EEVblog Electronics Community Forum
Products => Thermal Imaging => Topic started by: Fraser on February 02, 2020, 02:55:25 pm
-
Ultrapurple got in touch recently and made me a very generous offer. He had purchased an elderly (Circa early 1990's) AGEMA Thermovision 900 scanning thermal camera system. He said his tests suggested that the system was non-operational and would I like to carry out a teardown of the camera for the EEVBlog ? Of course I accepted his kind offer :) He also kindly gifted the system to me for my collection. What a Gentleman :) If anyone else wants to send me unwanted thermal camera technology, please feel free to get in touch :)
The AGEMA Thermovision 900 (THV900) thermal imaging camera is based upon the mechanical scanning mirror architecture that creates a raster image of the scene using a conventional primary lens, moving X and Y mirrors, fixed mirrors, conventional refractive optics and thermal detector electronics. These cameras are commonly called "Scanners" for obvious reasons. The scanner is very much a combination of precision engineering, optics and electronics and this did not come cheaply when these were sold in the 1980's and 90's. Such a camera plus lenses could easily cost more than £125K in 1992 when a 3 bedroom modern house in Northampton (UK) was around £50K at the time. Kind of puts things into perspective!
The THV900 is a late model of "Scanner" that uses a mini Stirling Cooler to lower the temperature of a single pixel MCT detector to 77K (-196C). This was a significant improvement in convenience for the user when compared to using Liquid Nitrogen to cool the detector. Note that I stated that a single pixel MCT detector is used. This is illuminated by the combination of scanning mirrors, fixed mirrors and refractive optics. The mirror servo control electronics create a raster pattern X-Y scan and the thermal scene is effectively read as a series of horizontal scans by the MCT pixel and sent to the electronics package that reassembles the raster scan into a format that may be processed on a PC or presented on a monitor.
All this mechanical cleverness may sound very unusual to those who are familiar with conventional FPA based visible light and Thermal imaging cameras. These scanning cameras were created out of necessity rather than choice. The limitations on producing suitable thermal imaging pixels in any significant numbers on a die meant that very low pixel count detectors were all that were available to build a camera. The mechanical scanning just made a camera feasible with a detector that would otherwise be suitable only for a single point cooled IR thermometer ! Once Cooled staring arrays with a decent pixel count were perfected, the scanning type thermal cameras were quickly consigned to the history books. That is not to say that decent imaging and science could not be done with them. They could produce very useful thermal images. They were just too expensive to produce compared to a cooled staring array based design and the imagery from those cooled FPA's was superior in all respects. An issue that remained, even with the cooled staring array based cameras, was the service life of the Stirling Cooler. The THV900 contains a well known K506 cooler that is rated for only 5000 Hours before requiring a very expensive service (£5K+). Stirling Cooler life, or rather, lack thereof, lead to the scrapping of many an otherwise functional cooled thermal camera, both scanning and staring array types. Modern Stirling Coolers have benefitted from decades of development since the old K506 and have a far longer predicted service life. Some trivia regarding the K506.... NASA uses Stirling Coolers in science probes sent into space. A K506 was used aboard the Clementine Mission:
“Way back in 1994, our first model, K506, was launched in the Clementine program,” says Zur. This NASA mission was intended to test sensors and spacecraft components under extended exposure to the space environment.
The later, improved, RICOR K508 model cooler was installed aboard the Curiosity Mars Rover. They had to be well designed and reliable for deployment on Space missions. Failure could ruin a science mission. You cannot exactly RTB a Mars Rover for a cooler replacement ! It was due to such reliability requirements that the Stirling Coolers have continued to be refined. They are still needed by NASA and other science users.
https://www.israel21c.org/israeli-fridge-lands-on-mars/ (https://www.israel21c.org/israeli-fridge-lands-on-mars/)
Now back to the topic of this thread........
The AGEMA Thermovision 900 scanning camera needs its own dedicated computer to create and analyze the images that it captures. The computer is bespoke and is not based on the common PC architecture or interfaces. As such this computer added great cost to the system. The computer does use a common VGA monitor for its output. I will not be going into depth on that computer in this teardown as it is not really very thermal imaging related, being basically a bespoke industrial computer with bespoke data interface to receive the data from the Scanning Camera head. I will be happy to document the computer in this thread if there is readership interest, but it really is not very exciting stuff. AGEMA produced their own software named "ERIKA" for the computer that enabled the user to capture, process and analyze thermal images. It could also create Thermography reports. In its day, it was a powerful Thermography tool for Science, Industry and the Military (Research).
The Scanning camera head could be fitted with a good range of optional lenses to meet the customers needs. These lenses included both refractive and reflective optics designs and all were made using very high quality Germanium lens elements or 1st surface mirrors. As such the lenses were very expensive. To reduce lens cost AGEMA used a systems where the rearmost Germanium lens element resided inside the Scanners lens mount. All the other lens elements were in the removable lenses. This reduced the number of Germanium lens elements in each lens by one. A significant cost saving when such large diameter diamond cut precision Germanium lenses are involved. This design feature does mean that lenses intended for use with an AGEMA 900 (or the 800 series) will not illuminate a FPA in another camera as the rear lens element is absent. Ultrapurple owns a large AGEMA lens that was designed for the AGEMA 800/900 series and so he discovered that, unlike many other "surplus" primary lenses, this one did not work as expected. It was with this in mind that I understand Ultrapurple sourced this AGEMA THV900 with a view to harvesting its lens mount and integrated lens. It was highly unlikely that the cooler in this elderly camera would still be operational so such harvesting is not an act of wanton vandalism ;) One the AGEMA THV900 lens mount and integrated lens id attached to the large lens that he has, it will be a complete primary lens, albeit designed to illuminate a scanning mirror system and not an FPA. A step in the right direction for him though :)
Upon receiving the AGEMA THV900 system I decided to establish exactly what state it was in. It goes against my instinct to destroy a thermal camera that is operational. If the system was working, I would give Ultrapurple the option to use it "as-is" or I could extract just the required lens mount so that the camera could be re-assembled again later if desired. A complete teardown would only be undertaken if the camera was beyond reasonable repair. I quickly established that the computer part of the system was not booting from its SCSI Hard Disk as its battery backed configuration memory in the RTC was forever lost. The DALLAS 1287 battery was exhausted. I suspect the computer would require reconfiguration using a AGEMA system utility for it to ever work again. Not a great start but I had a power supply and special PC interface that is capable of controlling the AGEMA THV900 Scanner directly, without the need for the AGEMA Computer. This was a later development by AGEMA. Once connected and powered I tried to wake the Stirling Cooler from its long sleep. The scanning system for the mirrors worked fine, but the Stirling Cooler motor remained silent and did not stir from its slumber. After some further tests I came to the conclusion that, as Bones would say in Start Trek, "It's dead Jim" :(
Now that I had established that the Scanning Camera was beyond recovery, my conscience allowed me to proceed with a complete disassembly and pictorial record for release on this forum. When looking at the pictures, please do study the mechanical engineering contained within these scanning type cameras. They are marvellous designs and very well constructed. It is the sort of fine engineering that I love to work on :)
Ultrapurple has very kindly offered me the Stirling Cooler for my collection. I will be removing that part intact and will look to produce a teardown of it in due course. The optical scanner block was disassembled only to a point where its contents could be viewed. Ultrapurple has plans for that complete scanning module so I did not want to damage it or misalign any mirrors. To help the reader understand the optical path through the optical scanner block I am including the optical path diagram from the AGEMA User manual. I am also including the simplified Block Diagram of the scanner head from the same manual.
I will include comments with the pictures to highlight items of interest along the way as I dismantle the unit.
Some specifications for the particular THV900 that is being dismantled in this thread:
Model: THV900 LW-ST
Detector: Single MCT pixel
Cooler type: Closed cycle Mini Stirling Cooler type K506 (5000 Hours)
Spectral Response: 8-12 micron
Temperature Range: -30C to +1500C (+2000C with filter)
Sensitivity: 0.08C at 30C
Accuracy: +-1C(Range 1) +-1% (Range 2-4)
Repeatability: +-0.5C (Range 1) +-1% (Range 2-40
Spatial Resolution: 230 Elements per line (50% modulation)
IR Line frequency: 2.5kHz
Frame frequency: 15Hz/30Hz
IR Lines per frame: 136
Samples per line: 272
Note that in the pictures that follow, my 'Supervisor', Bailey, makes an appearance. Fear not, he would not be allowed near ESD sensitive electronics in my lab. This camera is scrap and being disassembled in 'his' play area (the conservatory) so it was only right that he carry out an 'inspection'.
Fraser
-
Views of the AGEMA Thermovision 900 cooled scanner head including detail of the filter wheel access port and the Germanium Lens integrated into the lens mount.
-
Side panel removed for access to the Servo Board for removal
-
Side panel removed for access to the Processor Board for removal
-
Removal of the forced air cooling module to reveal the Stirling coolers heat-sink. IN the forced air cooling module, one fan blows while the other sucks so that they operate in unison.
-
After the cameras main side case sections have been removed, the inner chassis is revealed as a combination of distinct "Blocks" that are screwed together.
Bailey was carrying out a "Cat Scan" on the "patient" ;D
-
The AGC/AD and Pre-Amplifier Boards are screwed to either side of the chassis containing the Stirling Cooler and integrated MCT detector. This keeps lead lengths nice and short.
-
The "Hours" meter for the Stirling Cooler and views of the K506 Stirling Cooler.
-
Removal of the K506 Stirling Cooler, with integrated MCT detector, from the camera chassis.
In the last picture, note that the special soft metal sealing gasket material is visible. This is not normally seen so could indicate a "blown" panel gasket. The Helium gas "fill" is at a pressure of around 200psi and would escape past a blown seal with ease.
-
Removing the base plate to access the power board and underside of Optical scanner block and Filter/Aperture Block
-
Optical scanner block and Filter/Aperture Block
-
Bailey inspecting my work on the Optical scanner block and Filter/Aperture Block ;D
-
Separation of Optical scanner block from Filter and Aperture Block.
-
Optical scanner block
-
Lens mount with integrated Germanium lens element
-
Optical scanner block
-
Optical scanner block
-
Filter and Aperture Block with relay optics
-
Motorized Filter and Aperture selection wheels
-
Servo Board
-
Processor Board
-
AGC and AD board
-
Pre-Amplifier Board
-
Power Board
-
Final post for the moment. Phew :phew: That took a while.
I will add comments to the picture sets in slower time as I need a rest :=\
I will be going into more detail on the K506 Stirling Cooler in a new post in the future.
Hope this has been of some interest
Fraser
-
Thanks a million, Fraser - you've done a brilliant job, as usual.
Apparently the previous owner's last words before scrapping the system were "Can anyone smell a slight trace of helium...?"
:-DD
-
My pleasure :)
Some breaking news.... I have been studying the "Hours Run" meter under magnification and I can see the little bubble on the far left of the tube. There are some bodge wires changing the polarity applied across the tube and I just confirmed that the little bubble is nearest to the negative input. If my understanding of these electrochemical timers is correct, that means this camera had very low hours on it. I saw no evidence of the cooler having been changed so it is the original fitment. She died young :'(
It may be that the camera was used very infrequently and lost its Helium Charge over time. The cost of a service on the cooler was around £5K in 1994 when this camera was made. The price went up over time and I have a written quote from FLIR provided in the early 2000's for £9K for a cooler service ! Such costs were often deemed too high and the camera scrapped in favour of an uncooled microbolometer based unit.
Fraser
-
Another look at the hours meter and I am starting to doubt that it is working. The bubble is so far off the left mark of the scale that it just does not look right. I think I will ignore that meters indication. I enjoy investigating the history and failure of thermal cameras but I believe this one will be a simple case of "Years Vs the Stirling Cooler".
Would I recommend anyone buy one of these older Sterling cooled scanning type cameras to use and expecting a good few years life out of it? In short no, these cameras were great in their day, but the poor Cooler is long past its best. Hours run and any Helium leakage place the odds against such a camera lasting very long in service.
The Peltier cooled scanning cameras like the AGEMA Thermovision 450, 470 and 480 are more likely to give reasonable service as the only gas related concern is the quality of the vacuum in the Vacuum Dewar around the detector die. That vacuum is often very long lived, just as in Thermionic Vacuum Tubes.
Fraser
-
I have decided to upload all the pictures of the K506 Stirling Cooler for those interested.
This is a component that interests me greatly as I want to understand the design and potentially refilling with UHP Helium Gas.
There is likely to be a teardown of this cooler in the near future ;)
The date code (Motor: 9438/Compressor: 9444) on the cooler indicates that it was built in 1994 so it is now approx 26 years old.
Fraser
-
Pictures continued....
-
Pictures continued....
-
Pictures continued....
-
My Lily doing an impression of many who have read this thread ;D
[attachimg=1]
I can see from the picture views that this sort of camera is a specialist interest. I am both a Mechanical and Electronics Engineer so this sort of tech appeals to me. It is not everyones 'cup of tea' though.
Fraser
-
I for one find it fascinating, although I freely admit that in this case I have a vested interest, of sorts.
This is one of those posts whose value will appreciate with time. There can't be very many of these cameras left (fewer still in working order!) and it's nice to see the engineering that went into them before the last one goes to the Great Skip.
-
Looking at the pics of the lens mount it appears that the element is a meniscus lens - does it have a negative (as previously assumed) or positive power? I look forward to hearing how it helps get the big mirror lens going!
-
Hydron,
It is indeed different to the lens installation found in the AGEMA 870 and 880 that can use the same lenses.
Images of the 870 are to be found here (with 360 degree rotation video)
http://wll.kr/bbs/board.php?bo_table=06&wr_id=795 (http://wll.kr/bbs/board.php?bo_table=06&wr_id=795)
The 870 and 880 have a concave lens (exact type of lens not known) surface facing the removable lens and the 900 has a convex surface of a Meniscus lens facing the removable lens. The scanner section differs between the 800 series and 900 and this may explain the difference. In the 900 the beam output of the removable lens is refracted down to the the required dimensions for the first (tilting) mirror. The beam converges behind the lens element.
It is not known whether the large lens that Ultrapurple has will work with the lens taken from the THV900 but we are hopeful of compatibility. There remains the ugly possibility that the removable lenses are slightly different between the 800 and 900 series cameras and Ultrapurple mat have a lens incompatible with the 900 series integrated lens. We shall see.
Fraser
-
I forgot to say, the lens in the THV900 is a Positive Meniscus lens,
Fraser
-
A final comment on the THV900 lens mount and rev able lenses....
The THV900 lens mount is not the same as that used on the THV800 series. Whilst the lens barrels may be the same, or similar, the actual bayonet mount is different. This means that lenses cannot be shared between THV900 and 800 series cameras :( The THV900 lens mount bayonet fitting is effectively the inverse of that found on the THV800 series. What nasty, scheming, marketing pencil pusher thought of that I wonder ?
Fraser