REV. B

Information furnished by Analog Devices is believed to be accurate and
reliable. However, no responsibility is assumed by Analog Devices for its
use, nor for any infringements of patents or other rights of third parties
which may result from its use. No license is granted by implication or
otherwise under any patent or patent rights of Analog Devices.

10-Bit 20 MSPS 160 mW

CMOS A/D Converter

AD876

FEATURES
CMOS 10-Bit 20 MSPS Sampling A/D Converter
Pin-Compatible 8-Bit Option
Power Dissipation: 160 mW
+5 V Single Supply Operation
Differential Nonlinearity: 0.5 LSB
Guaranteed No Missing Codes
Power Down (Standby) Mode
Three-State Outputs
Digital I/Os Compatible with +5 V or +3.3 V Logic
Adjustable Reference Input
Small Size: 28-Lead SOIC, 28-Lead SSOP, or 48-Lead

Thin Quad Flatpack (TQFP)

PRODUCT DESCRIPTION

The AD876 is a CMOS, 160 mW, 10-bit, 20 MSPS analog-to-
digital converter (ADC). The AD876 has an on-chip input
sample-and-hold amplifier. By implementing a multistage pipe-
lined architecture with output error correction logic, the AD876
offers accurate performance and guarantees no missing codes
over the full operating temperature range. Force and sense con-
nections to the reference inputs minimize external voltage drops.

The AD876 can be placed into a standby mode of operation
reducing the power below 50 mW. The AD876's digital I/O
interfaces to either +5 V or +3.3 V logic. Digital output pins
can be placed in a high impedance state; the format of the out-
put is straight binary coding.

The AD876's speed, resolution and single-supply operation
ideally suit a variety of applications in video, multimedia, imag-
ing, high speed data acquisition and communications. The
AD876's low power and single-supply operation satisfy require-
ments for high speed portable applications. Its speed and reso-
lution ideally suit charge coupled device (CCD) input systems
such as color scanners, digital copiers, electronic still cameras
and camcorders.

The AD876 comes in a space saving 28-lead SOIC and 48-lead
thin quad flatpack (TQFP) and is specified over the commercial
(0

C to +70

C) temperature range.

PRODUCT HIGHLIGHTS
Low Power

The AD876 at 160 mW consumes a fraction of the power of
presently available 8- or 10-bit, video speed converters. Power-
down mode and single-supply operation further enhance its
desirability in low power, battery operated applications such
as electronic still cameras, camcorders and communication
systems.

Very Small Package

The AD876 comes in a 28-lead SOIC, 28-lead SSOP, and 48-
lead surface mount, thin quad flat package. The TQFP package
is ideal for very tight, low headroom designs.

Digital I/O Functionality

The AD876 offers three-state output control.

Pin Compatible Upgrade Path

The AD876 offers the option of laying out designs for eight
bits and migrating to 10-bit resolution if prototype results
warrant.

FUNCTIONAL BLOCK DIAGRAM

A/D

D/A

A/D

D/A

A/D

D/A

A/D

CORRECTION LOGIC

AIN

REFTF

REFTS

REFBS

REFBF

STBY

THREE-

STATE

(MSB)
D9

D0
(LSB)

DRV

CLK

DRV

CML

AD876

SHA

GAIN

SHA

GAIN

OUTPUT BUFFERS

One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A.
Tel: 781/329-4700

World Wide Web Site: http://www.analog.com

Fax: 781/326-8703

© Analog Devices, Inc., 1998

AD876SPECIFICATIONS

REV. B

AD876JR-8

AD876

Parameter

Min

Typ

Max

Min

Typ

Max

Units

RESOLUTION

Bits

DC ACCURACY

Integral Nonlinearity (INL)

0.3

1.0

LSB

Differential Nonlinearity (DNL)

0.1

0.75

0.5

LSB

No Missing Codes

GUARANTEED

Offset Error

0.1

0.4

% FSR

Gain Error

0.1

0.2

% FSR

ANALOG INPUT

Input Range

V p-p

Input Capacitance

5.0

REFERENCE INPUT

Reference Top Voltage

3.5

4.0

4.5

3.5

4.0

4.5

Reference Bottom Voltage

1.6

2.0

2.5

1.6

2.0

2.5

Reference Input Resistance

250

Reference Input Current

8.0

Reference Top Offset

Reference Bottom Offset

DYNAMIC PERFORMANCE

Effective Number of Bits

= 1 MHz

7.8

9.0

Bits

= 3.58 MHz

7.4

7.8

8.2

9.0

Bits

= 10 MHz

7.5

8.2

Bits

Signal-to-Noise and Distortion (S/N+D) Ratio

= 1 MHz

= 3.58 MHz

= 10 MHz

Total Harmonic Distortion (THD)

=1 MHz

= 3.58 MHz

=10 MHz

Spurious Free Dynamic Range

Full Power Bandwidth

150

MHz

Differential Phase

0.5

Degree

Differential Gain

POWER SUPPLIES

Operating Voltage

+4.5

+5.25

+4.5

+5.25

Volts

+4.5

+5.25

+4.5

+5.25

Volts

DRV

+3.0

+5.25

+3.0

+5.25

Volts

Operating Current

IAV

IDV

IDRV

0.1

POWER CONSUMPTION

160

190

160

190

TEMPERATURE RANGE

Specified

+70

NOTES

and DV

must be within 0.5 V of each other to maintain specified performance levels.

3.58 MHz Input Frequency.

Specifications subject to change without notice. See Definition of Specifications for additional information.

MIN

to T

MAX

with AV

= +5.0 V, DV

= +5.0 V, DRV

= +3.3 V, V

REFB

= +4.0 V, V

REFB

+2.0 V, f

CLOCK

= 20 MSPS, unless otherwise noted)

AD876

REV. B

DIGITAL SPECIFICATIONS

AD876

Parameter

Symbol

DRV

Min

Typ

Max

Units

LOGIC INPUT

High Level Input Voltage

3.0

2.4

5.0

4.0

5.25

4.2

Low Level Input Voltage

3.0

0.6

5.0

1.0

5.25

1.05

High Level Input Current

5.0

+10

Low Level Input Current

5.0

+50

Low Level Input Current (CLK Only)

5.0

+10

Input Capacitance

LOGIC OUTPUTS

High Level Output Voltage

= 50

3.0

2.4

5.0

3.8

= 0.5 mA)

5.0

2.4

Low Level Output Voltage

= 50

3.6

0.7

5.25

1.05

= 0.6 mA)

5.25

0.4

Output Capacitance

OUT

Output Leakage Current

Specifications subject to change without notice.

TIMING SPECIFICATIONS

Symbol

Min

Typ

Max

Units

Maximum Conversion Rate

MHz

Clock Period

Clock High

Clock Low

Output Delay

Pipeline Delay (Latency)

3.5

Clock Cycles

Aperture Delay Time

Aperture Jitter

NOTE

Conversion rate is operational down to 10 kHz without degradation in specified performance.

OUT

DATA N-4

DATA N-3

DATA N-2

DATA N-1

DATA N

SAMPLE N

SAMPLE N+1

SAMPLE N+2

AIN

CLK

Figure 1. Timing Diagram

MIN

to T

MAX

with AV

= +5.0 V, DV

= +5.0 V, DRV

= +3.3 V, V

REFT

= +4.0 V, V

REFB

= +2.0 V,

CLOCK

= 20 MSPS, C

= 20 pF unless otherwise noted)

REV. B

AD876

SOIC

TQFP

Symbol

Pin No.

Type

Name and Function

D0 (LSB)

Least Significant Bit.

D1D4

Data Bits 1 through 4.

D5D8

811

Data Bits 5 through 8.

D9 (MSB)

Most Significant Bit.

THREE-

THREE-STATE = LOW

THREE-STATE = HIGH

STATE

or N/C
Normal Operating Mode

High Impedance Outputs

STBY

STBY = LOW or N/C

STBY = HIGH

Normal Operating Mode

Standby Mode

CLK

Clock Input.

CML

Bypass Pin for an Internal Bias Point.

REFTF

Reference Top Force.

REFBF

Reference Bottom Force.

REFTS

Reference Top Sense.

REFBS

Reference Bottom Sense.

AIN

Analog Input.

+5 V Analog Supply.

Analog Ground.

+5 V Digital Supply.

14, 19, 20

17, 27, 28

Digital Ground.

DRV

+3.3 V/+5 V Digital Supply. Supply for digital
input and output buffers.

DRV

+3.3 V/+5 V Digital Ground. Ground for digital

input and output buffers.

Type: AI = Analog Input; AO = Analog Output; DI = Digital Input; DO = Digital Output; P = Power.

PIN CONFIGURATIONS

SOIC/SSOP

TQFP

PIN FUNCTION DESCRIPTIONS

PINS D0 AND D1 ARE LEFT OPEN
FOR THE AD876JR-8

DRV

AIN

REFBF

REFTF

CML

REFBS

REFTS

STBY

DRV

THREE-STATE

CLK

TOP VIEW

(Not to Scale)

AD876

NC = NO CONNECT

REFBS

REFTF

REFBF

REFTS

TOP VIEW

(Not to Scale)

AD876

DRV

CLK

THREE-STATE

STBY

DRV

AIN

CML

AD876

REV. B

ORDERING GUIDE

Temperature

Package

Model

Range

Description

Options

AD876JR

C to +70

28-Lead SOIC

R-28

AD876JST-Reel

C to +70

48-Lead TQFP

ST-48

(Tape and Reel 13")

AD876JR-8

C to +70

28-Lead SOIC

R-28

AD876AR

C to +85

C 28-Lead SOIC

R-28

AD876ARS

C to +85

C 28-Lead SSOP

RS-28

AD876JRS

C to +70

28-Lead SSOP

RS-28

AD876JRS-8

C to +70

28-Lead SSOP

RS-28

ABSOLUTE MAXIMUM RATINGS*

Parameter

With Respect to Min

Max Units

0.5

+6.5

Volts

, DRV

0.5

+6.5

Volts

, DRV

0.5

+0.5

Volts

AIN

0.5

+6.5

Volts

REFTS, REFTF
REFBS, REFBF

0.5

+6.5

Volts

Digital Inputs, CLK

, DRV

0.5

+6.5

Volts

Junction Temperature

+150

Storage Temperature

+150

Lead Temperature

(10 sec)

+300

*Stresses above those listed under Absolute Maximum Ratings may cause perma-

nent damage to the device. This is a stress rating only; functional operation of the
device at these or any other conditions above those indicated in the operational
sections of this specification is not implied. Exposure to absolute maximum
ratings for extended periods may effect device reliability.

CAUTION
ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily
accumulate on the human body and test equipment and can discharge without detection.
Although the AD876 features proprietary ESD protection circuitry, permanent damage may
occur on devices subjected to high energy electrostatic discharges. Therefore, proper ESD
precautions are recommended to avoid performance degradation or loss of functionality.

Figure 2. Equivalent Circuits

REFTF

REFTS

REFBS

REFBF

INTERNAL
REFERENCE
VOLTAGE

DRV

c) CLK

d) AIN

a) D0D9

b) Three-State, Standby

WARNING!

ESD SENSITIVE DEVICE

REV. B

AD876

CODE OFFSET

128

256

384

512

640

768

896

960

192

320

448

576

704

832

0.5

DNL LSBs

Figure 3. AD876 Typical DNL

10

1000

100

FREQUENCY MHz

GAIN dB

Figure 4. Full Power Bandwidth

INPUT FREQUENCY MHz

Figure 5. SINAD vs. Input Frequency
(f

CLK

= 20 MSPS, AIN = 0.5 dB)

20

70

10

40

30

60

50

FREQUENCY MHz

80

90

THD

2ND

3RD

Figure 6. THD vs. Input Frequency 2nd, 3rd Harmonics

CLOCK FREQUENCY MHz

Figure 7. SINAD vs. CLK Frequency (AIN = 0.5 dB)

CLOCK FREQUENCY MHz

180

170

120

160

150

140

130

110

100

Figure 8. Power Consumption vs. Sample Rate

Typical Performance Characteristics

AD876

REV. B

DEFINITIONS OF SPECIFICATIONS

INTEGRAL NONLINEARITY (INL)

Integral nonlinearity refers to the deviation of each individual
code from a line drawn from "zero" through "full scale". The
point used as "zero" occurs 1/2 LSB before the first code transi-
tion. "Full scale" is defined as a level 1 1/2 LSB beyond the last
code transition. The deviation is measured from the center of
each particular code to the true straight line.

DIFFERENTIAL NONLINEARITY (DNL, NO MISSING
CODES)

An ideal ADC exhibits code transitions that are exactly 1 LSB
apart. DNL is the deviation from this ideal value. It is often
specified in terms of the resolution for which no missing codes
(NMC) are guaranteed.

OFFSET ERROR

The first transition should occur at a level 1/2 LSB above
"zero." Offset is defined as the deviation of the actual first code
transition from that point.

GAIN ERROR

The first code transition should occur for an analog value 1/2 LSB
above nominal negative full scale. The last transition should
occur for an analog value 1 1/2 LSB below the nominal positive
full scale. Gain error is the deviation of the actual difference
between first and last code transitions and the ideal difference
between the first and last code transitions.

PIPELINE DELAY (LATENCY)

The number of clock cycles between conversion initiation and
the associated output data being made available. New output
data is provided every clock cycle.

REFERENCE TOP/BOTTOM OFFSET

Resistance between the reference input and comparator input
tap points causes offset errors. These errors can be nulled out
by using the force-sense connection as shown in the Reference
Input section.

THEORY OF OPERATION

The AD876 implements a pipelined multistage architecture to
achieve high sample rate with low power. The AD876 distrib-
utes the conversion over several smaller A/D subblocks, refining
the conversion with progressively higher accuracy as it passes
the results from stage to stage. As a consequence of the distrib-
uted conversion, the AD876 requires a small fraction of the 1023
comparators used in a traditional flash type A/D. A sample-and-
hold function within each of the stages permits the first stage to
operate on a new input sample while the second and third stages
operate on the two preceding samples.

APPLYING THE AD876

DRIVING THE ANALOG INPUT

Figure 11 shows the equivalent analog input of the AD876, a
sample-and-hold amplifier (SHA). Bringing CLK to a logic low
level closes Switches 1 and 2 and opens Switch 3. The input
source connected to AIN must charge capacitor C

during this

time. When CLK transitions from logic "low" to logic "high,"
Switch 1 opens first, placing the SHA in hold mode. Switch 2
opens subsequently. Switch 3 then closes, connects the feed-
back loop around the op amp, and forces the output of the op
amp to equal the voltage stored on C

. When CLK transitions

from logic "high" to logic "low", Switch 3 opens first. Switch 2
closes and reconnects the input to C

. Finally, Switch 1 closes

and places the SHA in track mode.

The structure of the input SHA places certain requirements on
the input drive source. The combination of the pin capacitance,
C

, and the hold capacitance, C

, is typically less than 5 pF.

The input source must be able to charge or discharge this ca-
pacitance to 10-bit accuracy in one half of a clock cycle. When
the SHA goes into track mode, the input source must charge or
discharge capacitor C

from the voltage already stored on C

(the previously captured sample) to the new voltage. In the
worst case, a full-scale voltage step on the input, the input
source must provide the charging current through the R

(50

)

of Switch 2 and quickly settle (within 1/2 CLK period). This
situation corresponds to driving a low input impedance. On the
other hand, when the source voltage equals the value previously
stored on C

, the hold capacitor requires no input current and

the equivalent input impedance is extremely high.

Adding series resistance between the output of the source and
the AIN pin reduces the drive requirements placed on the
source. Figure 12 shows this configuration. The bandwidth of
the particular application limits the size of this resistor. To
maintain the performance outlined in the data sheet specifica-
tions, the resistor should be limited to 200

or less. For appli-

cations with signal bandwidths less than 10 MHz, the user may
increase the size of the series resistor proportionally. Alterna-
tively, adding a shunt capacitance between the AIN pin and

HARMONICS (dBc)

2ND 68.02
3RD 72.85
4TH 70.68
5TH 78.09

6TH 77.74
7TH 75.62
8TH 75.98
9TH 81.20

THD = 64.12
SNR = 48.73
SINAD = 48.61
SFDR = 68.02

Figure 9. AD876JR-8 Typical FFT (f

= 3.58 MHz,

AIN = 0.5 dB, f

CLOCK

= 20 MSPS)

HARMONICS (dBc)

2ND 68.91
3RD 73.92

4TH 68.67
5TH 73.26

6TH 80.55
7TH 82.02
8TH 81.02
9TH 88.94

THD = 64.24
SNR = 55.71
SINAD = 55.14
SFDR = 68.67

Figure 10. AD876 Typical FFT (f

= 3.58 MHz, AIN = 0.5 dB,

CLOCK

= 20 MSPS)

REV. B

AD876

20 kHz. At a sample clock frequency of 20 MHz, the dc bias
current at 3 V dc is approximately 30

A. If we choose R2 equal

to 1 k

and R1 equal to 50

, the parallel capacitance should

be a minimum of 0.008

F to avoid attenuating signals close to

20 kHz. Note that the bias current will cause a 31.5 mV offset
from the 3 V bias.

In systems that must use dc-coupling, use an op amp to level-
shift a ground-referenced signal to comply with the input
requirements of the AD876. Figure 14 shows an AD817
configured in inverting mode with ac signal gain of 1. The dc
voltage at the noninverting input of the op amp controls the
amount of dc level shifting. A resistive voltage divider attenu-
ates the REFBF signal. The op amp then multiplies the attenu-
ated signal by 2. In the case where REFBF = 1.6 V, the dc
output level will be 2.6 V. The AD817 is a low cost, fast settling,
single supply op amp with a G = 1 bandwidth of 29 MHz. The
AD818 is similar to the AD817 but has a 50 MHz bandwidth.
Other appropriate op amps include the AD8011, AD812 (a dual),
and the AD8001.

AD817 OR

AD818

0.1 F

= 4.99k

AIN

AD876

= 4.99k

14.7k

REFBF

2V p-p

0Vdc

Figure 14. Bipolar Level Shift

An integrated difference amplifier such as the AD830 is an
alternate means of providing dc level shifting. The AD830
provides a great deal of flexibility with control over offset and
gain. Figure 15 shows the AD830 precisely level-shifting a
unipolar, ground-referenced signal. The reference voltage,
REFBS, determines the amount of level-shifting. The ac gain
is 1. The AD830 offers the advantages of high CMRR, precise
gain, offset, and high-impedance inputs when compared with a
discrete implementation. For more information regarding the
AD830, see the AD830 data sheet.

AD876

AIN

12V

+2V

REFBS

0.1

+12V

AD830

Figure 15. Level Shifting with the AD830

REFERENCE INPUT DRIVING THE REFERENCE
TERMINALS

The AD876 requires an external reference on pins REFTF and
REFBF. The AD876 provides reference sense pins, REFTS
and REFBS, to minimize voltage drops caused by external and
internal wiring resistance. A resistor ladder, nominally 250

connects pins REFTF and REFBF.

analog ground can lower the ac source impedance. The value
of this capacitance will depend on the source resistance and the
required signal bandwidth.

The input span of the AD876 is a function of the reference
voltages. For more information regarding the input range, see
the DRIVING THE REFERENCE TERMINALS section of
the data sheet.

AD876

AIN

Figure 11. AD876 Equivalent Input Structure

AIN

200

Figure 12. Simple AD876 Drive Requirements

In many cases, particularly in single-supply operation, ac-
coupling offers a convenient way of biasing the analog input
signal at the proper signal range. Figure 13 shows a typical
configuration for ac-coupling the analog input signal to the
AD876. Maintaining the specifications outlined in the data
sheet requires careful selection of the component values. The
most important concern is the f

-3 dB

high-pass corner that is a

function of R2, and the parallel combination of C1 and C2.
The f

-3 dB

point can be approximated by the equation

3 dB

× ×

( R2) Ceq ]

where Ceq is the parallel combination of C1 and C2. Note that
C1 is typically a large electrolytic or tantalum capacitor that
becomes inductive at high frequencies. Adding a small ceramic
or polystyrene capacitor on the order of 0.01

F that does not

become inductive until negligibly higher frequencies maintains
a low impedance over a wide frequency range.

AIN

AD876

BIAS

Figure 13. AC-Coupled Inputs

There are additional considerations when choosing the resistor
values. The ac-coupling capacitors integrate the switching
transients present at the input of the AD876 and cause a net dc
bias current, I

, to flow into the input. The magnitude of this

bias current increases with increasing dc signal level and also
increases with sample frequency. This bias current will result in
an offset error of (R1 + R2)

. If it is necessary to compen-

sate this error, consider making R2 negligibly small or modify-
ing V

BIAS

to account for the resultant offset.

As an example, assume that the input to the AD876 must have
a dc bias of 3 V and the minimum expected signal frequency is

AD876

REV. B

Figure 16 shows the equivalent input structure for the AD876
reference pins. There is approximately 5

of resistance between

both the REFTF and REFBT pins and the reference ladder. If
the force-sense connections are not used, the voltage drop
across the 5

resistors will result in a reduced voltage appear-

ing across the ladder resistance. This reduces the input span of
the converter. Applying a slightly larger span between the REFTF
and REFBF pins compensates this error. Note that the tem-
perature coefficients of the 5

resistors are 1350 ppm. The

user should consider the effects of temperature when not using
a force-sense reference configuration.

AD876

REFTF

REFTS

REFBS

REFBF

DACS

C (V

)

CLK

250

LADDER

Figure 16. AD876 Equivalent Reference Structure

Do not connect the REFTS and REFBS pins in configurations
that do not use a force-sense reference. Connecting the force
and sense lines together allows current to flow in the sense lines.
Any current allowed to flow through these lines must be negligi-
bly small. Current flow causes voltage drops across the resis-
tance in the sense lines. Because the internal D/As of the
AD876 tap different points along the sense lines, each D/A
would receive a slightly different reference voltage if current
were flowing in these wires. To avoid this undesirable condition,
leave the sense lines unconnected. Any current allowed to flow
through these lines must be negligibly small (<100

A).

The voltage drop across the internal resistor ladder determines
the input span of the AD876. The driving voltages required at
the V1 and V2 points are respectively +4 V and +2 V. Calculate
the full-scale input span from the equation

Input Span (V )

REFTS REFBS

This results in a full-scale input span of approximately +2 V
when REFTS = +4 V and REFBS = +2 V In order to maintain
the requisite 2 V drop across the internal ladder, the external
reference must be capable of providing approximately 8.0 mA.

The user has flexibility in determining both the full-scale span of
the analog input and where to center this voltage. Figure 17
shows the range over which the AD876 can operate without
degrading the typical performance.

2.5

3.0

3.5

4.0

4.5

1.0

1.5

2.0

2.5

3.0

REFBF, REFBS

REFTF, REFTS

(1.6, 4.5)

(2.5, 4.5)

(1.6, 3.5)

(2.5, 3.5)

Figure 17. AD876 Reference Ranges

While the previous issues address the dc aspects of the AD876
reference, the user must also be aware of the dynamic imped-

ance changes associated with the reference inputs. The simpli-
fied diagram of Figure 16 shows that the reference pins connect
to a capacitor for one-half of the clock period. The size of the
capacitor is a function of the analog input voltage.

The external reference must be able to maintain a low imped-
ance over all frequencies of interest in order to provide the charge
required by the capacitance. By supplying the requisite charge,
the reference voltages will be relatively constant and perfor-
mance will not degrade. For some reference configurations,
voltage transients will be present on the reference lines; this
is particularly true during the falling edge of CLK. It is impor-
tant that the reference recovers from the transients and settles to
the desired level of accuracy prior to the rising edges of CLK.

There are several reference configurations suitable for the
AD876 depending on the application, desired level of accuracy,
and cost trade-offs. The simplest configuration, shown in Fig-
ure 18, utilizes a resistor string to generate the reference volt-
ages from the converter's analog power supply. The 0.1

bypass capacitors effectively reduce high-frequency transients.
The 10

F capacitors act to reduce the impedances at the

REFTF and REFBF pins at lower frequencies. As input fre-
quencies approach dc, the capacitors become ineffective, and
small voltage deviations will appear across the biasing resistors.
This application can maintain 10-bit accuracy for input frequen-
cies above approximately 200 Hz. 8-bit applications can use this
circuit for input frequencies above approximately 50 Hz.

+5V

AD876

10 F

0.1 F

10 F

0.1 F

250 ( 1%)

140 ( 1%)

250
( 15%)

REFTS

REFTF

REFBF

REFBS

NC = NO CONNECT

Figure 18. Low Cost Reference Circuit

This reference configuration provides the lowest cost but has
several disadvantages. These disadvantages include poor dc
power supply rejection and poor accuracy due to the variability
of the internal and external resistors.

The AD876 offers force-sense reference connections to elimi-
nate the voltage drops associated with the internal connections
to the reference ladder. Figure 19 shows a suggested circuit
using an AD826 dual, high speed op amp. This configuration
uses 3.6 V and 1.6 V reference voltages for REFT and REFB,
respectively. The connections shown in Figure 19 configure the
op amps as voltage followers.

REV. B

AD876

+5V

1/2

AD826

1/2

AD826

REFT

REFB

REFTS

REFTF

REFBS

REFBF

0.1 F

Figure 19. Kelvin Connected Reference Using the AD826

By connecting the op amp feedback through the sense connec-
tions of the AD876, the outputs of the op amps automatically
adjust to compensate for the voltage drops that occur within
the converter. The AD826 has the advantage of being able to
maintain stability while driving unlimited capacitive loads. As a
result, 0.1

F capacitors C1, C2, and C3 can connect directly

to the outputs of the op amps. These decoupling capacitors
reduce high frequency transients. Capacitors C4 and C5 shunt
across the internal resistors of the force sense connections and
prevent instability.

This configuration provides excellent performance and a mini-
mal number of components. The circuit also offers the advan-
tage of operating from a single +5 V supply. While alternative
op amps may also be suitable, consider the stability of these op
amps while driving capacitive loads.

The circuit shown in Figure 20 allows a wider selection of op
amps when compared with the previous configuration. An

AD876

1/2

OP-295

10 F

0.1 F

REFT

REFTS

REFTF

47nF

20k

1/2

OP-295

10 F

0.1 F

REFB

REFBS

REFBF

47nF

20k

22 F

Figure 20. Kelvin Connected Reference Using the OP295

OP295 dual, single-supply op amp provides stable 3.6 V and
1.6 V reference voltages. The AD822 dual op amp is also suit-
able for single-supply applications. Each half of the OP295 is
compensated to drive the 10

F and 0.1

F decoupling capaci-

tors at the REFTF and REFBF pins and maintain stability.

Like any high resolution converter, the layout and decoupling of
the reference is critical. The actual voltage digitized by the
AD876 is relative to the reference voltages. In Figure 21, for
example, the reference return and the bypass capacitors are
connected to the shield of the incoming analog signal. Distur-
bances in the ground of the analog input, that will be common-
mode to the REFT, REFB, and AIN pins because of the

common ground, are effectively removed by the AD876's high
common-mode rejection.

High frequency noise sources, V

and V

, are shunted to

ground by decoupling capacitors. Any voltage drops between
the analog input ground and the reference bypassing points will
be treated as input signals by the converter via the reference
inputs. Consequently, the reference decoupling capacitors
should be connected to the same analog ground point used to
define the analog input voltage. (For further suggestions, see
the "Grounding and Layout Rules" section of the data sheet.)

REFTF

REFBF

AIN

AD876

Figure 21. Recommended Bypassing for the Reference
Inputs

CLOCK INPUT

The AD876 clock input is buffered internally with an inverter
powered from the DRV

pin. This feature allows the AD876

to accommodate either +5 V or +3.3 V CMOS logic input sig-
nal swings with the input threshold for the CLK pin nominally
at DRV

/2.

The AD876's pipelined architecture operates on both rising and
falling edges of the input clock. To minimize duty cycle varia-
tions the recommended logic family to drive the clock input is
high speed or advanced CMOS (HC/HCT, AC/ACT) logic.
CMOS logic provides both symmetrical voltage threshold levels
and sufficient rise and fall times to support 20 MSPS operation.
The AD876 is designed to support a conversion rate of 20 MSPS;
running the part at slightly faster clock rates may be possible,
although at reduced performance levels. Conversely, some
slight performance improvements might be realized by clocking
the AD876 at slower clock rates.

The power dissipated by the correction logic and output buffers
is largely proportional to the clock frequency; running at reduced
clock rates provides a reduction in power consumption. Figure
8 illustrates this trade-off.

DIGITAL INPUTS AND OUTPUTS

Each of the AD876 digital control inputs, THREE-STATE and
STBY, has an input buffer powered from the DRV

supply

pins. With DRV

set to +5 V, all digital inputs readily inter-

face with +5 V CMOS logic. For interfacing with lower voltage
CMOS logic, DRV

can be set to 3.3 V, effectively lowering

the nominal input threshold of all digital inputs to 3.3 V/2 =
1.65 V.

The format of the digital output is straight binary. Table I shows
the output format for the case where REFTS = 4 V and REFBS
= 2 V.

AD876

REV. B

For DRV

= 5 V, the AD876 output signal swing is compat-

ible with both high speed CMOS and TTL logic families. For
TTL, the AD876 on-chip, output drivers were designed to
support several of the high speed TTL families (F, AS, S). For
applications where the clock rate is below 20 MSPS, other TTL
families may be appropriate. For interfacing with lower voltage
CMOS logic, the AD876 sustains 20 MSPS operation with
DRV

= 3.3 V. In all cases, check your logic family data

sheets for compatibility with the AD876 Digital Specification
table.

THREE-STATE OUTPUTS

The digital outputs of the AD876 can be placed in a high im-
pedance state by setting the THREE-STATE pin to HIGH.
This feature is provided to facilitate in-circuit testing or
evaluation. Note that this function is not intended for enabling/
disabling the ADC outputs from a bus at 20 MSPS. Also, to
avoid corruption of the sampled analog signal during conversion
(3.5 clock cycles), it is highly recommended that the AD876
outputs be enabled on the bus prior to the first sampling. For
the purpose of budgetary timing, the maximum access and float
delay times (t

, t

shown in Figure 15) for the AD876 are

150 ns.

THREE-STATE

ACTIVE

HIGH IMPEDANCE

D0D9

Figure 22. High-Impedance Output Timing Diagram

Table I. Output Data Format

Approx.

THREE- DATA

AIN (V)

STATE

A low power mode feature is provided such that for STBY =
HIGH and the clock disabled, the static power of the AD876
will drop below 50 mW.

GROUNDING AND LAYOUT RULES

As is the case for any high performance device, proper ground-
ing and layout techniques are essential in achieving optimal
performance. The analog and digital grounds on the AD876
have been separated to optimize the management of return
currents in a system. It is recommended that a printed circuit
board (PCB) of at least 4 layers employing a ground plane and
power planes be used with the AD876. The use of ground and
power planes offers distinct advantages:

1. The minimization of the loop area encompassed by a signal

and its return path.

2. The minimization of the impedance associated with ground

and power paths.

3. The inherent distributed capacitor formed by the power

plane, PCB insulation, and ground plane.

These characteristics result in both a reduction of electro-
magnetic interference (EMI) and an overall improvement in
performance.

It is important to design a layout which prevents noise from
coupling onto the input signal. Digital signals should not be run
in parallel with the input signal traces and should be routed
away from the input circuitry. Separate analog and digital
grounds should be joined together directly under the AD876. A
solid ground plane under the AD876 is also acceptable if the
power and ground return currents are managed carefully. A
general rule of thumb for mixed signal layouts dictates that the
return currents from digital circuitry should not pass through
critical analog circuitry. For further layout suggestions, see the
AD876 Evaluation Board data sheet.

DIGITAL OUTPUTS

Each of the on-chip buffers for the AD876 output bits (D0D9)
is powered from the DRV

supply pins, separate from AV

. The output drivers are sized to handle a variety of logic

families while minimizing the amount of glitch energy gener-
ated. In all cases, a fan-out of one is recommended to keep the
capacitive load on the output data bits below the specified 20 pF
level.

REV. B

AD876

Figure 23. AD876 Evaluation Board Schematic

P1 2

P1 4

P1 6

P1 8

P1 10

P1 12

P1 14

P1 16

P1 18

P1 20

P1 22

P1 24

P1 28

P1 30

P1 32

P1 34

P1 36

P1 38

P1 40

P1 26

CLK

3_STATE

STBY

SUBST

REFTS

REFTF

REFBF

REFBS

CML

AIN

DRV

DB9

DB8

DB7

DB6

DB5

DB4

DB3

DB2

DB1

DB0

DRV

+5VD

JP1

C62

0.1 F

STBY

VP6

C50

10 F

REFTS

VP3

REFBF

VP2

VP4

REFBS

10 F

R17

1.13k

R16

R14

100

DC_IN

REFTF

VP1

AIN

TP2

R15
51

0.1 F

JP5

TP18

EXT_CM 2

INT_CM

R18

+5VD

JP4

VP8

REFTS

REFTF

REFBF

REFBS

C64

0.1 F

C56

0.1 F

10 F

+5VA

G1
G2
A7
A6
A5
A4
A3
A2
A1
A0

Y7
Y6
Y5
Y4
Y3
Y2
Y1
Y0

G1
G2
A7
A6
A5
A4
A3
A2
A1
A0

Y7
Y6
Y5
Y4
Y3
Y2
Y1
Y0

74ALS541

AD876

5
6
7
8
9

9
8
7
6
5
4
3
2

11
12
13
14
15
16
17
18

D5
D6
D7
D8
D9

0
1
2

9
8
7
6

4
3
2

11
12
13
14
15
16
17
18

D0
D1
D2
D3
D4

74ALS541

74F04

JP2

TP4

CLK_IN

TP1

3ST
VP5

47 F

+5VD

C49

10 F

C54

0.1 F

R5*

R6*

R7*

R9*

R10*

R11

R12

P1 1

P1 3

P1 5

P1 7

P1 9

P1 11

P1 13

P1 15

P1 17

P1 19

P1 21

P1 23

P1 27

P1 29

P1 31

P1 33

P1 35

P1 37

P1 39

P1 25

TP3

R2R12 = 20

REV. B

AD876

OUTLINE DIMENSIONS

Dimensions shown in inches and (mm).

R-28

28-Lead Wide Body (SOIC)

0.7125 (18.10)

0.6969 (17.70)

0.4193 (10.65)

0.3937 (10.00)

0.2992 (7.60)

0.2914 (7.40)

PIN 1

SEATING

PLANE

0.0118 (0.30)

0.0040 (0.10)

0.0192 (0.49)

0.0138 (0.35)

0.1043 (2.65)

0.0926 (2.35)

0.0500

(1.27)

BSC

0.0125 (0.32)

0.0091 (0.23)

0.0500 (1.27)

0.0157 (0.40)

0.0291 (0.74)

0.0098 (0.25)

x 45

ST-48

28-Lead Plastic Thin Quad Flatpack (TQFP)

TOP VIEW

(PINS DOWN)

0.276

0.004

(7.0

0.1)

0.354

0.008 (9.00

0.2) SQ

0.039 (1.00)
REF

0.02

0.003

(0.50

0.08)

0.007 +0.003 0.001

(0.18 +0.08 0.03)

SEATING

PLANE

0.055

0.002

(1.40

0.05)

0.059 +0.008 0.004

(1.50 +0.2 0.1)

0.02

0.008

(0.5

0.2)

0.005 +0.002 0.0008

(0.127 +0.05 0.02)

0.004

0.002

(0.1

0.05)

MIN

(3.5

3.5

)

RS-28

28-Lead Shrink Small Outline Package (SSOP)

0.407 (10.34)

0.397 (10.08)

0.311 (7.9)

0.301 (7.64)

0.212 (5.38)

0.205 (5.21)

PIN 1

SEATING

PLANE

0.008 (0.203)

0.002 (0.050)

0.07 (1.79)

0.066 (1.67)

0.0256

(0.65)

BSC

0.078 (1.98)

0.068 (1.73)

0.015 (0.38)

0.010 (0.25)

0.009 (0.229)

0.005 (0.127)

0.03 (0.762)

0.022 (0.558)

PRINTED IN U.S.A.

C1991a01/98

Ð­Ð»ÐµÐºÑ‚Ñ€Ð¾Ð½Ð½Ñ‹Ð¹ ÐºÐ¾Ð¼Ð¿Ð¾Ð½ÐµÐ½Ñ‚: AD876

ÐÐ»ÐµÐºÑ‚Ñ€Ð¾Ð½Ð½Ñ‹Ð¹ ÐºÐ¾Ð¼Ð¿Ð¾Ð½ÐµÐ½Ñ‚: AD876