TLV2553

SLAS354B – SEPTEMBER 2001 – REVISED SEPTEMBER 2002

6

www.ti.com

external reference specifications

PARAMETER

TEST CONDITIONS

MIN

TYP†

MAX

UNIT

Reference input voltage REF

VCC = 4.5 V to 5.5 V

–0.1

0

0.1

V

Reference input voltage, REF–

VCC = 2.7 V to 3.6 V

–0.1

0

0.1

V

Reference input voltage REF+

VCC = 4.5 V to 5.5 V

2

VCC

V

Reference input voltage, REF+

VCC = 2.7 V to 3.6 V

2

VCC

V

External reference input voltage difference,

VCC = 4.5 V to 5.5 V

1.9

VCC

V

g,

(REF+) – (REF–)

VCC = 2.7 V to 3.6 V

1.9

VCC

V

External reference supply current

CS at 0 V

VCC = 4.5 V to 5.5 V

0.94

mA

External reference supply current

CS at 0 V

VCC = 2.7 V to 3.6 V

0.62

mA

VCC =5V

Static

1

M

Ω

Reference input impedance

VCC = 5 V

During sampling/conversion

6

9

k

Ω

Reference input impedance

VCC =2 7V

Static

1

M

Ω

VCC = 2.7 V

During sampling/conversion

6

9

k

Ω

† All typical values are at TA = 25°C.

NOTE: Add a 0.1-

µF capacitor between REF+ and REF– pins when external reference is used.

PARAMETER

TEST CONDITIONS

MIN

TYP†

MAX

UNIT

INL

Integral linearity error (see Note 3)

–1

1

LSB

DNL

Differential linearity error

–1

1

LSB

Offset error (see Note 4)

See Note 2

–2

2

mV

Gain error (see Note 4)

See Note 2

–3

3

mV

Total unadjusted error (see Note 5)

±1.5

LSB

Address data input = 1011

2048

Self-test output code (see Table 2 and Note 6)

Address data input = 1100

0

Address data input = 1101

4095

† All typical values are at TA = 25°C.

NOTES:

2. Analog input voltages greater than the voltage applied to REF+ convert as all ones (111111111111), while input voltages less than

the voltage applied to REF– convert as all zeros (000000000000).

3. Linearity error is the maximum deviation from the best straight line through the A/D transfer characteristics.

4. Gain error is the difference between the actual midstep value and the nominal midstep value in the transfer diagram at the specified

gain point after the offset error has been adjusted to zero. Offset error is the difference between the actual midstep value and the

nominal midstep value at the offset point.

5. Total unadjusted error comprises linearity, zero-scale, and full-scale errors.

6. Both the input address and the output codes are expressed in positive logic.