v13n4pg30.pdf

(411 KB) Pobierz

A Comparison of Various Bipolar

Transistor Biasing Circuits

An up-to-date review of bias techniques

By Al Ward and Bryan Ward

Agilent Technologies

he bipolar junction transistor (BJT) is

often used as a low noise amplifier in cel-

lular, PCS and pager applications because

of its low cost. With a minimal number of exter-

nal matching networks, the BJT can produce an

LNA with RF performance considerably better

than an MMIC. Of equal importance is the DC

performance. Although the device’s RF perfor-

mance may be quite closely controlled, the vari-

ation in device DC parameters can be quite sig-

nificant because of normal process variations.

It is not unusual to find a 2 or 3 to 1 ratio in

device

. Variation in

from device to

device will generally not appear as a difference

in RF performance. In other words, two devices

with widely different

can have similar RF

performance, as long as the devices are biased at

the same

and

. The primary purpose of

the bias network is to keep

and

constant

as DC parameters vary from device to device.

The bias circuitry is often overlooked because

of its apparent simplicity. With a poorly

designed fixed bias circuit, the variation in

from lot to lot can have the same maximum to

minimum ratio as the

variation from lot to

lot. If there is no compensation

will double

when

is doubled. It is the task of the DC

bias circuit to maximize the circuit’s tolerance

variations. In addition, transistor para-

meters can vary over temperature causing a

drift in

at temperature. The low power supply

voltages typically available for handheld appli-

cations also make it more difficult to design a

temperature stable bias circuit.

One solution to the biasing dilemma is the

use of active biasing. Active biasing often makes

use of an IC or a PNP transistor and a variety of

resistors, effectively setting

and

regard-

less of variations in device

. Although the

technique of active biasing would be the best

choice for control of device to device variability

and over temperature variations, associated

costs are usually high.

Other biasing options include various forms

of passive biasing. This article discusses various

passive biasing circuits, including their advan-

tages and disadvantages.

Various BJT passive bias circuits

Passive biasing schemes usually consist of

two to five resistors properly arranged about the

transistor. Various passive biasing schemes are

shown in Figure 1. The simplest form of passive

biasing is shown as Circuit #1 in Figure 1. The

collector current

is simply

times the base

current

. The base current is determined by

the value of

. The collector voltage

determined by subtracting the voltage drop

across resistor

from the power supply volt-

age

. As the collector current is varied, the

will change based on the voltage drop across

. Varying

will cause

to vary in a fairly

direct manner. For constant

and constant

will vary in direct proportion to

. For

example, as

is doubled, collector current,

will also double. Bias circuit #1 provides no

compensation for variation in device

Bias circuit #2 provides voltage feedback to

the base current source resistor

. The base

current source is fed from the voltage

, as

opposed to the supply voltage

. The value of

the base bias resistor

is calculated based

upon nominal device

and the desired

Collector resistor

has both

and

flowing

through it.

The operation of this circuit is best explained

30 ·

PPLIED

ICROWAVE

& W

IRELESS

Figure 1(a). Circuit #1: nonsta-

bilized BJT bias network.

Figure 1(b). Circuit #2: voltage

feedback BJT bias network.

Figure 1(c). Circuit #3: voltage feed-

back with current source BJT bias net-

work.

exception that the series current source

resistor

is omitted. This circuit is

seen in bipolar power amplifier design

with resistor

replaced by a series sil-

icon power diode providing temperature

compensation for the bipolar device. The

current flowing through resistor

shared by both resistor

and the base

emitter junction

. The greater the

current through resistor

, the

greater the regulation of the base emit-

ter voltage

Figure 1(d). Circuit #4: voltage

Figure 1(e). Circuit #5: emitter

Bias circuit #5 is the customary text-

feedback with voltage source BJT

feedback BJT bias network.

book circuit for biasing BJTs. A resistor

bias network.

is used in series with the device emitter

lead to provide voltage feedback. This

circuit ultimately provides the best con-

as follows. An increase in

will tend to cause

to trol of h

variations from device to device and over tem-

increase. An increase in

causes the voltage drop perature. The disadvantage of this circuit is that the

across resistor

to increase. The increase in voltage emitter resistor must be properly bypassed for RF. The

across

causes

to decrease. The decrease in

typical bypass capacitor often has internal lead induc-

causes

to decrease because the potential difference tance which can create unwanted regenerative feedback.

across base bias resistor

has decreased. This topolo- The feedback can create device instability. Despite the

gy provides a basic form of negative feedback which problems associated with using the emitter resistor

tends to reduce the amount that the collector current technique, this biasing scheme generally provides the

increases as

is increased.

best control on

and over temperature variations.

Bias circuit #3 has been discussed in past literature

The sections that follow begin with a discussion of the

but predominately when very high

(> 15 volts) and BJT model and its temperature dependent variables.

(> 12 volts) were used [1]. The voltage divider net- From the basic model, various equations are developed

work consisting of

and

provide a voltage divider to predict the device’s behavior over

and tempera-

from which resistor

is connected. Resistor

then ture variations. This article is an update to the original

determines the base current.

times

provides

. article written by Kenneth Richter of Hewlett-Packard

The voltage drop across

is determined by the collec- [2] and Hewlett-Packard Application Note 944-1 [3].

tor current

, the base current

and the current con-

sumed by the voltage divider, consisting of

and

BJT modeling

This circuit provides similar voltage feedback to that of

The BJT is modeled as two current sources, as shown

bias circuit #2.

in Figure 2. The primary current source is

. In par-

Bias circuit #4 is similar to bias circuit #3 with the allel is a secondary current source

CBO

(1+

) that

V cc

32 ·

PPLIED

ICROWAVE

& W

IRELESS

+ I

CBO

(1+h

)

Figure 2. Gummel-Poon model of BJT with voltage feed-

back and constant base current source network.

describes the leakage current flowing through a reverse

biased PN junction.

CBO

is typically 1

–7

A at 25

degrees C for an Agilent Technologies HBFP-0405 tran-

sistor.

is the internal base emitter voltage with

representing the equivalent Hybrid PI input impedance

of the transistor.

is also equal to

/λI

where

40 at +25 degrees C.

will be defined as measured

between the base and emitter leads of the transistor. It

is equivalent to

is approximately 0.78

volts at 25 degrees C for the HBFP-0405 transistor.

The device parameters that exhibit the greatest

change as temperature is varied are

, and

CBO

These temperature dependent variables have character-

istics which are process dependent and fairly well under-

stood.

typically increases with temperature at the

rate of 0.5 percent/degrees C.

has a typical negative

temperature coefficient of –2 mV/degrees C. This indi-

cates that

decreases 2 mV for every degree increase

in temperature.

CBO

typically doubles for every 10

degree C rise in temperature. Each one of these para-

meters contributes to the net resultant change in collec-

tor current as temperature is varied.

For each bias network shown in Figure 1, several sets

of simplified circuit equations have been generated to

allow calculation of the various bias resistors. These are

shown in Figures 3, 4, 5, 6 and 7. Each of the bias resis-

tor values are calculated based on various design para-

meters such as desired

, power supply voltage

and nominal

CBO

and

are assumed to be zero for

the basic calculation of resistor values.

Additional information, usually provided by the

designer, is required for the three circuits that use the

voltage divider consisting of

and

. For the bias

network that uses voltage feedback with current source,

the designer must choose the voltage across

RB2

)

and the bias current through resistor

, which will be

termed

RB2

. If

RB2

, then a suggested

RB2

34 ·

PPLIED

ICROWAVE

& W

IRELESS

Figure 3. Equations for nonstabilized bias network.

−

= h

(

−

)

CBO

)

(

)

Figure 4. Equations for voltage feedback bias network.

would be 1.5 volts and a suggested

RB2

would be 10 per-

cent of

, or 0.5 mA.

The voltage feedback with a voltage source network

and the emitter feedback network also require that the

designer choose

RB2

. The ratio of

RB2

can play a

major role in bias stability.

An equation was then developed for each circuit that

calculates collector current,

, based on nominal bias

resistor values and typical device parameters, including

CBO

, and

. MATHCAD 7 was used to help

develop the

equation. Although the

equation

begins simply, it develops into a rather lengthy equation

−

Designer must choose I

and V

RB2

such that V

RB2

–

(

)

−

[

CBO

)

−

]

(

)

(

)

(

)

CBO

)

Figure 5. Equations for voltage feedback with current source bias network.

for some of the more complicated circuits. MATHCAD

helped to simplify this task.

Design example using the Agilent HBFP-0405 BJT

The HBFP-0405 transistor is used as a test example

for each of the bias circuits. The HBFP-0405 is described

in [4] as a low noise amplifier for 1800 to 1900 MHz

applications. The HBFP-0405 will be biased at a

2.7 volts and a drain current

of 5 mA. A power supply

voltage of 3 volts will be assumed. The nominal

the HBFP-0405 is 80. The minimum is 50 while the max-

imum is 150. The calculated bias resistor values for each

bias circuit are described in Table 1.

With the established resistor values,

is calculated

based on minimum and maximum

. The perfor-

mance of each bias circuit with respect to

variation

is shown in Table 2. Bias circuit #1 clearly has no com-

pensation for varying

, allowing

to increase 85

percent as

is taken to its maximum. Circuit #2 with

very simple collector feedback offers considerable com-

pensation due to

variations allowing an increase of

only 42 percent. Circuit #3 offers very little improve-

ment over circuit #2. Circuit #4 provides considerable

improvement in

control by only allowing a 9 percent

increase in

. Circuit #4 offers an improvement over

the previous circuits by providing a stiffer voltage source

across the base emitter junction. As we will see later,

this circuit has worse performance over temperature as

compared to circuits #2 and #3. However, when both

and temperature are considered, circuit #4 will

appear to be the best performer for a grounded emitter

configuration. As expected, circuit #5 provides the best

control on

with varying

allowing only a 5.4 per-

cent increase in

. Results are power supply dependent,

and with higher

, results may vary significantly.

BJT performance over temperature

Since all three temperature dependent variables

CBO

and

) exist in the

equation, differenti-

ating the

equation with respect to each of the para-

meters provides insight into their effect on

. The par-

tial derivative of each of the three parameters repre-

sents a stability factor. The various stability factors and

their calculation are shown in Table 3. Each circuit has

three distinctly different stability factors which are then

multiplied times a corresponding change in either

, or

CBO

and finally summed. These changes or

deltas in

, and

CBO

are calculated based on

R B2

V BE

I B2

V CE

I B2 . R B2

I B2

V CC

V CE

I B2

R B1

Designer must choose IB2

hFE

C B O RC I C B O

R B2

.V·

RB 2 . h F E

.h .I

ie C B O

R B1

.h .I

ie C B O

C B O RB 1 I

C B O

R B2

hF E

RB 2 F E

R B1

hF E

R B1

R B2

.V·

R B1

RB 2 . hF E

.h .I

ie C B O

R B1

1 . .

.h .I

h ie I C B O h ie . IC B O V C C

ie C B O V· B E

RB2

hFE

R B1

RB 2 . hF E

1 .

h ie

hF E

Figure 6. Equations for voltage feedback with voltage source bias network.

36 ·

PPLIED

ICROWAVE

& W

IRELESS

Pick

to be 10 percent of

, which is equal to .0005 A

V·B E

1 . .

hi e I

h i e. I

hF E

RE .I

1 .

hF E

. V·

R2 . hF E

.h .I

R1 RE

R 2 hF E

hF E

.R .I

hFE

R 1. I

VC C

R2. h F E

hF E

R1 RE R1

R2 hF E R 2

Figure 7. Equations for emitter feedback bias network.

variations in these parameters based on the manufac-

turing processes.

A comparison of each circuit’s stability factors will

certainly provide insight as to which circuit compen-

sates best for each parameter. MATHCAD was again

used to calculate the partial derivatives for each desired

stability factor. The stability factors for bias circuit #1

are shown in Table 4. The stability factors for the

remaining circuits are shown in the Appendix.

The change in collector current from the nominal

design value at 25 degrees C is then

Resistor

Non-

Voltage

Emitter

calculated by taking each stability

Stabilized

Feedback

Feedback Feedback

factor and multiplying it times the

Bias

Bias Network with Current with Voltage

Bias

corresponding change in each para-

Network

Source Bias Source Bias Network

meter. Each product is then summed

Network

to determine the absolute change in

140

Ω

138

Ω

126

Ω

126

Ω

collector current.

30770

Ω

19552

Ω

11539

Ω

As an example, the collector cur-

889

Ω

2169

Ω

2169

Ω

rent of the HBFP-0405 will be ana-

3000

Ω

1560

Ω

2960

Ω

lyzed as temperature is increased

138

Ω

from +25 degrees C to +65 degrees

Table 1. Bias resistor values for HBFP-0405 biased at

= 2 volts,

= 2.7

C. For the HBFP-0405,

CBO

is typi-

volts,

= 5 mA,

= 80 for the various bias networks.

cally 100 nA at +25 degrees C and

typically doubles for every 10

degrees

temperature

rise.

Bias Circuit

Non-

Voltage

Emitter

Therefore,

CBO

will increase from

Stabilized

Feedback

Feedback Feedback

100 nA to 1600 nA at +65 degrees C.

Bias

Bias Network with Current with Voltage

Bias

The difference or

∆

CBO

will be 1600

Network

Source Bias Source Bias Network

– 100 = 1500 nA. The 1500 nA will

Network

then be multiplied times its corre-

(mA)@

3.14

3.63

3.66

4.53

4.70

sponding

CBO

stability factor.

minimum

at 25 degrees C was measured

(mA)@

5.0

at 0.755 volts for the HBFP-0405.

typical

Since

has a typical negative tem-

(mA)@

9.27

7.09

6.98

5.44

5.27

perature coefficient of –2 mV per

maximum

degree C,

will be 0.675 volts at

Percentage

+85%

+42%

+40%

+9%

+5.4%

+65 degrees C. The difference in

change in

–37%

–27%

–9%

–6%

will then be 0.675 – 0.755 = –0.08

from nominal

volts. The –0.08 volts will then be

Table 2. Summary of

variation versus h

for various bias networks for the

multiplied by its corresponding

HBFP-0405,

= 2.7 volts,

= 2 volts,

= 5 mA,

= +25 degrees C.

stability factor.

38 ·

PPLIED

ICROWAVE

& W

IRELESS

Plik z chomika:

Amiga7878

Inne pliki z tego folderu:

A Textbook of Electrical Tech. Vol.1_Basic electrical engineering BL.THERAJA.pdf (54936 KB)
1118880943.pdf (9852 KB)
6.how-why-electronics_text.pdf (17095 KB)
144mhzWidebandNoiseAmplifier_text.pdf (4912 KB)
13cm_PSK_Transceiver_for_1.2Mbit_Packet_Radio.pdf (4488 KB)

v13n4pg30.pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: