Plastic lenses for LWIR?

[Ultrapurple]

Has anyone tried using plastic Fresnel lenses for thermal imaging?

For a long time I've been mulling over the multi-zone lenses on the front of PIR security sensors but then found that a few manufacturers produce lightweight Fresnel lenses that are supposed to be able to "replace expensive germanium lenses in systems requiring moderate imaging quality".
See http://www.kube.ch/downloads/pdf/kube_single_fresnel_lenses.pdf from Kube Electronics AG, Switzerland.

On the face of it, a 13.5mm FL f/0.5 (!) lens made from 0.5mm thick HDPE with ~70% transmission at 10µm looks quite attractive - and a lot cheaper than germanium. But their idea of 'moderate imaging quality' may be different from mine.

I asked about prices of a couple of the lenses and the answers were in the 10 to 20 euro range, plus 24 euro for P&P. (All very reasonable for one-off sample sales, I reckon).

Thoughts?

[Fraser]

The use of thin fresnel and diffractive lenses made from LWIR transmissive material for thermal cameras has been considered over the years. There is at least one company that took the idea of using plastic fresnel lenses into production for relatively low resolution tasking. From what I know of the plastic fresnel lens product, it is very 'course' and created a poor thermal image on a 320 x 240 pixel core. Diffractive lenses using ZnSe and Silicon have produced affordable and good quality imaging lenses. The plastic lenses have some way to go in terms of technology before they can be considered a serous alternative to conventional silicon, ZnSe, ZnS and Germanium optics. There are diffractive Silicon and ZnSe lenses that work very well at limited resoloutions. The TAU2 has a diffractive ZnS rear lens element and the LEPTON cores use Silicon diffractive lenses.

I was looking at using the plain (non fresnel) PIR windows as lens protector material. It's transmission losses seemed quite high compared to the thin shrink film that I normally use. A PIR is looking for a 36C target against a approx 22C background. The two or four pyroelectric detector elements are huge compared to a cameras microbolometer pixels. Transmission losses can be quite high in such a design and it will still work.

I would like to track down more affordable ZnSe lenses of greater diameter than 20mm and greater FL than 100mm. They would be useful for supplementary lenses. For me, plastic fresnel lenses is not the way forward at the moment.

Fraser

[Vipitis]

Is there any source where I could get such a cheap lens?

My Lepton isn't high resolution anyways, but I want a telephoto lens to shoot the moon, the stock the fov is really wide.

[Ultrapurple]

I have some of these plastic lenses on order and will test them with my Therm-App, which has easy access to the sensor. I will report what I find, though I strongly suspect Fraser is right and the results will be at best mediocre.

[Fraser]

Ultrapurple,

I will be very interested to hear what you think of the lenses.

A source of these lenses is decent quality IR Thermometers where they are often used to focus the energy onto the thermopile sensor.

Thank you for spending your money investigating these lenses. It is a pity the company does not offer a free sample as they are not expensive to produce once the press mounds are made.

Fraser

[Fraser]

This may be of interest.....

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

$350 per lens in quantities of 5 though ! Ouch 

Fraser

[Ultrapurple]

Ouch indeed! I'll play with my 12 euro lenses first, I think.

Interesting that they say they work best with large-pixel (50µ) cameras, presumably because the MTF is poor.

[Fraser]

I believe it comes down to the resolution of the lens itself. A Fresnel does have resolution limits.

https://en.m.wikipedia.org/wiki/Optical_resolution

The large microbolometer pixels are more forgiving of such but that makes the sensor die and associated lens system larger for say a 320 x 240 pixel microbolometer. 50um pixels were found in BST sensors and early Microbolometers, as used in the Indigo Alpha camera etc. I suspect that there are not a lot of 50um Microbolometers in current production.

Fraser

[Fraser]

http://modulatedlight.org/optical_comms/fresnel_lens_comparison.html

https://www.edmundoptics.com/resources/application-notes/optics/advantages-of-fresnel-lenses/

[Ultrapurple]

The lenses have arrived, as of Monday 25 September. I have 13.5mm, 25mm and 59.4mm focal lengths. Initial experiments show that they will form images on the Therm-App sensor, but the efficiency (=contrast) is relatively low.

Whether the lenses can be contrived into something that's actually useful remains to be seen. I'm not in a position to do any proper experimentation during the day - lunchtimes are very short - but I will have a play and see what I can achieve over the next few days. I do wonder about making an x4.4 Keplarian telescope with the 59.4 and 13.5mm lenses, but suspect the losses would be prohibitive. It certainly wouldn't be a patch on my germanium x3 telescope!

[Fraser]

I bought an Inframetrics 48" FL Germanium close-up lens from the US that is currently on its way to me.
It was very cheap so I thought I would experiment with it. It is just the meniscus lens element but is the same as my Inframetrics 6" and 12" FL Close-up mounted lenses.the 48" FL is HUGE for a telescope but I thought I might try it for a laugh  the lens diameter is around 60mm so not a huge aperture. Mating it with a 15mm FL Germanium bi-convex lens that I have could be interesting though  The losses will likely be huge. A 80x magnification Keperian telescope using a 1200mm and 15mm FL lens....... I do not think it is that easy 😄 I will have to buy some sturdy 50" long drain pipe to find out 😆. I may even have to resort to aluminium or steel pipe due to telescope 'droop' 🙂

The lens is still available as the seller has more than one.....

http://www.ebay.co.uk/itm/COHERENT-FLIR-MIRROR-laser-optics-INFRAMETRICS-00863-201-CLOSE-UP-LENS-SALE-49/252763373674?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

$39 is great value. Shipping is a little expensive to the UK though

For information, I have Inframetrics professional lenses that employ diffraction ring optics and are made from either ZnSe or ZnS. They work well. I have also found diffraction lenses made from plastic as the rear element in Raytheon BST core cameras. I believe FLIR used them in the Autoliv Audi and BMW cameras. They were in the same rear lens position. They can work, but they need to be very good quality.

Fraser

[Ultrapurple]

That lens looks like an absolute bargain - please report back when you find out what to do with it! I do like the idea of a 1200mm FL lens on my Therm-App but, as you say, in the configuration you suggest the aperture would be way too small for an uncooled sensor unless you were looking at a particularly warm target such as a large fusion reaction. Mayhap it would work for a cooled sensor? On the other hand, a 1200mm f/1 lens would be quite a hernia-inducing conversation piece.

There are rings on the back of the Therm-App (Umicore) 19mm lens, but they don't seem to correspond with significant 'steps' in the smooth lens surface so unless they're only of tiny height I don't know what they're for.

[Fraser]

Those are the diffraction rings that I mentioned. The steps are on the opossite side of the lens element  (pointing towards the front of the lens). The lens material for that element is likely either ZnSe or ZnS.

Using such lens elements significantly reduces the cost of the overall lens assembly. The objective lens element is either Germanium (expensive) or Chalcogenide IR transmissive glass (cheaper)

Canon have an interesting page on these rings here...

http://cpn.canon-europe.com/content/education/infobank/lenses/multi_layer_diffractive_optical_element.do

Fraser

[Ultrapurple]

That link is interesting.

Just for the avoidance of doubt, the rings on the lens in my photo face the sensor. Its colour (presumably the coating) is purplish-gold, depending on viewing angle. The objective is, as far as I know, germanium.

Apologies for the poor quality photo - it was one of my 'out-takes' when doing the other image.

[Fraser]

Interesting. Lens designs do vary and I am no expert on such.

Attached is a picture of the objective lens on an Inframetrics PM280 camera. The diffraction rings are clearly visible. The rings look to be on the front surface but are on the rear of the element. The front surface is perfectly smooth and without any ridges.

Fraser

[Vipitis]

Hey, I wanted to chime in.


I got found a fresnel window that is magnifying visually but not in LWIR. It's not reflective and not transparent.

I am hoping for some results in forms of images because I would love to see what's possible on the cheap side.

Building a pretty crappy diy scope is what my current goal is.

[Fraser]

It all comes down to the plastic used for the lens. Some plastics are more transmissive of thermal energy wavelengths than others.

Fraser

[Ultrapurple]

Agreed - these plastic Fresnel lenses are translucent to visible light, but a lot more transparent to thermal radiation.

I have done a little more playing and found that I can form poor-quality images with the Fresnels but they have a 'ghostly' quality to them rather like this visible light image (by someone else) - only not as good. This may be because I haven't been mounting the lenses perfectly square to the sensor, or being sure that the lens itself is completely flat (they're quite flexible). My best results so far were when I mounted the 25mm FL lens using blu-tack, although I was scared of dropping the sticky substance onto the sensor...

[Fraser]

Ultrapurple,

You are breaking new ground on this forum  AFAIK no one else has experimented with this type of lens and detailed the results here. Great work 

For info, it is a requirement that the fresnel lens is rigidly mounted in a frame otherwise all manner of distortions and focus point errors occur.

Some trivia ..... Regarding the visible light image you linked to. That appears to be the fresnel lens from an overhead projector. Those lenses are very poor performers as they have neither the precision nor the specification required to be a photographic lens. I previously read about which fresnel lenses are good for photography, and which are not. The overhead projector lens was in the 'not' category 

Happy experimenting 

Fraser

[Ultrapurple]

Thank you for your kind words!

I concur about the OHP lenses being of dubious quality, having tried using them in my youth to form primitive imaging systems. They did work, sort-of, but not well. I always thought it was pretty amazing that they managed to produce such readable images on a projection screen.

The usually-floppy page magnifiers are another case in point. They work well enough to give some magnification of high-contrast print, an application that is tolerant of misalignment and flexing, but is quite different from trying to form a well-focused image on (eg) a camera sensor.

One thing I have found so far is that the Fresnel lenses can be used, after a fashion, as a macro close-up lens with an existing camera focused at infinity (rather like the laser-type germanium lenses that get put on fixed-focus Flir etc cameras for close-up work). My limited magnifying experiments so far have used the Fresnels close to the object rather than close to the camera.

I think I'll take the tip of using a temperature-controlled soldering iron as a target, at least for the first wee while: not because of its temperature regulation, which isn't likely to be that accurate, but because I can dial it down to something like 150°C or lower and keep it there indefinitely. As others have noted, a soldering iron also provides a well-defined target so that one can check focus and resolution with ease.

Experiments continue; I hope soon to be in a position to post some sample images.

[Vipitis]

I follow this thread every day hoping for new discoveries.

There is a handful of people who make this a great place.

I thought about getting a few of those fresnel sheets and making them into a large refractor with 3d printed hardware to hold them in place and distort into the correct position.

I am waiting for a few more results on here before I start the research into resources and maths for engineering.

This week I have a different project to focus on.

[Ultrapurple]

I have some results. But they're not that encouraging, even with a temperature-controlled soldering iron at 150°C as my target.

Using a Therm-App camera and Therm-App Plus in white hot mode (but with no fancy processing tricks) I first took a reference image using the Therm-App 13mm f/1.0 lens. The distance between camera and target was about 18" (about 450mm).

Next, I removed the germanium lens and contrived a mount for the 60mm Fresnel. I did my best to have the lens reasonably flat and well-aligned, and got a much-magnified but poor quality image that was very ghosty and low contrast.

Finally, I mounted the 25mm lens and tried that. It gave an image of higher magnification than the 13mm, as expected, and notably better contrast than the 60mm, but still not what I'd call good.

It wasn't really possible to test the 13mm Fresnel because its focal length is too short to be easily focused on the sensor (the Therm-App body gets in the way).

Overall I'd say it was an interesting exercise but I don't think we're all going to be throwing away our germanium anytime soon. If anyone else want to have a play, please let me know - I'll accept any reasonable offer for the Fresnel lenses!

[Vipitis]

Hey,

Thanks a lot for those.

My plan is to build a small refractor for the moon in an afocal setup for my phone based Lepton.

Do you mind sharing your setup with a drawing/picture?

[Fraser]

My Inframetrics 48" close-up lens arrived today. It's a bit of a beast. Large diameter, thick and heavy !

I tested it and can confirm that it works perfectly on a LWIR camera. It is coated with the correct AR coating for our microbolometer cameras.

Fraser

[Ultrapurple]

Sure Vipitis - I will be happy to take a photo of what I'm doing. The main reasons i haven't done so yet are that my workbench was embarrassingly untidy and I ran out of time. I had the choice of making and posting the thermal images or tidying up and not doing any experiments. I like experimenting more than I like tidying...

Based on what I have seen so far, I don't think it looks practical to make good quality thermal images using lenses lime these. I have had some success using a cheap (~100 euro) reflecting telescope with a lens-less thermal imager, as reported in the thermal astrophotography thread. You might even do OK using a cheap concave shaving mirror - have a play and see what happens.

If your thermal sensor has a lens firmly attached you might find the optics a bit more challenging. I'm lucky that the Therm-App is an interchangeable-lens device, which makes it easy to access the sensor directly (well, from about half an inch away, maybe, unless you disassemble the case or grind off some metalwork that really needs to remain in place).

[Ultrapurple]

Here is (more or less) the setup I used for the initial Fresnel lens experiments.

The photos were taken the day after, so aren't a 100% accurate representation of what I did; in particular, the 25mm FL length Fresnel lens shown was much closer to the Therm-App imager for focus, and I  didn't bother to ensure the lens was perfectly aligned (height, L/R, square-on to the imager, etc).