Products > Thermal Imaging

My decent in to Infrared abyss (IR microscopy and telescopy, on the cheap)

(1/2) > >>

So, I got bitten by the IR dongle bug not so long ago, after it dawned on me that modern thermal dongle are better than the thing Predator 1 and 2 was filmed on.
I always wanted one, but the limitations of hardware (and lack of my personal skills) kept me at arms length from dongles, until now.   
After I conquered the PC to HikMik Explored E20+ interface thanks to helpful hands in DYI Thermal Imaging Discord, I started to expand my repertoire of what i can do with this wonderful little device.

The limitation of this device are obvious. Its meant for outdoors, fast refresh rate and good contrast but more or less unreliable temperature readings. And that is fine, I got it for pushing limits, not finding grid faults.(tho i am thinking of putting a poster on local boards for "thermal scan for leaks, call XXX-...", we got -30 winters here in Scandoland, a  quick coldspot sweep out and in will find the worst faults, and handheld spotter can take accurate readings, but i digress.) Despite the dongle being meant for long distances, mostly, it has respectable macro capability thanks to the manual focus. In a pinch you can use it as a fault finder and spot that 2mm SMD resistor heating up on a laptop PCB no problemo. It has respectable range in spotting human-sized objects, on a winter backdrop, over a lake, I have managed to see humans roughly 500m away. Identifying humans, reliably, begins at 200m, now you can tell human from a large dog 99% of the time. Identifying gender start 100-70m and you can start distinguishing facial features from 20m, the mouth and smiling face will start to resolve.

So, now that I got a good feel of what this thing can do unaided, its time to see what i can do with some Aliexpres parts. I got a pair of CVD D12mm F38.1 meniscus ZnSe CO2 laser focusing lenses and a Acuter Newtony D50mm 200F f/4 Newtonian telescope and went to town. I tried to do some math and use this site's wonderful simulator for planning, but most calculations I did was more or less back of the napkin. The AR coating alone would need a more professional program to model. Needles to say, I struggled to get it all work for a while. Firstly, I tried Mittenzwey configuration, but ended up going with Ramsted as it boosted the FoV and cleared the image by removing most of the IR noise from the focus. Both configurations worked fine as a IR microscope, I could get amazing IR closeups, way beyond anything useful, as the heat will equalize in its surroundings and the use cases would... limited to something that had good surface for emission and low conduction to surroundings, small ceramic parts maybe? :-//
The telescope part was initially a hot mess, in all the meaning of the word. I had huge hot halo and clod center, indication of all sorts of stray light getting in to camera. After a long fight, trial and error, I finally managed to increase the image quality by adding apertures to cut off some of the telescopes own noise and limiting possible reflections between eyepieces. Tin foil worked the best, as it reflected most stray IR away and metal surfaces themselves are not as IR noisy unless you catch reflection from them. The magnification is roughly +-7x, I took a picture from 7m away at my hot kettle that is 20m in diameter and was about 2 degree in angular size. It was then magnified by the telescope to fill the whole FoV, that is roughly 14 degree. Not bad! I was also pleasantly surprised, that the mirror could still gather enough IR to show the paper towels against background, but very faintly.

Next step will be getting some first surface gold coated D50mm F100mm spherical mirror to replace the one that came from Acuter. Finding gold coated secondary mirror has proven difficult. I am also thinking of adding some baffles and maybe Peltier coolers with heat-pipes to cool the optics, holders and mirror to sub ambient (some sensors+ logic to keep  temp above dewpoint). If any one of you fine gents have tips or ideas, I am all ears!

So, here are some pictures from my toil and labor.   
picture 1 and 2: Eyepieces
picture 3 and 4, the telescope
picture 5 is camera's own macro, 6 and 7 are the microscopic pictures with eyepiece
and the moneyshot 8 and 9: the pot and pot trough telescope

Here is more pictures with two different configuration, two identical D12mm F38mm ZnSe lenses forming the eyepiece.

The target image is a kettle, heated to 50C, room temperature is 21C. Distance is the same 7 meters. Different configuration magnification ranges from ~5x to ~8x.

Picture 1: kettle from 7 meters away. notice how the ceramic tiles reflect some IR.
Picture 2: Kettle from 40 cm away, roughly the same size as picture I got from my my better images from telescope, for comparison.
Picture 3: kettle from telescope with my clearest eyepiece configuration (Ramsted), here you can see how much of the details get lost trough the additional optics. Magnification is 8x
Picture 4: Eyepiece in the Mittenzwey configuration, you can see the eyepiece elements contributing to the image, the magnification is 5x.
Picture 5: same as above but a bit shorter distance from field lens to eyepiece and added 3mm aperture roughly between the two lenses. Magnification went up and image cleared a bit.

Now that I am more or less recovered from my cold, I will go outside today and find a good spot to test this on real world distances, 500m+ and see how it goes.

I've attempted a little bit of thermography through a telescope and it's neat to see others with the same idea!  Like you, my results weren't as picture perfect as I had hoped, but it would be interesting to see what could be optimized.

I do think the aluminum based coatings of the mirrors are a limiting factor, but IR lenses tend to have higher F ratios than standard ones anyways, so especially given the price of a gold mirror, it may be a cheaper option to get a larger aperture fast newtonian (I was trying using an 6" F/4, but even 8 and 10 inch scopes aren't too expensive, though they get progressively harder to support.)  Commercially available fairly inexpensive hobby telescopes also have pretty good optics, so while you'll see some coma around the edges, their optics for visible light tend to be relatively low aberration and thanks to the longer wavelength, should be that much better with IR use.  Another thing is that recoating mirrors on telescopes is a pretty common practice - maybe not cheap, but especially for larger mirrors something that's done every 15 or 20 years - so maybe you can find a place that is willing to recoat an existing telescope's optics with gold (or another less expensive material that still has better IR reflectivity.)  If zooming in at objects up close, you'll need several spacers in the eyepiece tube before your camera core, but for my 6" f/4, looking less than 10m out required something like 200mm of spacer tubes to find focus.

Personally, I think the peltier cooling of the mirrors probably wouldn't get you much performance and the dewing over of mirrors would be a hassle.  Unless your core is also cooled, you'd be running into the same limitations as looking through a lens in terms of heat noise of the optics, and if your mirror coating is reflective to IR, a lot of the heat from the glass of the mirror itself isn't going to be contributing much noise.

@DaJMasta: you brought up some interesting points, thank you for that. I would be also interested to see your results and gear you used in your experiments with IR telescopy.

As for my goals with this: the general idea is to keep the telescope inside the handheld dimensions. The limiting factor is the IR optics, the price for them jumps like it found a spider on toilet ring, when you start going to diameters over 3cm or even worse, parts that are not mass-produced for some industrial need. This was the main reason I went with the mirror as primary collector. Good, <95% transmittance, 5cm diameter 8-12um IR lens or window cost as much as my dongle. While 5cm mirror telescope is 35 euros, even the gold parts are 100 euros or less.  The point is not to create an astronomical device, just a good long range scouting optic with affordable, off the shelf, parts. The gold mirrors I've been able to find are well inside my meager budged for this project. I am planning to get more Cloudray's CVD ZnSe (the quality seems to be good enough for this job) lenses and IR window or two to eventually seal it all up. As you pointed out, the mirror themselves emit far less IR than they reflect, so while I am interested to test some sub ambient temperature performance for them, its not a hill I want to stick my flag on. The eyepiece for the dongle is far more noisier, but also something you can seal up without breaking the bank. And you will have only one surface to worry about getting moist if you can make sealed compartment for the camera dongle. However, talking about cooling is premature, as I have not yet even made proper parts for the optics. And so far I am making by with whatever I can fast prototype with and find at home or in local shops.

For my immediate goal I will need to make better fixture for the camera and lenses, as filming video, pointing telescope and keeping everything aligned turned out to be a big hassle. Anyhow, here is some tests in the wild with my current setup. The weather was very challenging for the camera: most surfaces were wet and glossy, so while anything hotter than background stood out like a neon sign, terrain itself was almost featureless and hard to resolve.

The images are screencaps from video, so they are not as sharp as the pictures but they look nice in the video, I swear!
Image 1 and 2: two people on the buss stop 100m away.
Image 3: a woman with a bag, 300m away.
Image 4 and 5: downrange image with clean dongle, in picture 5 the signature numbered in 1 is woman with bag (300m), signature 2 is person entering human identification range(~200m) of the dongle and signature 3 is man waiting for bus 100m away. The bike rider is inside the range (10m) where you can tell from his face if he approves my weird hobbies.
Image 6: a general view of my local area, the crane cabin looks warm and cozy (~150m).

All in all, i consider todays test a success as proof of concept.

Time for some updates!

So, after tinkering with different lens arrangements and building some mounting parts. All with time honored tradition of using optical-grade plumbing PVC  :-/O ... I have created something that only its mother creator could love, but it works... after a fashion.

This project has been a huge learning experience in optics for me, and what have I learned?
1# In optics, reflections inside your system are lava.
2# In IR optics, lenses are lava, fixtures and tubes are lava, your camera is lava, you holding the blasted thing is lava, lava is lava and internal reflections are giga-lava.
3# Having your optics aligned helps, allot. Minimize the source of errors. So don't skimp on secure mounting and worry about how it looks later. This is prototyping, not production run.

I managed to take some good shots at my towns frozen lake side. While the temperature was +1C, wind was rather gusty and biting, and everything was wet as usual.
Despite magnifications hovering around 10x, better alignment of the system as a whole permits much more of the IR signal to reach the sensor than in my previous freehand tests.

Here is youtube link for the video I made.

Some images to go along with the video.
1. The map showing the average distances and names of locations (I have them marked in video too, mostly)
Rest are my jury-rigged setup: it is as ghetto as they come, but hey, it works! (notice my patented viewfinder reticle put over vis camera image, makes it possible to point at things you want to see...or close enough 8) )


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod