Most people, Netflix and chill. GreatScott, Hot glue and Screw.

Once you have such a camera, you automatically start to find more applications for it. Like finding leaks in the insulation of your house in winter. Or finding where your cat is hiding in your bedroom.

Personally a thermal camera + ZnSe lens has been invaluable to me in troubleshooting shorted boards. This made a thermal camera well worth it for me.

As limited as the DIY option may be, I actually think I could use it for diagnosing hot spots on PCBs, though the drench it in alcohol method might work better.

If the surface is shiney according to FLIR, just put a piece of painters/masking tape on and you can then measure the objects temp.

I suggest: Lead acid battery desulfator, DIY or BUY.

something you didn't think of: you can simply overlay a LCD without a backlight, over the thermal image, it may need another esp32, but might be worth trying, using another esp or controller would also mean you could lock it to the same framerate as the thermal camera, without needing one controller to do too much work

The DIY will always have advantages with his skill.

Adding a simple regular camera and adding edge enhanced overlay would only be a few $$ more, and would provide an experience that feels visually more higher-res. Also emissivity can be easily programmed into the software. These would be fun additions but given the price per pixel is much higher DIY in this thermal sensor, BUY is a good conclusion for a goal of a practical device. I like your DIY vs. BUY series.

The Flir E4 can be "upgraded" to the 320x240 E8 with a firmware replacement. Huge value, and you can Google it easily.

Very cool video! My husband is the big fan of thermal imaging technology and has bought an infiray t2 to detect some thermal leak in the house, i will ask him to watch this video.

Ayyy, another DIY or BUY great! :))

I have the FLIR TG167. It does a good job. I've used it to find thermal "leak" around the house, monitored wheel bearing temps on my travel trailer as well as using it for monitoring my electronic circuits.

You could calibrate the DIY camera for use at a set distance and use it stationary. That way you could set something down in front of it and depending on the size and how far away it was, you could adjust the parameters to measure the X, Y of where the hot pixels are and know very accurately where the component is. It just wouldn't be a handy handheld tool for quick observations.

I think DIYing makes sense most of the time because you learn electronics while building!

It’s uncanny how well your videos sync up with what I’m working on. I feel like just a few days ago I was thinking about this exact same problem.

Very helpful video , I just joined a thermal camera company one month ago as calibration engineer 😀👍👍👍 And from the bottom we start from 80*80 sensor and at last we have 640 sensor of Ulis (a French company) very expensive 😂😂

Me : thinking thermal camera a really complicated you can never diy them Great scott : hold my beer

Bought a Flir One 1 year ago for my Android SmartPhone, it was only 199$ and is a Gen-III Thermal Imaging Sensor, it's very sensitive and has an extra 1080p camera built in as overlay so you can see details as well as the heat spots. Totally with 199$, but it does have a few drawbacks. Even though it auto-calibrates, calibrating can take some time. And it's prone to lock-ups and freezes, for example if you walk around with it from an cold area to a hot area, it can freeze up. Also since it plugs into the phones USB-C connector, it can fall right out of it. They have fixed the notorious dying-battery issues (the earlier models usually lasted a few months, then the battery just flat out died), but for 199$ you get a cheap mid-range resolution sensor, and that ain't all bad. I use it to find heat leaks in my house, check for pipe-leaks in the houses piping, and fire risks in the chimney. I also use it as Great Scott does, to find overheating components. It's also useful to detect thieves in the yard, something suspicious? Well they won't be able to hide from you - you can easily spot the hot body either of an animal (maybe your cat is missing?) or anything big and living in any dark areas. Here is an example image I took of my 3090 card to see how hot it gets during idle process: i.imgur.com/WpuLjkK.jpg

I cannot understate the usefulness of thermal cameras in electronics design. At work, we have a design we've been making since 1985, and for some reason the output channels periodically just die on it. The why had never been figured out until we got a thermal camera recently. I was curious to see if there were any issues, and one of the components was rapidly cycling between room temp and nearly 250F when the output would go into the positive cycle of its waveform. On paper, the components used were perfect for the application, in reality we swapped a 110 ohm resistor to a 1k resistor and it fixed the thermal issue.

Thank you for listening! Epic project!

For all those interested in acquiring a thermal camera: UNI-T has some budget options starting at around 130€ with a resolution of 120x90 (10800 pixels). However, they lack the handy Flir feature where it superimposes edges of the visual world (FLIR feature a normal camera next to the thermal one in its thermal imaging units). Aside from that, you can get Flir usb addons for smartphones starting at around 220€ that feature this image mix, tho the cheaper models do come with a lower thermal resolution (80x60) than UNI-T's. I can also recommend such a tool for renovations and thermal isolation of pipes, windows, and so on.

I like your perseverence with getting a good solution with your diy projects! Not an easy task to beat manufacturers in having great products!

my favorite serie

Your English has improved so much, bud. I'm so proud of you.

Get hold of the ESP32 camera module, remove the IR filter, revamp the colors to give IR representation.

Everytime Me: "I want to buy it (anything) but it's very expensive". 🤧 GreatScott!: "Here is a cheaper DIY Solution" 😁

Great video. So valuable for my learning. Does Flir manufactur their own sensor? It seems a thermal sensor should be available that has performance closer to the Flir one.

You can always use a laser pointer to show where it's taking the readings or will that interfear with the reading, example laser points at the component imager reads temperate that way you don't need high resolution

Go for cat62 GSM. For 600 bucks you get telephone with pretty decent flir camera. My wife uses it to generate heatmaps of buildings, but I've tested it for electronics and works great. It's more expensive than flir you have, but you get added bonus of very rugged telephone which can survive extremely harsh conditions

Cost on these is still high, so the sensors must be expensive, but not as much as the Webb telescope which will have the ability to look back in time into a specific IR wavelength for doppler shift.

You're really talented! But I prefer to just buy one.😆 I've just bought Infiray P2 Pro thermal camera. It's pretty amazing what I can get at such an affordable price nowadays!

Why not get one of those "seek thermal camera" which attaches to your smartphone

Others : Netflix and chill Me: Greatscott and learning a new skill

Why not use a visible spectrum camera module and overlay the thermal map over it?

I just bought an Hti301 from aliexpress. Cost was 600usd. 384*288 pixels. And with the app thermapp it is friggin' amazing!

Maybe you should have used a raspberry pi zero instead and added a cheap 3M pi cam to overlay over the thermal image. it would have been a bit more expensive but would make it perfectly suitable for your use case.

For anyone wondering, it's possible to get a high refresh rate off this camera when using i2C you just need to set the clock speed on the ESP32 to 1MHz and the refresh rate to 32Hz on the camera.

Well, if you put a glove on your hand, you thermally insulate your hand from the surroundings, and its surface temperature will also increase. So it will be a combined effect (I would say the insulation might be more important)...

Those cameras are long-wave IR (LWIR) cameras, they are sensible in range of 7-14 um, but not from 700 nm as you stated.

I also have some mosfet overheating problem, its working at around 500khz, but the current is only 1 amp, even though the mosfets are rated for 50 amps each, what could be the problem? Anyways great video,love how made its body too, 🙂

I don't disagree with your conclusion, but there are some reasons why you might want to DIY a thermal camera. If for instance you need a thermal camera with a high refresh rate commercial solutions will be very expensive compared to DIY. Or if you're interested in a specific temperature difference, a DIY solution will make it easy to focus on a specific gradient. I made a thermal camera based on the AMG8833 in combination with a regular camera. And it worked really well for tracking down leaks in the insulation of my dad's house. The nice thing was that I could set the threshold for a colour change to what ever I wanted, so I just took a red to blue scale and made everything that was below the ambient 18 degrees red. It made it super obvious where the major leaks where. I also used it to find hotspots on pcb's and it worked well for that too, although the resolution wasn't high enough to use it on small size components like smd-components. For the actual temperature of a hotspot a thermocouple is always the most accurate solution. I really hope a higher resolution open source sensor will become available, so I can build an improved version 😁

The DIY videos are the best! And more appealing to a wide range of viewers 👌🏻

Two other buy options: you can get USB C and Lightning FLIR attachments for your phone, which cost way less, but then your phone is the screen, you can easily save and share images and video, and eat mor control. You can also get smartphones with built in thermal cameras, although they’re typically targeted towards construction/engineering folks and while they’re typically very rugged,, they often lack other features for regular phones in that price point

FLIR used to make some cheaper version of their thermal camera that actually used their high resolution sensor, just artificially limited. The full sensor resolution could be restored with a firmware hack!

I needed an aluminium board and had no idea where to get it, so glad that they offer that now

scott you always amaze me whenever i get new project i always find something helpfull on your channel this time i abolutelly got what i looked for im working with cars and i try to diy a lot of things like year ago i made conversion from distributor to coil on plug for my honda now im trying to make tire surface temp monitor and literally all i need is simple small temp camera that will just give me temp across tire and its exactly what this camera u made is suited for as for what u said in the end about diy camera not having emissity coeficient i think its all down to code i can think of linear table in which u only need to put 2 values of real temp messured and 2 from camera reading to calculate it and correct it i guess it might be a lot for arduino but i guess if im about to only use about 8 pixels per tire because more isnt needed it could work, the refresh rate is a bit worrying but maybe camera with less pixels could give me better refresh rate i guess 12hz is like lowest i could go but perfectly it would want to be about 20 so i can monitor the data with precission there is alot happening when u race a car and having a lot of data to check after session is never a bad thing

GreatScott is the best at all time... learned a lots of things from you. thank you so much.

Actually you can also buy a USB OTG thermal camera which can be plugged into a mobile phone. But anyways, this was a good and informative video! 👍

This is the exact question I've been facing. I want one that's good enough to find hot components in cell phones or computer boards but not more expensive than it needs to be. If you buy one too cheap you waste your money since you have to throw it out and start over. Thanks so much. Sorry this did not work.

I was always curious how these worked. Very cool. Thank you for the high quality explanation!

Such a cool video! You are so brilliant! But for people like me who are not good at doing things, I choose to buy. Because now 200 can also buy only love that very good thermal imaging camera. I have infiray p2pro only more than 200, pixels have 256 * 192, motion capture is 25Hz.

Dont have a degree in maths , did the calculations,it was super hard, proud of myself ;-)

This is a great project. The outcome might seem a little disappointing, but it does show you what will be possible in a couple of years when higher resolution sensors become available at a more affordable price. Perhaps some of the limitations of the sensor you noticed will be possible to overcome with more sophisticated software or faster processors (like a Raspberry Pi or similar). They could certainly have multiple cameras. Although until the cost of the sensors come down anyone with a practical use for one is going to be better off buying a commercial unit. I was interested to find out, however, that IR sensors were available.

Scot, could you try to spend the same amount of money for DIY projects as you would for the "buy" option? This would give a real answer if DIY vs buy. Or a small modification: How much money do you need to spend to match the "buy" option? Anyway you doing a great job, I'm from a completely different industry but I'm having a good time while watching your videos.

Could you cheaply buy a bunch of low res sensors and use software to overlap them for better res and refresh rate?

Hello Scott it is far IR that is based on Temperatur other IR cam using near IR. Like secure cam..

Is there a reason nobody seems to mount the thermal and visible sensors physically aligned with a hot/cold mirror so there's no need for all the fancy software image alignment? It worked for 3ccd color cameras.

One option you didn't explore was the hacking that was done with the SEEK thermal imaging cameras. They are very affordable IR modules with a USB interface, but it's all proprietary. I think they managed to crack the older models but the newer models only produce a raw unprocessed image last time I looked into it.

Thsnks for a reduction in the pain of doing your own ir camera

I'm now curious about the extra cost of a higher resolution sensor and display. How much more resolution would you need to replicate FLIR, or make it usable for electronics diagnostics?

There's a pretty solid smartphone with built-in thermal camera that performs quite well considering the price (camera is made by FLIR). Look into Ulefone Armor 9 (older, around 450$) or Ulefone Armor 11T (newer, around 500$) if you're interested. Both models use FLIR Lepton 2.5 sensor. Just thought I'd mention it in case anyone has use for it.

A better sensor with the same resolution as the camera is the flir lepton 3.5. There are also breakout boards for it.

распечатать на принтере корпус и вкорячить начинку на термоклей!- шедеврально

why is thermal so expensive

This. This is why I would want thermal equipment, essentially large batteries.

You've got to change the battery in that S7. Had the same problem with it drawing ≈0.2A and when I replaced the battery it was all fine.

I don't see if anyone else has already suggested this, but many of the (cheaper) commercial solutions will mix the thermal image with a normal higher-res mono image to realize a composite image that looks MUCH better than the raw thermal. The human eye is very forgiving of low res color channels in a video signal, especially on such a small screen, provided the luminance portion is even half decent.

Interesting. Not sure if you're aware or not, but digital camera sensors are sensitive to IR. They're all fitted with an IR blocking filter... There was someone removing these filters to get night vision. And guys were adding filters to filter out daylight. If you have an old digital camera, you should be able to experiment and feed the image into the software you borrowed.

Changing the screen or the screens library for a faster one would help and a better looking image can be had with bilinear filtering, I've even had quite good results with the amg8833 using this technique. I just need to pull my finger out and build one with the mlx90640 which I've had sitting around for 2 years! Great video though and I'd love to get a flir but can't justify the money for the real use I'd get.

Bought the new UNI-T from aliexpress 300$... much higher resolution that this one too!

Well that's a interesting idea. What about trying to convert a simple normal camera sensor in a termal one . Just a simple question

Really a great idea. A shame to have to work with so low rez. I'm wondering, since a brand new is 450€+, what precision could you get by spending let's say 250 on that kind of project ?

Amazing as always. Thank you for explaining how to troubleshoot with it. I am going to try that (because my circuits rarely work first time). It is obvious that the components you tested for your DIY version are not granular enough for the kind of work you mentioned. In my mind, that just means you haven't found a better camera module yet with a DIY price tag. If you could bring the DIY version under $250, then it would still be a good DIY win assuming getting better resolution. It seems as if finding a module with DMA transfer and hardware DSP to handle the correlational computations would probably require an embedded MCU or PIC, so the price would much higher. You kinda missed out for this video because tearing down the expensive camera could have shown what model components they use.

This is the reason i love Sunday's. GreatScott is gonna post something cool

Video suggestion: Rp2040 (raspberry pi pico) as a good arduino alternative.

You made a thermal camera. That's awesome

You should have compared your camera with a commertial camera with the same resolution, for instance FLIR One (plug in for the smartphone). Actually, can you make a DIY plugin thermal camera for a smarphone instead of a handy one? Anyway, great video. Very inspiring!!

I got myself a cheap off-brand version. The resolution is about half way between your DIY and the FLIR; I've used it now and then for troubleshooting and/or augmenting my videos with stills - as it cannot record video. But then, the price was also half way between your options: about 250$ IIRC.

I loved writings and also project

The only issue I see is that an upgraded DIY unit would probably tip the scales around $200, and still be vastly inferior to a commercial one. Taking a quick look, I see the FLIR ONE Gen3 (80x60)is available for $200 (iOS or Android phone required, but who doesn't have one of those?). I guess that settles the debate. I also see 160x120 unit for just $350 instead of $530. Sticking with the DIY solution, a RPi could easily be used so that you could overlay the thermal sensor data image over the standard RPi camera. Image manipulation is definitely intimidating for the average programmer, but easily doable. Seems to me like 99% of the time you just want to visualize temperature differentials so I really wouldn't worry about the emissivity coefficient, but it's basically just a scaling factor for that data represented by either a constant or a variable as far as the code goes so that would be easy to extend.

Does the thermal camera show any effect if the distance is changed while measuring the temperature of an object?

that diy version is fricking cool great job dude

I was literally just looking at some of these other day and now I'm trying to invent a problem to solve so I can order some from them 🤣

I'll save the $530 and stick with the wet finger technique I've been using for years thank you.

Awesome work, dude! Thanks a lot! 😃 Even though it didn't work out, it's a fantastic starting point! I bet we're going to see a fully functional IR camera pretty soon! 😃 Anyway, stay safe and creative there! 🖖😊

If the budget would be around 250$, would it make worth building one? Around 20% of the retail price of the buy one is a little bit wide to compare.

Great 👍🏼 Scott !!!! I purchased a cheap module but the quality is really poor. 8X8 Is it possible to add normal camera along side thermal sensor , and merge the video , this way thermal image will be overlaying normal photo. And resolution could be higher ( that’s what I think , might not be possible)

I’ve been waiting for this vid for a longgg time since I saw it’s preview on the patreon

If you want to use a thermal camera on any compact circuit board "Smasung Galaxy board", you will need a "MACRO LENS". Buy a ZnSe Meniscus focusing lens for Co2 Laser, put it in front of your thermal camera which ever way you know how and bam you can do circuit diagnostics with your low res thermal camera. Also Flir's Double image is next to useless. Seek Compact Pro is what you want, you do need a smart phone but you can always dedicate an old smart phone for this task.

Last conversation from the Thermal camera store... GreatScott: The FLIR TG267 with a 160x120 resolution... Store owner: That´s a great thermal camera, you know your stuff GreatScott: The GridEye 2000 360° HD with a 400m range and 1080x1920 resolution Store owner: guy, guy, only what you see on the shelf GreatScott: The GridEye 8x8 pixel

If you want it for PCBs only i would go with the chinese high quality brands these days. Qianli Supercam X for example has a very good resolution and good focus abilities (from 2mm to infinitiy). It has a 260x200 pixel sensor which is by far the highest out there as far as i know (in that budget range). Due to the mix of the nice focus lens and the high resolution you can even see the heat distribution of the single PCB traces. For a cheap entry a used or refurbished Flir One Pro or Seek is nice (i'm using one too), but if you spend about 500 bucks for a new one i would rather check the alternatives with much higher resolutions and better focus.

The DIY may come out on top with a higher budget sensor and switching to rasp-pi for a higher refresh rate. But at that point you are probably spending 200-300$. So I guess it depends on you budget and at what point its just worth it to buy.

Another great video from you sir Scott! :)

Hello, kindly try diffrent design . use mobile as output and convert it to usb device or via bluetooth .it may solve ur refresh rate problem.

Excellent video! I use my camera just to point out the problem area and pay no attention to the true temperature of the hot spot.

You should try with FLIR lepton sensor. Its more expensive, but still would be cheaper than FLIR camera.

So what Teledyne did was 1) Recruit a buch of quarter Greatscotts 2) Develop a working prototype spending about 200USD. 3) Polish the design and functionality of the prototype to a 500USD FLIR camera by spending 2 Million USD.

I know you're trying to keep costs low but with such a price gap I wonder if there are slightly more expensive components that can make this DIY version a bit more useful.

cam overlay could be implemented for a few extra bucks and some tinkering, but the resolution and other downsides make buy the winner i agree.