Datasheet LT1008 (Analog Devices) - 8

LT1008
U U W U APPLICATIO S I FOR ATIO Achieving Picoampere/Microvolt Performance
The LT1008 is specified over a wide range of power supply In order to realize the picoampere—microvolt level accu- voltages from ±2V to ±18V. Operation with lower supplies racy of the LT1008, proper care must be exercised. For is possible down to ±1.2V (two Ni-Cad batteries). example, leakage currents in circuitry external to the op
Test Circuit for Offset Voltage and Its Drift with Temperature
amp can significantly degrade performance. High quality insulation should be used (e.g., TeflonTM, Kel-F); cleaning 50k* of all insulating surfaces to remove fluxes and other 15V *RESISTORS MUST HAVE LOW residues will probably be required. Surface coating may be 2 – 7 THERMOELECTRIC POTENTIAL necessary to provide a moisture barrier in high humidity 6 THIS CIRCUIT IS ALSO USED AS 100Ω* LT1008 VO THE BURN-IN CONFIGURATION environments. 3 + FOR THE LT1008 WITH SUPPLY VOLTAGES INCREASED TO 4 ±20V Board leakage can be minimized by encircling the input 50k* VO = 1000VOS –15V circuitry with a guard ring operated at a potential close to 1008 AI02 that of the inputs: in inverting configurations the guard ring should be tied to ground, in noninverting connections
Noise Testing
to the inverting input at Pin 2. Guarding both sides of the printed circuit board is required. Bulk leakage reduction The 0.1Hz to 10Hz peak-to-peak noise of the LT1008 is depends on the guard ring width. Nanoampere level leak- measured in the test circuit shown. The frequency re- age into the compensation terminals can affect offset sponse of this noise tester indicates that the 0.1Hz corner voltage and drift with temperature. is defined by only one zero. The test time to measure 0.1Hz to 10Hz noise should not exceed 10 seconds, as this time COMPENSATION limit acts as an additional zero to eliminate noise contribu- tions from the frequency band below 0.1Hz. V+ A noise voltage density test is recommended when mea- suring noise on a large number of units. A 10Hz noise 8 OUTPUT 7 1 voltage density measurement will correlate well with a 6 0.1Hz to 10Hz peak-to-peak noise reading since both 5 2 results are determined by the white noise and the location 4 3 of the 1/f corner frequency. Current noise is measured in the circuit shown and calcu- INPUTS lated by the following formula where the noise of the V– source resistors is subtracted. GUARD 1008 AI01 1/2
REFERENCE ONLY—OBSOLETE PACKAGE
e ⎡ 2 2 ⎤ no – n 820 V ( ) i = ⎣⎢ ⎦⎥ Microvolt level error voltages can also be generated in the n M 40 Ω × 100 external circuitry. Thermocouple effects caused by tem- 10k perature gradients across dissimilar metals at the con- tacts to the input terminals can exceed the inherent drift of 10M* 10M* 2 the amplifier. Air currents over device leads should be – 6 minimized, package leads should be short, and the two 100Ω LT1008 eno 10M* 10M* 3 input leads should be as close together as possible and + maintained at the same temperature. *METAL FILM 1008 AI04 1008fb 8