Active Load
Figure 9 shows a BJT di amp with an active load formed by a current mirror with base current
compensation. Similar circuits are commonly seen as the input stages of operational ampliers and
audio ampliers. The object is to solve for the open-circuit output voltage v
oc
, the short-circuit
output current i
sc
, and the output resistance r
out
. By Thvenins theorem, these are related by the
equation v
oc
= i
sc
r
out
. It will be assumed that the current mirror consisting of transistors Q
3
Q
5
is perfect so that its output current is equal to its input current, i.e. i
c4
= i
c1
. In addition, the r
0
approximations will be used in solving for the currents. That is, the Early eect will be neglected
except in solving for r
out
. For the bias solution, it will be assumed that the tail bias current I
Q
splits equally between Q
1
and Q
2
so that I
E1
= I
E2
= I
Q
/2.
Figure 9: Di amp with active current-mirror load.
Because the tail supply is assumed to be a current source, the common-mode gain of the circuit is
zero when the r
0
approximations are used. In this case, it can be assumed that the two input signals
are pure dierential signals that can be written v
i1
= v
id
/2 and v
i2
= v
id
/2. For dierential input
signals, it follows by symmetry that the signal voltage is zero at the node above the tail current
8
supply I
Q
. Following the analysis above, the small-signal collector currents in Q
1
and Q
2
are given
by
i
c1(sc)
=

r
ie
+R
E
v
id
2
i
c2(sc)
=

r
ie
+R
E
v
id
2
where
r
ie
=
R
B
+r

1 +
The short-circuit output current is given by
i
sc
= i
c4
i
c2
With i
c4
= i
c3
= i
c1
and i
c2
= i
c1
, this becomes
i
sc
= 2i
c1(sc)
=

r
ie
+R
E
v
id
=

r
ie
+R
E
(v
i1
v
i2
)
The output resistance is given by
r
out
= r
04
kr
ic2
where r
ic2
is given by
r
ic2
= r
0

1 +
(2R
E
+r
ie
)
R
B
+r

+ 2R
E
+r
ie

+ (R
B
+r

) k (2R
E
+r
ie
)
By Thvenins theorem, the small-signal open-circuit output voltage is given by
v
oc
= i
sc
r
out
=
(r
04
kr
ic2
)
r
ie
+R
E
(v
i1
v
i2
)
Example 2 For I
Q
= 2 mA, R
B
= 100 , R
E
= 51 , V
+
= 15 V, V

= 15 V, V
T
= 0.025 V,
r
x
= 50 , = 99, = 0.99 V
BE1
= V
BE2
= 0.65 V, V
EB3
= V
EB4
= V
EB5
= 0.65 V, V
C2
=
V
C4
= 13.7 V, and V
A
= 50 V, calculate i
sc
, r
out
, and v
oc
. Because r
x
> 0, we add it to R
B
in the
equations above.
Solution.
r
e1
= r
e2
=
2V
T
I
Q
= 25 r
ie1
= r
ie2
=
R
B
+r
x
1 +
+r
e
= 26.5
R
te2
= 2R
E
+r
ie1
= 128.5 i
sc
=

r
ie
+R
E
(v
i1
v
i2
) = 0.0128 (v
i1
v
i2
)
r
02
=
V
A
+ (V
C2
+V
BE
)
I
Q
/2
= 65 k r
ic2
= r
0

1 +
R
te
R
B
+r

+R
te

+ (R
B
+r

) kR
te
= 362.7 k
r
04
=
V
A
+ (V
+
V
C4
)
I
Q
= 51.82 k r
out
= r
04
kr
ic2
= 45.34 k
v
oc
= i
sc
r
out
= 579.2 (v
i1
v
i2
)
This is a dB gain of 55.3 dB.
9