### Q

For AD7713: Is it possible to deduct the actual value of the constant currentsource (set value = 200µA) from the zero and full scale calibration

coefficients ?

Example

Given: zero reference resistance 100 Ohms system calibration

step one

gain reference resistance 5590.50 Ohms system

calibration step two

Gain = 1

Ref Voltage 2.5 Volts

Zero calibration coefficient 0x3BEADB (dec 3926747)

FS calibration coefficient 0xE4288E (dec 14952590

)

the measured actual value of the constant current is 220.19 µAmps.

Following the data sheet rev. C p 17 the result is obtained by

subtracting the zero cal.

value from the measured value and then multiplying with the FS cal

coefficient.

Result= (measured value[counts] - ZERO cal) * FS

000000<>((100Ohms * 220.19µA)/2.5 Volts* 16777215- 3926747) * 14952590

FFFFFF<>((5590,5 Ohms * 220.19µA)/2,5 Volts * 16777215 - 3926747) *

14952590

Do we have erroneous assumptions or is this simply not possible ?

### A

The 200uA current sources included on the AD7713 have quite wide tolerances onthem at +/-20% and the matching between them is about +/-1%. For most

applications it is not necessary to know the absolute value of the current

source. For example, if you have a sensor whose output is a change in

resistance, you can use a low TC precision resistor in series with the sensor,

then use the two differential channels of the ADC to measure the voltage across

the precision resistor, and the voltage across the sensor. Since the same

current is flowing through both the precision resistor and the sensor, the

whole thing is ratiometric and the absolute value of the current source is not

required.

I think that what you want to do is compensate for the tolerance of the current

sources by adjusting the calibration co-efficients. In most cases this is not

necessary, as often you can make the system ratiometric by looking again at the

measurement that you wish to make.

It is possible to adjust the cal co-efficients but this is not as straight

forward as you describe.