Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~ ~ 2 ~ t ~
- 1 - GECAN 3058
SWITC~ED WIDE DYNAMIC RANGE AMPLIFIER
The present invention relates to an amplifier for
use in a receiver of an optical electric circuit. More
specifically, the present invention relates to an
amplifier switchable between transimpedance modes and a
follower mode for improving the dynamic range of the
amplifier.
BACRGRO~ND OF THE INVENTION
Transimpedance amplifiers are commonly used in the
amplification of electrical signals produced by
photodiodes. The bandwidth of the amplifier is inversely
proportional to the input capacitance and the load
resistance of the transducer. It is desirable to reduce
the effective input capacitance as well as the load
resistance to improve the operating bandwidth of the
ampli~ier. However, at low power signals the load
resistance should be as high as possible to match the
photodiode low noise.
U.S. Patent No. 4,764,732 issued August 16, 1988 to
Bruno Dion discloses a switched mode amplifier that
operates in a first mode as a transimpedance amplifier
2 ~ 2 c~ ~ ~ J
- 2 - GECAN 3058
having good sensitivity over broad bandwidth operation.
This patent teaches switching the mode of operation by
adding transducer load impedance to alter the DC bias of
the amplifier. This causes the amplifier to switch from
05 a transimpedance mode to a follower mode of operationO
The amplifier disclosed operates well to amplify low
power signals detected by the photodiode and does not
saturate for high power signals since the amplifier is
switched into a follower mode. In the follower mode the
sensitivity of the amplifier is reduced and the power
level required to saturate the amplifier is increased.
U.S. Patent No. 4,564,818 issued January 14, 1986
to Timothy R. Jones discloses a transimpedance amplifier
having a transimpedance feedback path that includes a
first and second resistor. This patent discloses a
capacitor having one terminal connected between the first
and second resistors, and having its other terminal
connected to a further resistor which in turn is
connected to ground. The patent teaches that the purpose
of the capacitor and further resistor is to compensate
for parasitic capacitances across the first and second
feedback paths that would otherwise effect the gain
bandwidth product.
U.S. Patent No. 4,574,249 issued March 4, 1986 to
Gareth F. Williams discloses several circuits adapted to
control the dynamic range of an amplifier. In
particular, Figure 39 illustrates using an FET as a
variable shunt where the FET is controlled by an
automatic gain control. A feedback current source is
also employed.
S~MMARY OF THE INVEN~ION
It is therefor an object of the present invention
to provide an amplifier suitable for use with a
photodiode or transducer wherein the dynamic range of the
amplifier is improved by switching the amplifier between
s~ F~
- 3 - GECAN 3058
a first transimpedance mode, a second transimpedance mode
and a follower mode of operation.
In accordance with a broad aspect of the present
invention there is provided an amplifier circuit,
switchable between transimpedance modes and a follower
mode of operation, for changing the dynamic range of the
amplifier. The amplifier circuit includes a transducer,
an amplifier stage, and first and second feedback paths.
The transducer is adapted to convert wave energy signals
to electrical signals. The amplifier stage has an input
connected to the transducer to amplify the electrical
signals and an output. The first, resistive, feedback
path is coupled between the output and the input of the
amplifier. The second feedback path is responsive to the
transimpedance gain of the electrical signals at the
output of the amplifier stage and is selectively coupled
and uncoupled across the input and output of the
amplifier stage. The second feedback path has a
resistance less than the resistance of the first feedback
path and has a capacitance coupled to ground from at
least a portion of the resistance to determine a mode
cross-over frequency. When the second feedback path is
uncoupled, the amplifier operates in a first
transimpedance mode where the transducer load impedance
is that of the first resistance feedback path. When the
second feedback path is coupled in circuit, the
transducer load impedance becomes that of the second
feedback path and the amplifier operates in its second
transimpedance mode for electrical signals having a
frequency below the mode cross-over frequency and in a
follower mode for electrical signals having a frequency
above the mode cross-over frequency.
The amplifier of the present invention operates to
switch the transimpedance between a first relatively
higher resistance and a second relatively lower
.r ~
- 4 - GECAN 3 058
resistance. The first relatively higher resistance
provides a high input transducer load resistance that
increases the sensitivity of the amplifier to low power
transducer signal levels. When the power level of
5 signals received from the transducer rises above a
predetermined power level, the second feedback path
dominates the transducer load impedance. By reducing the
transducer load impedance to that of the second feedback
path, the power range of the amplifier before it
saturates is increased thereby increasing the dynamic
range of the amplifier. The capacitor functions to open
circuit the feedback path for higher frequencies by short
circuiting these higher frequencies signals to ground.
This in effect switches the amplifier to a follower mode
15 of operation. Moreover, the capacitor in the s~cond
feedback path also functions to stabilize the amplifier
by introducing a phase shift to higher frequency signals.
The second feedback path may be coupled in and out
of circuit with the first feedback path by means of a
switch such as an field effect transistor (FET). The
gate of the transistor is connected to an automatic gain
control circuit (AGC) which controls the conduction of
the FET transistor in response to changes in the
transimpedance gain of the amplified signal at the output
of the amplifier stage.
BRIEF DE8CRIPTION OF T~E DRAWINGS
For a better understanding of the nature and
objects of the present invention reference may be had by
way of example to the accompanying diagrammatic drawings
in which:
Figure 1 is a schematic circuit drawing of the
amplifier circuit of the present invention; and,
Figure 2 is a graph of the relative transimpedances
of the first and second feedback paths.
- S - GECAN 3058
DETAILED DE8CRIPTION OF TRE PREFERRED EMBODINENT
Referring to Figure 1, the amplifier circuit of the
present invention is shown at 10. The amplifier circuit
includes a transducer or photodiode 12 having one
5 terminal connected to a DC biasing potential -V and its
other ter~inal connected to the input 14 of the amplifier
stage shown within broken box 16.
The amplifier stage comprises an FET transistor 18
having its gate 20 connected to the input 14. The source
22 of the FET is connected to ground while the drain 24
of the FET is connected through biasing resistor 26 to DC
bias potential +V. FET 18 provides a high input
impedance.
The drain 24 is connected to the emitter of pnp
transistor 28 whose collector is connected to the base of
pnp transistor 30. Transistor 28 has its base connected
to biasing capacitors 32 and resistors 34 and 36
connected to a +V biasing potential. The gain of the
amplifier stage is determined by the transconductance
current of FET 18 and the collector resistance 38 of
transistor 28. The gain of the amplifier stage lies in
the range of 10 to 200 and is preferably about 100.
Transistor 30 is biased by resistor 40 and acts a buffer
to provide the output at 42. The output of the amplifier
stage is passed through a further buffer transistor 44.
A first feedback path comprising resistor 46 is
connected between output 42 and input 14 of the amplifier
stage 16. This resistor provides a load impedance to
photodiode 12 that effects the bandwidth. This resistor
may have a value in the range of 1 kilo-ohm to 10
Mega-ohms and is preferably about 330 kilo-ohms.
A second feedback path shown within broken box 48
is coupled between the output 42 and the input 14 of the
amplifier stage 16. The second feedback path 48 includes
an FET transistor 50 whose gate 52 is connected to an AGC
.r~ f~
- 6 - GECAN 3058
circuit 54. The drain of the FET 58 is connected to the
input 14 while the source of the FET is connected to the
resistor 60. Resistor 60 is also connected in series
with resistor 62 to the output 42 of the amplifier stage.
A capacitor 64 is connected between the resistors 60 and
62 to ground. Resistor 62 has a low value of a few
hundred ohms while resistor 60 has a value of a few
kilo-ohms. The ratio of resistors 60 and 62 is chosen
such that a flat frequency response can be obtained from
DC up to the desired amplifier bandwidth (illustrated by
F in Figure 2).
It should be understood that the sensitivity and
bandwidth requirements of the amplifier are determined by
the value of resistor 46 when the second feedback path 48
is uncoupled. The value of resistor 46 (R46) in this
preferred embodiment is 330 kilo-ohms which sets the
transimpedance gain of the amplifier from dc up to the
bandwidth. Referring to Figure 2 this transimpedance is
shown by upper line 70 to be approximately the value of
R46. In order to obtain a predetermined tran~impedance
reduction when the second feedback path 48 is coupled in
circuit, a reduction factor of, for example X = -22dB or
X = 0.082 is chosen. This results in the second feedback
path having a transimpedance chosen to be in the order of
27 kilo-ohms (R46 x 0.082). This reduced transimpedance
is shown in Figure 2 by lines 72 and 74. Line 72 is
determined by the primarily by the value of resistor 62
(R62) for frequencies below the mode cross-over
frequency. (It should be understood that the actual
transimpedance will be the sum of r~sistors 62 and
60(R60)in parallel with R46). At frequencies above the
mode cross-over frequency, the transimpedance becomes
that of R60 times the open gain of the amplifier which is
in the order of 100. As a result a relatively flat
frequency response along lines 72 and 74 is achieved. In
~ ~3 ! J i ~
- 7 - GECAN 3058
the preferred embodiment R60 is chosen to be 270 ohms for
a gain in the amplifier stage of 100. It should be
understood that the above teachings may only provide an
approximation of the values of resistors R46, R60 and R62
and that final adjustments of the component values may be
required.
The mode cross-over frequency (illustrated by FMoDE
in Figure 2) is determined by R60 and C64. Capacitor 64
and resistor 62 determine a pole for frequencies above
which the feedback signal will be shorted to ground.
Above these frequencies, the amplifier effectively
operates in a follower mode. The introduction of
capacitor 64 provides a phase compensation that keeps the
amplifier stable when the second feedback path 48 is
switched in circuit. If a lower feedback resistance is
switched without phase compensation (ie: without C64~,
then the amplifier would have a tendency to oscillate.
In operation, the photodiode provides electrical
signals of varying frequency to the input 14 of the
amplifier stage. These signals are amplified by the
amplifier stage 16 and taken at the output of transistor
44. For low power signals, the AGC 54 provides a low
bias voltage to the drain of FET 50 maintaining YET 50
"off". As a result, the second feedback path 48 is
uncoupled from the feedback circuit and the feedback is
that of resistor 46. Since the resistance of resistor 46
is relatively high, the load impedance as seen by the
diode 12 and its sensitivity to low power signals is
enhanced. As the gain of the electrical signals from
photodiode 12 increases in the power, at a given power
level, the AGC increases its output bias to the gate of
FET 50 resulting in FET 50 conducting and coupling the
second feedback path 48 in parallel with resistor 46.
Since the resistance of resistors 60 and 62 are
considerably less than that of resistor 46, the DC
t'~ t~
- % - GECAN 3058
feedback current passes through the second feedback path
48. As a result, the load transimpedance (Output
Voltage/Input Current) of the diode 12 will be that of
resistor 60 seen in series with the parallel combination
of resistor 62 and capacitor 64. The load transimpedance
will be effectively reduced by a factor of 10 allowing
the power to be increased before saturation and thereby
improving the dynamic range of the amplifier.
An amplifier as described in the preferred
embodiment has been tested to have a sensitivity in the
first transimpedance mode of -51 dBm of optical power at
25 Megabits per second. When the amplifier is switched
to its second transimpedance and follower modes, the
saturation power level becomes -4dBm which results in an
overall dynamic range of 47 dB.
