Thermal Camera teardown - Raytheon Cadillac DeVille Thermal Nightvision camera

[Fraser]

This is the sort of video that is commonly seen from owners of these old cameras..... no that is not the image quality that I would expect to see from this well made BST core. It is a little sick.

To cover your points.......

The GRP casing can start to degrade due to age and exposure to weather. None of my cameras have the issue that you describe so yours is seriously weathered.

The chopper wheel should spin up to operating rpm and only be audible as a whir if you move your ear close to the camera. A noisy chopper wheel suggests either a ripped membrane (likely) or a failed motor bearing (unlikely unless very high hours or water ingress has occurred)

The FOV was intended to cover the road ahead for approximately the width of the high beam headlights but to a greater detection range.

High current draw at initial start is normal for these cores but 4A appears high to me. The current draw is increased at start by the microbolometer heater that raises the microbolometer temperature to around 32C. Once at this temperature, the drive to the heater is reduced. There is also a front window heater that is controlled via an  external ambient temperature measuring circuit. That heater, when activated, will also draw significant current but in your test case, may be ignored.

Yes your image is not up to the expected standard. I suggest a failed chopper wheel membrane. It will likely have rips in it. The membrane does discolour with age but this is not an issue. It is tearing of the membrane that is the issue. A new opaque type chopper wheel may be 3D printed and this works OK for many applications. Bill W has new chopper wheels but you need to compare the chopper wheel membrane spiral to determine if they are the same.

When you have the camera disassembled it would be a good idea to fit all new tantalum capacitors of the same, or better specification. They are subjected to a hard life in these cameras and can degrade over time. Complete failure can lead to a ‘burn-up’ that damages the PCB.

Fraser

[ArsenioDev]

Fraser, I dealt with this exact problem on one of my cameras during installation into my friend's FJ cruiser as a NAVFLIR system, above poster has an alright camera, it's just decently out of focus. Unscrewing that gasketed hex screw at the back, then feeding an allen key up the port opened to turn the gear train to move internal focus mechanism while camera is running to nice sharp video. Definitely looks like he's running the 2x zoom software option inside the camera, not sure how to disable that.

[bostwickenator]

So THAT is what that screw is for! I hadn't thought too much about the mechanical assembly up front figured it was aperture control from a glance at Fraser's photos but now I think about that it's pretty obvious that doesn't make sense. I'll try adjusting the focus shortly. edit: Imperial hex!

Follow up I didn't know there was a digital zoom feature on these. There are only the two output lines aside from the power. Without any service IO port I suspect any options would be by either be set at firmware load time or controlled by a PCB jumper. I'll have a look around when I open it up.

Fraser I think you are probably right. It's likely got damage to at least one quadrant of the chopper from the fluttering sound. I should be able to print a replacement if that is true. I'll throw some CAD models up here if I do. edit: Well that will save me some CAD time https://www.eevblog.com/forum/thermal-imaging/raytheoncadillac-deville-camera-3d-printable-focusing-mod-and-chopper-wheel/

[ArsenioDev]

One of the first things I figured out was what that screw did.
Also, it's BST so your SNR won't be that great in comparison to more modern VOx systems. My 08 and 10 vintage L3 QVGA sensors beat the absolute PANTS off my '02 and the BSTs, even the customized ref design version that has digizoom disabled.

[Bill W]

The chopper wheel should spin up to operating rpm and only be audible as a whir if you move your ear close to the camera. A noisy chopper wheel suggests either a ripped membrane (likely) or a failed motor bearing (unlikely unless very high hours or water ingress has occurred)

The CD motor bearings used on later models are not great, and the mounting is not ideal with the wheel hanging off the front (although Argus was far worse).  Only the earlier 200 analogue cores had the Swiss motors.  later digital cores used cheap CD spindle motors, but of course horizontally mounted.

I've loads of the motors (correct voltage) if anyone needs one.

Bill

[Bill W]

I wonder if the text I quoted about an interlock was added after Cadillac stopped shipping these cameras. From the descriptions of the Volvo/BMW thermal units they use a software interlock.

The text has been present since the major 2007 edit of Wassenaar (when all the 9Hz / silicon sensors clauses were changed), so predates a lot of BST production.  The concerns and need for change however came from the use of BST in civilian vehicles.
2007 wording:

c.   Where the camera is specially designed for installation into a civilian passenger land vehicle of less than three tonnes (gross vehicle weight) and having all of the following:
1.   Is only operable when installed in any of the following:
a.    The civilian passenger land vehicle for which it was intended; or
b.    A specially designed, authorized maintenance test facility; and
2.   Incorporates an active mechanism that forces the camera not to function when it is removed from the vehicle for which it was intended.
Note   When necessary, details of the item will be provided, upon request, to the appropriate authority in the exporter’s country in order to ascertain compliance with the conditions described in Note 3.b.4. and Note 3.c. above.

It is up to the origin state to determine if the clauses were adequately met by any given design.  Proir to 2007 BST was controlled anyway, so did Cadillac not ship TIC equipped outside the US/Canada ?  As it would be under EAR, the US would likely make an exception for Canada anyway as Canadians cannot drive anywhere else.

Still does not cover what should have happened at the US-Mexico border.

Bill

[bostwickenator]

Built a test housing for it and I figured out some of the sound. The position sensor for the chopper wheel has a tiny plastic spring clip holding it in place. Or it's supposed to, half of the clip was rattling around inside the camera and the position sensor was loose enough it touched the wheel when you moved the camera around. There is still a little chattering noise. Also gave the sensor a quick wipe down. Oddly I'm getting really low power output on the composite line. My mini LCD will work correctly on it but that's the only monitor that will. All the front side caps test sane values. I don't really want to be poking around on the PCB. It might be easier to add an opamp to boost the output.

Also, it's BST so your SNR won't be that great in comparison to more modern VOx systems.

This really surprises me. The sensor size is so much larger I thought that the noise would be orders lower. The pyroelectric effect must be massively less efficient than I supposed.

[Fraser]

Nice housing 

If it is 3D printed, would you be willing to share the STL file please ?

Fraser

[ArsenioDev]

On the low voltage output, I saw this as an issue on several of my units, it was a challenge of putting it in my best friend's FJ as a NAVFLIR. A nice video amp did the trick perfectly and made locking onto the signal easy

[Bill W]

This really surprises me. The sensor size is so much larger I thought that the noise would be orders lower. The pyroelectric effect must be massively less efficient than I supposed.

The BST is specified as NETD=80mK, current ASi are 60mK and VOX usually around 40mK, if using an f/1 lens and the factory test bench electronics. 

Most of the sensor improvements that could have led to lower NETD have instead gone into pixel size reduction (reduced sensor and lens costs) and less fussy electronics requirements especially on bias voltage noise.  With VOx in particular the lower NETD has simply resulted in the tendency to fit smaller aperture lenses, so a 40mK sensor with an f/1.4 lens is effectively back to being 80mK in the image.

BST has the 'halo' effect from the chopper wheel but will also always be free of fixed pattern noise, something that does not get factored into NETD figures.  The relatively crude automatic gain control does not do the BST any favours.

Bill

[bostwickenator]

Thanks @Bill . Maybe I'll pick up a more performant VOx device one of these days. Hopefully lenses get faster if these new reverse vulcanized Sulphur plastics become a thing.

@ArsenioDev thanks again for the help. It's great to hear that I'm not alone in that. Oddly it definitely had a stronger signal when I first turned it on. Evidenced by the video I recorded and included on previously. Makes me wonder if it's dying on me or if my wiring has such bad impedance mismatch I'm losing all the power. Regardless I ordered some highspeed opamps and I'll try adding a booster to the output signal. I included a BNC video output on my case design and as such I feel obligated to make that work lol.

@Fraser Well I picked up some replacement caps because of the above concern something is dying. I need to check the underside of the board. How bad was getting the connector off the PCB am I in for a fight?

Including an image of the position sensor bodge repair

[Fraser]

Removing the PCB from the surface soldered connector pins was pretty simple. Just use plenty of good sticky flux and decent desoldering braid/wick and it will wick away the solder. Then press a little on the PCB next to the connector as a soldering iron is applied to each pin and joint opens. The PCB then just tilts up and out of the casing.

Fraser

[bostwickenator]

Ah cheers, I haven't worked with this style of connector before and was expecting to have to pull remove the retaining pin and pull the connector out of the housing after desoldering it. If I can pivot the board out with the connector in place that will be much better. I'll update after my caps come in.

[bostwickenator]

Update I am an idiot. I'd wired the video ground to the system ground. I thought I had bench tested that being ok but apparently I misremembered my collection of alligator clips . Now that I separated the two ground planes and the video output works perfectly again.

[Fraser]

 

[simply]

Built a test housing for it and I figured out some of the sound.

That housing looks awesome - are you able to share the design for it?

Just rescued a Thermal Eye 4000M from the scrap bin. After som cleanup it works like a charm, and with a nice case, it could be put up for some good use too!
I've got a SmallHD screen that will make a perfect "portable" combo.

Rgds
/Andreas

[autroguy]

Hi All,

I wanted to share the info I was able to obtain form L-3 on these Raytheon units. I have a 2000B camera that came form a marine night vision navigation system made by NVTi. I believe mine has a 100mm electric focus lens. Photos attached. I made an aluminum housing for it and have it running off a Lithium Ion battery pack.
I was also able to obtain the factory interface software for it called ControlIR Manger. I have attached a few PDF's which may be helpful. I am not sure how to share the software and I also have one PDF too large to share here with more technical info. If anyone is interested in the unit I have or the software send me an email.

Fred

[Bill W]

Many thanks for finding the normalisation process, saves me trying to remember it.  My copy was binned along with the jig when some idiot threw out the storage cabinet it was kept in.

Apparently 'Must be kept phone x ****'
translates into
'As you are too thick to use this or know why it is useful feel free to throw it away'. 

Stupid 5S 

Bill

[bostwickenator]

Amazing info already. I dropped you a PM. I wonder if the software works with the Cadillac units. I remember there were a bunch of additional pins brought to the connector but no obvious labeling.

[Bill W]

There were two versions of the Raytheon BST hardware, although each appeared in several shapes !

The early analogue version has an on board pot, and the OEM PCB has a card edge connector at the far end does not AFAIK do RS232.

The later digital version (no pot, OEM PCB with pinned header both ends) has the RS-232/485 connection and functionality.
The RS-232/485 is on both connectors of the OEM PCB, so the Cadillac PCB must have RS232 somewhere too.

I have pictures of each OEM PCB type on Fire-TICS  http://www.fire-tics.co.uk/spares.htm

The Control-IR software can also do the recalibration in 'Cadillac' 320x120 mode. See step 2.4 in the 12142104.pdf document posted above where there is an option button.

I have not tried but it might allow a Cadillac electronics set to be converted into a 320x240 standard camera, but that would likely need the different chopper wheel and maybe a different lens.  Possible however that the Cadillac PCB build had reduced memory for the half size image.

Bill

[ArsenioDev]

I would be VERY interested in tinkering with this controlir manager software, as I have both the modern digital version and the old cadillac version, can do testing to see if I can unlock the cadillac into full resolution mode and do some recal.
Upload it onto something like Uploadfiles.io or drive?

[Bill W]

ControlIR has been released onto the forum via a dropbox link, and I have a copy of that file archived on Fire-TICS as part of the Argus3 tearup project

http://www.fire-tics.co.uk/project3/index.htm

Difficulty might be on how much got stripped out of the hardware going from the OEM PCB design to the Cadillac PCB. Might be some unpleasant defaults you cannot change ?

Bill

[Fraser]

Not forgetting that the Cadillac version also uses a very different chopper wheel spiral. So that would need to be changed to the standard type fitted in the 300A , 300D , Argus 2 and Talisman WASP BST cameras etc.

A suitable chopper wheel can be 3D printed relatively easily though  The required STL file is on this forum. It is a solid chopper spiral so you lose the dark halo effect around warm objects.

Fraser

[autroguy]

The unit pictured above is for sale if anyone is interested. Please PM me.

Fred