Comments: MLX90614 IR Thermometer Hookup Guide

The explanation of radiation implies that only infrared light warms things. That is false.

Original: "Per the Stefan-Boltzman law, any object that isn't below absolute zero (0°K) emits (non-human-eye-visible) light in the infrared spectrum that is directly proportional to its temperature. The special infrared thermopile inside the MLX90614 senses how much infrared energy is being emitted by materials in its field of view, and produces an electrical signal proportional to that."

What I think would be less roughshod: "Molecules emit light in proportion to their temperature—This is called blackbody radiation. For earthly temperatures, most of that is infrared. The thermopile is covered by a filter that lets through a wide band of the infrared part of the spectrum, but not visible light. In our common environments, there are many visible lights that aren't from "blackbody radiation" but rather from narrow-band fluorescent or LED lamps; so, we don't want to let those freaky lights affect our readings. The thermopile itself warms, compared to the material in which it's embedded, by receiving more radiation from warmer objects than it emits, or it cools by receiving less radiation from cooler objects than it emits. That changes its electrical properties, and those properties are compared to an unexposed internal thermometer, which is really a diode whose bandwidth changes with temperature. So, the device outputs a signal that is a comparison between its known temperature and its exposed temperature. For example, the device knows its internal temperature is 30°C, but the thermopile on the device is conducting electricity as if it were exposed to something 80°C hotter than itself, so the device reports that it is exposed to some source that is 110°C. Well, it's not quite that simple; the thermopile does not reach the same temperature as the object in its field of view, so we need to calibrate."

"The thermopile itself is a pile of two different metals, where the interface between each layer of metals on one side is exposed and the interface on the other side is unexposed—Imagine cooking a triple-decker sandwich on its side. The electrical potential difference between the metals depends on temperature. It's a very small difference normally, so many layers are piled on top of each other to multiply the effect; and even then the small current still needs careful amplification for our purposes."

Hi, I find that the temperature measurement is highly sensitive to distance. As such, it is difficult to get an accurate reading off any object. Is there a way to improve the accuracy of this module? Thanks.

Hi,

I am using the MLX90614 infra-red thermopile to read temperature. The code and library found here works just fine on the uno and mega, but not on the arduino due.

I've tried modifying the library as suggested on this [url=https://forums.adafruit.com/viewtopic.php?f=19&t=73118&p=418612#p418323]page[/url] (which refers to i2c on the due in general) but am an unsure as to what lines of code to change.

I've tried the i2c scanner which picks up the device address just fine.

Can anybody advise on how to resolve this? or perhaps there are other ways in which to read the SDA pin?