Principles of Analog Electronics

477

Amplifiers: Variants

Here, we report some variants of the congurations already discussed and

analyzed to improve the performance of some of the ampliers.

In particular, we will deal with the methods used to increase the amplier’s

input resistance value (Section 10.2.2 in Chapter 10), to increase the known low

efciency of the class A topology (Section 10.2.4), to increase the bandwidth

(Section 10.2.6) and, nally, to increase the signal to noise ratio (Section 10.2.9)

with the rejection of unwanted signals that can couple into the input.

13.1 INCREASED INPUT RESISTANCE

We already discussed and evaluated (particularly, but not exclusively, in

Section10.2.2 of Chapter 10) the importance of designing proper values of input

and output resistances (or, in general, impedances) of an arbitrary amplier. These

resistance values can greatly affect the overall gain of an amplier, so the best

matching conditions (Section 6.9 in Chapter 6) have to be assured with respect to

the type of amplier (Section 10.3 in Chapter 10) to be realized. Here, we will deal in

particular with the input resistance of the C.C. and swamped C.E. congurations.

We demonstred for C.C. (Section 11.6.3.2 in Chapter 11) that

R

hhRh

hR

hR

iBJT CC

ie oe EL re fe EL

oe EL

(..)

=

+

()

+−

()

+

()

+

111

1

13

K18911_Book.indb 477 27/12/13 6:32 PM

Principles of Analog Electronics

478

and in output open loop conditions considered when the load is unknown, the

term R

EL

can be replaced by R

E

.

It was demonstrated for swamped C.E. (Section 11.5.2.1) that

R

hhhhhR

hR R

hR

iBJT sCE

ie ie oe re fe CL

oe CE

oe CL

(..)−

=

+−

()

++

()

+

+

1

1L

hhh hh hhR

hR R

ie oe re fe fe re E

oe CE

−

()

++ −









++

()

1

1

1

1L

and the emitter resistance R

E1

can be renamed here for simplicity as R

E

without

losing generality.

As we have already seen (Section 11.5.2.2), in many practical cases the BJT’s

inner feedback effect can be considered zero, that is h

re

≌ 0, and the Early effect

as well, that is h

oe

≌ 0. These conditions are quite often considered valid when

hR

fe E



01

.

which can be true but only admitting a certain approximation.

In these occurrences, the two previous equations become

RR Rh

hR

iBJT CC iBJT sCEiBJTi

efeE

(..) (..)

===++

()

−

1

Therefore to increase the input resistance, proportional to the emitter resis-

tance, we have to increase the value of the latter:

R

i

R

E

→

→



∞

∞

But if we hypothetically admit that R

E

→∞, the hypothesis h

fe

R

E

≪0.1 is no

longer valid, but

R

hh hh h

h

hh

h

1

1

iBJT CC

R

ie oe re fe fe

oe

fe

oe

(..)

E



()

≅

+−+

=

+∆ +

→∞

R

hR hh hh hh

h

hR

iBJT sCE

R

oe Cieoerefefere

oe

oe

E

(..)−

→

≅

+−

()

++ −

=



∞

1

C

Cfer

e

oe

hhh

h

++

+−

∆ 1

where

∆= −hhhhh

ie oe re

fe

.

These two equations give an idea of the upper limits of the input resistances

for the two congurations. Without considering the impracticable hypothesis

R

E

→∞, let’s admit to “simply” having a very high value of the emitter resistance.

In this case, the DC voltage drop across R

E

would be very high so that, to keep

the BJT working within its forward active region, the required DC voltage supply

K18911_Book.indb 478 27/12/13 6:32 PM