鈥?/div>
1+j
This equation suggests that the frequency response is strictly
due to the feedback network. This does not explain why
transimpedance amplifiers are prone to oscillate. Figure 2
provides more insight into the stability problem. The photo-
diode is replaced with an ideal current source in parallel with
its equivalent resistance, R
D
, and capacitance, C
D
. The op amp
input capacitance cannot be considered insignificant and
should be included as part of C
D
.
The noise gain (i.e., the noninverting closed-loop gain) of this
configuration determines the stability of the circuit. The
reason for this is that any noise signal, no matter how small,
can trigger an unstable circuit into oscillation. From inspec-
tion, the transfer function can be determined to be:
C
F
R
F
The dc gain is set solely by the resistors. The pole frequency,
f
P
, is set by the feedback network, just as in the transimpedance
function. The zero frequency, f
Z
, is determined by (a) the sum
of the feedback and the diode capacitances and (b) the parallel
combination of the feedback and the diode resistances.
Typically, the feedback resistor is much smaller than the
photodiode鈥檚 equivalent resistance. This makes the dc resis-
tive gain unity. The value of the parallel combination is
essentially equal to the feedback resistor alone. Therefor, f
Z
will always be lower than f
P
, as shown in Figure 3.
Log Av
1
V
OUT
f
Z
f
P
Log f
FIGURE 3. Bode Plot of Noise Gain.
FIGURE 1. Typical Photodiode Transimpedance Amplifier.
C
F
R
F
C
D
R
D
V
OUT
Log Av
A
OL
1
f
Z
f
P2
f
P3
f
P1
f
GBW
Log f
FIGURE 2. Photodiode Modelled with Ideal Elements.
FIGURE 4. Various Feedback Responses Intersecting Op Amp
Open-loop Gain.
AB-050
Printed in U.S.A. March, 1993
漏
1993 Burr-Brown Corporation
1
2
