About Me

My Photo

I previously worked on Virtual Reality and other hardware at Valve.  I currently work at Google[x].

Prior to starting at Valve, I built computer peripherals such as keyboards, mice, and joysticks that were designed to be used inside MRI machines.  My company, Mag Design and Engineering, sold these devices directly to researchers at academic institutions who used them to publish scientific papers in peer-reviewed journals.

After work, I spend time on many different types of projects that usually involve circuit design, machining, material selection, and general fabrication/hacking.  My favorite place to be is my home workshop.

ben dot krasnow at gmail

http://www.youtube.com/user/bkraz333

http://www.linkedin.com/pub/ben-krasnow/4/6a9/679

http://www.twitter.com/BenKrasnow

Thursday, October 7, 2010

Low-cost DIY thermal imaging -- liquid crystal paint testing

I am trying to develop a low-cost DIY thermal imaging device. The commercially available thermal imaging cameras still cost well over $1000 because of high production costs and low demand. Many hobbyists would like to have a cheap thermal imaging camera even if performance is not as good as commercial or military units. My goal is to build such a camera.

In this video, I am testing one possible approach: Using a very thin projection screen that is painted with thermochromic liquid crystals. These liquid crystals change color in the temperature range 77*F to 86*F. Ideally, the projection screen housing would be heated (or cooled) to 77*F, so that all incident thermal radiation would raise the screen temperature higher than this, and immediately cause a color change.

23 comments:

  1. Wow, I might have to do some work on this myself :)

    A few things you might want to try:
    Larger IR collector
    A smaller (and more focused?) thermochromic region, then optically magnified
    Air-chilled super-low-thermal-mass mylar (like yours, but they have down to 1 micron that I can find)

    ReplyDelete
  2. Pete, thanks for the comment and good suggestions. Do you have a source for the 1-micron-thick mylar? I searched the web and found some "XRF sample holders", but they are brutally expensive and sold only in large quantity.

    I was originally planning to use a color camera with lens focused on the screen to discern small color variations and amplify them. So far, the project hasn't quite gotten far enough to warrant the effort yet.

    The focal length of my germanium lens is about .75" and the diameter is about 1.3", so it has an f ratio of about f/0.6, which is outstandingly low for optical systems. Germanium has a very high refractive index, and so it concentrates a lot of IR.

    ReplyDelete
  3. http://www.freeflightsupplies.co.uk/
    Check their price list, and you want "OS Film" which is .5 micron and £28.00 per 25' roll... A bit pricey, but it has more uses than you'd think :)

    The reason for a smaller, more focused imaging area is to make quicker response time to IR heat, and then active cooling (and lower thermal mass) can cool the area off quicker.

    A small webcam with the thermochroic paint put directly on the CCD (or whatever chip) might also work well. I did this myself with some Cu-activated ZnS for some a-radiation measurements.

    Super-small USB webcams from laptop LCD bezels can be had for less than $25 on Ebay.

    ReplyDelete
  4. I have made the same test using a silicium lens from ebay and some thermocromic plastic and I get similar result to yours ,it's pretty cool ,I know that the minimum temp difference that this type of ink can detect is 0.2 c and effectively ,the maximum sensibility in archive when the lc is just at the right temperature (begin of transition)I could detect my soldering iron at about a meter of distance and my kitchen stove at about 2 meter .It might be possible to increase the sensibilities by using dielectric filter on the light you use to watch the lc as the temperature change induce a variation in the wavelink of the reflected light . for the idea of puting some lc on a ccd directly ,I don't think it would work because the lc itself is almost transparent ,only good result are visible on a dark surface ,on the other hand ,if your dark room is dark enough you may be able not to use paint witch would reduce the thermal inertia of the membrane...

    ReplyDelete
  5. Anonymous, I am glad to hear that someone else has tried this method as well. I think it's pretty cool, but even with substantial improvements, it will never be comparable to microbolometer performance for thermal imaging. I ordered the .5 micron film that was mentioned in the comments, and it improved response time and spatial resolution, but not quite as much as I had hoped.

    ReplyDelete
  6. Maybe, to collect more infrared you need a larger lens (but a germanium lens, of course, is expensive)

    You might consider a plastic fresnel lens:

    http://www.fresneltech.com/thermalimaging.html

    fresnel lenses in polyethylene are used in low cost passive infrared detectors.

    http://en.wikipedia.org/wiki/Passive_infrared_sensor

    http://www.3dlens.com/infraredfresnellens.htm

    ReplyDelete
  7. angros, these are good ideas. I think the major problem with this thermal imaging idea is that the sensitivity of the liquid crystal film is just not high enough to make a decent image. My germanium lens already has a very low f number, meaning it captures a lot of light for its diameter. Getting a larger lens would certainly be possible, but I don't think the f number could be lowered much further without major image distortions. I hadn't thought about imaging with polyethylene. Cool!

    ReplyDelete
  8. Imho, film thickness is not the main problem; a very thin film could even be transparent to heat, so thermal infrared rays would pass through it, being not detected.

    The ideal film should be not transparent and not reflective to thermal infrared: it must absorb all of it.

    Liquid crystals are not sensitive to thermal infrared: they are sensitive to temperature. So, you need a material that will change its temperature in response to heat, and not all materials have the same response (they have different heat capacity); I believe that a material with a low heat capacity is needed (it will have a greater temperature change, at the same amount of heat).

    Here is a table:

    http://www.engineeringtoolbox.com/specific-heat-solids-d_154.html

    Maybe gold would be a good detector? A bit expensive, but you can do a very thin film. Also lead could work, maybe.

    Just my 2 cents.

    ReplyDelete
  9. angros, using a thick film will decrease the temporal response of the device because the film will have a larger thermal mass and will take longer to change temperature. Another problem is that a thick film will have better thermal conduction in the x and y directions, thus lowering the spatial resolution. Imagine a "film" an inch thick -- there would be no image at all because the temperature would be almost perfectly equalized across the surface. Commercial microbolometers are built so that each temperature sensing segment is mechanically isolated from its neighboring segments and is also very thin for fast temporal response.

    You are right about IR absorption being a problem for very thin films. In my testing, I have been using black paint to increase the IR absorption of the film. The paint is also necessary to be clearly see the liquid crystals. It's entirely possible that metal-coated plastic would be a better solution, but then metal has good thermal conductivity, and would produce problems with spatial resolution again.

    ReplyDelete
  10. Black paint can absorb normal light, I'm not sure it can absorb infrared. You'd need something that is black to thermal infrared.

    About resolution: the problem is the same of old film camera (for normal light): some film had an higher resolution, but they required more light: to take a photo in a dark room you could use a more sensitive film, but the price was a lower resolution. Also, a longer pose time allowed to capture more light... but with less time resolution.

    For the spatial resolution, the obvious solution is a larger detector (you'd need a bigger lens, too)

    Also, I've found these links:

    http://www.redshiftsystems.com/Site/TechnologyProducts/OpticalThermalImaging/tabid/54/Default.aspx

    http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6VM5-4KWJP53-1K&_user=10&_coverDate=08%2F31%2F2006&_rdoc=1&_fmt=high&_orig=search&_origin=search&_sort=d&_docanchor=&view=c&_searchStrId=1651777168&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=a6e48e73807b8751745ce271283a6e4c&searchtype=a


    The TLV detector seems to be based on the same principle of your liquid crystal detector... maybe you could find some useful ideas.

    ReplyDelete
  11. angros, I contacted Redshift Systems when I first started on this project. It seems they are planning to make a lower cost alternative to microbolometers. Unfortunately, they are not able to sell any samples of their product (at any price), and were generally unwilling to help in any way. Either their idea doesn't work, or their company ran out of money and is essentially on-hold until they can get moving again.

    Note how the TLV detector is built: with many cells that are isolated from each other to prevent thermal transfer. I was trying to think of a way to build a liquid crystal screen/film with that same cellular pattern. Maybe you have an idea?

    ReplyDelete
  12. That cellular pattern made me think about the collimator of a gamma camera (I use it at work):

    http://blog.imabug.net/archives/2009/01/museum-collimator-core.php

    On a gamma camera, the collimator is needed, because no lenses are available (no material can refract gamma rays)
    The collimator is made of lead (since it has to block gamma rays), but for thermal infrared, lead is not needed: the structure of a collimator is similar to the corrugated fiberboard (http://en.wikipedia.org/wiki/Corrugated_fiberboard) .... maybe a board made of transversal sections of corrugated fiberboard could be a base?

    Having many cells will, of course, decrease camera resolution; the camera will work like the compound eye of an insect. So, another idea could be borrowed from insect's eyes: we could have a lens for every pixel.

    It wouldn't be hard to achieve: for passive infrared detectors (PIR), there are already compounded lenses:

    http://www.3dlens.com/infraredfresnellens.htm

    They are built to focus all infrared rays on the same point (in a PIR, there is only one sensor) , but, putting the detector closer to the lens, it should be possible to get different pixels.

    I'm brainstorming.

    ReplyDelete
  13. I'm wondering that would it be possible to use inkjet technology to apply the heat sensitive material. You do not show the process but your description suggest that it might be possible.

    This method would allow to create small separated islands of the material.

    It would be probably quite challenging because film to print on is very thin and printer has to be modified (though there is a lot of information available from efforts to create a inkjet printed PCB).

    Anyway your material is very inspiring.

    ReplyDelete
  14. Kros, that is a great idea! I think it should be possible to dispense thermochromic dots with inkjet tech. The problem is the substrate onto which the dots are printed will still conduct heat and provide spatial and temporal smoothing problems.

    However, I think you are on to something, and using inkjets may be a way to increase the performance of the system significantly.

    ReplyDelete
  15. Ben, Thanks you so much for actually making this thing! I had (almost) all the parts to build one several years back, but I lost my belief in the project due to overestimating the thermal mass of the microscopic beads that carry the liquid crystals in the paint. Foolish me... You have shown a more than useful effect with even the thicker film. I wish I would have been watching this thread a while back, I could have sent you some of the ultra-thin film. A friend of a friend who makes indoor planes sent me about a square foot of the stuff. The next step for me was to wash the binder out of the liquid crystal pain, leaving the beads, dry the beads, then crush the beads and filter the resulting liquid crystal goo. Then put a very small droplet between two (relatively) tightly-stretched films, one with an absorptive coating or lampblack (deposited on one side while the other side had water or other heat-sink keeping the film from going "fup!". I ran out of funds/time/housing and ended up selling my IR Fresnel lenses, paint, and other bits and pieces so I could eat. You have cured me of the 'bummer' I have been carrying around. It is so encouraging to see the thing get legs! Here is a post I made about it on cr4 (not much interest, but I think my manic tone and complete lack of proper (industry) lingo may have scared off the engineers...: http://cr4.globalspec.com/thread/45667/Return-of-the-50-Thermal-Camera-Challenge-Liquid-Crystals
    Thanks Again Ben!
    :)
    RR

    ReplyDelete
  16. My, sorry for not re-reading my comment before posting... (blush)
    "Thanks" = Thank
    "pain" = paint
    Oh-my...
    RR

    ReplyDelete
  17. Ryan, thanks for the comment. I think fresnel lenses might distort the image severely.

    I like your idea of isolating the crystals from the thermochromic paint -- I will give that a shot. Good luck!

    ReplyDelete
  18. I hear ya', I thought so too, but my budget was way below what I needed for even a cracked germanium lens.
    What surprised me though Ben, was that the entire selection of IR fresnels purchased from EO were in fact not the old/crappy faceted things I had been exposed to up until that point. They were actually much more like the 'ideal' fresnel, where the center was a nicely curved/formed/polished(actually cast? not pressed?) round lens, surrounded by varying widths of equally nice doughnut slices of the classical lens shape.
    I was used to a simple pressed/cast sheet based on a metal mother mold where each minute ring was just an angular cut. Wonderful for melting small rocks, but always a disappointment when trying to pull-off imaging.
    Far from that experience, these pretty little sheets focused crisp images even in the visible, defying my low expectations looking at their milky appearance. I wish I had held onto one... It may have surprised you too. From $23 to $28, it is a cheap test to make, with a potential payoff that (if it performs well at all in some applications) shifts the price point down into the 'disposable' range.
    No clue about using blank silicon wafers as visible light filters, never got that far in the project, but it seems like the absorptive coating on the film puts out the light fire. :)
    Thanks Again,
    RR

    ReplyDelete
  19. Quick update Ben, found some still viable LC thermal paint in storage, smeared it on a glass slide, dried it, & put it under the microscope @ very low power. The beads are variable in size, & crush easily under the tip of an x-acto knife. Far more easily than expected. I had a vague notion that they were thick-skinned little things, but not so. Whether washed & sorted (or not) before dried & crushed may not matter, even with impurities of shell fragments the contrast should be fine for a test of the film/LC/film idea. Who knows, may even provide a way to maintain semi consistent LC thickness (by propping against collapse). By the same token... Maybe size-sorting the beads under fluid would give you the raw material for a 1-bead-thick array of nested beads. But I have taken up enough space on your blog already.
    A hearty Thanks in the new year Ben, for being a playful discoverer & sharer of your doings.
    You are an inspiration.
    RR

    ReplyDelete
  20. What about putting a glass over the thin layer coated with thermocromic paint?

    In fact, if the layer is too thin, some of the heat will transpass it without being detected; the glass will reflect it, so it will hit the layer from backward, heating it anyway. And, since the glass is still transparent to normal light, you will still be able to see the change of color.

    The glass might work like the "tapetum lucidum" (in the eyes of nocturnal animals), increasing sensitivity.

    ReplyDelete
  21. Dear Ben Krasnow I love thermal imaging cameras
    I really love yours
    comment posted by Peter Smale

    ReplyDelete
  22. Another approach in building a cheap thermal imaging device:
    http://hackaday.com/2012/11/12/building-a-thermal-imaging-sensor-from-scratch

    ReplyDelete
  23. May I ask where you purchased the thermochromic liquid crystal paint?? My email is dghebner@verizon.net. Thanks in advance! I'm looking to test it in nail polish making :-)

    ReplyDelete