AD627

–12–

REV. A

2 V or 1 V to maximize the available gain and output swing.

Note that in most cases, there is no advantage to increasing the

single supply to greater than 5 V (the exception being an input

range of 0 V to 1 V).

5

–6

4

3

2

1

0

–1

–2

–3

–4

–5

–5

–4

–3

–2

–1

0

1234

Figure 35. Reference Input Voltage vs. Negative Input

5

–0.5

4

3

2

1

0

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Figure 36. Reference Input Voltage vs. Negative Input

Output Buffering

The AD627 is designed to drive loads of 20 k

Ω or greater but

can deliver up to 20 mA to heavier loads at lower output voltage

swings (see Figure 7). If more than 20 mA of output current is

required at the output, the AD627’s output should be buffered

with a precision op amp such as the OP113 as shown in Figure

37 (shown for the single supply case). This op amp can swing

from 0 V to 4 V on its output while driving a load as small as

600

Ω.

0.1 F

REF

0.1 F

0.1 F

0.1 F

AD627

OP113

Figure 37. Output Buffering

INPUT AND OUTPUT OFFSET ERRORS

The low errors of the AD627 are attributed to two sources,

input and output errors. The output error is divided by G when

referred to the input. In practice, the input errors dominate at

high gains and the output errors dominate at low gains. The

total offset error for a given gain is calculated as:

Total Error RTI = Input Error + (Output Error/Gain)

Total Error RTO = (Input Error

× G) + Output Error

RTI offset errors and noise voltages for different gains are shown

below in Table III.

Table III. RTI Error Sources

Max Total

Max Total

RTI Offset Error

RTI Offset Drift

Total RTI Noise

V

V

V/ C

V/ C

nV/

Gain AD627A

AD627B AD627A

AD627B AD627A & AD627B

5

450

250

5

3

95

10

350

200

4

2

66

20

300

175

3.5

1.5

56

50

270

160

3.2

1.2

53

100

270

155

3.1

1.1

52

500

252

151

3

1

52

1000 251

151

3

1

52

Make vs. Buy: A Typical Application Error Budget

The example in Figure 38 serves as a good comparison between

the errors associated with an integrated and a discrete in amp

implementation. A

± 100 mV signal from a resistive bridge

(common-mode voltage = +2.5 V) is to be amplified. This ex-

ample compares the resulting errors from a discrete two op

amp in amp and from the AD627. The discrete implementation

uses a four-resistor precision network (1% match, 50 ppm/

°C

tracking).

The errors associated with each implementation are detailed in

Table IV and show the integrated in amp to be more precise,

both at ambient and over temperature. It should be noted that

the discrete implementation is also more expensive. This is pri-

marily due to the relatively high cost of the low drift precision

resistor network.

Note, the input offset current of the discrete in amp implemen-

tation is the difference in the bias currents of the two op amps,

not the offset currents of the individual op amps. Also, while the

values of the resistor network are chosen so that the inverting

and noninverting inputs of each op amp see the same impedance

(about 350

Ω), the offset current of each op amp will add an

additional error which must be characterized.

Errors Due to AC CMRR

In Table IV, the error due to common-mode rejection is the

error that results from the common-mode voltage from the

bridge 2.5 V. The ac error due to nonideal common-mode

rejection cannot be calculated without knowing the size of the ac

common-mode voltage (usually interference from 50 Hz/60 Hz

mains frequencies).

A mismatch of 0.1% between the four gain setting resistors will

determine the low frequency CMRR of a two op amp in amp.

The plot in Figure 38 shows the practical results, at ambient

temperature, of resistor mismatch. The CMRR of the circuit in

Figure 39 (Gain = 11) was measured using four resistors which