-
FLIR SC3000 cooled GaAs Quantum Well FPA LWIR camera teardown by Fraser
Posted by Fraser on 30 Sep, 2023 21:02 -
Now for something a little different !
I have owned a FLIR SC3000 science camera for quite some time but have not previously detailed it on this forum. This camera in notable for several reasons....
1. It is a Science Camera, so is very capable
2. It uses a Stirling Cooler to cool the FPA to 77K (-196 Celsius)
3. The FPA uses GaAs Quantum Well photodetectors
4. The camera is a cooled LWIR unit sensitive to the 8um-9um wavelengths (yes 8um to 9um)
5. The QWIP FPA provides a thermal sensitivity of better than 20mK
FLIR were very proud of their QWIP science camera and made great claims for it, including it being the future of high sensitivity, high measurement accuracy LWIR thermal imaging. It was certainly a new technology but it required the use of a Stirling Cooler so that made it very expensive. Modern LWIR microbolometers claim excellent thermal sensitivities but the SC3000 offered amazing sensitivity for its time and is still a very low noise imaging device today. Cooled cameras do not tend to exhibit the 'next curtain / chicken mesh' effect that is prevalent in microbolometer technology cameras.
The SC3000 is a static camera that is controlled by either a handheld remote control or a computer. A digital data output is included for passing 14 bit image data to a PC for analysis. Radiometric images may be stored on the internal PCMCIA ATA flash card (CF card compatible with an adapter)
As this is an unusual cooled LWIR camera I thought I would share its teardown with the forum. It has much in common with the Cooled Agema Thermovision 550 that I detailed in another teardown on this forum. The FPA is where the main physical differences exist. The optics are designed for LWIR operation covering the 8um to 9um wavelengths used by the GaAs QWIP FPA.
To the documents and pictures ....
I attach the brochures and Patent for the SC3000 QWIP camera.
-
Pictures begin.....
Note the SC3000 lens needs a clean but is otherwise in perfect condition
-
Continued.....
The two halves of the casing showing the main electronics package in one side and the Cooler/FPA/Optics package in the other side.
The Stirling Cooler is of the rotary type. It is connected to the QWIP FPA via a cold finger that is within a vacuum Dewar. The output of the FPA is connected to a small PCB that, in turn, connects to the FPA processor PCB mounted on the Main computer PCB.
The optics block in this camera is a motorised focus internal design that is not removable to change the FOV. Optional supplementary lenses are mounted on the front of the camera when a different FOV is required. The supplementary lenses have magnets mounted in them that communicate to the camera which lens is fitted. Automatic calibration offsets are then employed for measurements. The optical block also incorporates a flat field flag and solenoid but this is different to that of a microbolometer FFC flag as it does not operate unless commanded to do so or upon changes of range. There is a motor driven high temperature filter in this camera to enable the measurement of temperatures up to +2000 Celsius.
-
Continued......
Pictures show the Main computer PCB, FPA Detector processor PCB, Power Supply PCB and TAXI 14 bit digital communications PCB
-
Continued......
-
Comparison of Agema Thermovision 550 SWIR cooled camera and the FLIR SC3000 LWIR cooled camera........
-
Excellent info and nice to see the teardown. Little bit of correction on the QWIP's. They have to run colder to maximize performance and hence the SC3000 runs at 70K. The data sheets call that out but can't remember if the camera could tell you. Seen other QWIPs running a little colder.
-
IR_Geek,
Thank you for the correction on QWIP detector operating temperature
Sadly this camera will never reach that operating temperature again as its Stirling Cooler is dead and beyond repair. The cooler had been dismantled prior to my owning the camera and there is something loose in the QWIP detector capsule. I have photographed the disassembly of the cooler as I know that some forum members are interested in what resides within a coolers casing. I will upload the pictures soon.
Fraser -
Whats the point to cover from 8um to 9um only?
-
I am no expert on Quantum Wells but I understand the GaAs Quantum Well dimensions set the peak sensitivity of the detector and this results in the wavelength coverage of only 8um to 9um in the SC3000 camera. In other words, the wavelength coverage is determined by the physics of Quantum Well detector technology.
Fraser -
The teardown of this cameras Stirling cryo-cooler is to be found here…….
https://www.eevblog.com/forum/thermal-imaging/micro-stirling-mechanical-cryo-cooler-teardown-by-fraser/
Fraser -
Whats the point to cover from 8um to 9um only?
The other side of this choice is that the QWIP is achieving very good sensitivity compared to standard LWIR microbolometers WITHOUT using much of the 'ambient temperature' emissions of everyday objects.
One use is gas imaging, where gas absorptions are fairly narrow band and would be swamped by a full spectrum of LWIR -
QWIP detectors are old hat, I sat down and spent a few hours talking to Dr Murzy Jhabvala learning about SLS detectors and the next generations of wide response detectors and on-detector filters.
-
Whats the point to cover from 8um to 9um only?
Fraser basically nailed it. The detector material, structure, and bias voltage determines wavelength. Rogalski's book "Infrared Detectors" is an expensive yet excellent source. He cite's 100's of papers per chapter.
interesting video from NASA
-
I was also lucky enough to get a FLIR SC3000. My machine ran well, the cooler reached the working temperature in about 5 minutes, but the imaging effect was not very ideal. I will share the video I shot with you. Or I can shoot for anyone who needs it
-
I attach video, I don't know how this device uses computer to save video
https://youtu.be/AyQiMnjhBrU -
Fraser,
Thanks for doing another teardown. Very interesting and entertaining, as usual. It’s always interesting to see designs from a prior era of technology. Reminded me a bit of the old pevicons. Just one of those caps looks as large as some of today’s camera cores. And, the camera housing is made out of a machined block of aluminum!
In about 1995 or so, there was a company in the US named SafetyScan. They manufactured a firefighting camera that had a housing also machined out of a block of aluminum. It was heavy and did not do well in the market. Just another company that came and went in this industry. The man who ran it was Mark Stroze. I don’t remember much about him other than he didn’t seem to like me very much. I think the camera’s core was a Lockheed Martin SIM100 or 150. Big boards; big housing. Nice images, too heavy.
Dunno if the EEV alumni on this board also remember that one.
Nice work Fraser! Thanks again!
David
-
David,
I am pleased that your enjoyed the journey into the past where electronics were less heavily integrated and “heavy metal” was considered a sign of quality and some justification for a units high cost ! When you spent the sort of money that these cameras cost, you kinda expected something nice and solid feeling in your hands
I bought this camera as a project in 2015 but that was the time that I had to leave my job due to M.E and this camera just got boxed up to await my attention. I thought it had a power supply fault and always knew that it was a gamble with regard to the coolers serviceability. I was not aware that someone had already ‘got at’ the Stirling Cooler though. I cannot remember what I paid for the camera but it was not a great deal of money at the time as thermal cameras were not as well known on eBay and ‘exotics’ like the SC3000 often attracted little attention.
The SC3000 camera shares many similarities to my AGEMA Thermovision 550 units that were sold 10 years earlier. It just goes to show that in the World of thermal imaging Circa 2005, the older, proven, electronics packages were still valid and were adapted to the newer thermal sensor technologies. The AGEMA single board computer in these cameras is based upon the Motorola MC68340 micro-processor and this IC integrated additional interfacing that is not found in the simpler MC68000 version. I studied these Motorola micro-processors when reverse engineering the AGEMA/FLIR PM570 microbolometer based camera. I was not surprised to find that the original AGEMA single board computer design from the THV550 continued to be tweaked and used in all the later industrial cameras in the PM series as it was a very versatile controller board. It had an imaging sensor input that was fed data from another processor/interface board that was unique to the sensor technology in use at the time. This meant that it was relatively simple to deploy a new sensor technology or version with an updated image processor/interface board that connected to the same main single board computer design. Some new firmware and hey-presto, a new camera was born ! The same basic SBC design was to be found in the PM695 that marked the end of the PM series. That thermal camera had a visible light camera added to it in a ‘carbuncle’ bolted to the front of it (I never liked that design decision !). To support the addition of a visible light camera, FLIR took the standard SBC and added an additional digital input. The visible light camera had its own composite video to digital data converter (Video decoder and ADC) mounted in the ‘Carbuncle’ housing. The data from the video to Digital data converter was fed to the new digital port on the SBC and the firmware was modified to support the VL camera. FLIR were responding to market demand for visible light cameras on thermal imaging systems to assist in documenting thermal imagery in reports and also to provide the user with thermal scene context. The old Motorola MC68K based SBC was so versatile that it could be adapted to designers needs with relative ease. The introduction of the Microbolometer FPA required a separate image processing PCB that used four physically large CPLD’s and another microprocessor (another Motorola IC I think) with its own Flash ROM and RAM, to deal with the needs of the Microbolometer and pass the resultant processed data to the SBC. This design did provide FLIR with the ability to develop the Microbolometer FPA image processing side of things without needing to respond the main SBC design. A nice situation for the designers but I do not know if this was a consideration whe the SBC was originaly designed for the THV550. For me, it is great as I can open any thermal camera from the THV550, through to the later PM series and find my way around the SBC with ease. It is like finding an old friend whenever I open one of these cameras. It took me 3 months of evenings to reverse engineer the PM570 camera design and friends at both my workplace and FLIR were both impressed and somewhat shocked at my efforts. FLIR were horrified to see a 3rd party produced schematic of their PM570 At the time I was doing that reverse engineering, the SBC PCB was a £5K spare part that needed programming by FLIR before it could be used. I had no choice but to reverse engineer the camera in order to trace the faulty component. I thought the fault lay on the SBC as the camera was halting during boot. It turned out to be a failed output on a humble 74ALS245 IC located on the Microbolometer processor board and not the SBC. I found that out by diagnostics on the SBC that proved it to be in perfect working order, but awaiting a response from a sub-system.That sub-system was the whole microbolometer processor PCB that was being held in RESET by the failed buffer gate. A £1 IC that effectively paralysed the whole camera. I could possibly have found the fault with completely reverse engineering the whole cameras electronics package, but I enjoyed the challenge. It has held me in good stead ever since as I repair PM series cameras using the knowledge gained from that PM570.
I may produce a teardown and commentary on the PM570 camera and it’s descendants as it is an old camera now, and no longer “cutting edge” technology. I promised FLIR that I would not release the schematics that I created for the PM570 but I can still describe the cameras design without breaking that promise Watch this space
Fraser