I have long wanted to reach the thermal imager - in my profile I’ve probably had an ad for a year. Finally, the imager was found - and I wanted to share what I was able to remove and a story about how they work.

The temperature of objects can be determined because the amount of IR light emitted by objects primarily depends on the temperature and less on the material ( emissivity ; in correct pyrometers and practically all thermal imagers, you can adjust the emissivity correction to get a fairly accurate temperature.).

With increasing temperature - radiation becomes shorter-wave, from 700-800 degrees already capturing an area of ​​visible light. Both thermal imagers and pyrometers (infrared thermometers) and motion sensors work on this principle.
')
Thermal imagers, like many interesting things in this world, were originally invented for the military. Any warm objects inevitably glow in the IR light (day and night), the air freely transmits IR light in the range of 7-14µm and finally - fog, dust - retains IR light much less, that is, in conditions of poor visibility it can be seen farther.

Transparency of the atmosphere:


Of course, optics require a special one — ordinary optical glass after 1-1.5 microns is no longer transparent, so you have to use germanium, silicon, zinc selenide, salt (but the optics from it require coating with a protective film because it is destroyed by moisture in air) or mirror optics ( coated with gold, aluminum, copper or molybdenum). Also, some plastics are transparent to infrared radiation - for example, plastic optics can be found in pyrometers and motion sensors.

You can see infrared radiation by measuring the heating of the micro thermometer matrix (this is called a microbolometer). Each pixel of a microbolometer has a size of the order of 17-45 micrometers (there is no place less, the wavelength of IR light is in fact 7-14 μm).

The material from which resistances are made can most often be vanadium oxide (requires heating to the phase transition temperature, maximum sensitivity) or amorphous silicon (cheaper and easier to manufacture, but less sensitive). In motion sensors, the sensor is ferroelectric, it is inconvenient for measuring the absolute temperature (as it responds to changes in the radiation flux)

The military also has infrared cameras for a range of 3-4 micrometers - there is usually a semiconductor sensor, with much greater sensitivity, but requires cryogenic cooling (in the best case, Peltier core elements, at worst - liquid nitrogen). In the civilian sector you will not see this much.

Thermal imagers are rapidly becoming more accessible - if previously they didn’t allow 25 kilobax and no threshold, now - the simplest Flir i3 with a matrix of 60x60 pixels can already be found for ~ 50 thousand rubles. The thermal imager that came to me - UlirVision Ti384 (384 x 288 matrix) is slightly more expensive, about 220 thousand. The most advanced - with a matrix of 640x480 - 3 times more expensive. And for more simple industrial applications - there is already MLX90620 - a $ 16x4 sensor for $ 65, recently lit up in the Habré .

Practical application in the civilian sector - control of heat leaks in homes and industries, control of electrical equipment (if there is poor contact, this will be seen long before the breakdown), production processes. Well, recently there was an article about the thermal imager in the data center .

Imager itself: UlirVision Ti384

Telephoto lens, 53mm, similar to germanium:


The matrix itself. It looks like it is covered with a protective cover that is transparent only to IR, so it’s useless to look at the optical microscope:

Photos on request

Kitchen

Fresh fritters:


They are from the fridge:



Street (most part - filmed on a 53mm telephoto, standard lens - 14mm)


Bath water


Windows, batteries, etc.


Electricity
Dashboard, 2.35kW consumption:


Guard, consumption 4.25kW:


Socket from which 1.5kW is burned:


Pilots:


Wire mash:




Iron all
AMD 6990 outdoor:




Dog

Video

Unfortunately, this thermal imager does not know how to write a video “to itself”, only to output to the analog video output, so I had to blow the dust off the TV tuner.

Some videos are very long and tedious - there that boldly use rewind. Sounds and music will have to include their own.

Drama

In general, as in the well-known picture , having a 40W CO2 laser (emits 11 µm) and a thermal imager (sees a radiation of 7–14 µm), it is quite natural to try to apply them together :–)

I turn on the laser at minimum power - I see a point and diffraction rings ... But the beam is not visible. (By itself, only the reflected light, the laser itself in the lens is not a candle, so the matrix can be cut). Well, you need to add power ... 5W, 15W, 25W ... And then suddenly I notice that the spot at the point where the laser shines leaves a cable on the thermal imager ... With a characteristic bang, a couple of bricks fall out.

I turn everything off, after 5 minutes I turn on the imager again - the “loop” is still on the screen. Here bricks fell out weakly. The "train" of course decreased, but it remained all the same (below in the center):

Following the golden rule of laser workers “not to look at the radiation with the remaining eye,” the laser decided not to photograph.

Fortunately, the train on the screen gradually passed with time, so everything worked out.
If you have both a CO2 laser and a thermal imager, I recommend not to do as I do.

Update: Dave on EEVBlog made a video review in time: