[Damien] wanted to build a thermal camera. He was dismayed about how much a microbolometer costs so he salvaged one from a dead FLIR he picked up on eBay for 75 pounds. That’s about $100, and less than half what a new sensor costs. He selected one that didn’t turn on, which he hoped meant the Lepton 3 160×120 pixel microbolometer would not be the reason the camera failed.
Once it arrived, he pulled the pricey module, connected it to a breakout board and a Raspberry Pi. His gamble paid off; it worked fine. That wasn’t the end of the project, though. He went on to make a portable, self-contained camera with a rechargeable battery and an LCD screen.
The ESP32 software is sophisticated, using code ported from Linux and FreeRTOS. Topped off with a 3D printed case, [Damien] wound up with a great looking camera that is highly personal.
We saw another good-looking thermal camera that used an older sensor in one of our 2017 competitions. Once you’ve tired of looking at your soldering iron, you can always examine yourself for injuries.
Cool use of the breakout board.
http://www.pureengineering.com/projects/lepton
You got lucky that the needed part stilled worked — but there was an equal chance that it wouldn’t.
Those were very risky odds (50:50) — I personally wouldn’t have tried it because it would amount to a coin toss.
For me to take a gamble like this (I would have to know that said part, the bolometer) worked ahead of time.
https://www.alibaba.com/showroom/microbolometer.html
The cost of these:
$400-$1000 and up.
So given the 50:50 odds of getting a working microbolometer from a dead FLIR camera — you could buy 4 of them at $100 before hitting the minimum cost of $400 (at least based on alibaba prices.
This would increase the odds of finding a good one (by three)….
Eh, it was a gamble admittedly, but I struggled to think of a failure that would both affect the sensor itself and also cause the symptoms the seller described (absolutely dead, wouldn’t boot even with the Lightning connector connected to a phone). If you have a look on eBay, it seems quite a lot have this failure mode.
I agree, the chance that the sensor was bad was unlikely. Since it didn’t turn on a bad battery was more likely. Before lepton 3 was available on GroupGets the only way to get the sensor was pulling them from the FLIR one units.
In the hundreds of leptons I have fired up I have yet to have one not work. (Except for bad shutters, and perhaps an uncalibrated bad row(fixed via FFC).
Eh, it was a gamble admittedly, but I struggled to think of a failure that would both affect the sensor itself and also cause the symptoms the seller described (absolutely dead, wouldn’t boot even with the Lightning connector connected to a phone). If you have a look on eBay, it seems quite a lot have this failure mode.
You can also relist the item if it does not work out. I have done the same on several projects and was able to get 85%-150% of my cost. Yes, sometimes I made a profit.
do you know how many other parts in the camera could have gone bad before this one? it is not a 50/50 chance….
The senor Will work or it wont. But that wont make it 50:50. To know the odds we would need to have statistics on on how these devices fail.
In My experince it i would think the odds of the sensor being broken in a device that wont even power on to be pretty small unless the device have some design flaws.
IF we use your method of odds i must say that, given when i get back home from work sombody could have left a million bucks om my door step or they wont have done that. Given those 50:50 odds i must say that i have bad luck since i never found a million bucks on my door step comming back from work.
Dizzey, great comment. It reminds me of the joke that 100% of the people who quit smoking die. It says something about how statistics can be horribly abused which is a very good lesson.
It’s not 50:50 by any means, nearer 95% chance that it will work. I have bought a good few non workers off ebay and they are all easily mendable so I still haven’t got the Lepton module I want because it makes more sense to mend them and sell them again than it does to rob the module from one that works.
I’ve got one of my guys does something similar. Buys dead junk from ebay, fixes it, cleans it up and re-sells it. Does OK out of it.
Works/Doesn’t work isn’t a 50/50 chance. You’d have to know what percentage of faults would affect the sensor module itself. Given it’s the most expensive part of the instrument and it’s a high cost instrument, I would think it’s very well protected and most faults would be due to the power supply.
I’d say it’s a pretty good chance of finding a good sensor in one of these.
Great work!
Some valid points.and.worth noting… but I noticed very FEW of the MANY which said much about rhe great idea and work he did. But thinkiog of a dozen sub-systems in it that could easily fail, the sensor gets hit by mere photons, not milliamps… which weight a lot more and move almost as fast. Fair odds… And GREAT work!