AB-057 [ETC]

AB-057 - COMPARISON OF NOISE PERFORMANCE BETWEEN A FET TRANSIMPEDANCE AMPLIFIER AND A SWITCHED INTEGRATOR ; AB - 057 - 间比较一个FET跨导放大器和噪声性能 - 开关集成商\n


型号：	AB-057
厂家：	ETC
描述：	AB-057 - COMPARISON OF NOISE PERFORMANCE BETWEEN A FET TRANSIMPEDANCE AMPLIFIER AND A SWITCHED INTEGRATOR AB - 057 - 间比较一个FET跨导放大器和噪声性能 - 开关集成商\n 开关放大器
文件：	总7页 (文件大小：80K)
中文：	中文翻译
下载：	下载PDF数据表文档文件

AP P LICATION BULLETIN

Mailing Address: PO Box 11400 • Tucson, AZ 85734 • Street Address: 6730 S. Tucson Blvd. • Tucson, AZ 85706

Tel: (602) 746-1111 • Twx: 910-952-111 • Telex: 066-6491 • FAX (602) 889-1510 • Immediate Product Info: (800) 548-6132

COMPARISON OF NOISE PERFORMANCE BETWEEN A

FET TRANSIMPEDANCE AMPLIFIER AND A

SWITCHED INTEGRATOR

By Bonnie C. Baker

Low-input current FET operational amplifiers are univer-

sally used to monitor photodetector, or more commonly

photodiode currents. These photodetectors bridge the gap

(a)

200pF

R_EQ

between a physical event, light, and electronics. There are a

R₃

R₄

variety of amplifier configurations to select from and the

choice is based on noise, bandwidth, offset, and linearity.

The most popular design approach is shown in Figure 1. A

considerable amount has been written on the performance of

this traditional transimpedance amplifier. This topology has

dominated applications such as CT scanners, star-tracking

instruments, electron microscopes, etc., where a light-to-

voltage conversion is required.

10MΩ

100kΩ

I_IN

R₅

HP5082-4204

e_O

OPA128

R_EQ= R₃+ R₄+ (R₃R₄/R₅)

= 1000MΩ

To prevent gain peaking

C₂

R₂

1000MΩ

0.01µF

R₂

(b)

+15V

Guard

A₁A₂: 1/2 OPA2111

A₁

I_IN1

INA105

I_IN

Output

OPA124

R₂

(1)

C_IN

R_IN

50MΩ

25kΩ

I_IN

–15V

e_O

A₃

D₁

Circuit must be well shielded.

R₂

NOTE: (1) R_INand C_INused to reduce DC and AC errors caused by input

bias currents, however, R_INalso increases noise at the output by a factor

of √4KTRB times the noise gain of the circuit.

50MΩ

25kΩ

I_IN2

FIGURE 1. Most Popular Design Approach to Gain Precise

Low Level Currents from a Photodetector.

A₂

Until now, the only feasible solution to the high precision,

current-to-voltage design problem has been an op amp

network with a resistor in the feedback loop. Variations such

as using resistor T-networks or an instrumentation amplifier,

as shown in Figure 2, still use the fundamental concept of a

resistive feedback loop to perform the I/V conversion func-

tion. In these circuits, the fundamental transfer function is:

e_O= (I_IN1– I_IN2)R₂

R_EQ= R₂

e_O= 2 I_INR₂

R_EQ= 2R₂

FIGURE 2 (a) Using a T-Network to Design the Feedback

Resistor for a Transimpedance Circuit.

(b) An Instrumentation Amplifier Topology Can

Be Used to Convert Low-Level Photodiode Cur-

rents to a Voltage Output.

AB-057A

Printed in U.S.A. January, 1994

V_OUT= R_EQ• I_IN,

(Signal)

OUT

where V_OUT= Output voltage

R_EQ= Equivalent resistive feedback element

I_IN= Current generated by the photodetector

C_A

C_B

C_C

C_D

An alternative design method, a switched integrator, is

shown in Figure 3. With this topology, the capacitor in the

feedback loop of the amplifier dominates the transfer func-

tion. The switches perform the functions of removing the

excitation signal from the input of the amplifier (S₁) and

resetting the output of the amplifier to ground (S₂). The

fundamental transfer function of this circuit is:

Logic

V₁

V₂

C_Band C_Dform a voltage divider. Output charge is due

to differences in C_Band C_D.Similar divider action on

input pin.

V₁

V₂

C₂

V_OUT= –

I_INdt,

∫

Charge Out

where V_OUT= Output voltage

FIGURE 4.Topology Used for the Switches in the ACF2101

Switched Integrator to Reduce Charge Injection

Errors.

C₂= Capacitive feedback element

I_IN= Current generated by the photodetector

The discrete design of the switched integrator is impractical

for low noise, precision applications because of the switch-

ing noise of S₁and S₂. The switching noise is caused by the

injection of charge across the parasitic gate-to-source, gate-

to-drain, and source-to-drain capacitances of the FET

switches. The ACF2101 (block diagram shown in Figure 3)

implements the switched integrator design on a monolithic

chip and uses a charge injection cancellation design for S₁

and S₂(see Figure 4) to reduce the switch contribution to DC

offset and noise.

parasitic capacitors C_A, C_B, C_Cand C_Dare carefully matched,

the total charge injection caused by switching at the SIG-

NAL IN node and the SIGNAL OUT node is zero.

The comparison of the noise performance of the traditional

resistor feedback transimpedance amplifier and the switched

integrator starts with the analysis of the input sensor, the

photodetector.

PHOTO DETECTOR CHARACTERISTICS

C₂

Photodiodes generate low level currents that are propor-

tional to the level of illumination. An equivalent circuit for

the photodiode is shown in Figure 5. The value of the

junction capacitor, C₁, can have a wide range of values

dependent of the diode junction area and bias voltage. A

value of 50pF at zero bias is typical for small area diodes.

The value of the shunt resistor, R₁, is usually in the order of

10⁸Ω at room temperature and decreases by a factor of two

every 10°C rise in temperature. The range of the shunt

resistor, R₁, can be as high as 100GΩ and low as 10kΩ at

room temperature. There is no direct correlation between the

values of C₁and R₁. C₁can usually be found in the product

data sheet of the photodiode. The value of R₁is not always

published by the manufacturer, however, its effect on noise

100pF

S₂

Reset

I_IN

1/2

S₁

Hold

ACF2101

S₃

Select

NOTE: S₁: Hold Switch – When closed connects photodiode to the input of

the ACF2101 op amp. S₂: Reset Switch – When closed discharges C₂,

output goes to ground. S₃: Select Switch – Used to multiplex several

ACF2101’s outputs.

R_S< < R_I

FIGURE 3.Block Diagram of 1/2 of the Dual ACF2101

Switched Integrator.

Incident

Light

As illustrated in Figure 4, when the LOGIC node changes

from low to high the voltage change (V₁) across the capaci-

tors, C_Aand C_B, pull charge out of the SIGNAL IN node and

SIGNAL OUT node. The inverted LOGIC signal at V₂

pushes an equal amount of charge through C_Cand C_Dback

into the SIGNAL IN node and SIGNAL OUT node. If

I_IN

R_I

C_I

Ideal

Diode

Photo

Current

≈ 10kΩ

to 100GΩ

FIGURE 5. Equivalent Circuit for a Photodiode.

in both the traditional transimpedance amplifier and the

switched integrator occur at lower frequencies. The overall

noise contribution in the lower frequencies is usually very

small compared to the contribution at higher frequencies,

therefore, knowing the exact value of R₁is not critical.

The first step in designing the transimpedance amplifier is

selecting the feedback resistor, R₂. By knowing the maxi-

mum expected I_IN, R₂is selected to optimize the signal-to-

noise ratio with the formula:

V_OUT(max)

R₂=

I_IN(max)

NOISE ANALYSIS OF TRADITIONAL

TRANSIMPEDANCE AMPLIFIER

where V_OUT(max)= maximum output voltage

of the op amp

I_IN(max)= maximum current from the

photodiode based on maximum

expected light intensity

For the noise comparison between the transimpedance am-

plifier and the switched integrator refer to Figure 6 for a

more complete circuit diagram. The optimum amplifier

would have infinite bandwidth, zero voltage noise, zero

input bias current, and zero offset voltage. The optimum

amplifier does not exist; however, several op amps come

close to meeting one or a few of these general requirements.

Table I summarizes key specifications of the FET amplifiers

OPA111, OPA124, OPA128, OPA404, OPA2111, OPA2107

and OPA627, which are usually used in transimpedance

applications.

Typical values for R₂would be between 10kΩ and 100MΩ.

It is possible that optimum noise performance can be ob-

tained with a V_OUT(max)that is less than the full output swing

of the amplifier, in which case the above equations are not

applicable. The above equation is designed to optimize the

signal-to-noise ratio at the output of the amplifier.

C₂

R₂

Equivalent Circuit for photodiode

I_IN

R₂

i_N

e_n

C₁

I_IN

R₁

C₁

OPA124

e_O

(a)

|A| (dB)

NOTE: In+ shorted in this configuration.

Open-Loop Gain of Op Amp

FIGURE 6.Complete Circuit Diagram Used for the Evalua-

tion of the Noise and Bandwidth Performance of

the Classical Transimpedance Amplifier and the

Switched Integrator.

Signal Gain

NOISE at

10kHz

(nV/√Hz)

INPUT

INPUT BIAS

CURRENT

(pA, max)

f_p=1/(2πR₂C₂)

BANDWIDTH CAPACITANCE

2πR

PRODUCT

(MHz, typ)

(pF, typ)

f_z=1/(2π(R₁|| R₂) (C₁+ C₂))

OPA111BM

OPA124BP

OPA627BP

OPA404G

OPA128BM

OPA2111BM

OPA2107BM

1.6

6.4

0.15

1 + C₁/C₂

High Frequency

Noise Gain

DC Noise Gain

1 + R₂/R₁

12⁽¹⁾

15⁽¹⁾

(b)

f_z

f_p

Log f

4.5

8⁽¹⁾

NOTE: (1) Denotes typical values.

FIGURE 7. Noise And Signal Response of the Classical

Transimpedance Amplifier.

TABLE I. Low Noise FET Input Op Amps Typically Used

In Transimpedance Amplifier Applications. In

transimpedance applications, the input capaci-

tance of the amplifier equals input common-

mode capacitance plus input differential capaci-

tance.

The model in Figure 7 shows the overall noise gain response

of the transimpedance circuit. The signal bandwidth is deter-

mined by a pole generated by R₂and C₂. The noise band-

width is determined by the open loop gain roll off of the op

amp. To maximize the signal bandwidth and insure an

approximate 45° phase margin with a 25% step function

overshoot, C₂is selected using the formula below:

100k

10k

R₂= 1MΩ

Noise ≈ 20µVrms

R₂= 10MΩ

Noise ≈ 35µVrms

C₁

C₂=

2πR₂• f_u

0.1

where f_u= op amp unity gain bandwidth

100

200

300

400

500

600

Here the signal bandwidth and noise bandwidth are identical

and equal to:

Photodiode and Op Amp Capacitance, C₁(pF)

FIGURE 8. Signal Bandwidth and Output Noise Change

With Changes in Input Capacitance, C₁, of a

Transimpedance Amplifier Using the OPA111

as the Op Amp.

BW =

2πR₂C₂

In some applications, an overshoot of 25% may be too much.

A more conservative 5% overshoot can be designed with a

phase margin of 65° by using the formula below to select C₂:

Burr-Brown macromodel for the OPA627 is the

OPA627E.MOD. The PSpice Probe command needed to

calculate the cumulative rms noise is SQRT(S(V(ONOISE)

• V(ONOISE))). Figure 8 shows how a transimpedance

amplifier’s (designed using the OPA124) signal bandwidth

and output noise change with values of input capacitance.

Note that the output noise is unexpectedly flat across changes

in C₁. This is because the signal and noise bandwidth

decrease with increases of C₂. A lower noise bandwidth

yields lower rms noise at the output of the amplifier. The

signal-to-noise ratio improves.

= 2 •

2πR • f

where f_u= op amp unity gain bandwidth

Typical calculated values for C₂would be from sub-pico

farads to 20 or 30pF. Actual minimum circuit values for C₂

are dependent on the stray capacitance of R₂and PC board

layout. Typically, a resistor has 0.5pF of stray capacitance.

Using the C₂value calculated above (for a 65° phase mar-

gin), the effective noise bandwidth is equivalent to the noise

gain 3dB bandwidth times π/2 or:

To improve the signal-to-noise ratio of a transimpedance

amplifier, the designer can select a lower noise amplifier,

reduce the (1+C₁/C₂) noise gain, reduce the feedback resistor

value, or reduce the signal bandwidth of the system with an

additional filter or a slower op amp. Lower noise FET

amplifiers usually have a wider bandwidth and higher input

capacitance than the higher noise FET amplifiers. If a lower

noise, wider bandwidth amplifier is selected as the op amp

for the transimpedance amplifier, the increase in bandwidth

and input capacitance may cause more noise in the system

than the original op amp. A filter can be used to reduce the

overall bandwidth and reduce the noise. Additionally, the

noise gain of the transimpedance amplifier can be reduced

by increasing C₂or decreasing C₁. The photodetector can be

selected in order to reduce C₁. Sometimes this is not possible

because of the design constraints of photodetector vs the

signal source. C₂can be increased at the expense of reduced

bandwidth. More elaborate techniques can be used to reduce

noise, such as a more complex feedback network around the

amplifier or boot-strapping the photodetector. These tech-

niques are beyond the scope of this application note. Refer

to the reference articles for more depth.

BW_{effective noise}

2R₂C₂

and the signal bandwidth is:

BW_signal

2πR₂C₂

Usually it is necessary to follow the transimpedance ampli-

fier with a low pass filter to further reduce the wideband

noise beyond the signal bandwidth. A single pole, low-pass

filter with a bandwidth at twice the signal bandwidth of the

transimpedance amplifier can easily improve the dynamic

range of the transimpedance amplifier by 4 or 5 dB.

Brute force calculations should be performed to understand

the noise contributions of the regions illustrated in Figure 7

(see OPA101 data sheet). For instance, R₁contributes to

overall noise gain in the lower frequencies and can be mostly

ignored. This insight is valuable when considering design

options to further improve the circuit. Once the details are

understood, an easier approach is to use a macromodel and

simulate the results. The macromodel must be able to simu-

late the noise performance of the op amp. The appropriate

One fundamental performance difference between the tradi-

tional transimpedance amplifier and the switched integrator

is that the amplifier gives a real time representation of the

light excitation at the output of the amplifier, and the

switched integrator gives a time-averaged representation of

100

C₂

R₂

C₁

I_IN

R₁

1/2

ACF2101

e_O

(a)

100 200 300 400 500 600 700 800 900 1000

C₁(µF)

|A| (dB)

FIGURE 10. Total Output Noise vs C₁and C₂of the ACF2101

Switched Integrator.

Open-Loop Gain of Op Amp

tor amplifier transfer function is dominated by the ratio of

the feedback capacitor, C₂, and time. Referring to Figure 3,

when S₁is closed, the input current flows past the inverting

node of the op amp (which is held at virtual ground) and

charges C₂. The input current consequently causes the volt-

age at the output of the op amp to change in the negative

direction over time. The voltage output of the switched

integrator represents the average current input signal over a

specified time, as opposed to the real time signal of the

previous example. The ACF2101 switched integrator, shown

in Figure 3, has a maximum input current specified at

100µA. The input current is restricted by the internal capaci-

tor of the ACF2101 (100pF) and the 1V/µs slew rate of the

amplifier. If an external capacitor is used, input currents can

exceed 100µA as long as the following ratio is true:

DC Noise Gain

R₂

= (1 + —)

R₁

1/(2πR₂C₂)

High Frequency

C₁

Noise Gain = (1 + —)

C₂

Noise Gain

1/(2πR₁|| R₂(C₁+ C₂))

Region 3

Log f

(b)

f_p

f_z

Region 1

Region 2

C₂

≥ 10⁶

FIGURE 9. Noise and Signal Response of a Switched

Integrator.

farad/sec

(

)

(integration time)

where the integration time equals the amount of time be-

tween samples.

the input information from the photodetector. The real time

solution is limited in bandwidth by the selected amplifier,

the settling time of the amplifier, and the required feedback

capacitor and resistor (C₂and R₂). Additionally, a filter is

usually used following the output of the transimpedance

amplifier to further reduce noise at higher frequencies.

If that ratio is less than 10⁶(farad/sec), the ACF2101 may

loose accuracy at the output. Whenever possible, the internal

capacitor should be used with the ACF2101 to insure greater

gain and linearity accuracy. Figure 9 is used to evaluate the

noise contribution of the op amp, gained by the feedback

network of the ACF2101 and the photodiode. Here the reset

switch, (S₂as illustrated in Figure 3) is modeled as a

noiseless resistor (R₂). The typical resistance of S₂, when it

is open, is 1000GΩ. The switched integrator, ACF2101, has

an internal feedback capacitor, C₂, of 100pF. The user may

choose to use an external capacitor instead of the internal

capacitor provided. Typical values can range up to 2000pF.

If an external capacitor is used, care must be taken to choose

an integration capacitor with a low voltage coefficient,

temperature coefficient, memory, and leakage current. Suit-

able types include NPO ceramic, polycarbonate, polysty-

rene, and silver mica.

This approach is optimal for low and medium bandwidth

applications where information about the amplitude and

shape of the input signal is critical. The design problem is

complex because of the trade-offs between noise and band-

width and the abundance of op amp choices. Also, the

capacitor and resistor accuracy requirements make this de-

sign difficult and sometimes expensive to manufacture in a

production environment.

NOISE ANALYSIS

OF SWITCHED INTEGRATOR AMPLIFIER

Where the traditional transimpedance transfer function is

dominated by the feedback resistor, R₂, the switched integra-

The total noise contribution of the op amp and the feedback

network of the switched integrator is equivalent to the

tor, the slew rate of the amplifier, the settling time of the

switches and amplifier. The bandwidth of the switched

integrator can be improved by decreasing the integration

capacitor, C₂. As shown in Figure 11, the settling time of the

reset switch is increased with increases in C₂.

square root of the sum of the squares of three regions as

shown in Figure 9. The noise in the first region is equal to

the average op amp noise over that region times the square

root of the region bandwidth. The pole in the noise gain of

the switched integrator circuit generated by R₂and C₂is in

the sub-Hertz region. For example, if C₂= 100pF and R₂=

1000GΩ the pole generated by the RC pair is equal to

1.59mHz. To calculate the noise contribution of this region

the average noise over the region is multiplied times the

square root of 1.59mHz which is equal to 39.87E^–3. Quick

calculations show that any noise gained by the DC gain

(1+R₂/R₁) of the switched integrator is negligible. In addi-

tion, the zero generated by the (R₁|| R₂)(C₁+C₂) combination

is also in the low frequency range. Using a typical value of

100MΩ for R₁, 50pF for C₁, 1000GΩ for R₂and 100pF for

C₂, the zero is located at 10.6Hz. Again, the noise contribu-

tion from this region is negligible. Consequently, the noise

contribution from the op amp and its feedback network in

conjunction with the photodiode is dominated by the op amp

noise times (1 + C₁/C₂).

100 200 300 400 500 600 700 800 900 1000

C₂(pF)

In addition to the gained op amp noise mentioned above,

charge injection and ^KT/_C(capacitor noise) also contribute to

the total noise figure of the switched integrator. The switch

network shown in Figure 4 is used for the switches of the

ACF2101 to reduce charge injection noise. Figure 10 illus-

trates the total output noise of the switched integrator with

various C₁and C₂values.

FIGURE 11. Reset Time C₂vs the ACF2101 Switched

Integrator.

The switched integrator requires external digital support

circuitry to drive the hold, reset and select switches. If an

external integration capacitor (C₂) is used, gain accuracy can

be compromised due to the accuracy of the capacitor. The

select switch allows the user of the ACF2101 switched

integrator to eliminate a sample hold amplifier.

The ACF2101 switched integrator is a sampled system

controlled by the sampling frequency (f_S), which is usually

dominated by the integration time. The fastest feasible

sampling frequency for the ACF2101 is 42.55kHz. This

assumes that the internal capacitor (100pF) is used. The full

scale output is –10V and the settling time requirements are

to 0.01% accuracy. Input signals should be below the Nyquist

frequency (f_S/2) to avoid aliasing errors. The bandwidth of

the ACF2101 is determined by the slew rate of the amplifier,

settling times of the reset (S₂) and select (S₃) switches as well

as the on-to-off and off-to-on switching speeds. The slew

rate of the amplifier is guaranteed a minimum of 1V/µs. The

output node requires at least 10µs to reach full scale. This

time restraint can be reduced if the full 10V swing capability

of the ACF2101 is not used.

COMPARING THE TRADITIONAL

TRANSIMPEDANCE AMPLIFIER

TO THE SWITCHED INTEGRATOR

Two examples are selected for comparison of the

transimpedance amplifier and the switched integrator ampli-

fier. In both examples, optimum noise solutions and opti-

mum bandwidth solutions are considered.

OPTIMUM NOISE PERFORMANCE OPTIMUM BANDWIDTH PERFORMANCE

The settling time of the reset switch (S₂) is 5µs to 0.01%

accuracy, which is limited by slew rate of the amplifier and

the R₂|| C₂time constant. The reset switch settling time

increases with larger values of C₂; however, if the user also

reduces the full scale signal output as described above, this

time is reduced proportionally to the output swing maxi-

mum. The select switch (S₃) has a 2µs settling time to 0.01%

accuracy for loads ≤ 100pF and the delay between switching

should be about 0.5µs.

TRANS-

SWITCHED

TRANS-

SWITCHED

IMPEDANCE

INTEGRATOR

IMPEDANCE

INTEGRATOR

Device

OPA124

1pA

ACF2101

1pA

OPA627

5pA

ACF2101

1pA

I_B

C₂

0.5pF

100pF

0.5pF

1.5pF

Signal

Bandwidth

3.2kHz

50Hz

3.2kHz

3.35kHz

Noise

Bandwidth

5kHz

200µVrms

94dB

250kHz

35µVrms

109dB

79kHz

400µVrms

88dB

250kHz

100µVrms

100dB

Noise

SNR

The signal-to-noise ratio of the switched integrator can be

reduced by selecting a higher value integration capacitor, C₂.

The switched integrator is limited in bandwidth by the

amplifier and the nature of the transfer function. The output

signal of the integrator is time averaged. The sampling

frequency is restricted by the size of the integrating capaci-

TABLE II. Transimpedance and Switched Integrator De-

sign Comparison Using a 100pF Photodiode

with Maximum Output Current of 100nA. Val-

ues of R were calculated assuming a 65° phase

margin. ²

The first design problem uses an average sized photo diode

(C₁= 100pF) with a maximum output current of 100nA. As

shown in Table II, the OPA124 was selected for the low

noise comparison and the OPA627 was selected for the

wide-bandwidth comparison. The fullscale output of the

transimpedance amplifier is designed to be –10V. R₂is

selected to be 100MΩ. C₂of the OPA124 low noise circuit

is restricted by the parasitic capacitance of the feedback

resistor, 0.5pF. The OPA124 has a maximum input bias

current of 1pA, low drift of 1µV/°C, and low offset voltage.

The signal-to-noise ratio of this design is 94dB with a signal

bandwidth of 3.2kHz.

OPTIMUM NOISE PERFORMANCE OPTIMUM BANDWIDTH PERFORMANCE

TRANS-

SWITCHED

TRANS-

SWITCHED

IMPEDANCE

INTEGRATOR

IMPEDANCE

INTEGRATOR

Device

OPA124

1pA

ACF2101

1pA

OPA627

5pA

ACF2101

1pA

I_B

C₂

18.2pF

500pF

6.76pF

210pF

Signal

Bandwidth

87kHz

10kHz

235kHz

24kHz

Noise

Bandwidth

275kHz

31µVrms

110dB

250kHz

15µVrms

116dB

740kHz

108µVrms

99dB

250kHz

20µVrms

114dB

Noise

SNR

TABLE III.Transimpedance and Switched Integrator De-

sign Comparison Using a 100pF Photodiode

With Maximum Output Current of 100µA. Val-

ues of R were calculated assuming a 65° phase

margin. ²

In contrast, the ACF2101 is configured to minimize noise by

using the on-chip capacitor C₂= 100pF. As shown in Figure

10, the rms noise performance is typically 35µVrms. This

noise performance can be improved by using a higher value

external capacitor, but, the bandwidth performance of the

circuit is compromised. The signal-to-noise ratio of the

switched integrator is better than the transimpedance con-

figuration at 109dB, but the bandwidth is only 50Hz.

The switched integrator is designed with C₂= 500pF (exter-

nal capacitor). A larger capacitor is used to improve the

noise performance of the circuit. The noise performance of

the switched integrator is improved over the transimpedance

amplifier, yet the bandwidth is considerably smaller.

In the second section of Table II, the two circuits are

optimized for bandwidth. Here the OPA627 is selected as

the preferred op amp in hopes of improving the bandwidth

of the transimpedance amplifier. The transimpedance design

using the OPA627 does indeed change the bandwidth of the

circuit, but in an undesirable way. C₂is still limited to 0.5pF

because of the stray capacitance of R₂, consequently the

signal bandwidth does not change from the OPA124 design.

The noise bandwidth, however, does change by a factor of

^~16, causing a significant increase in noise. The signal-to-

noise ratio of this circuit is 88dB. The designer would be

better off using the OPA124 as the amplifier instead of the

OPA627 for this application.

The bandwidth and noise performance of the transimpedance

amplifier can be improved by using the OPA627 in place of

the OPA124. As shown in Table III, the signal bandwidth is

nearly tripled. On the other hand, the switched integrator is

designed for improved bandwidth performance by decreas-

ing C₂to equal 210pF. Although the noise performance is

better than the transimpedance amplifier, the bandwidth is

restricted by the slew rate, settling times, and switching

times of the switched integrator.

IN CONCLUSION

In contrast, the ACF2101 is also configured to maximize

bandwidth. Here a feedback capacitor of 1.5pF is selected.

PC board layout precautions should be taken to reduce stray

capacitance. The signal bandwidth of the circuit is designed

to 3.35kHz with a signal-to-noise ratio of 100dB. In this

example, the noise and bandwidth performance of the

ACF2101 switched integrator is better than the

transimpedance configuration. The ACF2101 switched inte-

grator is best optimized for low input current, high input

capacitance applications.

This application note has taken a look at a few variables that

optimize the performance of circuits that amplify photo-

diode signals. In addition to the solutions presented, alterna-

tives should also be explored before the final design is

released to production.

REFERENCES

OPA101 Product Data Sheet, Burr-Brown PDS-434.

Graeme, Jerald, “FET Op Amps Convert Photodiode Out-

puts to Usable Signals”, EDN, October 29, 1987.

Table III illustrates the second design problem where the

photodiode has the same stray capacitance (C₁) of 100pF but

a maximum current signal of 100µA. The OPA124 is se-

lected for the low noise transimpedance amplifier. In this

case, R₂is selected to be 100kΩ and C₂equal to 18.2pF. The

noise performance of this amplifier is 31µVrms with a

signal-to-noise ratio of 110dB.

ACF2101 Product Data Sheet, Burr-Brown PDS-1078.

Graeme, Jerald, “Circuit Options Boost Photodiode Band-

width”, EDN, May 21, 1992.

Graeme, Jerald, “Phase Compensation Optimizes Photo-

diode Bandwidth”, EDN, May 7, 1992.

The information provided herein is believed to be reliable; however, BURR-BROWN assumes no responsibility for inaccuracies or omissions. BURR-BROWN assumes

no responsibility for the use of this information, and all use of such information shall be entirely at the user’s own risk. Prices and specifications are subject to change

without notice. No patent rights or licenses to any of the circuits described herein are implied or granted to any third party. BURR-BROWN does not authorize or warrant

any BURR-BROWN product for use in life support devices and/or systems.

AB-057 [ETC]

相关型号：

SI9130DB

SI9135LG-T1

SI9135LG-T1-E3

SI9135_11

SI9136_11

SI9130CG-T1-E3

SI9130LG-T1-E3

SI9130_11

SI9137

SI9137DB

SI9137LG

SI9122E