Wideband, Ultra-Low Noise, Voltage-FeedbackOPERATIONAL AMPLIFIER with Shutdown
APPLICATIONSHIGH DYNAMIC RANGE ADC PREAMPSLOW NOISE, WIDEBAND, TRANSIMPEDANCEAMPLIFIERSWIDEBAND, HIGH GAIN AMPLIFIERSLOW NOISE DIFFERENTIAL RECEIVERSULTRASOUND CHANNEL AMPLIFIERSIMPROVED UPGRADE FOR THE OPA687,CLC425, AND LMH6624
FEATURESHIGH GAIN BANDWIDTH: 3.9GHzLOW INPUT VOLTAGE NOISE: 0.85nV/√HzVERY LOW DISTORTION: –105dBc (5MHz)HIGH SLEW RATE: 950V/µsHIGH DC ACCURACY: VIO < ±100µVLOW SUPPLY CURRENT: 18.1mALOW SHUTDOWN POWER: 2mWSTABLE FOR GAINS ≥≥≥≥≥ 12
Ultra-High Dynamic RangeDifferential ADC Driver
DESCRIPTIONThe OPA847 combines very high gain bandwidth and largesignal performance with an ultra-low input noise voltage(0.85nV/√Hz) while using only 18mA supply current. Wherepower savings is critical, the OPA847 also includes anoptional power shutdown pin that, when pulled low, disablesthe amplifier and decreases the supply current to < 1% of thepowered-up value. This optional feature may be left discon-nected to ensure normal amplifier operation when no power-down is required.
The combination of very low input voltage and current noise,along with a 3.9GHz gain bandwidth product, make theOPA847 an ideal amplifier for wideband transimpedanceapplications. As a voltage gain stage, the OPA847 is opti-mized for a flat frequency response at a gain of +20V/V andis stable down to gains as low as +12V/V. New externalcompensation techniques allow the OPA847 to be used atany inverting gain with excellent frequency response control.Using this technique in a differential Analog-to-Digital Con-verter (ADC) interface application, shown below, can deliverone of the highest dynamic-range interfaces available.
OPA847
SBOS251C – JULY 2002 – REVISED OCTOBER 2003
www.ti.com
Please be aware that an important notice concerning availability, standard warranty, and use in critical applications ofTexas Instruments semiconductor products and disclaimers thereto appears at the end of this data sheet.
PRODUCTION DATA information is current as of publication date.Products conform to specifications per the terms of Texas Instrumentsstandard warranty. Production processing does not necessarily includetesting of all parameters.
Power Supply ............................................................................... ±6.5VDC
Internal Power Dissipation ........................ See Thermal Analysis SectionDifferential Input Voltage .................................................................. ±1.2VInput Voltage Range ............................................................................ ±VS
Storage Temperature Range: D, DBV .......................... –40°C to +125°CLead Temperature (soldering, 10s) .............................................. +300°CJunction Temperature (TJ ) ............................................................ +150°CESD Rating (Human Body Model) .................................................. 1500V
NOTE: (1) Stresses above these ratings may cause permanent damage.Exposure to absolute maximum conditions for extended periods may degradedevice reliability. These are stress ratings only, and functional operation of thedevice at these or any other conditions beyond those specified is not implied.
PACKAGE/ORDERING INFORMATIONSPECIFIED
PACKAGE TEMPERATURE PACKAGE ORDERING TRANSPORTPRODUCT PACKAGE-LEAD DESIGNATOR(1) RANGE MARKING NUMBER MEDIA, QUANTITY
OPA847 SO-8 D –40°C to +85°C OPA847 OPA847ID Rails, 100" " " " " OPA847IDR Tape and Reel, 2500
OPA847 SOT23-6 DBV –40°C to +85°C OATI OPA847IDBVT Tape and Reel, 250" " " " " OPA847IDBVR Tape and Reel, 3000
NOTE: (1) For the most current specifications and package information, refer to our web site at www.ti.com.
Top View SOT
1
2
3
4
8
7
6
5
NC
Inverting Input
Noninverting Input
–VS
DIS
+VS
Output
NC
NC = No Connection
1
2
3
6
4
+VS
Inverting Input
Output
–VS 5 DIS
Noninverting Input
OATI
1 2 3
6 5 4
Pin Orientation/Package Marking
ELECTROSTATIC DISCHARGE SENSITIVITYElectrostatic discharge can cause damage ranging fromperformance degradation to complete device failure. TexasInstruments recommends that all integrated circuits be handledand stored using appropriate ESD protection methods.
ESD damage can range from subtle performance degrada-tion to complete device failure. Precision integrated circuitsmay be more susceptible to damage because very smallparametric changes could cause the device not to meetpublished specifications.
OPA847 3SBOS251C www.ti.com
OPA847ID, IDBV
TYP MIN/MAX OVER TEMPERATURE
0°C to –40°C to MIN/ TESTPARAMETER CONDITIONS +25°C +25°C(1) 70°C(2) +85°C(2) UNITS MAX LEVEL(3)
ELECTRICAL CHARACTERISTICS: VS = ±5VBoldface limits are tested at +25°C.RL = 100Ω, RF = 750Ω, RG = 39.2Ω, and G = +20 (see Figure 1 for AC performance only), unless otherwise noted.
NOTES: (1) Junction temperature = ambient for +25°C specifications. (2) Junction temperature = ambient at low temperature limit: junction temperature = ambient +23°Cat high temperature limit for over temperature specifications. (3) Test Levels: (A) 100% tested at 25°C. Over temperature limits by characterization and simulation.(B) Limits set by characterization and simulation. (C) Typical value only for information. (4) Current is considered positive out of node. VCM is the input common-modevoltage. (5) Tested < 3dB below minimum specified CMRR at ±CMIR limits.
AC PERFORMANCE (see Figure 1)Closed-Loop Bandwidth G = +12, RG = 39.2Ω, VO = 200mVPP 600 MHz typ C
G = +20, RG = 39.2Ω, VO = 200mVPP 350 230 210 195 MHz min BG = +50, RG = 39.2Ω, VO = 200mVPP 78 63 60 57 MHz min B
Gain Bandwidth Product (GBP) G ≥ +50 3900 3100 3000 2800 MHz min BBandwidth for 0.1dB Gain Flatness G = +20, RL = 100Ω 60 40 35 30 MHz min BPeaking at a Gain of +12 4.5 7 10 12 dB max BHarmonic Distortion G = +20, f = 5MHz, VO = 2VPP
2nd-Harmonic RL = 100Ω –74 –70 –69 –68 dBc max BRL = 500Ω –105 –90 –89 –88 dBc max B
3rd-Harmonic RL = 100Ω –103 –96 –91 –88 dBc max BRL = 500Ω –110 –105 –100 –90 dBc max B
2-Tone, 3rd-Order Intercept G = +20, f = 20MHz 39 37 36 35 dBm min BInput Voltage Noise Density f > 1MHz 0.85 0.92 0.98 1.0 nV/√Hz max BInput Current Noise Density f > 1MHz 2.5 3.5 3.6 3.7 pA/√Hz max BPulse Response
Rise-and-Fall Time 0.2V Step 1.2 1.75 2.0 2.2 ns max BSlew Rate 2V Step 950 700 625 535 V/µs min BSettling Time to 0.01% 2V Step 20 ns typ C
0.1% 2V Step 10 12 14 18 ns max B1% 2V Step 6 8 10 12 ns max B
DC PERFORMANCE(4)
Open-Loop Voltage Gain (AOL) VO = 0V 98 90 89 88 dB min AInput Offset Voltage VCM = 0V ±0.1 ±0.5 ±0.58 ±0.60 mV max AAverage Offset Voltage Drift VCM = 0V ±0.25 ±0.25 ±1.5 ±1.5 µV/°C max BInput Bias Current VCM = 0V –19 –39 –41 –42 µA max AInput Bias Current Drift (magnitude) VCM = 0V –15 –15 –40 –70 nA/°C max BInput Offset Current VCM = 0V ±0.1 ±0.6 ±0.7 ±0.85 µA max AInput Offset Current Drift VCM = 0V ±0.1 ±0.1 ±2 ±3.5 nA/°C max B
INPUTCommon-Mode Input Range (CMIR)(5) ±3.3 ±3.1 ±3.0 ±2.9 V min ACommon-Mode Rejection Ratio (CMRR) VCM = ±0.5V, Input Referred 110 95 93 90 dB min AInput Impedance
Differential VCM = 0V 2.7 || 2.0 kΩ || pF typ CCommon-Mode VCM = 0V 2.3 || 1.7 MΩ || pF typ C
OUTPUTOutput Voltage Swing ≥ 400Ω Load ±3.5 ±3.3 ±3.1 ±3.0 V min A
100Ω Load ±3.4 ±3.2 ±3.0 ±2.9 V min ACurrent Output, Sourcing VO = 0V 100 60 56 52 mA min ACurrent Output, Sinking VO = 0V –75 –60 –56 –52 mA min AClosed-Loop Output Impedance G = +20, f = < 100kHz 0.003 Ω typ C
POWER SUPPLYSpecified Operating Voltage ±5 V typ CMaximum Operating Voltage ±6 ±6 ±6 ±6 V max AMaximum Quiescent Current VS = ±5V 18.1 18.4 18.7 18.9 mA max AMinimum Quiescent Current VS = ±5V 18.1 17.8 17.5 17.1 mA min APower-Supply Rejection Ratio
+PSRR, –PSRR |VS| = 4.5V to 5.5V, Input Referred 100 95 93 90 dB min A
POWER-DOWN (disabled low) (Pin 8 on SO-8; Pin 5 on SOT23-6)Power-Down Quiescent Current (+VS) –200 –270 –320 –370 µA max AOn Voltage (enabled high or floated) 3.5 3.75 3.85 3.95 V min AOff Voltage (disabled asserted low) 1.8 1.7 1.6 1.5 V max APower-Down Pin Input Bias Current (VDIS = 0) 150 190 200 210 µA max APower-Down Time 200 ns typ CPower-Up Time 60 ns typ COff Isolation 5MHz, Input to Output 70 dB typ C
THERMALSpecification ID, IDBV –40 to +85 °C typ CThermal Resistance, θJA Junction-to-Ambient
D SO-8 125 °C/W typ CDBV SOT23 150 °C/W typ C
OPA8474SBOS251Cwww.ti.com
TYPICAL CHARACTERISTICS: VS = ±5VTA = 25°C, G = +20V/V, RG = 39.2Ω, and RL = 100Ω, unless otherwise noted.
6
3
0
–3
–6
–9
–12
–15
NONINVERTING SMALL-SIGNALFREQUENCY RESPONSE
Frequency (MHz)
Nor
mal
ized
Gai
n (d
B)
1 10 100 1000
G = +50
See Figure 1
VO = 0.2VPPRG = 39.2ΩRL = 100ΩRF Adjusted
G = +30
G = +12
G = +20
6
3
0
–3
–6
–9
–12
–15
INVERTING SMALL-SIGNALFREQUENCY RESPONSE
Frequency (MHz)
Nor
mal
ized
Gai
n (d
B)
1 10 100 1000
G = –50See Figure 2
VO = 0.2VPPRL = 100ΩRG = RS = 50ΩRF Adjusted
G = –40
G = –30G = –20
29
26
23
20
17
14
11
8
NONINVERTING LARGE-SIGNALFREQUENCY RESPONSE
Frequency (MHz)
Gai
n (d
B)
10 100 1000
VO = 2VPP
See Figure 1
RG = 39.2ΩRL = 100Ω
G = +20V/V
VO = 5VPP
VO = 1VPP
VO = 200mVPP
35
32
29
26
23
20
17
14
INVERTING LARGE-SIGNALFREQUENCY RESPONSE
Frequency (MHz)
Gai
n (d
B)
10 100 1000
VO = 2VPP
See Figure 2
RL = 100ΩRG = RS = 50ΩG = –40V/V
VO = 5VPP
VO = 0.2VPPVO = 1VPP
0.25
0.20
0.15
0.10
0.05
0
–0.05
–0.10
–0.15
–0.20
–0.25
1.25
1.00
0.75
0.50
0.25
0
–0.25
–0.50
–0.75
–1.00
–1.25
NONINVERTING PULSE RESPONSE
Time (5ns/div)
Out
put V
olta
ge (
50m
V/d
iv)
Out
put V
olta
ge (
250m
V/d
iv)
Small Signal ± 100mV
See Figure 1
G = +20V/V
Left Scale
Large Signal ± 1V
Right Scale
0.25
0.20
0.15
0.10
0.05
0
–0.05
–0.10
–0.15
–0.20
–0.25
1.25
1.00
0.75
0.50
0.25
0
–0.25
–0.50
–0.75
–1.00
–1.25
INVERTING PULSE RESPONSE
Time (5ns/div)
Out
put V
olta
ge (
50m
V/d
iv)
Out
put V
olta
ge (
250m
V/d
iv)
Small Signal ± 100mV
See Figure 2G = –40V/VRG = RS = 50ΩRL = 100Ω
Left Scale
Large Signal ± 1V
Right Scale
OPA847 5SBOS251C www.ti.com
TYPICAL CHARACTERISTICS: VS = ±5V (Cont.)TA = 25°C, G = +20V/V, RG = 39.2Ω, and RL = 100Ω, unless otherwise noted.
–70
–75
–80
–85
–90
–95
–100
–105
–110
–115
5MHz HARMONIC DISTORTION vs LOAD RESISTANCE
Load Resistance (Ω)
Har
mon
ic D
isto
rtio
n (d
Bc)
100 150 200 250 300 350 400 450 500
See Figure 1
G = +20V/VVO = 2VPP
2nd-Harmonic
3rd-Harmonic
–75
–80
–85
–90
–95
–100
–105
1MHz HARMONIC DISTORTION vs LOAD RESISTANCE
Load Resistance (Ω)
Har
mon
ic D
isto
rtio
n (d
Bc)
100 150 200 250 300 350 400 450 500
See Figure 1
G = +20V/VVO = 5VPP
2nd-Harmonic
3rd-Harmonic
–65
–75
–85
–95
–105
–115
HARMONIC DISTORTION vs FREQUENCY
Frequency (MHz)
Har
mon
ic D
isto
rtio
n (d
Bc)
0.1 1 10 100
3rd-Harmonic
2nd-Harmonic
G = +20V/VVO = 2VPPRL = 200Ω
See Figure 1
–75
–80
–85
–90
–95
–100
–105
–110
–115
HARMONIC DISTORTION vs OUTPUT VOLTAGE
Output Voltage Swing (VPP)
Har
mon
ic D
isto
rtio
n (d
Bc)
0.1 1 10
See Figure 1
G = +20V/VF = 5MHzRL = 200Ω
2nd-Harmonic
3rd-Harmonic
–75
–80
–85
–90
–95
–100
–105
–110
HARMONIC DISTORTION vs NONINVERTING GAIN
Gain (V/V)
Har
mon
ic D
isto
rtio
n (d
Bc)
15 20 25 30 35 40 45 50 55 50
See Figure 1
VO = 2VPPRL = 200ΩF = 5MHzRF = 750ΩRG Adjusted
2nd-Harmonic
3rd-Harmonic
–70
–75
–80
–85
–90
–95
–100
–105
–110
HARMONIC DISTORTION vs INVERTING GAIN
Gain –V/V
Har
mon
ic D
isto
rtio
n (d
Bc)
20 25 30 35 40 45 50
See Figure 2
VO = 2VPPRL = 200ΩF = 5MHzRG = 50ΩRF Adjusted
2nd-Harmonic
3rd-Harmonic
OPA8476SBOS251Cwww.ti.com
TYPICAL CHARACTERISTICS: VS = ±5V (Cont.)TA = 25°C, G = +20V/V, RG = 39.2Ω, and RL = 100Ω, unless otherwise noted.
10
1
0
INPUT VOLTAGE AND CURRENT NOISE
Frequency (Hz)
Vol
tage
Noi
se (
nV/√
Hz)
Cur
rent
Voi
se (
pA/√
Hz)
101 102 103 104 105 106 107
2.7pA/√HzCurrent Noise
0.85nV/√HzVoltage Noise
50
45
40
35
30
25
20
2-TONE, 3RD-ORDER INTERMODULATION INTERCEPT
Frequency (MHz)
Inte
rcep
t Poi
nt (
+dB
m)
5 10 15 20 25 30 35 40 45 50
G = +20V/V20dB to matched load.
750Ω
50ΩOPA847PI
PO
50Ω
50Ω
39.2Ω
0.5
0.4
0.3
0.2
0.1
0
–0.1
–0.2
–0.3
–0.4
–0.5
NONINVERTING GAIN FLATNESS TUNE
Frequency (MHz)
Dev
iatio
n fr
om 2
1.58
dB G
ain
(0.1
dB)
1 10 100 1000
NG = 12
NG = 14
NG = 20
NG = 18
NG = 16
VO = 200mVPPAV = +12V/VNG = Noise Gain
External CompensationSee Figure 8
1
0
–1
–2
–3
–4
–5
–6
–7
–8
–9
LOW GAIN INVERTING BANDWIDTH
Frequency (MHz)
Nor
mal
ized
Gai
n (1
dB)
1 10 100 1000
G = –8
G = –4G = –2
G = –1
VO = 0.2VPPRF = 750Ω
External CompensationSee Figure 6
100
10
1
RECOMMENDED RS vs CAPACITIVE LOAD
Capacitive Load (pF)
RS (
Ω)
1 10 100 1000
G = +20V/V29
26
23
20
17
14
FREQUENCY RESPONSE vs CAPACITIVE LOAD
Frequency (MHz)
Nor
mal
ized
Gai
n to
Cap
aciti
ve L
oad
(dB
)
1 10 100 1000
C = 22pF
C = 47pFC = 100pF
C = 10pFRS adjusted for capacitive load.
750Ω
RS
OPA847VI
VO
50Ω
1kΩCL
39.2Ω(1kΩ is optional.)
OPA847 7SBOS251C www.ti.com
TYPICAL CHARACTERISTICS: VS = ±5V (Cont.)TA = 25°C, G = +20V/V, RG = 39.2Ω, and RL = 100Ω, unless otherwise noted.
120
110
100
90
80
70
60
50
40
30
20
COMMON-MODE REJECTION RATIO AND POWER-SUPPLY REJECTION RATIO vs FREQUENCY
Frequency (Hz)
CM
RR
and
PS
RR
(dB
)
102 104 105103 106 107 108
CMRR +PSRR
–PSRR
120
100
80
60
40
20
0
–20
0
–30
–60
–90
–120
–150
–180
–210
OPEN-LOOP GAIN AND PHASE
Frequency (Hz)
Ope
n-Lo
op G
ain
(dB
)
Ope
n-Lo
op P
hase
(°)
102 104 105103 106 107 108 109
20log (AOL)
∠AOL
4
3
2
1
0
–1
–2
–3
–4
OUTPUT VOLTAGE AND CURRENT LIMITATIONS
IO (mA)
VO (
V)
–150 –100 –50 0 50 100 150
RL = 100Ω
RL = 25Ω
RL = 50Ω
10
1
0.1
0.01
0.001
CLOSED-LOOP OUTPUT IMPEDANCE vs FREQUENCY
Frequency (Hz)
Out
put I
mpe
danc
e (Ω
)
103 104 105 106 107 108
G = +20V/V
750Ω
OPA847
ZO
VDIS
39.2Ω
10
8
6
4
2
0
–2
–4
–6
–8
–10
0.5
0.4
0.3
0.2
0.1
0
–0.1
–0.2
–0.3
–0.4
–0.5
NONINVERTING OVERDRIVE RECOVERY
Time (40ns/div)
Out
put V
olta
ge (
V)
Inpu
t Vol
tage
(m
V)
See Figure 1
G = +20V/VRL = 100Ω
OutputLeft Scale
InputRight Scale
10
8
6
4
2
0
–2
–4
–6
–8
–10
0.25
0.20
0.15
0.10
0.05
0
–0.05
–0.10
–0.15
–0.20
–0.25
INVERTING OVERDRIVE RECOVERY
Time (40ns/div)
Out
put V
olta
ge (
V)
Inpu
t Vol
tage
(m
V)
See Figure 2
G = –40V/VRG = 50Ω
RL = 100Ω
OutputLeft Scale
InputRight Scale
OPA8478SBOS251Cwww.ti.com
TYPICAL CHARACTERISTICS: VS = ±5V (Cont.)TA = 25°C, G = +20V/V, RG = 39.2Ω, and RL = 100Ω, unless otherwise noted.
0.25
0.20
0.15
0.10
0.05
0
–0.05
–0.10
–0.15
–0.20
–0.25
SETTLING TIME
Time (ns)
Per
cent
of F
inal
Val
ue (
%)
0 5 10 15 20 25 30 35 40
G = +20V/VRL = 100Ω
VO = 2V Step
See Figure 1
89
86
83
80
77
74
71
PHOTODIODE TRANSIMPEDANCEFREQUENCY RESPONSE
Frequency (MHz)
Tra
nsim
peda
nce
Gai
n (d
BΩ
)
1 10 100
CD = 100pF
RF = 20kΩCF Adjusted
CD = 50pF
CD = 20pF
CD = 10pF[20log 20kΩ]
20kΩ
OPA84720kΩ VO
CDIODE[CD]
CF
IO
0.01µF
0.2
0.1
0
–0.1
–0.2
25.0
12.5
0
–12.5
–25.0
TYPICAL DC DRIFT OVER TEMPERATURE
Ambient Temperature (°C)
Inpu
t Offs
et V
olta
ge (
mV
)
Inpu
t Bia
s an
d O
ffset
Cur
rent
(µ
A)
–50 –25 0 25 50 75 100 125
100 x IOS
VIO
Ib
100
90
80
70
60
50
20
18
16
14
12
10
SUPPLY AND OUTPUT CURRENT vs TEMPERATURE
Ambient Temperature (°C)
Out
put C
urre
nt (
mA
)
Sup
ply
Cur
rent
(m
A)
–50 –25 0 25 50 75 100 125
Sourcing Output Current
Supply Current
Sinking Output Current
5
4
3
2
1
0
–1
–2
–3
–4
–5
COMMON-MODE INPUT RANGE AND OUTPUT SWINGvs SUPPLY VOLTAGE
Supply Voltage (±V)
Vol
tage
Ran
ge (
V)
2.50
2.75
3.00
3.25
3.50
3.75
4.00
4.25
4.50
4.75
5.00
5.25
5.50
5.75
6.00
Positive Output
RL = 100Ω
Negative Output
Negative Input
Positive Input
107
106
105
104
103
102
COMMON-MODE AND DIFFERENTIALINPUT IMPEDANCE
Frequency (Hz)
Inpu
t Im
peda
nce
(Ω)
102 104 105103 106 107 108
Common-Mode(2.3MΩ, DC)
Differential(2.7kΩ, DC)
OPA847 9SBOS251C www.ti.com
TYPICAL CHARACTERISTICS: VS = ±5VTA = 25°C, GD = 40V/V, RG = 50Ω, and RL = 400Ω, unless otherwise noted.
The OPA847 provides a unique combination of a very lowinput voltage noise along with a very low distortion outputstage to give one of the highest dynamic range op ampsavailable. Its very high gain bandwidth product (GBP) can beused to either deliver high signal bandwidths at high gains, orto deliver very low distortion signals at moderate frequenciesand lower gains. To achieve the full performance of theOPA847, careful attention to PC board layout and compo-nent selection is required, as discussed in the followingsections of this data sheet.
Figure 1 shows the noninverting gain of a +20V/V circuit usedas the basis for most of the Typical Characteristics. Most ofthe curves are characterized using signal sources with a50Ω driving impedance and with measurement equipmentpresenting a 50Ω load impedance. In Figure 1, the 50Ωshunt resistor at the VI terminal matches the source imped-ance of the test generator, while the 50Ω series resistor atthe VO terminal provides a matching resistor for the mea-surement equipment load. Generally, data sheet voltageswing specifications are at the output pin (VO in Figure 1)while output power specifications are at the matched 50Ωload. The total 100Ω load at the output combined with the790Ω total feedback network load presents the OPA847 withan effective output load of 89Ω for the circuit of Figure 1.
Voltage-feedback op amps, unlike current-feedback designs,can use a wide range of resistor values to set their gain. Thecircuit of Figure 1, and the specifications at other gains, use anRG set to 39.2Ωand RF adjusted to get the desired gain. Usingthis guideline ensures that the noise added at the output dueto the Johnson noise of the resistors does not significantlyincrease the total over that due to the 0.85nV/√Hz input
voltage noise for the op amp itself. This RG is suggested asa good starting point for design. Other values are certainlyacceptable, if required by the design.
WIDEBAND, INVERTING GAIN OPERATION
There can be significant benefits to operating the OPA847 asan inverting amplifier. This is particularly true when a matchedinput impedance is required. Figure 2 shows the invertinggain of a –40V/V circuit used as a starting point for theTypical Characteristics showing inverting mode performance.
Driving this circuit from a 50Ω source, and constraining the gainresistor (RG) to equal 50Ω, gives both a signal bandwidth and a noiseadvantage. RG, in this case, acts as both the input termination resistorand the gain setting resistor for the circuit. Although the signal gainfor the circuit of Figure 2 is double that for Figure 1, their noise gainsare nearly equal when the 50Ω source resistor is included. This hasthe interesting effect of approximately doubling the equivalent GBP forthe amplifier. This can be seen by observing that the gain of –40bandwidth of 240MHz shown in the Typical Characteristics implies again bandwidth product of 9.6GHz, giving a far higher bandwidth ata gain of –40 than at a gain of +40. While the signal gain from RG tothe output is –40, the noise gain for bandwidth setting purposes is1 + RF/(2 • RG). In the case of a –40V/V gain, using an RG = RS =50Ω gives a noise gain = 1 + 2kΩ/100Ω = 21. This inverting gain of–40V/V therefore has a frequency response that more closelymatches the gain of a +20 frequency response.
If the signal source is actually the low impedance output ofanother amplifier, RG should be increased to be greater thanthe minimum value allowed at the output for that amplifierand RF adjusted to get the desired gain. It is critical for stableoperation of the OPA847 that this driving amplifier show avery low output impedance through frequencies exceedingthe expected closed-loop bandwidth for the OPA847.
WIDEBAND, HIGH SENSITIVITY,TRANSIMPEDANCE DESIGN
OPA847
+5V
–5V–VS
+VS
50ΩVO
VDIS
VI 50Ω
+0.1µF
+6.8µF
6.8µF
RG39.2Ω
RF750Ω
50Ω Source
50Ω Load
0.1µF
FIGURE 1. Noninverting G = +20 Specification and Test Circuit. FIGURE 2. Noninverting G = –40 Specification and Test Circuit.
OPA847
+5V
–5V
+VS
–VS
95.3Ω50ΩVO
VI
+6.8µF0.1µF
+6.8µF0.1µF
0.01µF
RF2kΩ
RG50Ω
50Ω Source
50Ω LoadVDIS
OPA847 11SBOS251C www.ti.com
The high GBP and low input voltage and current noise for theOPA847 make it an ideal wideband transimpedance ampli-fier for low to moderate transimpedance gains. Very hightransimpedance gains (> 100kΩ) will benefit from the lowinput noise current of a JFET input op amp such as theOPA657. Unity-gain stability in the op amp is not required forapplication as a transimpedance amplifier. Figure 3 showsone possible transimpedance design example that would beparticularly suitable for the 155Mbit data rate of an OC-3receiver. Designs that require high bandwidth from a largearea detector with relatively low transimpedance gain willbenefit from the low input voltage noise for the OPA847. Theamplifier’s input voltage noise is peaked up over frequencyby the diode source capacitance, and can (in many cases)become the limiting factor to input sensitivity. The key ele-ments to the design are the expected diode capacitance (CD)with the reverse bias voltage (–VB) applied, the desiredtransimpedance gain (RF), and the GBP for the OPA847(3900MHz). With these three variables set (including theparasitic input capacitance for the OPA847 added to CD), thefeedback capacitor value (CF) can be set to control thefrequency response.
To achieve a maximally flat 2nd-order Butterworth frequencyresponse, set the feedback pole as shown in Equation 1.
Equation 2 gives the approximate –3dB bandwidth thatresults if CF is set using Equation 1.
( )HzCR2
GBPfDF
dB3 π=− (2)
The example of Figure 3 gives approximately 104MHz flatbandwidth using the 0.18pF feedback compensation capaci-tor. This bandwidth easily supports an OC-3 receiver withexceptional sensitivity.
If the total output noise is bandlimited to a frequency lessthan the feedback pole frequency, a very simple expressionfor the equivalent input noise current is shown as Equation 3.
( )3FC2e
Re
RkT4ii
2DN
2
F
N
F
2EQ N
π+
++= (3)
where:
iEQ = Equivalent input noise current if the output noise isbandlimited to F < 1/(2πRFCF)iN = Input current noise for the op amp inverting inputeN = Input voltage noise for the op amp
CD = Total Inverting Node Capacitancef = Bandlimiting frequency in Hz (usually a post filter priorto further signal processing)
Evaluating this expression up to the feedback pole fre-quency at 74MHz for the circuit of Figure 3 gives an equiva-lent input noise current of 3.0pA/√Hz . This is slightly higherthan the 2.5pA/√Hz input current noise for the op amp. Thistotal equivalent input current noise is slightly increased bythe last term in the equivalent input noise expression. It isessential in this case to use a low-voltage noise op amp. Forexample, if a slightly higher input noise voltage, but other-wise identical, op amp were used instead of the OPA847 inthis application (say 2.0nV/√Hz), the total input referredcurrent noise would increase to 3.7pA/√Hz . Low input volt-age noise is required for the best sensitivity in these widebandtransimpedance applications. This is often unspecified fordedicated transimpedance amplifiers with a total outputnoise for a specified source capacitance given instead. It isthe relatively high input voltage noise for those componentsthat cause higher than expected output noise if the sourcecapacitance is higher than specified.
The output DC error for the circuit of Figure 3 is minimizedby including a 12kΩ to ground on the noninverting input.This reduces the contribution of input bias current errors (fortotal output offset voltage) to the offset current times thefeedback resistor. To minimize the output noise contributionof this resistor, 0.01µF and 100pF capacitors are included inparallel. Worst-case output DC error for the circuit of Figure3 at 25°C is:
Adding the common-mode and differential mode input ca-pacitance (1.2 + 2.5)pF to the 1pF diode source capacitanceof Figure 3, and targeting a 12kΩ transimpedance gainusing the 3900MHz GBP for the OPA847 requires a feed-back pole set to 74MHz to get a nominal Butterworth fre-quency response design. This requires a total feedbackcapacitance of 0.18pF. That total is shown in Figure 3, butrecall that typical surface-mount resistors have a parasiticcapacitance of 0.2pF, leaving no external capacitor requiredfor this design.
FIGURE 3. Wideband, High Sensitivity, OC-3 TransimpedanceAmplifier.
RF12kΩ
12kΩ0.1µF100pF
Power-supplydecoupling not shown.
λ
OPA847
+5V
–5V
–VB
CF0.18pF
1pFPhotodiode
VDIS
OPA84712SBOS251Cwww.ti.com
Even with bias current cancellation, the output DC errors aredominated in this example by the offset current term. Im-proved output DC precision and drift are possible, particu-larly at higher transimpedance gains, using the JFET inputOPA657. The JFET input removes the input bias currentfrom the error equation (eliminating the need for the resistorto ground on the noninverting input), leaving only the inputoffset voltage and drift as an output DC error term.
Included in the Typical Characteristics are transimpedancefrequency response curves for a fixed 20kΩ gain overvarious detector diode capacitance settings. These curvesare repeated in Figure 4, along with the test circuit. As thephotodiode capacitance changes, the feedback capacitormust change to maintain a stable and flat frequency re-sponse. Using Equation 1, CF is adjusted to give theButterworth frequency responses shown in Figure 4.
Considering only the noise gain (which is the same as thenoninverting signal gain) for the circuit of Figure 5, the low-frequency noise gain (NG1) is set by the resistor ratio, whilethe high-frequency noise gain (NG2) is set by the capacitorratio. The capacitor values set both the transition frequenciesand the high-frequency noise gain. If the high-frequencynoise gain, determined by NG2 = 1 + CS/CF, is set to a valuegreater than the recommended minimum stable gain for theop amp, and the noise gain pole (set by 1/RFCF) is placedcorrectly, a very well controlled 2nd-order low-pass fre-quency response results.
LOW-GAIN COMPENSATION FOR IMPROVED SFDR
Where a low gain is desired, and inverting operation isacceptable, a new external compensation technique can beused to retain the full slew rate and noise benefits of theOPA847, while giving increased loop gain and the associ-ated distortion improvements offered by a non-unity-gainstable op amp. This technique shapes the loop gain for goodstability, while giving an easily controlled 2nd-order low-passfrequency response. This technique is used for the circuit onthe front page of this data sheet in a differential configurationto achieve extremely low distortion through high frequencies(< –90dBc through 30MHz). The amplifier portion of thiscircuit is set up for a differential gain of 8.5V/V from adifferential input signal to the output. Using the input trans-former shown improves the noise figure and translates froma single-ended to a differential signal. If the source is differ-ential already, it can be fed directly into the gain settingresistors. To set the compensation capacitors (CS and CF),consider the half circuit of Figure 5, where the 50Ω sourceis reflected through the 1:2 transformer, then cut in half, andgrounded to give a total impedance to the AC ground for thecircuit on the front page equal to 200Ω.
FIGURE 4. Transimpedance Bandwidth vs CD.
89
86
83
80
77
74
71
PHOTODIODE TRANSIMPEDANCEFREQUENCY RESPONSE
Frequency (MHz)
Tra
nsim
peda
nce
Gai
n (d
BΩ
)
1 10 100
CD = 100pF
RF = 20kΩCF Adjusted
CD = 50pF
CD = 20pF
CD = 10pF
20kΩ
OPA84720kΩ VO
CD CF
IO
0.01µF
[20 log(20kΩ)]
To choose the values for both CS and CF, two parameters and onlythree equations need to be solved. The first parameter is the targethigh-frequency noise gain (NG2), which should be greater than theminimum stable gain for the OPA847. Here, a target of NG2 = 24 isused. The second parameter is the desired low-frequency signalgain, which also sets the low-frequency noise gain (NG1). To simplifythis discussion, we will target a maximally flat 2nd-order low-passButterworth frequency response (Q = 0.707). The signal gain shownin Figure 5 sets the low-frequency noise gain to NG1 = 1 + RF/RG
(= 5.25 in this example). Then, using only these two gains and theGBP for the OPA847 (3900MHz), the key frequency in the compen-sation is set by Equation 4.
−−
−=
2
1
2
121
O NGNG21
NGNG1
NGGBPZ (4)
Physically, this ZO (4.4MHz for the values shown above) isset by 1/(2πRF(CF + CS)) and is the frequency at which therising portion of the noise gain would intersect the unity gainif projected back to a 0dB gain. The actual zero in the noisegain occurs at NG1 • ZO and the pole in the noise gain occursat NG2 • ZO. That pole is physically set by 1/(RFCF). SinceGBP is expressed in Hz, multiply ZO by 2π and use to get CF
by solving Equation 5.
CF =1
2πRFZONG2= 1.76pF( ) (5)
FIGURE 5. Broadband, Low-Inverting Gain ExternalCompensation.
RF850Ω
CS39pF
OPA847
+5V
–5V
VO
VI
CF1.7pF
RG200Ω
VDIS
OPA847 13SBOS251C www.ti.com
Finally, since CS and CF set the high-frequency noise gain,determine CS using Equation 6 (solving for CS by using NG2 = 24):
F2S C)1NG(C −= (6)
which gives CS = 40.6pF.
Both of these calculated values have been reduced slightlyin Figure 5 to account for parasitics. The resulting closed-loop bandwidth is approximately equal to Equation 7.
GBPZf OdB3– •≅ (7)
For the values shown in Figure 5, f–3dB is approximately131MHz. This is less than that predicted by simply dividingthe GBP product by NG1. The compensation network controlsthe bandwidth to a lower value, while providing the full slewrate at the output and an exceptional distortion performancedue to increased loop gain at frequencies below NG1 • ZO.
Using this low-gain inverting compensation, along with thedifferential structure for the circuit shown on the front page ofthis data sheet, gives a significant reduction in harmonicdistortion. The measured distortion at 2VPP output does notrise above –95dB until frequencies > 20MHz are applied.
The Typical Characteristics show the exceptional bandwidth con-trol possible using this technique at low inverting gains. Figure 6repeats the measured results with the test circuit shown.
The compensation capacitors, CS and CF, are set by targeting ahigh-frequency noise gain of 21 and using equations 4 through6. This approach allows relatively low inverting gain applicationsto use the full slew rate and low input noise of the OPA847.
LOW-NOISE FIGURE,HIGH DYNAMIC RANGE AMPLIFIER
The low input noise voltage of the OPA847 and its very high2-tone, 3rd-order intermodulation intercept can be used togood advantage as a fixed-gain amplifier. While input noise
figures in the 10dB range (for a matched 50Ω input) areeasily achieved with just the OPA847, Figure 7 illustrates atechnique to reduce the noise figure even further, whileproviding a broadband, high-gain HF amplifier stage usingtwo stages of the OPA847.
FIGURE 6. Low-Gain Inverting Performance.
1
0
–1
–2
–3
–4
–5
–6
–7
–8
–9
LOW GAIN INVERTING BANDWIDTH
Frequency (MHz)
Nor
mal
ized
Gai
n (1
dB)
1 10 100 1000
G = –8
G = –4G = –2
G = –1
VO = 0.2VPP
RF750Ω
CS
OPA847 VO
VI
CF
RG
0Ω Source
VDIS
+5V
–5V
FIGURE 7. Very High Dynamic Range HF Amplifier.
OPA847
10pF
PI
1.6pF
6.19kΩ
750Ω 1.5kΩ
200Ω
30.1Ω
420Ω
4.3dBNoiseFigure
1:250Ω Source
PO
> 55dBm interceptto 30MHz
OPA847
46pF
Input matchset by this
feedback path
Overall Gain = 35.6dBPO
PI
+5V
–5V+5V
–5V
OPA84714SBOS251Cwww.ti.com
This circuit uses two stages of forward gain with an overallfeedback loop to set the input impedance match. The inputtransformer provides both a noiseless voltage gain and asignal inversion to retain an overall noninverting signal pathfrom PI to PO. The second amplifier stage is inverting toprovide the correct feedback polarity through the 6.19kΩresistor. To achieve a 50Ω input match at the primary of the1:2 transformer, the secondary must see a 200Ω loadimpedance. At higher frequencies, the match is provided bythe 200Ω resistor in series with 10pF. The low-noise figure(4.3dB) for this circuit is achieved by using the transformer,the low-voltage noise OPA847, and the input match set bythe feedback at lower frequencies intended for this HFdesign. The 1st-stage amplifier provides a gain of +15V/V.The very high SFDR is provided by operating the outputstage at a low signal gain of –2 and using the invertingcompensation technique to shape the noise gain to hold itstable. This 2nd-stage compensation is set to intentionallybandlimit the overall response to approximately 100MHz. Foroutput loads > 400Ω, this circuit can give a 2-tone SFDR thatexceeds 90dB through 30MHz. In narrowband applications,the 3rd-order intercept exceeds 55dBm. Besides offering avery high dynamic range, this circuit improves on standardHF amplifiers by offering a precisely controlled gain and avery flexible output interface capability.
NONINVERTING GAIN FLATNESS COMPENSATION
Decreasing the operating gain from the nominal design point of+20 decreases the phase margin. This increases Q for theclosed-loop poles, peaks up the frequency response, andextends the bandwidth. A peaked frequency response showsovershoot and ringing in the pulse response, as well as higherintegrated output noise. When operating the OPA847 at anoninverting gain < +12V/V, increased peaking and possiblesustained oscillations may result. However, operation at lowgains may be desirable to take advantage of the higher slewrate and exceptional DC precision of the OPA847. Numerousexternal compensation techniques are suggested for operatinga high-gain op amp at low gains. Most of these give zero/polepairs in the closed-loop response that cause long term settlingtails in the pulse response and/or phase nonlinearity in thefrequency response.
Figure 8 shows a resistor based compensation techniquethat allows the flatness at low noninverting signal gains to becontrolled separately from the signal gain. This approachretains the full slew rate to the output but gives up some ofthe low-noise benefit of the OPA847. Including the effect ofthe total source impedance (25Ω in Figure 8), tuning resistorR1 can be set using Equation 8.
R1 =RF + RSAVNG − AV
(8)
where,
AV = desired signal gain (+12V/V in Figure 8)
NG = target noise gain (adjusted in Figure 9)
RS = total source impedance
The effect of this noninverting gain flatness tune is shown inFigure 9. At an NG of 12, R1 is removed and only RF and RG
are present in Figure 8. The peaking is typically 4.5dB, asshown in the small-signal frequency response curves versusgain curves at this setting. As R1 is decreased, the operatingnoise gain (NG) increases, reducing the peaking and band-width until the nominal design point of +20 noise gain givesa non-peaked response.
FIGURE 8. Low Noninverting Gain Flatness Trim.
RF750Ω
RG66.5Ω
R150Ω OPA847
+5V
–5V
VI
VO50Ω
50ΩVDIS
FIGURE 9. Frequency Response Flatness with ExternalTuning Resistor.
0.5
0.4
0.3
0.2
0.1
0
–0.1
–0.2
–0.3
–0.4
–0.5
NONINVERTING GAIN FLATNESS TUNE
Frequency (MHz)
Dev
iatio
n fr
om 2
1.58
dB G
ain
(0.1
dB)
1 10 100 1000
NG = 12
NG = 14
NG = 20
NG = 18
NG = 16
VO = 200mVPPAV = +12V/VNG = Noise Gain
DIFFERENTIAL OPERATION
Operating two OPA847 amplifiers in a differential invertingconfiguration can further suppress even-order harmonic terms.The Typical Characteristics show measured performance forthis condition. These measurements were done at the relativelyhigh gain of 40V/V. Even lower distortion is possible operatingat lower gains using the external inverting compensation tech-niques, as discussed previously. For the distortion data pre-sented in Figure 10, the output swing is increased to 4VPP into400Ω to allow direct comparison to the single-channel data at2VPP into 200Ω. Comparing the 2nd- and 3rd-harmonics at20MHz in Figure 10 to the gain of +20, 2VPP, 200Ω data, showsthe 2nd-harmonic is reduced to –76dBc (from –67dBc) and the3rd-harmonic is reduced from –80dBc to –85dBc. Using the two
OPA847 15SBOS251C www.ti.com
OPA847 can support this mode of operation down to a singlesupply of +5V and up to a single supply of +12V. If shutdownis desired for single-supply operation, it is important torealize that the shutdown pin is referenced from the positivesupply pin. Open collector (drain) interfaces are suggestedfor single-supply operation above +5V.
DESIGN-IN TOOLSDEMONSTRATION BOARDS
Two PC boards are available to assist in the initial evaluationof circuit performance using the OPA847 in its two packagestyles. Both of these are available, free, as unpopulated PCboards delivered with descriptive documentation. The sum-mary information for these boards is shown in Table I.Contact your sales representative or go to the TI web site(www.ti.com) to request these evaluation boards.
amplifiers in this configuration has significantly reduced the2nd-harmonic, even after doubling the output voltage swing (to 4VPP)and the gain (to 40V/V).
SINGLE-SUPPLY OPERATION
The OPA847 can be operated from a single power supply ifsystem constraints require it. Operation from a single +5V to+12V supply is possible with minimal change in AC perfor-mance. The Typical Characteristics show the input andoutput voltage ranges for a bipolar supply range from ±2.5Vto ±6.0V. The Common-Mode Input Range and Output Swingvs Supply Voltage curve shows that the required headroomon both the input and output pins remains at approximately1.5V over this entire range. On a single +5V supply, forinstance, this means the noninverting input should remaincentered at +2.5V ± 1V, as should the output pin. Figure 11shows an example application biasing the noninverting inputat mid-supply and running an AC-coupled input to the invert-ing gain path. Since the gain resistor is blocked off for DC,the bias point on the noninverting input appears at the output,centering up the output as well as on the power supply. The
Computer simulation of circuit performance using SPICE isoften a quick way to analyze the performance of the OPA847in its intended application. This is particularly true for videoand RF amplifier circuits where parasitic capacitance andinductance can play a major role in circuit performance. ASPICE model for the OPA847 is available through the TI website (www.ti.com). These models do a good job of predictingsmall-signal AC and transient performance under a widevariety of operating conditions. They do not do as well inpredicting the harmonic distortion characteristics. Thesemodels do not attempt to distinguish between the packagetypes in their small-signal AC performance.
OPERATING SUGGESTIONSSETTING RESISTOR VALUES TO MINIMIZE NOISE
The OPA847 provides a very low input noise voltage whilerequiring a low 18.1mA of quiescent current. To take fulladvantage of this low input noise, careful attention to the otherpossible noise contributors is required. See Figure 12 for theop amp noise analysis model with all the noise terms included.In this model, all the noise terms are taken to be noise voltageor current density terms in either nV/√Hz or pA/√Hz.
The total output spot noise voltage is computed as thesquare root of the squared contributing terms to the outputnoise power. This computation adds all the contributing noisepowers at the output by superposition, then takes the square
OPA84716SBOS251Cwww.ti.com
practice, this only holds true when the phase margin ap-proaches 90°, as it does in high-gain configurations. At lowgains (increased feedback factors), most high-speed ampli-fiers exhibit a more complex response with lower phasemargin. The OPA847 is compensated to give a maximally flat2nd-order Butterworth closed-loop response at a noninvertinggain of +20 (see Figure 1). This results in a typical gain of+20 bandwidth of 350MHz, far exceeding that predicted bydividing the 3900MHz GBP by 20. Increasing the gain causesthe phase margin to approach 90° and the bandwidth to moreclosely approach the predicted value of (GBP/NG). At a gainof +50, the OPA847 very nearly matches the 78MHz band-width predicted using the simple formula and the typical GBPof 3900MHz.
Inverting operation offers some interesting opportunities toincrease the available GBP. When the source impedance ismatched by the gain resistor (see Figure 2), the signal gainis (1 + RF/RG), while the noise gain for bandwidth purposesis (1 + RF/2RG). This cuts the noise gain almost in half,increasing the minimum operating gain for inverting opera-tion under these condition to –22 and the equivalent gainbandwidth product to > 7.8GHz.
DRIVING CAPACITIVE LOADS
One of the most demanding, and yet very common, loadconditions for an op amp is capacitive loading. Often, thecapacitive load is the input of an ADC, including additionalexternal capacitance that may be recommended to improveADC linearity. A high-speed, high open-loop gain amplifierlike the OPA847 can be very susceptible to decreasedstability and may give closed-loop response peaking when acapacitive load is placed directly on the output pin. When theamplifier’s open-loop output resistance is considered, thiscapacitive load introduces an additional pole in the signalpath that can decrease the phase margin. Several externalsolutions to this problem are suggested. When the primaryconsiderations are frequency response flatness, pulse re-sponse fidelity, and/or distortion, the simplest and mosteffective solution is to isolate the capacitive load from thefeedback loop by inserting a series isolation resistor betweenthe amplifier output and the capacitive load. This does noteliminate the pole from the loop response, but rather shifts itand adds a zero at a higher frequency. The additional zeroacts to cancel the phase lag from the capacitive load pole,thus increasing the phase margin and improving stability.
The Typical Characteristics help the designer pick a recom-mended RS versus capacitive load. The resulting frequencyresponse curves show a flat response for several selectedcapacitive loads and recommended RS combinations. Para-sitic capacitive loads greater than 2pF can begin to degradethe performance of the OPA847. Long PC board traces,unmatched cables, and connections to multiple devices caneasily cause this value to be exceeded. Always consider thiseffect carefully and add the recommended series resistor asclose as possible to the OPA847 output pin (see the BoardLayout section).
root to get back to a spot noise voltage. Equation 9 showsthe general form for this output noise voltage using the termsillustrated in Figure 11.
Dividing this expression by the noise gain (NG = 1 + RF/RG)gives the equivalent input referred spot noise voltage at thenoninverting input, as shown in Equation 10.
(10)
( )NGkTR4
NGRIkTR4RIEE F2FBI
S2
SBN2NIN +
+++=
Putting high resistor values into Equation 10 can quicklydominate the total equivalent input referred noise. A 45Ωsource impedance on the noninverting input adds a Johnsonvoltage noise term equal to the amplifier’s voltage noise byitself. As a simplifying constraint, set RG = RS in Equation 10and assume an RS/2 source impedance at the noninvertinginput, where RS is the signal source impedance and anothermatching RS to ground is at the noninverting input. Thisresults in Equation 11, where NG > 12 is assumed to furthersimplify the expression.
( )
++=2R3kT4RI
45EE S2
SB2NIN (11)
Evaluating this expression for RS = 50Ω gives a total equiva-lent input noise of 1.4nV/√Hz . Note that at these highergains, the simplified input referred spot noise expression ofEquation 11 does not include the gain. This is a goodapproximation for NG > 12, as is typically required by stabilityconsiderations.
FREQUENCY RESPONSE CONTROL
Voltage-feedback op amps exhibit decreasing closed-loopbandwidth as the signal gain is increased. In theory, thisrelationship is described by the Gain Bandwidth Product(GBP) shown in the Electrical Characteristics. Ideally, divid-ing GBP by the noninverting signal gain (also called theNoise Gain, or NG) predicts the closed-loop bandwidth. In
FIGURE 12. Op Amp Noise Analysis Model.
4kTRG
RG
RF
RS
OPA847
IBI
EO
IBN
4kT = 1.6E – 20Jat 290°K
ERS
ENI
√4kTRS
√4kTRF
OPA847 17SBOS251C www.ti.com
The criterion for setting the RS resistor is a maximum band-width, flat frequency response at the load. For the OPA847operating in a gain of +20, the frequency response at theoutput pin is very flat to begin with, allowing relatively smallvalues of RS to be used for low capacitive loads. As thesignal gain is increased, the unloaded phase margin alsoincreases. Driving capacitive loads at higher gains requireslower RS values than those shown for a gain of +20.
DISTORTION PERFORMANCE
The OPA847 is capable of delivering an exceptionally lowdistortion signal at high frequencies over a wide range ofgains. The distortion plots in the Typical Characteristics showthe typical distortion under a wide variety of conditions. Mostof these plots are limited to a 110dB dynamic range. TheOPA847’s distortion driving a 200Ω load does not rise above–90dBc until either the signal level exceeds 2.0VPP and/orthe fundamental frequency exceeds 5MHz. Distortion in theaudio band is < –130dBc.
Generally, until the fundamental signal reaches very highfrequencies or powers, the 2nd-harmonic dominates the dis-tortion with a negligible 3rd-harmonic component. Focusingthen on the 2nd-harmonic, increasing the load impedanceimproves distortion directly. Remember that the total loadincludes the feedback network—in the noninverting configura-tion this is the sum of RF + RG, while in the invertingconfiguration this is only RF (see Figure 2). Increasing theoutput voltage swing increases harmonic distortion directly. A6dB increase in output swing generally increases the 2nd-harmonic 12dB and the 3rd-harmonic 18dB. Increasing thesignal gain also increases the 2nd-harmonic distortion. Finally,the distortion increases as the fundamental frequency in-creases due to the rolloff in the loop gain with frequency.Conversely, the distortion improves going to lower frequenciesdown to the dominant open-loop pole at approximately 80kHz.
The OPA847 has an extremely low 3rd-order harmonicdistortion. This also gives a high 2-tone 3rd-orderintermodulation intercept, as shown in the Typical Character-istics. This intercept curve is defined at the 50Ω load whendriven through a 50Ω matching resistor to allow directcomparisons to RF devices. This matching network attenu-ates the voltage swing from the output pin to the load by 6dB.If the OPA847 drives directly into the input of a high-impedance device, such as an ADC, this 6dB attenuation isnot taken. Under these conditions, the intercept as reportedin the Typical Characteristics increases by a minimum of6dBm. The intercept is used to predict the intermodulationspurious power levels for two closely spaced frequencies. Ifthe two test frequencies, f1 and f2, are specified in terms ofaverage and delta frequency, fO = (f1 + f2)/2 and ∆F = |f2 – f1| /2,the two 3rd-order, close-in spurious tones appear at fO ± 3 • ∆F.The difference between the two equal test-tone power levelsand these intermodulation spurious power levels is given by∆dBc = 2(IM3 – PO), where IM3 is the intercept taken from theTypical Characteristics and PO is the power level in dBm atthe 50Ω load for one of the two closely spaced test frequen-cies. For instance, at 30MHz, the OPA847 at a gain of +20has an intercept of 34dBm at a matched 50Ω load.
If the full envelope of the two frequencies needs to be 2VPP,this requires each tone to be 4dBm. The 3rd-orderintermodulation spurious tones will then be 2(34 – 4) =60dBc below the test-tone power level (–56dBm). If thissame 2VPP 2-tone envelope is delivered directly into theinput of an ADC without the matching loss or the loading ofthe 50Ω network, the intercept would increase to at least40dBm. With the same signal and gain conditions, but nowdriving directly into a light load, the spurious tones will thenbe at least 2(40 – 4) = 72dBc below the 4dBm test-tonepower levels centered on 30MHz. Tests have shown thatthey are in fact much lower due to the lighter loadingpresented by most ADCs.
DC ACCURACY AND OFFSET CONTROL
The OPA847 can provide excellent DC signal accuracy dueto its high open-loop gain, high common-mode rejection, highpower-supply rejection, and low input offset voltage and biascurrent offset errors. To take full advantage of its low ±0.5mVinput offset voltage, careful attention to the input bias currentcancellation is also required. The low-noise input stage forthe OPA847 has a relatively high input bias current (19µAtypical into the pins), but with a very close match between thetwo input currents—typically ±100nA input offset current.Figures 13 and 14 show typical distributions of input offsetvoltage and current for the OPA847.
FIGURE 13. Input Offset Voltage Distribution in µV.
1200
1000
800
600
400
200
0
µV
Cou
nt
< –
600
< –
540
< –
480
< –
420
< –
360
< –
300
< –
240
< –
180
< –
120
< –
60 0<
60
< 1
20<
180
< 2
40<
300
< 3
60<
420
< 4
80<
540
< 6
00>
600
Mean = 48µVStandard Deviation = 110µV
Total Count = 4040
FIGURE 14. Input Offset Current Distribution in nA.
900
800
700
600
500
400
300
200
100
0
nA
Cou
nt
< –
600
< –
540
< –
480
< –
420
< –
360
< –
300
< –
240
< –
180
< –
120
< –
60 0<
60
< 1
20<
180
< 2
40<
300
< 3
60<
420
< 4
80<
540
< 6
00>
600
Mean = 50nAStandard Deviation = 120nATotal Count = 4040
OPA84718SBOS251Cwww.ti.com
The total output offset voltage can be considerably reducedby matching the source impedances looking out of the twoinputs. For example, one way to add bias current cancella-tion to the circuit of Figure 1 is to insert a 12.1Ω series resistorinto the noninverting input from the 50Ω terminating resistor.When the 50Ω source resistor is DC coupled, this increasesthe source impedance for the noninverting input bias currentto 37.1Ω. Since this is now equal to the impedance lookingout of the inverting input (RF || RG) for Figure 1, the circuitcancels the gains for the bias currents to the output, leavingonly the offset current times the feedback resistor as aresidual DC error term at the output. Using the 750Ωfeedback resistor, this output error is now less than±0.85µA • 750Ω = ±640µV over the full temperature range forthe circuit of Figure 1, with a 12.1Ω resistor added as de-scribed. The output DC offset is then dominated by theinput offset voltage multiplied by the signal gain. For thecircuit of Figure 1, this is a worst-case output DC offset of±0.6mV • 20 = ±12mV over the full temperature range.
A fine-scale output offset null, or DC operating point adjust-ment, is sometimes required. Numerous techniques areavailable for introducing a DC offset control into an op ampcircuit. Most of these techniques eventually reduce to settingup a DC current through the feedback resistor. One keyconsideration to selecting a technique is to ensure that it hasa minimal impact on the desired signal path frequencyresponse. If the signal path is intended to be noninverting,the offset control is best applied as an inverting summingsignal to avoid interaction with the signal source. If the signalpath is intended to be inverting, applying the offset control tothe noninverting input can be considered. For a DC-coupledinverting input signal, this DC offset signal sets up a DCcurrent back into the source that must be considered. Anoffset adjustment placed on the inverting op amp input canalso change the noise gain and frequency response flatness.Figure 15 shows one example of an offset adjustment for aDC-coupled signal path that has minimum impact on thesignal frequency response.
In this case, the input is brought into an inverting gain resistorwith the DC adjustment as an additional current summedinto the inverting node. The resistor values setting this offsetadjustment are much larger than the signal path resistors.This ensures that this adjustment has minimal impact on theloop gain and, hence, the frequency response.
POWER SHUTDOWN OPERATION
The OPA847 provides an optional power shutdown featurethat can be used to reduce system power. If the VDIS controlpin is left unconnected, the OPA847 operates normally. Thisshutdown is intended only as a power saving feature. For-ward path isolation is very good for small signals. Largesignal isolation is not ensured. Using this feature to multiplextwo or more outputs together is not recommended. Largesignals applied to the shutdown output stages can turn onparasitic devices, degrading signal linearity for the desiredchannel.
Turn-on time is very quick from the shutdown condition,typically < 60ns. Turn-off time is strongly dependent on theexternal circuit configuration, but is typically 200ns for thecircuit of Figure 1. Using the OPA847 with higher externalresistor values, such has high-gain transimpedance circuits,slows the shutdown time since the time constants for theinternal nodes to discharge are longer.
To shutdown, the control pin must be asserted low. This logiccontrol is referenced to the positive supply, as shown in thesimplified circuit of Figure 16.
FIGURE 15. DC-Coupled, Inverting Gain of –20 with OutputOffset Adjustment.
RF1kΩ
±200mV Output Adjustment
Power-supply decouplingnot shown.
5kΩ
5kΩ
48Ω0.1µF
RG50Ω
VI
20kΩ100Ω
0.1µF
–5V
+5V
OPA847
+5V
–5V
VCC
VEE
VO
= – = –20V/VVO
VI
RF
RG
FIGURE 16. Simplified Shutdown Control Circuit.
17kΩ 120kΩ
8kΩ
ISControl –VS
+VS
VDIS
Q1
In normal operation, base current to Q1 is provided throughthe 120kΩ resistor, while the emitter current through the 8kΩresistor sets up a voltage drop that is inadequate to turn onthe two diodes in Q1’s emitter. As VDIS is pulled low,additional current is pulled through the 8kΩ resistor, even-tually turning on these two diodes ( ≈180µA). At this point,any further current pulled out of VDIS goes through thosediodes holding the emitter-base voltage of Q1 at approxi-mately 0V. This shuts off the collector current out of Q1,turning the amplifier off. The supply current in the shutdownmode is only that required to operate the circuit of Figure 16.
OPA847 19SBOS251C www.ti.com
The shutdown feature for the OPA847 is a positive-supplyreferenced, current-controlled interface. Open-collector (ordrain) interfaces are most effective, as long as the controllinglogic can sustain the resulting voltage (in open mode) thatappears at the VDIS pin. The VDIS pin voltage is one diodebelow the positive supply voltage applied to the OPA847 if thelogic voltage is open. For voltage output logic interfaces, theon/off voltage levels described in the Electrical Characteristicsapply only for a +5V supply. An open-drain interface isrecommended for a shutdown operation using a higherpositive supply and/or logic families with inadequate high-level voltage swings.
THERMAL ANALYSIS
The OPA847 does not require heatsinking or airflow in mostapplications. Maximum desired junction temperature sets themaximum allowed internal power dissipation, as describedhere. In no case should the maximum junction temperaturebe allowed to exceed 150°C.
Operating junction temperature (TJ) is given by TA + PD • θJA.The total internal power dissipation (PD) is the sum ofquiescent power (PDQ) and additional power dissipated in theoutput stage (PDL) to deliver load power. Quiescent power issimply the specified no-load supply current times the totalsupply voltage across the part. PDL depends on the requiredoutput signal and load but would, for a grounded resistiveload, be at a maximum when the output is fixed at a voltageequal to half either supply voltage (for equal bipolar sup-plies). Under this worst-case condition, PDL = VS
2/(4 • RL),where RL includes feedback network loading. This is theabsolute highest power that can be dissipated for a given RL.All actual applications dissipate less power in the outputstage.
Note that it is the power in the output stage and not into theload that determines internal power dissipation.
As a worst-case example, compute the maximum TJ using anOPA847IDBV (SOT23-6 package) in the circuit of Figure 1operating at the maximum specified ambient temperature of+85°C and driving a grounded 100Ω load. Maximum inter-nal power is:
PD = 10V • 18.9mA + 52/(4(100Ω || 789Ω)) = 259mW
Maximum TJ = +85°C + (0.26W • 150°C/W) = 124°C
All actual applications will operate at a lower junction tem-perature than the 124°C computed above. Compute youractual output stage power to get an accurate estimate ofmaximum junction temperature, or use the results shownhere as an absolute maximum.
BOARD LAYOUTAchieving optimum performance with a high-frequency am-plifier like the OPA847 requires careful attention to boardlayout parasitics and external component types. Recommen-dations that will optimize performance include:
a) Minimize parasitic capacitance to any AC ground for allof the signal I/O pins. Parasitic capacitance on the output andinverting input pins can cause instability: on the noninverting
input, it can react with the source impedance to causeunintentional bandlimiting. To reduce unwanted capaci-tance, create a window around the signal I/O pins in all of theground and power planes around these pins. Otherwise,ground and power planes should be unbroken elsewhere onthe board.
b) Minimize the distance (< 0.25") from the power-supplypins to high-frequency 0.1µF decoupling capacitors. At thedevice pins, the ground and power plane layout should notbe in close proximity to the signal I/O pins. Avoid narrowpower and ground traces to minimize inductance betweenthe pins and the decoupling capacitors. The power-supplyconnections should always be decoupled with these capaci-tors. Larger (2.2µF to 6.8µF) decoupling capacitors, effectiveat lower frequencies, should also be used on the main supplypins. These can be placed somewhat further from the deviceand can be shared among several devices in the same areaof the PC board.
c) Careful selection and placement of external compo-nents preserves the high-frequency performance of theOPA847. Use resistors that have low reactance at highfrequencies. Surface-mount resistors work best and allow atighter overall layout. Metal film and carbon compositionaxially leaded resistors can also provide good high-fre-quency performance. Again, keep their leads and PC boardtrace length as short as possible. Never use wirewound typeresistors in a high-frequency application. Since the output pinand inverting input pin are the most sensitive to parasiticcapacitance, always position the feedback and series outputresistor, if any, as close as possible to the output pin. Othernetwork components, such as noninverting input terminationresistors, should also be placed close to the package. Wheredouble-side component mounting is allowed, place the feed-back resistor directly under the package on the other side ofthe board between the output and inverting input pins. Evenwith a low parasitic capacitance shunting the external resis-tors, excessively high resistor values can create significanttime constants that can degrade performance. Good axialmetal film or surface-mount resistors have approximately0.2pF in shunt with the resistor. For resistor values > 2.0kΩ,this parasitic capacitance can add a pole and/or zero below400MHz that can effect circuit operation. Keep resistor val-ues as low as possible, consistent with load driving consid-erations. It has been suggested here that a good startingpoint for design would be to set RG to 39.2Ω. Doing thisautomatically keeps the resistor noise terms low, and mini-mizes the effect of their parasitic capacitance. Transimped-ance applications can use much higher resistor values. Thecompensation techniques described in this data sheet allowexcellent frequency response control, even with very highfeedback resistor values.
d) Connections to other wideband devices on the boardcan be made with short, direct traces or through onboardtransmission lines. For short connections, consider the traceand the input to the next device as a lumped capacitive load.Relatively wide traces (50mils to 100mils) should be used,preferably with ground and power planes opened up aroundthem. Estimate the total capacitive load and set RS from the
OPA84720SBOS251Cwww.ti.com
plot of Recommended RS vs Capacitive Load. Low parasiticcapacitive loads (< 4pF) may not need an RS, since theOPA847 is nominally compensated to operate with a 2pFparasitic load. Higher parasitic capacitive loads without anRS are allowed as the signal gain increases from +20V/V(increasing the unloaded phase margin). If a long trace isrequired, and the 6dB signal loss intrinsic to a doubly-terminated transmission line is acceptable, implement amatched impedance transmission line using microstrip orstripline techniques (consult an ECL design handbook formicrostrip and stripline layout techniques). A 50Ω environ-ment is normally not necessary onboard and, in fact, a higherimpedance environment improves distortion, as shown in thedistortion versus load plots. With a characteristic board traceimpedance defined based on board material and trace di-mensions, a matching series resistor into the trace from theoutput of the OPA847 is used, as well as a terminating shuntresistor at the input of the destination device. Rememberalso that the terminating impedance is the parallel combina-tion of the shunt resistor and the input impedance of thedestination device; this total effective impedance should beset to match the trace impedance. If the 6dB attenuation ofa doubly-terminated transmission line is unacceptable, along trace can be series-terminated at the source-end only.Treat the trace as a capacitive load in this case and set theseries resistor value as shown in the plot of RecommendedRS vs Capacitive Load. This does not preserve signal integ-rity as well as a doubly-terminated line. If the input imped-ance of the destination device is low, there will be somesignal attenuation due to the voltage divider formed by theseries output into the terminating impedance.
e) Socketing a high-speed part like the OPA847 is notrecommended. The additional lead length and pin-to-pincapacitance introduced by the socket can create an ex-
tremely troublesome parasitic network that can make italmost impossible to achieve a smooth, stable frequencyresponse. Best results are obtained by soldering the OPA847onto the board.
INPUT AND ESD PROTECTION
The OPA847 is built using a very high-speed complementarybipolar process. The internal junction breakdown voltages arerelatively low for these very small geometry devices. Thesebreakdowns are reflected in the Absolute Maximum Ratingstable. All device pins are protected with internal ESD protec-
tion diodes to the power supplies, as shown in Figure 17.
These diodes provide moderate protection to input overdrivevoltages above the supplies as well. The protection diodescan typically support 30mA continuous current. Where highercurrents are possible (e.g., in systems with ±15V supply partsdriving into the OPA847), current limiting series resistorsshould be added into the two inputs. Keep these resistorvalues as low as possible, since high values degrade bothnoise performance and frequency response.
FIGURE 17. Internal ESD Protection.
ExternalPin
+VCC
–VCC
InternalCircuitry
OPA847 21SBOS251C www.ti.com
PACKAGE DRAWINGS
D (R-PDSO-G**) PLASTIC SMALL-OUTLINE PACKAGE8 PINS SHOWN
8
0.197(5,00)
A MAX
A MIN(4,80)0.189 0.337
(8,55)
(8,75)0.344
14
0.386(9,80)
(10,00)0.394
16DIM
PINS **
4040047/E 09/01
0.069 (1,75) MAX
Seating Plane
0.004 (0,10)0.010 (0,25)
0.010 (0,25)
0.016 (0,40)0.044 (1,12)
0.244 (6,20)0.228 (5,80)
0.020 (0,51)0.014 (0,35)
1 4
8 5
0.150 (3,81)0.157 (4,00)
0.008 (0,20) NOM
0°– 8°
Gage Plane
A
0.004 (0,10)
0.010 (0,25)0.050 (1,27)
NOTES: A. All linear dimensions are in inches (millimeters).B. This drawing is subject to change without notice.C. Body dimensions do not include mold flash or protrusion, not to exceed 0.006 (0,15).D. Falls within JEDEC MS-012
OPA84722SBOS251Cwww.ti.com
PACKAGE DRAWINGS (Cont.)
DBV (R-PDSO-G6) PLASTIC SMALL-OUTLINE
0,10
M0,200,95
0 –8
0,25
0,550,35
Gage Plane
0,15 NOM
4073253-5/G 01/02
2,603,00
0,500,25
1,501,70
46
31
2,803,00
1,450,95
0,05 MIN
Seating Plane
6X
NOTES: A. All linear dimensions are in millimeters.B. This drawing is subject to change without notice.C. Body dimensions do not include mold flash or protrusion.D. Leads 1, 2, 3 may be wider than leads 4, 5, 6 for package orientation.
IMPORTANT NOTICE
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, modifications,enhancements, improvements, and other changes to its products and services at any time and to discontinueany product or service without notice. Customers should obtain the latest relevant information before placingorders and should verify that such information is current and complete. All products are sold subject to TI’s termsand conditions of sale supplied at the time of order acknowledgment.
TI warrants performance of its hardware products to the specifications applicable at the time of sale inaccordance with TI’s standard warranty. Testing and other quality control techniques are used to the extent TIdeems necessary to support this warranty. Except where mandated by government requirements, testing of allparameters of each product is not necessarily performed.
TI assumes no liability for applications assistance or customer product design. Customers are responsible fortheir products and applications using TI components. To minimize the risks associated with customer productsand applications, customers should provide adequate design and operating safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any TI patent right,copyright, mask work right, or other TI intellectual property right relating to any combination, machine, or processin which TI products or services are used. Information published by TI regarding third-party products or servicesdoes not constitute a license from TI to use such products or services or a warranty or endorsement thereof.Use of such information may require a license from a third party under the patents or other intellectual propertyof the third party, or a license from TI under the patents or other intellectual property of TI.
Reproduction of information in TI data books or data sheets is permissible only if reproduction is withoutalteration and is accompanied by all associated warranties, conditions, limitations, and notices. Reproductionof this information with alteration is an unfair and deceptive business practice. TI is not responsible or liable forsuch altered documentation.
Resale of TI products or services with statements different from or beyond the parameters stated by TI for thatproduct or service voids all express and any implied warranties for the associated TI product or service andis an unfair and deceptive business practice. TI is not responsible or liable for any such statements.
Following are URLs where you can obtain information on other Texas Instruments products and applicationsolutions:
