Author Topic: Thermal Camera teardown - Raytheon Cadillac DeVille Thermal Nightvision camera  (Read 4946 times)

0 Members and 1 Guest are viewing this topic.

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
I dug my two Raytheon Cadillac DeVille cameras out of the 'awaiting repair' pile today.

Both cameras need a new chopper wheel as detailed in a previous post on these thermal cameras. The membrane of the chopper wheel can split over time in these cameras due to the constant heat to which they are exposed. This was also an opportunity to inspect the PCB's for any potential issues. IT is just s well I did as there was an issue with both. The problem took the form of tantalum capacitors that are splitting ! They have not burned-up as is their usual form of 'suicide' but cracked cases cannot be good so they are going to be replaced. Only one value shows this issue on both cameras, 330uf @10V rated. Both cameras were built in late 2000. I have marked the cracked capacitors with a blue marker to identify them but I will replace all the 330uF Tantalums as a precaution. There are quite a few used in these cameras.

By way of introduction to these cameras, they are designed and built by Raytheon for Cadillac and based on the Raytheon Thermal Eye 300A. The sensor is a Peltier temperature stabilised BST FPA. A chopper wheel is required for this sensor and is the standard Thermal Eye part. All electronic components are located on a single PCB with the rear of the case acting as a heatsink for the Peltier element. The lens is mounted in the plastic front section of the case and comprises a large diameter Germanium element combined with some form of plastic lens at the rear. An unusual lens design likely to be due to BoM cost restrictions for the camera. The fragile Germanium lens is protected from stones and dirt by a Germanium window that may be changed by the owner if it is damaged. In an automotive deployment, such a lens protector is essential if no other form of lens impact protection is employed. The Germanium window is heated in very cold weather by a built in heating element ring and thermostatically controlled.

The important specs now, the camera has a 320 x 240 pixel BST array and the output is 60fps making it suitable for use on a fast moving platform. The BST camera does not need the FFC flag and event found on many microbolometer cameras so there is no frozen frame distraction for the driver. The BST FPA was and still is an excellent choice for automotive deployments. As a 60fps 320 x 240 pixel thermal camera made by Raytheon, these are liable to ITAR regulations if being shipped between countries.

EDIT - Correction on import/export rues by Bill W who knows a lot about such......

"A slight correction on export controls, these are not ITAR but are Wassenaar dual use controlled so need a license to cross borders, apart from internally within the EU."


Hope you enjoy the pictures.

Fraser
« Last Edit: August 10, 2016, 02:42:32 am by Fraser »
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
Inside the rear case
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
More
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
More
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
Failed 330uF 10V tantalum capacitors. Marked with blue marker pen.
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
The Patent detailing the design of this camera is attached.

The automotive deployment Patent US5763882 of 1998 is a duplicate in many areas but describes a 'low cost night vision camera'

http://www.google.co.uk/patents/US5763882

It should be noted that the term 'low cost' was relative to the industrial thermal cameras of that era costing more the $40K. These cameras were not a cheap option from Cadillac, that is for sure.

The Cadillac DeVille cameras are still popular and to be found on ebay. They range in price depending upon condition but expect to pay between $150 and $500. Their popularity stems from their rugged construction and ease with which they may be deployed on a vehicle for navigation in the dark and hunting. You just feed the camera with 12V DC and it produces NTSC composite video. No adjustments or commands, it is fully automatic and simplicity itself. People on You Tube have shown these being used in Pan Tilt Lamp housings an bolted behind the grille of their cars. They cost less than the Audi and BMW Autoliv FLIR cameras and, unlike those nightmarish units, the Raytheon is easy to deploy without any hacking needed.

Note that many of the cameras on the market are circa 2000/2001, will have broken lens protectors and split chopper wheel membranes. As such these should be dismantled, cleaned and serviced if you want the best out of them. Many You Tube videos using these cameras show signs of a poorly camera in need of some TLC.

As a point of interest, these Raytheon BST sensors and electronics design are used in EEV(E2V) ARGUS 2 and ISG Talisman firefighting thermal cameras.

Fraser
« Last Edit: August 10, 2016, 02:39:17 am by Fraser »
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
Interesting Kemet Tantalum failure analysis document. It even shows a thermal camera being used to detect failed Tantalums.

Fraser
 

Offline nidlaX

  • Frequent Contributor
  • **
  • Posts: 653
  • Country: us
Nice teardown and info Fraser!

Do these sensor modules have any type of temperature read-back capability built in?
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
NidlaX

No sadly the BST sensor array in this camera is non radiometric and does not easily provide precise temperature output for radiometric tasks. There have been clever designs using  the BST sensor that enable it to be used in radiometric applications, but for such tasks the microbolometer is a better choice.

Fraser
 

Offline Bill W

  • Frequent Contributor
  • **
  • Posts: 277
  • Country: gb
    • Fire TICS
The temperature measuring options on fire cameras that used the Raytheon BST (EEV Argus2, Marconi Argus3, ISG Talisman/K90) all used Raytek spot temperature sensors aligned with the camera view.
The BST is a 'ferroelectric/pyroelectric' detector so responds to changes in energy, hence the chopper, so is unsuited to measuring temperature by itself.  The only way is if the chopper is both solid and its' temperature is known when it becomes a bolometer that shutters at video rate.  The standard chopper is translucent and acts as an unsharp mask filter, which gives the BST image its' classic dark halo effect.


http://www.fire-tics.co.uk/images/A2R_TI.jpg

A slight correction on export controls, these are not ITAR but are Wassenaar dual use controlled so need a license to cross borders, apart from internally within the EU.


Bill

www.fire-tics.co.uk

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
Removing the chopper wheel from a Raytheon Thermal Eye camera .......

The Chopper wheel is a pretty fragile assembly that is attached to its drive motor using a press-fit. That is to say, no screw or adhesive is used.

The frame of the chopper wheel would break if any attempt was made to use it in the removal process. The removal requires care and an understanding of the 'hard points' on the wheel. There is only one hard point on the chopper wheel, the central boss.

If the membrane is totally destroyed and of no use, it would be feasible to use the standard knob removal technique where a strong nylon cord is threaded under the hard point and used to pull the assembly off axially. There IS a risk of breaking the central boss however. I opted for the leverage approach as that is safer and preserved the membrane for the purposes of templating a new one.

There must be no levering on the fragile frame, only the central boss. In order to achieve this I borrowed a serrated edge bread knife from the kitchen. This has a strong but thin blade and the serration permits force either side of the motor shaft rather than just on one side. To assist in the process of removal I heated the boss with a hair dryer for around 15 seconds on high heat. The knife is then twisted carefully to prise the boss off of the shaft. On my cameras the boss popped off easily. Take care and take your time. Rotate the boss 90 or 180 degrees if it is being stubborn and this will work it off of the shaft.

When the chopper wheel is off of the shaft you can see a reflective indexing disk on the rear. this works with an IR reflection sensor on the motor deck to provide chopper wheel position sensing for the synchronous chopper spiral timing circuits. It is very important that this index is correctly positioned with respect to the spiral start point.

Fraser
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
The chopper wheel mask is made up of translucent dots so is hard to photograph. I found a blue background helped so i have taken some pictures that i hope will help others to make their own replacement spiral chopper shape for use with the original wheel frame. Opaque spriral's will work for most peoples applications of these cameras.

There is another chopper wheel in one of the pictures. It is from an EEV Argus 2 camera. I have yet to establish the compatibility of this slightly different chopper wheel with the Cadillac camera.

Fraser
« Last Edit: August 10, 2016, 02:26:47 am by Fraser »
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
Higher resolution chopper wheel image
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
Raytheon Cadillac camera Chipset

Altera FLEX EPF10K100EFC256-3    10K series FPGA

ISI 61LV6416   64K x 16 SRAM 3V3

AMD L400BB55RI   256K x 16 Boot sector Flash Memory

ADV7127    240 MHz 10 Bit High Speed DAC

Analogue Devices M 01   Tiny RS232 Transceiver

NAT Semi 01AB CLC5602IM  Dual high output video amplifier 

TI HB244 98K EELD    Tri-State Octal buffer

OP279 G 943 G3332  Dual high output current Op-Amp

LT 945 112913     Low dropout regulator 1V3

4412   Mosfet

AD9220 ARS 9931 AP6958.1   12 Bit ADC

Altera EPC2 TI32 Cad489949   Altera 10K configuration Flash RAM

H 1135IB    360Mhz video Op-Amp

NAT Semi JM86AB LMC6084IM   Precision CMOS quad Op-Amp

Zetex T6790   Diode array

OP279 G009 G8986     Dual high output current Op-Amp   

NAT Semi 02AB LMC 6082IM  Precision CMOS dual Op-Amp

LT  012 1624I    Switching Regulator controller

LT 002 111825     Low dropout regulator 2V5

Raytheon 2270-35  BST 320 x 240 Pixel Array
« Last Edit: August 10, 2016, 04:12:39 am by Fraser »
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 14846
  • Country: za
Very nice Fraser. From your linked paper it looks like those caps were cracked by improper storage before soldering which started the cracking.
 

Offline AKMac

  • Contributor
  • Posts: 6
  • Country: us
For someone also attempting a repair on one of these cameras, with I'm guessing a burnt voltage regulator and several failed capacitors, how do you determine information about these tantalum capacitors? I notice that the 330uF caps say 337 on top; are those numbers related?
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 7460
  • Country: gb
AKMac,

Just Google "Tantalum capacitor markings" and all will be explained. There is a voltage rating as well  ;)

http://www.radio-electronics.com/info/data/capacitor/capacitor-markings.php?sa=X&sqi=2&ved=0ahUKEwjjm5381aPZAhUnBsAKHUPPBoUQ9QEIDjAA
« Last Edit: February 14, 2018, 06:57:01 am by Fraser »
 

Offline Bill W

  • Frequent Contributor
  • **
  • Posts: 277
  • Country: gb
    • Fire TICS
A common failure mode is that of tantalum capacitors used  on power input or SMPS smoothing being hit with surge current.  That is more a total burn out than the cracks Fraser shows.  The Raytheon BST Core used standard tants (for example AVX TAJ series etc) in a few places where it would have been better to have used surge rated (for example AVX TPS).

From Frasers' pictures it looks like a one size fits all design approach with the biggest Tant they could find (330uF).

The 200 / 205D core for OEM use had 220uF for the actual sensor biases etc, and I don't recall any 330uF on the SECCA (processing PCB) either.  It was the SECCA ones that burnt up.

The Cadiallac PCB looks like both those OEM PCB combined, but without some of the OEM interfaces.

Bill

Offline Bill W

  • Frequent Contributor
  • **
  • Posts: 277
  • Country: gb
    • Fire TICS
Had a check, and the 'newer' SECCA from the 205D OEM core has one or more 330uF/10V/0.2 ESR on all the switch mode supply voltages - of which there are quite a few.
That dated from 2003.

The older '200' series from 1996 to 2002 used 220uF's and a lot less voltage rails pretty much all on 5V.

Bill


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf