**DIY Thermal Imaging Camera: A Budget-Friendly Alternative**

In this episode of DIY or Buy, we explore the world of thermal imaging cameras and create our own DIY version that costs a fraction of the commercial price. With the help of affordable components and some creative coding, we'll show you how to build a functional thermal camera that can be used for various applications.

**The Need for Thermal Imaging Cameras**

Thermal imaging cameras are handy tools for electronics enthusiasts, engineers, and makers. They allow us to detect temperature anomalies in circuits, which can be crucial for identifying potential issues or hotspots. However, commercial thermal cameras can be expensive, with prices ranging from $500 to over $1,000.

**The Theory Behind Thermal Imaging Cameras**

To understand how thermal imaging cameras work, let's dive into the theory behind them. These cameras use infrared radiation to measure temperature differences on a surface. The sensor reacts to infrared light, which is emitted by all objects, and converts it into a heat map. However, there's a catch – the emissivity coefficient can affect the accuracy of the readings.

**The Emissivity Coefficient: A Challenge**

The emissivity coefficient is a value that represents how well an object emits radiation compared to a perfect blackbody. Different materials have varying emissivity coefficients, which can lead to inaccurate temperature readings if not accounted for. Commercial thermal cameras often come with adjustable emissivity coefficients, but our DIY version will have to make do without this feature.

**The Components: A Rundown**

For our DIY thermal imaging camera, we'll be using the following components:

* MLX90640 sensor (32x24 pixels)

* ESP32 microcontroller

* ILI9341 screen

* 3D printed enclosure

* Li-Po supercharger circuit with a Li-Po battery

**The Code: A Collaboration**

We'll be using the Adafruit library to interface with the MLX90640 sensor and create a temperature map. However, we won't have fine-tuned emissivity coefficients, which will affect the accuracy of our readings. To make up for this, we'll use the code provided by stoppi, who has created an impressive DIY thermal imaging camera project.

**The Assembly: A Hands-On Experience**

Once we've assembled all the components and written the necessary code, we'll have a functional thermal imaging camera that can display temperature maps on an ILI9341 screen. This camera will be capable of outputting max, min, and centre point temperatures, as well as converting temperature values into color differences.

**The Results: A Comparison**

After testing our DIY thermal imaging camera alongside a commercial one, we found that the low resolution and missing picture from a real camera make it difficult to determine hot components on a circuit board. The emissivity coefficient also affects the accuracy of our readings, making them less reliable than those from a commercial thermal camera.

**Conclusion: A Verdict**

While our DIY thermal imaging camera was an interesting project, it's not suitable for the intended purpose. The low resolution and missing picture make it impractical for detecting temperature anomalies on a circuit board. However, this project serves as a great educational tool for explaining how thermal imaging cameras work.

As always, thank you for watching! Don't forget to like, share, subscribe, hit the notification bell, and maybe even support me through Patreon if you enjoy my videos and want me to produce more. Stay creative, and I'll see you next time!

WEBVTTKind: captionsLanguage: enSo I finally gave in and bought myself a thermalimaging camera which is a handy tool to havearound when working with electronics.For example you just created a new circuitthat you now want to test for the first time.Everything seems to work just fine but aftera few minutes you notice some odd behaviourand the reason is of course the missing coolingof certain components.But by using the thermal imaging camera rightfrom the get go you would have noticed thispossibly component destroying temperaturerise way faster.Or another practical example is that you gota busted smartphone that you would like torepair.After taking it apart you of course have noidea where the problems lays but by usingthe thermal imaging camera; there is a goodchance you might find a hotspot in the circuitand thus a clue which sadly was not the casewith my smartphone since it was only drawinga little bit of power.But I think it should be clear that such acamera can be very useful and the only reasonwhy I hesitated to get one for this long wasits price of roughly around $530.Of course not every maker wants to spend somuch money on such a tool which is why inthis episode of DIY or Buy we will have alook at more budget friendly thermal imagingcamera modules and use one of them in orderto create our own DIY thermal imaging camerathat only costs a fifth of the commercialones price.And in the end we will do some comparisonsbetween them in order to find out whetherDIYing such a tool makes sense or whetherwe should stick to the commercial solutioninstead.Let's get started!This video is sponsored by JLCPCB, which isa PCB manufacturer that I can highly recommend.And today I am happy to announce that JLCPCBnow offers aluminum boards.That means you are no longer restricted tocommon PCBs where thermal conductivity ismore or less pretty limited.Instead you can just order aluminum boardsfor your thermally demanding projects fora price of only $2.So feel free to visit JLCPCB to find out more.First off let's talk about how such a thermalimaging camera functions.As you would expect the eyes of the camerais obviously a camera sensor.In the case of my commercial one, it comeswith a resolution of 160x120 pixel so 19200in total.Now those pixel do not react to the visiblelight which comes with a wavelength between400 and 700nm like a normal camera would do.No; they only react to infrared light whichcomes with a wavelength of above 700nm andas some of you probably already guessed, ourhuman eye cannot see this wavelength.But every body emits such infrared light andbest of all the amount of radiation is proportionalto the temperature of the body.So what sensors do in order to measure suchradiation is firstly getting its pixels toa known constant temperature.Then as soon infrared radiation hits one pixel,the energy of the radiation heats it up andthus changes its resistance which the sensorcan measure due to an increased voltage drop.This voltage drop can then be calculated backinto a temperature and thus the sensor canpretty much measure and calculate a heat mapfor an entire surface and if you listenedcarefully you should understand that thisalso works in the dark.Of course this was only a super simplifiedversion of the functional principle whichactually sounds too good to be true; and yes,there is one big problem when it comes toinfrared temperature measurements.As you can see the camera measures the temperatureof my hand to be around 35 degrees Celsius.But when I put on a glove that obviously comeswith a different textured surface than myskin, you can see that the thermal cameranow measures a slightly higher temperatureof around 36 degrees Celsius.The reason is the emissivity coefficient whichdescribes how much heat radiation is givenoff in comparison to an ideal black body whichwould perfectly absorb and emit the heat radiation.Its coefficient would be 1, human skin wouldbe around 0.97 and glossy materials wouldgo as low as for example 0.1.You can always fine adjust this coefficientin a commercial thermal camera but that reallydoes not make it easier to guess this coefficientfor particular objects, especially when theyare reflective.That is why I would always recommend usinga contact thermometer for precise temperaturemeasurements and only a thermal imaging cameraif you need to observe a big area in orderto look for anomalies.And with that long theory out of the way let'sfinally get to the first IR thermal camerasensor which I actually had lying around foryears now.This is the AMG8832 that comes with 8x8 pixeland apparently it is nowadays obsolete andreplaced by the AMG8833 which you can geton such a lovely breakout board.My board however is a rather big evaluationboard which I really didn't feel like transforminginto a thermal camera.But just for fun I hooked it up to my computerand downloaded and opened the provided softwarein order to find out whether this sensor evenworks.And it seems like it does but there was somethingwrong with the serial connection which frozeafter 4 to 5 pictures.That was not a problem though because theresolution of 64 pixel made it clear for methat I really do not want to use this sensorfor a thermal camera or did you recognizemy finger or my head in the picture?So I put this sensor board away and insteadhad a look at this MLX90640 which is a bitmore expensive but comes with 32x24 pixelso 768 in total.Unfortunately you apparently need a degreein mathematics in order to calculate valueswith the sensor but luckily Adafruit is onceagain here to help us with a great library.And in case you have not noticed yet; usinga normal not so powerful Arduino with thesensor will not be easily possible which iswhy instead I went with this ESP32.After connecting its I2C pins to the sensorwe can simply use the Adafruit example codein order to output the temperature map overthe serial monitor.At this point though it was hard to see whetherthe sensor even works correctly.So the two things we can do to improve thatis to increase the refresh rate to 4Hz aswell as using an ASCII type format for thetemperatures.And I feel like now you should be able tosee my head in the serial monitor data, brilliant.So next it was time for the real deal by addinga small ILI9341 screen to the setup and creatinga bit of code which not only outputs the max,min and centre point temperature but alsoconverts the measured temperature values intocolour differences like a commercial thermalcamera would do.And after a couple of hours I was able tocalculate some important values but what Iwas able to output on the screen was prettymuch just a joke.Not only do the colours not make any sense,the refresh rate is also terrible.That was the moment though I stumbled upona project by stoppi which had the same goalas mine.So I reached out to him and asked whetherI could use his code and he said yes.So huge thanks to stoppi whose YouTube channeland website with various interesting projectsyou should definitely check out.Buy anyway after uploading his code, the LCDfinally greeted me with a more promising lookingheat map.Now in order to turn this hardware setup intoa crude DIY thermal imaging camera, I designeda rudimentary enclosure, 3D printed it, mountedall of the components inside it with hot glueand screws, added my Li-Po supercharger circuitwith a Li-Po battery to the mix for powerand closed everything up.And just like that you can make your own DIYthermal imaging camera for roughly around110$.So time to put it to the test in direct comparisonto my commercial camera and I have to saythat while stoppi did an awesome job withthe code, this DIY camera is not really suitablefor what I had in mind.The low resolution and missing picture froma real camera makes it pretty much impossibleto determine the hot components on a circuitboard.Also since you can not fine tune the emissivitycoefficient the temperature readings willbe even more off than usually with a commercialthermal camera.So all in all while such a DIY thermal imagingcamera initially sounded promising, it ispretty much only useful if you want to practicallyexplain how such a camera works.And that basically means that Buy is thistime for me the winner.But what do you think?Let me know in the comment section below.As always thanks for watching, don't forgetto like, share,subscribe, hit the notificationbell and maybe even support me through Patreonif you enjoy my videos and want me to producemore.Stay creative and I will see you next time.