Raytheon 2000AS core teardown - and back up again

[Bill W]

A bit of background.  The Raytheon 2000AS was the first generation of commercial IR cores let lose out of the military.  It was a 160 x 120 sensor and could put out 20Hz NTSC video (so 3 consecutive identical fields).

A bit of an anomaly as being Silicon it was (but is not now) completely free of export controls much as TGS detectors (Pevicons) had been.  It was therefore popular for commercial products including Firefighter cameras.  it was used in 'export' versions of Argus 2 & Argus 3 from Marconi (ex EEV) and also in products from Bullard.

It is a very simple volts-in video-out system, although Marconi did add on graphics and colour rather than simply display the core output.

[Bill W]

Now the sensor photos.

The optical block is a lump of aluminium with the solenoid / flag embedded in it.  It splits with the lens in one part and the sensor / flag in the other.  As a 'bolometer' it needed a flag.  Competitive products at the time were Raytheon's BST (use a wheel) and Lockheed's VOx (also flag)

The sensor itself is bonded to a flexi PCB / cable.  It is also vacuum wafer packaged, pretty novel for back in 2000.  There's still a lot of sensors made in vacuum ceramic cases.

[Bill W]

The first PCB, 'Processor PCB', is attached to the sensor and solenoid.

This looks like it dealt with sensor correction and has as main IC's TI TMS320VC5410 DSP and AD7472 12 bit 1.25MHz ADC and an unidentifiable clock generator IC (perhaps custom ?)

It is 55 x 100mm

[Bill W]

The second PCB, 'Reformatter PCB', looks like it dealt with producing the NTSC video image and has power supplies (MIC4680, LTC3401)
There is a 24MHz clock, TI TMS320VC5410 DSP, and AD9708 8 bit DAC.

It has the end user connections of J1 and J3.

1    Power in +       6 - 32V. DO NOT EXCEED 33V
2    0V                       Ground
3    Low Battery       Shows a low battery logo when grounded. Software version dependant.
4    RS485+    
5    RS485-    
6    Ground               Ground
7    Shutter               Ground to force a manual shutter / flat field correction
8    Video Polarity       Ground to select black hot imagery
9    Low Voltage in    2.5V - 4V an alternative to using pin 1, typically 2 x AA cells. DO NOT EXCEED 5V
10    Power Out       Connected to pin 1 via a diode to aid wiring.
11    3.4 - 3.7V out       To drive a display, 250mA maximum
12    Video Gnd       Ground
13    Video               NTSC video


This is all public domain as the user guide is available on the internet, and not marked commercially confidential.

[Bill W]

So to performance.  By current standards nothing great, it is a 160 x 120 after all, and video presentation was still being developed.  It is not that happy looking at a very cold sky, tending to lose a lot of contrast.  However it is a lot better than a FLIR one and is very simple to use.  The lens can be set free to focus down to a few inches, so good for PCB inspection without any secondary lenses.



A couple of still images, the respective videos are on www.fire-tics.co.uk/project1

[Bill W]

So to the 'tear up'.

With cores available both inside and outside the US this is a route to build your own camera, recovering them from firefighter cameras (Bullard T3 series or the 'Blue' Argus 3 ASi P4466).  Note that the more common yellow Argus 3 P4467 is a different core, Raytheon BST (for a future thread).
The cores are now export restricted as exceeding 9Hz after the 'silicon' loophole was closed off in Wassenaar.  However no ongoing US controls apply so control is by the countries/control zones where they reside.  So cores in the UK can go across the EU and a good few hundreds must be out in the wild.

A camera is a simple 'box it up' job with some batteries and a video display for me, but a much smaller camera could be made if fed power and viewed on a monitor or PC TV card, eg for PCB inspection.

Gory details including wiring on www.fire-tics.co.uk/project1


Hope this was informative and enjoyable

Bill

[Fraser]

Thank you Bill 

The venerable Raytheon Thermal Eye 2000AS must have been an exciting development for OEM's who wanted a compact and relatively simple to use core for a thermal camera design. For its age, it is a nicely designed and compact core. The 20fps update rate is a little disappointing compared to that of the Raytheon BST core but it is still decent enough.

Fraser

[Bratster]

My best thermal camera score on eBay was a batch of these guys. Seems to be a couple different revisions, but I got 10 of them for $350 us.

3 work perfect, 1 has a damaged latch on a zif connector but works fine. The rest were all parts units with a couple different kinds of lenses.

I have yet to go through and see if I can swap parts around and get them to work. I know that the circuit boards are somewhat tied to individual camera cores, but it's worth a shot.

Do you have a link handy to the public domain manual?

EDIT: nevermind, I see the manual on your site.

 I do have two different versions of manuals that I got from the seller, but I would like to see if there is any different information contained in that one.

Sent from my Nexus 6 using Tapatalk

[Bill W]

I haven't tried it, but would think there's a good chance the reformatter / power supply boards might swap.
The processor board is almost certainly tied to the sensor with calibration in the memory.

Yes the ZIF can be a problem, that is the fault with one of mine, at the sensor end.  I think it should be possible to put a pad on top held by the screws at the sensor end.

regards
Bill

[Uho]

Hey. Thanks for the info. I made a thermal imager from such a matrix.
https://youtu.be/KDzDS8_u-EM