﻿ Calibration and monitoring sensors_Circuit Diagram World ﻿
Position: Index > Monitor Circuit >

# Calibration and monitoring sensors

2015-01-20 02:16
Declaration:We aim to transmit more information by carrying articles . We will delete it soon, if we are involved in the problems of article content ,copyright or other problems.

Every sensor that measure real world things have real tolerances. There is no absolut accurate sensors as there are many factors that influence tolerances and reliability. If you ever tried to measure simple resistor resistance you noticed that 1k??resistor actually is between 990??to 1,1k??. This means that tolerance of resistor is 1%. if resistor tolerance would be 5% then this range would be much bigger. Also these values vary due to temperature.

Lets say wee want to measure temperature with 1?°C accuracy. When you buy temperature sensor you cant get this accuracy as its tolerance vary. Especially cheap sensors can’t be trusted. For instance you may get 98?°C for boiling water and so on. In this case ve need to adjust sensor to give 100?°C on boiling water. Of course this is simple example. As sensor is a part of whole measuring system where are board, other components. We need to adjust sensor to work as part of system. First thing you have to do isto exclude environment factors affecting sensor like board temperature or temperature of sensor itself. This is called compensation.

Compensation doesn’t calibrate sensor as it only exclude environment factors, but doesn’t adjust sensor itself. Calibration is an operation when sensor is placed to a known environment where sensor output is measured as X voltage for T temperature. Without calibration sensor produces Y voltage for T temperature. So measuring known temperatures we can build a reference table which can be stored in microcontroller memory. This table then can be used as look-up table for every read of sensor to determine actual temperature. This operation is called a calibration. Sometimes calibration can be done by calculating the real value if nature of inaccuracy is known. But principal is same.

Second important part is howcalibration and monitoring of sensorsis performed by microcontroller. There can be several cases:

Sensor is connected to MCU remotely via cable. MCU stores calibration values in EEPROM. This case has few disadvantages. First – calibrated sensor is tied to MCU board and on each part replace recalibration has to be done; Second a€“ MCU doesn’t know weather sensor were changed and readings are correct; Third a€“ If there are more sensors connected. Then each sensor has to be calibrated what means MCU has less resources for main tasks.

Second method is better because calibration data (EEPROM) is located on board where sensor is located. So now each sensors carry its own calibration data. This allows interchanging of MCU boards without affecting results. But more memory is required a€“ each sensor has its own memory IC.

And there is third concept a€“ add MCU to the sensor board (Smart Sensors). Calibrating and calibration data is near sensor. But more processors are needed so the costs may rise. But main processor doesn’t have to perform calibration so resources are free for main tasks. Third variant is popular in these days especially im tele-medicine, where sensors on human body are independent and can send reading in case there is something wrong or suspicious. Microcontrollers are cheap and powerful today that allows to shift part of calculation force towards sensors.

And the main advise is – try to automatecalibration and monitoring of sensorsin order to avoid human factors.