The story of an orphan thermal imaging core that no one seemed to want.......

[Fraser]

The story of an orphan thermal imaging core that no one seemed to want.......

Even though I have stopped actively hunting thermal cameras on eBay, I still come across the odd interesting unit whilst searching for thermal camera accessories for my cameras. This is the story of one such occasion.

I was looking for FLIR supplemental lenses and saw an unusual item for sale. There was the imaging core from a dismantled FLIR Ti20 for sale at £75. I initially ignored the auction as I thought it likely to be of little use on its own without the other electronics of the camera.

A day later I came upon the Ti20 core again. This time I wrote to the seller asking if the main board was available. The response was in the negative so once again I moved on and ignored the core.

So why was I ignoring this neat little thermal imaging Core ? Well there are several reasons really

1. The Ti20 is a 128 x 96 pixel thermal camera. Anything below 160 x 120 is of little interest to me these days.

2. The Core was of unknown make so finding information on it could be challenging.

3. With, or without the cores OEM known, it might not be possible to obtain the required data on the core to drive and control it. Thermal camera Core OEM's are infamously cagey about releasing such information to the public.

4. It was not clear that the imaging Core was complete in terms of being a viable imaging unit, even with the required command set.

So basically the core looked like it would be more trouble to me than it was worth, so I ignored it.

I kept ignoring that darned core every time it appeared in my searches but today a thought came to me. Maybe I already knew this cores OEM as FLUKE do not make their own cores. It was likely sourced from one of the big names in thermal imaging cores for integration into other products. There are not that many companies manufacturing such cores, so maybe with a bit of image study and research I could identify the Core or at least its OEM. Any information that I then found would tell me whether the core was still viable.

The pictures of the core on the auction were not great and I only found one other picture of the Fluke Ti20 core on the internet. This was not a great start but some key information was present on one of the auction pictures. Namely, a label listing four Patents applicable to the core. Now some times a Patent does not actually tell you who made a product, but it can suggest the technology that the equipment is based upon.

I found the four patents with ease. I recognised them and they were significant. What did I find ? That will appear in the next installment

For now, I leave you with pictures from the auction.......

[Fraser]

Part 2.......

OK, so we left off where I had researched the four Patents that were listed as applying to the Core. What was the significance of what I found then ?

The Patents were filed in the name of "Lockheed Martin Thermal Imaging Inc."
So what ? you might ask. Well I have come across Lockheed Martin Patents on thermal cameras previously I now had a very good idea who manufactured the mystery core I knew these Patents and I suddenly recognised them as having been associated with a core I had previously worked on. I also had an inkling that I knew the OEM from the PCB stack design. A picture was forming......

With these pieces of critical information I went back in my thermal camera teardown picture archive and quickly identified the likely core family. In fact better than that, I found the data sheet for the same core Luck was on my side this day.

So who made the core and is it complete ?

The core is made by the Venerable BAE Systems and it is their SCC500 model. I previously found that core type in the SCOTT Eagle fire fighting camera. The auction core is complete and comprises a stack of 4 PCB's.

So now I knew the identity of the thermal imaging core, could I actually drive it without the FLUKE main board ?

The simple answer is yes. The SCC500 is a self contained core that is remote controlled via RS485. It only requires configuration using simple commands and can be set up to operate in 'full auto' mode, or any other mode really, as soon as power is applied. I own the complete document set for the SCC500, including its command set

It is possible to treat the core as a remote controlled thermal camera using just a simple PC terminal program or other application that sends commands to the camera via the RS485 link. An Arduino could easily do such to make a remote control keypad. In its full auto mode the core is actually a complete headless thermal camera that outputs either digital or analogue video to a computer or composite video display. Quite a useful little core then.

The SCC500 was intended to be a highly configurable thermal imaging core that needs very little in the way of other support electronics. Its intended usage was Military weapon aiming sights and missile guidance ! It was popular in the thermal imaging Industry due to its high performance and ease of integration into a system.

The SCC500 core has a distinctive three point support system for its 4 PCB's and this is what quickly identified the mystery core for me. I had just needed to confirm that it was the same model of core.

Within minutes of identifying the core as an SCC500 I had returned to the auction and purchased it.

I had ignored this little imaging core for days and though keenly priced, no one else seemed interested in the poor thing. As already stated, this was likely because it presented a significant risk of being a paperweight unless identified and data on it acquired. Without clear OEM identification, the core was of little interest to many hobbyists. As stated, even I kept ignoring it.

Well now this poor little Orphan thermal imaging core is joining my family
Oh and the SCC500 cores true resolution is 160 x 120 pixels and this is settable in the configuration options so it is only FLUKE hobbling it to the lesser 128 x 96 via the firmware or factory configuration. 

Today was a good day for both me and the unloved Orphan imaging core

I attach the pictures of the SCC500 Core used in the SCOTT Eagle camera. Ignore the different lens and microbolometer mounting method. The SCC500 is very versatile and the microbolometer may be located remotely from the PCB 'stack'. I include a picture of the SCOTT core with its rear mounted accessory control PCB removed to reveal the cores rear PCB and its Patent stickers. I then include teh auction cores similar picture showing the same PCB ..... Core identity established

Fraser

[Fraser]

Now the SCC500 datasheet........

[Fraser]

Finally for tonight, the link to my teardown of the SCC500 core that was found inside my SCOTT Eagle thermal cameras.

https://www.eevblog.com/forum/thermal-imaging/scott-eagle-x-fire-fighting-camera-very-nice-unit-and-core-by-fraser/msg1308661/#msg1308661

Fraser

[Vipitis]

that is a great find I guess. I wouldn't be able to as I don't have this amount of knowledge.

but how does the sensor compare image wise to today's microbolometers?

and with configurable - could you do long exposure with this sensor or it is locked to 1/60 for those 60fps?

[Fraser]

The BAE slide show detailing their SCC500 product and its advantages in Military use.

https://slideplayer.com/slide/4037562/

Fraser

[Fraser]

Viputis,

The core uses microbolometer technology from Circa 2004. That does not, however, mean that it is 'old' technology so "no good". The microbolometer used in the SCC500 was intended to meet the requirements of the Mikitary applications for which it was originally intended. That means it's a decent piece of technology.

Modern Microbolometers come in varying levels of quality and performance. They have reduced in size, use smaller pixels and some use cost reduced manufacturing techniques to make them affordable in the consumer market place. Some offer excellent performance, whilst others leave much to be desired. Compromises are sometimes made that effectively spoil a Microbolometers performance.

So back to the Circa 2004 SCC500 MinIR core. It is a high speed core with decent resolution and sensitivity. With a decent lens fitted, performance is as good as can be expected from 160 x 120 pixels. The core is based on VOx technology so produces a cleaner image than A-Si based cores without the need for lots of noise reduction processing. A-Si has got better over the years however and image processing keeps getting better.

The FLIR imaging cores from Circa 2000 are still very capable pieces of technology. The BAE cores are no different. Can the SCC500 core compete with modern offerings from the likes of FLIR ? Maybe, depending upon which FLIR Microbolometers we compare it to. If would trounce the Lepton 3 but better FLIR cores use far more sophisticated image enhancement processing so they would out perform the SCC500 in terms of 'pretty pictures'. The Microbolometer FPA's have greatly improved over the past two decades in terms of cost and image noise reduction. I would always expect a modern Semi-Pro camera to outperform the equivalent standard of camera of 10 or 20 years ago. That is to be expected. Would I expect a consumer grade camera to outperform a Semi-Pro camera of 10 years ago ? Maybe not. So much depends upon the quality of components and image processing used in the modern consumer grade camera. Many modern 'budget' cameras suffer from relatively poor image processing software so are unable to make the most of the microbolometer creating the signals.

I do not think the SCC500 offers changes to its integration time. I would need to check the developers pack.

Fraser

[Fraser]

Further to my last.....

If we were to compare similar core formats of the same era as the SCC500, I suspect the BAE offering would offer stiff competition to the offerings from Raytheon, such as the Thermal Eye 2000AS. Why would this be the case ? Well, as stated, the SCC500 uses VOx and the TE2000AS and its descendants use A-Si.

As I have discovered, owning both VOx and A-Si based cameras, easily A-Si technology produced very useable image, but they tended to contain quite a lot of distracting column noise. This is due to the small signal produced by A-Si microbolometer technology. VOx in comparison produces far cleaner images thanks to the larger signal that IRS microbolometer can produce. In more recent years, the development of A-Si Microbolometers has improved their performance greatly. They can still suffer from some noise issues but these are dealt with in the image processing stages. The performance of an A-Si microbolometer core can now rival that if a decent VOx core.

Image processing has improved in all modern thermal cameras thanks to the availability of high performance DSP chips and more powerful mobile processors that can crunch the numbers with ease. Thermal imaging technology is still developing but there is also the commercial drive to just build cheaper rather than 'better'.

Fraser

[Fraser]

The topic of age Vs image quality reminds me of a recent conversation I had with a chap who normally specialises in the RF world.....

I have known this chap for some years and he provides support to his employer on all matters RF related. He was recently handed boxes of development equipment left over from a past project. Whilst sorting through the boxes he found a Raytheon Thermal Eye 2000AS development kit. It caught his eye so he wired it up to a monitor to see if it still worked.

In his words he was "blown away" by the thermal image that the core produced  He thought it amazing resolution, as good as CCTV he had seen and the smoothness of movement when panning was just like a camcorder. This was a chap experiencing his first direct exposure to a thermal imaging core and it made no difference to him that it was only 160 x 120 pixels and likely produced some noise in the image due to it's A-Si microbolometer technology. That 2000AS core was the pathfinder for Raytheons later thermal imaging cores and it still impresses people today in spite of its limitations.

It was so nice to hear a "thermal camera virgin" getting excited about the images produced by a relatively old core. He has not been spoilt with higher quality imaging as many of us have. The more you have, the more you want  It is true that once this chap is exposed to more modern thermal cameras, he will see how far the technology has developed since the venerable Thermal Eye 2000AS was created, but for the moment, he is more than happy and impressed that such a compact thermal core is creating such lovely thermal images. Good for him I say

I still own and use 160 x 120 pixel cameras but I prefer the 320 x 240 pixel images that I have used for so many years. The higher resolution just makes life easier. I have a 640 x 480 pixel DRS camera but it's greater resolution is offset by its wider field of view as it is a CCTV pan tilt head. I never cease to be impressed with how far this specialist technology has come over a relatively short period of time. Sure consumer digital cameras and mobile phones etc have advanced at a greater pace but thermal cameras are not so main stream. Compared to the single pixel detector scanning cooled cameras that I used to use, all of the microbolometer and BST starring FPA based cameras are amazing

Fraser

[Vipitis]

I have been discovering analog photography and the quality and detail in negatives. Even the build quality and detail of old cameras is astonishing.

For thermal cameras, I have only whistnessed my very own Lepton based phone as well as some older FLIR T350 or something at some events. I would be impressed by a 60fps 160x120 camera as well.


But I got a question. In the document, there are high resolution models available of the same stack... And they just lower the pixel size on the sensor... Doesn't this reduce the signal strength and needs higher amplification (= more noise)? I wonder becuase those pixels for the 160x120 model are 46um, which is huge compared to today's 17um or even 12um. That is a really high factor... But probably shows how much the noise reduction has improved. 

[Zucca]

Fraser, thanks for writing, I enjoyed it. Please consider to collect all your IR ebay adventures in a book.

[Fraser]

The diminutive BAE SCC500 core arrived today 

It is beautifully made with a nice solid aluminium lens block chassis. The PCB's are supported on the standard BAE SCC500 tri-point mounts and all is nice and rigid. Ribbon cables and connections are locked in place with hard black epoxy for vibration resistance.

The core looks to be in fine condition and stated as working. The Ti20 suffered a main board failure. The same seller is asking £100 for just the Ti20 LCD panel so I think I got myself a good deal on this core 

Fraser

[Fraser]

The cores microbolometer and confirmation that it is a MicroIR core (SCC500) on the ribbon cable.

Fraser

[eKretz]

Nice! Good deal for sure.

[Bill W]

The SCC500 is 'second generation' in commercial cores.  As such it is certainly ahead of the Raytheon AS2000, the natural comparator would be a AS3500.
Raytheon BST was always a bit different and so widely liked for that in 'pretty pictures'.

There was an earlier Lockheed VOx core, forget the product code,  used in several fire service cameras such as early MSA and also an Argus3 model
http://www.fire-tics.co.uk/datasheets/A3VOX_datasheet.pdf

Bill

[Bill W]

But I got a question. In the document, there are high resolution models available of the same stack... And they just lower the pixel size on the sensor... Doesn't this reduce the signal strength and needs higher amplification (= more noise)? I wonder becuase those pixels for the 160x120 model are 46um, which is huge compared to today's 17um or even 12um. That is a really high factor... But probably shows how much the noise reduction has improved.

Generally the pixels are getting better at the same amounts as getting worse due to smaller size so that device performance is fairly static. 
No one will bring in a new sensor pitch to give a worse picture in an equivalent product.
So better vacuums, better resonant structures and absorbers and also lower noise bias drives are necessary to get a lower pixel pitch.

The benefit of lower pitch is in cost, mainly dies per wafer, and also optics costs for the end user.


Bill