Thermal Cameras - Professional Telescope and close up lenses by Inframetrics

[Fraser]

I have often been asked about how to make a telescope auxiliary lens for thermal camera use. I have also assisted forum members with obtaining close up images of PCB's etc by use of a ZnSe or GaAs CO2 laser close focus lens.

During my current tidying of my 'toy cupboard' (garage) I found my Inframetrics telescope and close up lenses. I thought some forum members might like to see what professionals in industry use on their thermal cameras. It may give you ideas for your own lens projects as well.

Inframetrics were a producer of high performance thermal cameras and lenses up until being absorbed by FLIR. Their optics were, and are, superb quality with mainly highly expensive Germanium elements used in their construction. I say mainly because the Inframetrics PM2xx and 3xx series used lenses that also contained ZnS or ZnSe in their construction to lower costs.

I am fortunate to have several thermal camera lenses made by Inframetrics. Before I go any further, a warning for young players.... Germanium lenses are very expensive, even on the secondary market. There are three wavelengths of thermal camera and lenses can be matched to the required wavelength using the AR (anti reflective) coating. If you try to use a Short Wave lens with a Long Wave camera, you will see virtually nothing through it. Know your subject before buying a thermal camera lens. Failing to do so can lead to frustration and disappointment.
The lenses that I will present in the attached images fall into three types.

1.   Auxiliary Telescope
2.   Auxiliary Close-Up
3.   Auxiliary Right angle Close-Up

The term auxiliary is used to indicate that these lenses are designed to mount in front of the cameras objective and not replace it. The light coming out of the lenses is parallel and not focussed to a point behind the lens. Unlike an objective. Many industrial thermal cameras accept such auxiliary lenses and provide a bayonet or screw thread mount in front of the objective. The mount often includes lens type detection and ID via contacts or several Hall-effect detectors in the camera and small magnets in the lens.

The telescopes that I purchased for my cameras are the Inframetrics 3X model that was originally designed for the Inframetrics 5xx and 6xx series of scanning mirror cameras. As such the lens mount has to be adapted to fit my AGEMA/FLIR PM 5xx and 6xx microbolometer cameras (very different to the Inframetrics cameras). The lens design uses 4 large Germanium elements forming a Keplerian telescope with Collimator to shape the output.  Sadly this is an ongoing project that I have yet to complete.

Inframetrics also produced a series of Close-Up lenses that are designed to fit directly onto the Inframetrics 5xx and 6xx cameras objective using their 60mm threaded mount (thread pitch is 1mm and not the more common 0.75mm) These Close-Up lenses came in a variety of focus distance and contain only one large Germanium element. I have a 6” and a 12” close up lens. These Close-Up lenses were also designed to fit to the front of the 3X telescope to provide 3X magnification with a close focus capability. Great for electronics work.

Finally, I recently acquired an unusual Inframetrics lens. It is a right angle Close-Up lens containing two lens elements and a gold plated mirror set at 45 degrees. The focus distance of this lens is just 1.8”. It is almost impossible to find any information on this lens but it was likely used for industrial inspection with the camera horizontal and the lens looking down onto the DUT. Why have the camera horizontal you may wonder….. well the early thermal cameras contained a DEWAR that was filled with Liquid Nitrogen. Such cameras could not be mounted in the vertical plane as the Liquid Nitrogen would pour out ! A right angle lens is also used in applications where vertical height is limited or camera support is an issue. This was a VERY expensive lens and a rarity. The 1.8” Close Focus is useful, as is the ability to leave the camera horizontal when in use. The lens mount is 60mm, the same as the other Close-Up lenses. It is also designed to mount on the front of the 3X telescope to form a 3X close focus right angle lens assembly.

The lens elements used in all of these lenses are around 50 mm in diameter (except the front lens in the right angle CU lens) They are made from pure Germanium. I am often asked why Germanium lenses are so expensive.  Here is the answer…

Germanium lens elements are formed from a single crystal of Germanium that is grown for the purpose.  Germanium is a metal. The lens ‘blanks’ are cut to shape on a diamond lens lathe that costs a small fortune and takes time. The lens is then polished and an Anti-Reflective coating is applied to match its intended use. This process is not normal mass production and so costs a great deal of money for the raw material and the creation of the lens. This is why GASIR came into existence. It’s a material that does contain some Germanium particles but is capable of being moulded to shape rather than needing to be cut on a diamond lathe. GASIR does have some shortcoming however so Germanium is still the material of choice in professional thermal camera applications.

Also note that as Germanium is a metal, damp storage conditions will cause it to corrode and the AR coating to blister off of the surface.  Abused lenses show severe corrosion on the surface. This cannot be polished off and the loss of the AR coating renders the lens pretty much scrap. Ex Fire Brigade cameras often suffer some degree of lens corrosion due to constant exposure to water, but they often still perform OK in spite of this mild corrosion.

I attach some pictures of my Inframetrics lenses for your interest.

Fraser

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

For those curious about the cost of this small collection of lenses when new. From what I can discover, you would not get much change out of £40K or $60K for our American friends. And that was in the 1980's when prices were lower  Specialist lenses made from germanium are not cheap !

I paid around £100 each for them on ebay USA.

Fraser

[LWIR-Microscop-Man]

Hi Fraser,

You sound like a kindred spirit and it's a great pleasure to make your acquaintance! 

I've had an interest in the history and technological development of Thermal-IR imaging for ~15-years, since I discovered the first Raytheon amorphous silicon 160x120 FPA microbolometers, and became curious about how they work.

This eventually lead to me owning a FSI Prism-DS, which I purchased used in excellent condition.
I spent a considerable amount of time trying to reverse engineering the digital interface, in the hope of eventually being able to make full radiometric recordings.
Unfortunately, the camera suddenly stopped working, and I recently sold it to help recover some of the original purchase cost.

I had it mounted in a vertical orientation for examining some 10x10 mm electrodes I was forming via electrolysis.
Having the camera mounted vertically, may have caused an asymmetry in the operation of the piston in the linear Sterling Cooler, which I suspect was designed to normally operate in a horizontal orientation. Also, the camera's user interface had been flashing a low battery warning, even though it was being operated from an AC power source.
The alternate hypothesis is that the back-up battery for the parameter RAM and NUC (Non-Uniformity-Calibration) storage may have failed.
However the case, I did learn a lot from using this camera, which was a marvel of design for its time!

I'm now in the process of getting a BAE SCC500H working, and have become interested in Thermal-IR microscopy.
I've got a 15x reverse Casegrain reflective microscope objective, and hope to adjust the close focus distance of the BAE SCC500H to test its use with this lens.

Note: The microscope objective is only specified to 7 microns, but is also available with a gold coating of its reflective surfaces, which will extend its response to 14-microns.

It maybe of interest that FLIR hopes to ship the 640x480 version of their new 12-micron Boson camera in April, and I'm consider the prospect of obtaining one.
I did consider the 320x240 version of the Boson, which is available now for $1400 USD, but FLIR is not yet supporting a Radiometry feature, and is not promising backwards compatibility via firmware or software upgrades.  The Boson development software has a manual NUC feature which will be very useful for enabling the use of alternate lens configurations.
The radiometry I'm interested in performing may eventually require measuring 5-micron features at temperatures of ~1600 degrees C.
As such, the expected LWIR attenuation of the Al surface coating of my microscope objective, may actually work in my favour if I can improvise a suitable blackbody source for calibration.

It may also be of interest, that the Boson maybe small enough for a pair of them to be used to form a stereo microscope.
A binocular view would be of great benefit for revealing the surface morphology on the electrodes I'm studying.

Considering my meagre budget, I can't afford to make expensively bad purchasing decisions when acquiring used gear :-( 
Going forward, I'd welcome the chance to review some of my design requirements with an expert such as yourself!

Please message me if interested and inclined!

Doug in Ottawa

[Fraser]

Hi Doug,

I will respond here just in case what I say is of use/interest to others on the forum. You sent me a PM with further details that I will also address in this response. Hope that is OK.

I should first state that I am a user and repairer of thermal cameras and not a designer of the electronics or optical systems used within them.
My knowledge of the optical blocks used in thermal cameras comes only from what I have read in reference books and seen in cameras that I have worked with. I may not be the best person to advise you on your project as a result.

My comments .......

1. You wish to build a stereo thermal microscope. I have a monocular thermal microscope and I am not familiar with a stereo version but it likely follows classic optical block practice for such. That is to say, you have two thermal camera cores that each have their own dedicated optical block. The two optical blocks are aligned adjacent to each other to produce a true stereo image. Viewing of the thermal camera outputs would need to be via a true stereo VR headset or similar device. A Germanium or ZnSe Barlow lens may be used with the stereo thermal microscope as per visible light types and this can set the working distance to that which you need.

2. A misnamed 'Stereo' microscope with just a single optical block feeding a splitter prism for the two outputs provides no advantage in your scenario. Such a design is just a Binocular microscope. Not what you need.

3. As is well known, it is possible to  create macro lenses using a reversed standard lens block or, close focus using either a back-focus spacer ring or auxilliary close-up lens element.

4. As I possibly mentioned in this thread already. It is essential to use thermal camera lenses that has good transmission at the wavelength of the sensor array. In your case you are using a Long Wave FPA so the lens AR coating must also be LW transmissive. A Short Wave AR coating will block most, if not pretty much all LW energy ! A reverse Casegrain reflective lens uses a combination of mirrors and lenses. If the lens is designed for use below 7um as you have suggested in your message to me, the optical elements will likely be coated for MW or SW thermal energy. The Mirrors may work at LW, but the lenses will not.

5. Modern Microbolometer Sensors contain a sun protection filter and their input window cut-off wavelength is 7um. They are therefore blind below 7um.

6. Inframetrics elected to build refracting Keplerian telescopes that were designed to be auxiliary lenses placed in front of a cameras normal optics.  As such the energy beams leave the lens virtually parallel ready for presentation to the cameras objective. The Keplerian telescope is inverting so be aware that the image will be flipped and reversed ! Not a problem on most cores as there is an option to 'reverse' the image via the camera  options. The Inframetrics Telescopes were able to be converted into microscopes by the addition of various FL close-up lens elements as detailed in this thread. My telescopes are X3 power in their original application and the close up lens sets the working distance. My right angle close up lens has a FL of only 1.8". The right angle lens enables use with Liquid Nitrogen cooled cameras that need to be kept horizontal.

7. You owned a FSI PRISM camera that failed. No surprises there at all. The PRISM is a very early Stirling cooled camera that is on borrowed time now. The cooler loses its Helium charge over time due to leakage past seals and even through the aluminium case of the cooler. A Helium Recharge is not really financially viable and mechanical failure is common. You did well to sell the camera to a new home. IIRC the PRISM was a SW camera so sadly its lenses are likely coated for only such and so incompatible with LW cameras. Stirling and Cryo cooled cameras produce a far superior noise performance to even modern microbolometers. Be aware that microbolometers tend to be noisy and so video processing is employed, (usually DSP based) to reduce noise content. Such noise reduction s a compromise solution that will degrade the image in ways that may or may not be a problem for you. Increases noise reduction often means loss of detail and/or slow image refresh at the camera output.

8. You provided details of a DIOP lens auction. That lens is way over priced at £500 for the current secondary market. It is a lens designed for use with a cryo-cooled SW FPA and as such will likely be AR coated for SW and useless at LW. No good at all for you. I attach a picture of the lens and a similar DIOP lens fitted to its original host, a cryo-cooled sensor array. Walk away from that lens. You will see some thermal camera lenses at crazy prices as some sellers are Surplus Dealers who pay virtually nothing for the items and hope that some exotic looking stuff is worth a fortune due to rarity. The sellers often advertise the items for several years and just hope for that elusive desperate customer who needs it at any price. Do not overpay. I know the lenses were expensive when new but most need adaption to work with modern cameras so are not plug-and-play. The market for such is very small. I pay around £100 to £200 for high quality large lenses.

9. If I were building a thermal microscope today, I would buy a 320 x 240 thermal camera or core and either add an auxiliary telescope lens with close up attachment, or change the cameras objective to work at near focus. If an option, choose the mast powerful lens option on the camera The FLIR TAU 320 largest lens was a 60mm assembly. The 60mm lens set to very close focus (using back focus spacers) or fitted with a close-up auxiliary lens would be an excellent little thermal microscope.

10. Stereo thermal microscope ? If you are feeling rich then maybe. Two FLIR TAU Gen2 320x240 cores with 60mm lens set at very close focus, aligned adjacent for stereo imaging (Each lens would need to be focused so not optimal). Composite video output is available so it could feed a pair of decent LCD screens in a VR headset. Such a headset is very expensive for true stereo vision.   Worth the effort and expense ? Not for me. I would rather spend the money on a top notch single lens block/core microscope.

11. Lastly, how much magnification is enough ? My Industrial AGA 680 Cryo-Cooled SW Microscope is fitted with a 15X right angle refractor lens assembly. You mention that you have a reflector microscope lens. That is a first for me. Reflectors usually have large diameter and are not generally suited to close-up or microscope work. I would be interested to see a picture of the lens.

Best Wishes

Fraser

[Fraser]

I just had a quick look for reflective microscope lenses and found them. I learned something today. I never knew such were used in microscopes !

If the lens is not specifically designed for thermal wavelengths you may have some problems using it. If there is any glass in the beam path it will be opaque to thermal energy. If the mirrors are not surface coated they will not work at thermal wavelengths. The aluminium coating can deteriorate with time or damp environs due to oxidation. All glass must be removed from the beam path and replaced with suitable Germanium or ZnSe lenses. Reflective lenses tend to have qiuite poor transmission numbers due to losses within the design. The Inframetrics X10 thermal camera reflective lenses commonly have a 50-60% transmission figure.

Is your lens specifically designed for thermal imaging ? If so, please will you post a picture as this is interesting to me. My reflective thermal camera lens is 6" in diameter and designed for long range use.It contains a Gemanium front window and at least two Germanium beam shaping lenses in its output path.

I attach the picture of a visible light reflective microscope lens for those who were also unaware of them.

Fraser

[Fraser]

Pictures of the Inframetrics 10X reflective telescope lens for the 600 series scanning cameras. Its a biggie 

My AGA 680 microscope refractive lens measures 50mm x 30mm and offers 15X magnification !

Fraser

[tokotak]

Dear Fraser,

Is it a PMXXX series adapter on the right hand side of telescopic lens seen on the image 0503 from the first topic?

[Fraser]

@tokotak,

No that is just the standard Inframetrics bayonet mount. LatervInframetrics cameras had the bayonet mounting on the scanner head whilst earlier models had a screw thread mounting. The Bayonet adapter is just screwed onto the lens as an adapter. A similar adapter can be made using thick wall aluminium tube and a lathe used to create a FLIR PM series compatible bayonet mount.

I will see if I can do some test pictures with the 3X lens for you on a PM695. I am dealing with the death of a loved one at the moment so please be patient.

Fraser

[eKretz]

Fraser, does that 3x telescope invert the image when placed in front of your FLIR Exx or does it provide an erect image?

[Fraser]

The 3x telescope is a Keplerian type so yes it does invert. It contains 4 lens elements. The first two are the Keplerain telescope section and the rear two are used to collimate the output beam to match into the large aperture scanning camera head image input port.

Fraser

[Hydron]

Thor labs page on reflective objectives is very interesting - take a look at the reflectance plot!
https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=6933

Fraser - remember that some of us purchased those IR macro lenses from a US based forum member - I've used mine in anger a few times and while I haven't done any rigorous and objective testing, they do perform pretty well looking at small PCB features, easily down to 0201 parts (not into the microscopic range though). There definitely is a sweet spot - too high or low magnification and the IQ decreases. Unfortunately not gonna help with PCB work for those who didn't get in on the deal though.

[Fraser]

Hydron,

Very interesting 

I have a few of those very nice close-up lenses 

The Germanium lens elements are of excellent quality and likely cost a small fortune new.
Out of interest, how do you deploy the close-up lenses ? Just in place of the cameras original lens or have you modified the close-up lens to act as Supplemental lens ? I was thinking of removing the rear lens element to see if the output beam would then match into a thermal camera that is still fitted with its original primary lens.

Fraser

[Hydron]

Stuck it straight in-front of the FPA, nothing in between. I have not tested with any other lenses in the middle, it is possible that it would then yield different performance.

I do have a few focal lengths I could try, I'll have a quick play (right now I have a couple of VGA cores setup on the bench with 19mm lenses for work) but probably don't have time right now to do a thorough investigation.

Good call on getting more than one of those lenses - wish I'd picked up a second!

[eKretz]

Ahh, reading back now I see I missed the earlier post that noted the 3x lens inverts the image. I was asking since I just managed to score the same lens for $40 on eBay. The seller had it mixed in with a couple bins of old television camera equipment and I just happened to spot it.

[eKretz]

Got that $40 3X lens today. Looks like a good score. Lenses are in excellent shape, images look fine. Here is a quick sample with two photos taken from the same location, lens just held in front of the FLIR E40 (upgraded).

[eKretz]

BTW Fraser I noticed in another thread you were investigating the possibility of flipping the image from the 3X lens via a firmware change on the FLIR Exx cameras. Did you ever find anything? I took the above image with the camera upside down.

[Fraser]

Hiya,

Sadly no. The Exx does not appear to offer image reversal and inversion, unlike the TAU cores. I am still considering a dual element image inverter to mount on the rear of the Inframetrics lens...... but that makes it a bit long ! A pair of short FL Bi-convex ZnSe or Germanium lenses would do the trick. A project for the a Winter maybe. The lens you have fits the Inframetrics 6xx series of cameras. The rearmost mount may be unscrewed to remove it from the lens barrel. That leaves a course thread on the lens barrel to mount the lens into a 3D printed adapter. Making an adapter with a thread is often easier than using a bayonet mount and the thread is further forward along the barrel which is an advantage. The threaded barrel directly mounts onto the Inframetrics 5xx series cameras as this lens was adapter from a 5xx to 6xx mounting.

Fraser

[eKretz]

It's already pretty long! Shame that there's no way to do the reversal, that would have been handy.

[Fraser]

Regarding its length, it may be possible to remove the two collimating lenses at the rear as they are used to match the beam into a scanning camera head first mirror. I have not experimented with this idea though. You could use the space released by those two lenses to position an image inversion pair within the current lens barrel ? These lenses are large beasts but contain lovely quality Germanium optics 

The first two elements are the actual Keplerian telescope and the image inverts between them. Focus is also set by adjusting the distance between the front two elements. I plot the beam path of this lens a long time ago. I X-Ray imaged the whole lens to establish the lens dimensions, curvature and profile etc. I shall see if Incan find it again.

Fraser