Where can I get a cheap SWIR camera?

[Ben321]

Fires, let me explain a bit about the different IR spectral bands. There's NIR (near infrared), SWIR (shortwave infrared), MWIR (midwave infrared), LWIR (longwave infrared), and FIR (far infrared). While SWIR (short wave infrared) sounds like it's just a synonym for NIR (near infrared), it's not. Despite its name, SWIR is actually of a longer wavelength than NIR (though there's actually some overlap between the NIR and SWIR regions). NIR is approximately 700nm to 1200nm. SWIR is approximately 900nm to 3000nm (with most SWIR cameras sensitive to the portion between 900nm and 1700nm). MWIR (midwave infrared) is then approximately, 3000nm to 7000nm (with most MWIR cameras sensitive to 3000nm to 5000nm). LWIR is 7000 to 14000nm (with most LWIR cameras sensitive between 8000nm and 14000nm). FIR is everything between LWIR and microwaves.

What I'm looking for is not a NIR camera (which I already have, by modifying a normal video camera by removing its infrared blocking filter). What I'm looking for is the much harder to find SWIR camera. Standard silicon CCD and CMOS chips have an absolute longest wavelength they are sensitive to that's about 1000nm to 1200nm (and getting within a few nm of this maximum wavelength, the sensitivity is dramatically decreases to about 1% of its maximum sensitivity, where the wavelength of its maximum sensitivity is usually around 750nm). Even if you assume a maximum wavelength of 1200nm (which is like an absolute maximum for any sensors made out of silicon, as it's simply a material property of silicon) this falls far short of 1700nm that is the upper limit of most SWIR cameras.

What is so special about these few hundred more nanometers of wavelength? Just as NIR shows you natural phenomena that can't be seen with visible light (such as green plants appearing white, because NIR is strongly reflected by chlorophyll), SWIR can show you other natural phenomena. For example, there's a very interesting phenomenon called "sky glow" or "night glow", which is best seen in the SWIR part of the spectrum (NIR is not much better than visible light, when it comes to sky glow, as sky glow is very dim in both visible and NIR light). Also, some very hot objects, that are too cool to emit visible light or even NIR, can still be hot enough to emit SWIR. For example, a hot soldering iron will be seen to glow in the SWIR part of the spectrum.


Now the problem is that these SWIR cameras cost about as much as LWIR cameras cost 10 years ago. A simple 160x120 SWIR camera will cost around $10000. This is strange, considering that SWIR cameras have far more in common with ordinary visible light CCD cameras than they do with exotic devices like LWIR cameras. SWIR cameras have charge wells, just like visible light CCD cameras, where photons striking the photosites build up electrons in the charge wells. The much more exotic technology of LWIR cameras involves a bolometer array (literally an array of microscopic electronic temperature sensors). So LWIR cameras don't use the photoelectric effect. However SWIR cameras depend on the photoelectric effect, just like visible light CCDs do. This means that the basic technology for making an SWIR camera is much more well established. Yet for some reason, the prices of SWIR cameras are insanely high. The only real difference between SWIR and visible light CCDs, is the material that chip's light sensors are is made from. With visible light, the material is silicon. With SWIR, the material is InGaAs (indium gallium arsenide). I don't think it's actually that much more expensive to manufacture InGaAs sensors than it is to manufacture silicon ones. Probably somewhat more expensive, but not orders of magnitude more expensive to manufacture an InGaAs chip than a silicon chip. A 640x480 visible light webcam will cost less than $100. A 640x480 SWIR camera will likely cost over $20000.  Even LWIR cameras now cost about $3000 at the cheapest for 640x480 (and LWIR cameras are a lot harder to make, and require more exotic manufacturing procedures, than SWIR cameras). The insane prices for SWIR cameras seem to be a perfect example of price gouging (high prices with no logical economic reason for setting the prices so high).


So this brings me to the topic of this thread, that's stated in the title. Is there an SWIR camera manufacturer out there that is underselling their competitors (setting the price lower than competitors, to attract more customers)? Is there a place where I can buy a truly cheap SWIR camera? I've always wanted to know what the world around me looks like in this wavelength. From what I've read about it, it sounds like a fascinating wavelength in which to examine the world around me, but I've only read about it and seen a couple pictures online. I've never had a chance to see it for myself using an SWIR camera. So does anybody know where I can buy a cheap (price < $1000) SWIR camera? Is anybody even selling SWIR cameras for such a cheap price?

[lukier]

I have no idea about brand new SWIR cameras and pricing, but sometimes one can find old Hamamatsu C2400/C2700 1800nm cameras or things like that:

http://www.ebay.com/itm/322260125875

these are usually not CCD but Vidicon tube based :/

I once found Hamamatsu C2400 camera head in the dumpster, but no control unit and the camera was probably broken anyway - the refurbishment and reverse engineering would be too difficult, not worth it.
Also, there was no lens and I have absolutely no idea if common machine vision/CCTV lenses can be used in SWIR. I doubt it, usually after 1400nm the transmission is only 50% (compared to 90% at 550nm).

[Fraser]

Used SWIR cameras do come along on ebay reasonably often. As Lukier stated, they are often Hamamatsu or Find-R unit with the extended range SWIR vidicon tubes. An excellent condition Find-R unit sold recently on ebay UK for just over £250. I bid on it but my interest was exceeded by the price.

I already have an extended range vidicon tube camera but I have not fired her up for ages and I have no idea what her wavelength limits are.I do know that she was a specialist laboratory unit that cost a small fortune new though. It may only have a Nuvicon in it though.

The Find-R auction was here:

http://www.ebay.co.uk/itm/Infra-Red-Camera-FIND-R-SCOPE-85400A-400nm-1800nm-with-view-finder-amp-video-out-/162367515157?_trksid=p2047675.l2557&ssPageName=STRK%3AMEDWX%3AIT&nma=true&si=i%252F4LbtG6%252FLD9UF0GHMsipNyck8E%253D&orig_cvip=true&rt=nc

Granted it was a 1800nm camera but others do come along.

As with a lot of specialist camera equipment, it is often more affordable to buy the previous technology rather than the latest. Vidicon technology based cameras are not bad and at least you get to see the wavelengths of interest.

Specialist optical materials are needed for SWIR. Standard CCTV camera lenses are not suitable. The optics can be a real challenge to find used so always try to buy a camera that is complete with its lens.

Also note that on ebay these cameras are often poorly described or incorrectly described as CCTV cameras. They can be hard to locate as a result.

Fraser

[mahony]

The cheapest way to go for a new unit might be something using a phosphours coated CCD: ~2k $/€
https://www.edmundoptics.de/cameras/nir-uv-cameras/1500-1600nm-nir-ccd-usb-2-0-camera/3599/

[-jeffB]

FThe only real difference between SWIR and visible light CCDs, is the material that chip's light sensors are is made from. With visible light, the material is silicon. With SWIR, the material is InGaAs (indium gallium arsenide). I don't think it's actually that much more expensive to manufacture InGaAs sensors than it is to manufacture silicon ones. Probably somewhat more expensive, but not orders of magnitude more expensive to manufacture an InGaAs chip than a silicon chip.

Are we anywhere close to a mature manufacturing pipeline for large-scale InGaAs chips? I was under the impression that it's still mostly a discrete device technology, with larger applications still only in small-scale production. Given silicon's massive economies of scale, I kind of would expect InGaAs to be orders of magnitude more expensive.

There's also the question of optics, and manufacturing scale for those.

Even "conventional" thermal IR enjoys a lot more economy of scale than this "neither fish nor fowl" band. Until someone comes up with a compelling reason to sell millions of SWIR cameras, I'd expect them to stay mostly at yesteryear's thermal IR prices.

Great, I never though much about that gap in the spectrum, but now you've got me wishing for equipment I'd never heard of before. 

[mikeselectricstuff]

I'm sure price difference is simply about markets - without  a volume application, there's no reason for anyone to invest in bringing the cost down.

[sam1275]

I agree to Mike, I was just about to say that.
BTW, I remember I saw there was a few second-hand Ingaas camera under $1000, they have better sensitivity and life compare to tube based ones, but they are very rare at low price.

[mikeselectricstuff]

If the wavelength you're interested in can be done with a phosphor, I wonder if DIY phosphor coating might be an option - either direct onto a die, or with an intermediate screen.
http://maxmax.com/phosphorsdyesandinks/infrared-phosphors-dyes-and-inks

[Ben321]

FThe only real difference between SWIR and visible light CCDs, is the material that chip's light sensors are is made from. With visible light, the material is silicon. With SWIR, the material is InGaAs (indium gallium arsenide). I don't think it's actually that much more expensive to manufacture InGaAs sensors than it is to manufacture silicon ones. Probably somewhat more expensive, but not orders of magnitude more expensive to manufacture an InGaAs chip than a silicon chip.

Are we anywhere close to a mature manufacturing pipeline for large-scale InGaAs chips? I was under the impression that it's still mostly a discrete device technology, with larger applications still only in small-scale production. Given silicon's massive economies of scale, I kind of would expect InGaAs to be orders of magnitude more expensive.

There's also the question of optics, and manufacturing scale for those.

Even "conventional" thermal IR enjoys a lot more economy of scale than this "neither fish nor fowl" band. Until someone comes up with a compelling reason to sell millions of SWIR cameras, I'd expect them to stay mostly at yesteryear's thermal IR prices.

Great, I never though much about that gap in the spectrum, but now you've got me wishing for equipment I'd never heard of before. 

Well there is a manufacturing pipeline for these, as they are standard items at the companies that sell them (at an insanely huge cost). It's not like an item that they have to custom-build for you.

[Ben321]

I'm sure price difference is simply about markets - without  a volume application, there's no reason for anyone to invest in bringing the cost down.

There's one volume application that I can think of. Because they can see airglow at night (also known as skyglow or nightglow), they have applications in military nightvision scopes. The military has 3 types of NV scopes: thermal (LWIR), image intensifier tube (visible and NIR), and airglow (SWIR and MWIR). Skyglow itself is a form of atmospheric phosphorescence. The upper atmosphere is excited during the daytime, due to bombardment by high intensity UV rays form the sun. At night, this energy is released in the form of very dim visible light, slightly brighter NIR radiation, and much brighter SWIR and MWIR radiation. SWIR or MWIR cameras are usually required to see airglow (though at shorter wavelengths such as NIR, and even visible light, it's possible to see with a long exposure, usually several minutes long). So yes, SWIR cameras most certainly do have a reason to be produced in large volumes. The military probably quite often places large quantity orders for these devices, as nightvision equipment. And yes, airglow in the SWIR range is bright enough at night to people, illuminated almost as if it were daylight. And this does not require any light amplification (like traditional NV scopes), as there is already plenty of light in the SWIR range coming from the airglow. It's just that you can't see it with your eyes, or even NIR modified conventional digital cameras (which use silicon CCD or CMOS image sensors). So yes indeed, there is quite a demand from the military for these, so there is absolutely a reason for companies to manufacture them in large quantities (hence the economy of scale should kick in, and these things should be quite cheap, but for some reason they aren't, and I can't figure that out).

[LaserSteve]

The Edmund camera adapter is probably made by Scintacore, using a anti-stokes phosphor. I doubt the phosphor method is sensitive enough for Skyglow.

Steve