ADR1581

Rev. 0 | Page 7 of 12

1.2488

1.2498

1.2500

1.2502

1.2504

1.2506

1.2508

1.2494

1.2496

1.2490

1.2492

SLOPE = TC =

SLOPE = TC =

–55

–35

–15

5

25

45

65

85

105

125

TEMPERATURE (°C)

600

300

0

500

400

200

100

–55

–35

–15

5

25

45

65

85

105

125

TEMPERATURE (°C)

Figure 12. Output Voltage vs. Temperature

Figure 13. Residual Drift Error

REVERSE VOLTAGE HYSTERESIS

For example, the ADR1581BRT initial tolerance is ±1.5 mV;

a ±50 ppm/°C temperature coefficient corresponds to an error

unit is guaranteed to be 1.250 V ± 5.6 mV over the operating

temperature range.

A major requirement for high performance industrial

equipment manufacturers is a consistent output voltage at

nominal temperature following operation over the operating

temperature range. This characteristic is generated by measuring

the difference between the output voltage at +25°C after operating

at +85°C and the output voltage at +25°C after operating at −40°C.

Duplication of these results requires a combination of high

accuracy and stable temperature control in a test system. Evaluation

of the ADR1581 produces curves similar to those in

Figure 14 displays the hysteresis associated with the ADR1581.

This characteristic exists in all references and has been minimized

in the ADR1581.

Figure 4

and Figure 12.

VOLTAGE OUTPUT NONLINEARITY VS.

TEMPERATURE

0

15

20

25

30

35

40

5

10

HYSTERESIS VOLTAGE (µV)

–400

–300

–200

–100

0

100

200

300

400

When a reference is used with data converters, it is important to

understand how temperature drift affects the overall converter

performance. The nonlinearity of the reference output drift

represents additional error that is not easily calibrated out of the

system. The usual way of showing the reference output drift is to

plot the reference voltage vs. temperature (see Figure 12). An

alternative method is to draw a straight line between the

temperature endpoints and measure the deviation of the output

from the straight line. This shows the same data in a different

format. This characteristic (see Figure 13) is generated by

normalizing the measured drift characteristic to the endpoint

average drift. The residual drift error of approximately 500 ppm

shows that the ADR1581 is compatible with systems that require

10-bit accurate temperature performance.

Figure 14. Reverse Voltage Hysteresis Distribution