The story of another thermal camera purchase by Fraser
If you are sitting comfortably, I shall begin
As many readers will already know, I collect and repair thermal cameras as my hobby.
On many occasions I have decided to stop buying these wonderful pieces of technology, but my will is weak in this area. Inevitably an unusual thermal camera will stray into my field of view, at a price that just begs my attention.....and my resolve will buckle under the strain !
Such an occurrence recently resulted in my becoming the proud owner of another unusual thermal camera. Before I go on, I have decided to not identify the camera or its manufacturer at this time. Such details are not essential to the story yet could cause some issues as you may see as the story progresses.
So what is unusual about this thermal camera ? Well its specs are nothing earth shattering at 384 x 288 pixels producing a 60fps output to the user. The NETD is specified as <70mK but in high sensitivity mode this drops to <35mk. None too shabby really. This improvement is made possible by carefully set bias voltages and the use of a thermo-electrically temperature stabilised microbolometer array. The camera outputs are many. It has Firewire, RS232, Composite Video (NTSC or PAL), S-Video (Y & C) and VGA. Plenty to be meet the varied needs of a user.
This camera has no control panel, no LCD display and no viewfinder...... making it not very user friendly eh ? Well it was never intended to operate in a stand-alone mode. It is a specialist industrial thermal camera module that is intended to be incorporated into a larger machine. It is not just a 'Core' as it is a complete thermal camera with a full gambit of on board capabilities. It just lacks the normal human interfaces. The camera is controlled from a computer that sets up the modes of operation and the image is either streamed to the computer or displayed on a suitable video monitor. Control may be via Firewire or RS232 and the user interface is a bespoke PC program. It is possible to command the camera over RS232 using a simple terminal program but this has its complications as I shall shortly detail.
The camera actually appears very boring from the outside. It is just a 'box camera' and, to those unfamiliar with thermal camera optics, it looks very much like the usual heavy duty industrial visible light cameras that are so common in industry. It build shouts 'ruggedized' and such cameras are often very expensive. The seller knew this so its price was set quite high. In truth the seller likely knew nothing of the camera, beyond, "it looks shiny and expensive" !
In communications with the seller I established that the camera was purchased at general auction without any other accessories or paperwork. I received the serial number and sent that to the manufacturer to check that it was legitimate. It was and it was originally sold in 2014 ! It was virtually new compared to most used industrial thermal cameras sold. That may be explained later though, so hang in there
The seller could not test the camera or state that it was working. This is one of those occasions when you have to decide whether the risk is worth taking. My thought process was as follows:
Pro's:
1. Very nice cosmetic condition industrial thermal camera
2. Decent resolution and frame rate
3. Lemo connectors shouted quality
4. It is fitted with a large 50mm close up lens !
5. It is an auto and manual focus camera
6. The camera has an SD card recording option
7. The camera is made by a reputable company and I found data on it.
8. The price might be negotiable
9. The camera is only around 2 years old
10. It is unusual, so I like it
Con's
1. It is an industrial modular camera, so not totally self contained.
2. The camera is in unknown, untested condition and may be faulty or degraded.
3. No leads or software come with the camera. Both are expensive from the OEM
4. The camera needs bespoke software to control it as the command set is not public
5. The OEM is unlikely to provide much assistance
6. The camera is not cheap if it turns out to be a paperweight !
After some thought and communications with the seller of the camera, I decided to risk a sum of money with which I was comfortable. This is essential when considering the purchase of an unknown condition thermal camera. They can end up as expensive paperweights. I knew the lens would be useful even if the camera was scrap. I offered the seller £200 and he happily accepted. To him, this was a total unknown and he was happy to get some decent money for it.
The camera arrived in due course and I quickly inspected it for signs of abuse. The Lemo sockets were all in good order, the Aluminium case had some bumps on the corners and one mounting lug on the lens was badly bent. Hmmm it looked like the camera had suffered impacts to both its case and the lens. There were no dents in the lens barrel but there was a witness mark from a blunt metallic impact on the side with the bent lug. These are Industrial cameras so such damage is not that worrying as they are tough and designed to take a beating.
I had already procured the required LEMO connectors ..... total cost £45 for the two from Germany, and that was the cheapest source ! that is a consideration when buying a camera without cables.... if LEMO or Fischer connectors are used, they can be very expensive. It is sometimes the case that it is better to change the connector on the camera to LEMO from Fischer as the Fischer types are both more expensive and harder to source in the UK.
Now to the challenging part. I had been in communications with the OEM and had provided the serial number so that they could check their records. I was up front about buying the camera used and not being the original purchaser. They seemed happy to talk to me, which was a bonus, but it was not all good news. I had asked them if they were willing to supply me with the pin-out of the cameras connectors, any documentation on the command set and a user manual for my model. The response was polite but disappointing. The pin-out was part of an extended documentation set that was chargeable. The documentation set also includes the command set for the camera. Neither could be supplied to me for free. The software was also a chargeable item. It looked like I had hit the proverbial brick wall with the OEM as everything was chargeable, and at high prices. We are talking thousands of Pounds here and not pennies
This is the risk you take when buying specialist thermal cameras or equipment that requires bespoke software, or a command set that is not in the public domain. It was looking as if I had myself a £200 paperweight.
I am not one to give up so easily though. I asked the OEM if they would be willing to send me just the user manual. This they were happy to do and I received a 20MB PDF by return. The user manual made interesting reading but sadly no remote control information is provided, except to refer the reader to the 'Interface and control' document that may be purchased from the agent or OEM.
I knew I could reverse engineer the cameras interfaces so was not concerned about not having the connector pin-outs. What I needed was either the command set or the software. Both seemed out of reach but if there is one thing I know about the internet, it has all manner of information and software squirreled away in dark corners out of the normal public gaze. I am experienced in searching out such assets on legitimate servers and not hacking sites. A long and comprehensive search brought fruit in the form of an official Agent who was happily providing a download of the software that I was searching for. It is the original software from when the camera was released. This has been replaced with a brand new software package so maybe the old version was made available for download as a legacy support decision ? The OEM was offering me the latest software and made no mention for the earlier software so maybe it is considered obsolete and no longer offered by them ? I downloaded the software from the dealers support page and tried to install it on my Windows 7 laptop It installed without the need for any keys or passwords. It runs perfectly and has no signs of being a demo, or limited use version. Happy days
So I now had the essential software to configure and use the camera. I could intercept the RS232 communications between the PC and camera now and so reverse engineer the command set for the camera
So where are we at this point then.....
I have a camera that remains untested
I have the required LEMO connectors
I have the required OEM software
I have the user manual for the camera and software
I do not have proof that the camera is in an operational state or 'safe' to apply power
I do not know if damage or degradation exists within the camera or microbolometer
I do not know the pin-out of the two LEMO connectors
Next obvious step ..... open her up
And that is what I did
Opening the cameras case was very straight forwards with just two hidden screws. One under a rubber foot, and one under the ID label. Pretty common stuff really.
With one side of the metal case removed, I could inspect the internal parts for evidence of damage. One thing you learn when doing forensic investigations is to use your eyes first and not your hands. I visually inspected the camera and notes some interesting facts about its interior.
1. We already know that there are impact witness marks on the case and lens barrel
2. We also know that one lens retaining lug is bent as the lens was forced sideways
3. There were small flakes of matt black paint floating around inside the cameras case
4. A microprocessor daughter board was only partially retained in its edge connector (similar to a SIMM connector). One end was completely out of the socket. If I had applied power to the camera, all manner of issues could have been caused if the edge connector was shorted by miss-aligned connections.
5. There was no sign of where the mat black paint was being used in the cameras chassis or optical path.
6. The camera looked in good physical condition with no evidence of previous entry or distortion of the mechanical elements.
I know the camera had suffered some form of impact to its lens and case but thankfully it seems to have survived.
I made another discovery whist inspecting the camera. There is a small micro switch positioned behind one of the SD Card cover retaining screws. Both the screws are too short to operate the micro switch. The SD Card recording capability is a chargeable option on this camera. There are two possibilities for the micro switch that come to mind.
1. The SD Card option is enabled by fitting longer screws to the SD card cover ! It would not be the first time that equipment 'upgrades' and options were a very simple physical change like a jumper link or push button sequence etc.
2. The micro switch could actually be a 'hard reset' button concealed behind the sealed SD Card Cover. This protects the camera from accidental reset and dirt ingress if a hole was provide for reset button access. Saying that, such a hole could easily be sealed with a blind grommet. Maybe the designer though this a more elegant solution?
At this point in time I do not know the function of the micro switch. It is connected to an I/O board expander that talks I2C. when I get the camera running I can quickly work out its function though
I was able to quickly and easily dismantle the cameras internal parts down to module and PCB level. Each was then visually inspected for damage or interesting elements.
It was immediately obvious that the cameras main board comes from one of the OEM's camcorder style industrial thermography cameras. In fact I think I know which model. There are several unused ribbon connectors and headers present. The same is true on the microbolometer PCB. The camera does have a hardware option that is not present on my camera but this does not explain all the unused connectors. The camcorder style camera would have the keyboard and viewfinder connected to this PCB.
Upon examination of the main PCB, the architecture of the camera became pretty obvious. No surprises here at all. Thankfully the OEM used COTS IC's and the datasheets are freely available to download. The date codes on the IC's suggest that the camera was built in 2009 which is earlier than I expected for a camera sold in 2014. Maybe they had a lot of stock in the warehouse ? I will likely add pictures of the PCB later but it is nothing unusual or very exciting really. The use of a daughter board for a microprocessor proved a vulnerability as the sprung retainers were unable to keep a grip on the PCB during the impact event. Judicial use of RTV silicone might have addressed this vulnerability, as it does on rugged fire fighting camera internal header connectors.
Another daughter board provides the required major supply rails for the camera. It is a very neat design and easily reverse engineered. It uses a pair of LM5642 IC's to take the input supply voltage and generate 5V0, 5v2, 2V5, and 3V3. No negative rails are present at this boards output. The voltages are annotated on the PCB which is kind of the designer. How often I have wished other designers would do the same. It makes it so much easier to check for correct supply rails.
The 'optical block' with associated microbolometer PCB is interesting in that the whole microbolometer PCB, and its associated chassis, is moved forwards and backwards on linear ball bearing runners that fasten the microbolometer chassis to the optical block chassis. Two stepper actuators are present. One drives a circular plate that contains three different apertures and the other drives the microbolometer chassis back and forth. There is a challenge to the designer when the microbolometer moves. The microbolometer in this camera needs a heat-sink on its rear against which the Peltier temperature stabiliser may operate. How do you attach the microbolometer to the case such that it acts as the heat-sink ? Well this cameras designer has used an impressive form of flexible 'heat pipe'. The microbolometer has a special 'retainer' attached to its rear. This retainer holds four very large copper braided cables. These braided cables are then attached to the case via a bracket. The flexed braided cables permit the microbolometer chassis/sled to move whilst maintaining the required thermal path to the metal case. A neat solution to a challenging problem. FLIR take a different approach in the PM series cameras. They fix the microbolometer to a heat-sink that is in turn attached solidly to the metal case. The focus is accomplished by moving the middle lens in its optical block. The PM series do not have a removable lens however so they do not need to move the microbolometer. Professional thermal cameras are often a combination of high performance electronics and clever mechanical engineering. This camera is no different.
Whilst investigating the optical block I found more flakes of matt black paint. Even the microbolometer had matt black paint specs on its window. It was time to track down the source of the paint flakes.
Where is matt black paint commonly found in a thermal camera ?
Matt black paint is often an excellent black body surface and has excellent emissivity properties. It is anti reflective in a thermal camera optical path. The inside of lens barrels, areas around the microbolometer and the FFC flag are common areas where you will find such paint.
I had inspected the lens and could find no source of the paint flakes. Two areas remained tat I needed to access, the area around the microbolometer and the rotating aperture plate. This was not going to be easy however as I established that I would need to completely dismantle the optical block chassis. Not a job I take lightly as all screws are locked, and for good reason. Dismantling the chassis presents unknown risks such as alignment loss as I have no service manual to help me. You have to be very careful and methodical when undertaking such a task. I would not recommend such to the beginner and in all cases I take lots of pictures so that I can retrace the disassembly steps and see 'what goes where' !
I disassembled the optical block and inspected the parts. I could see where the matt black paint flakes had originated before spalling away and into the cameras interior.
There is a small insignificant metal plate mounted adjacent to the microbolometer. It is likely present to prevent some imaging artefact due to internal reflections or off axis energy from the lens. It may be small but the fact that it was shedding paint flakes in to the cameras optical block was very significant and very bad for the cameras imaging health.
Why is the paint flaking off the small metal plate ?
I cannot be certain as I do not have enough evidence, but I suspect that the impact event that the camera suffered put a shock wave through the optical block via the lens mount, and through the rotating aperture plate shaft and rear bearing. The plate rests on the rear bearing and this flexed the thin metal plate. The paint on it is very brittle so spalled easily. There is no evidence of a bent chassis. the camera received impact energy from two directions. To the side of the lens barrel, and the rear of the case. There may have been quite a lot of energy passing through the camera and its internal parts.
To repair the paint damage to the small plate, I removed all loose paint, roughened the surface with very fine 'wet & dry' paper before cleaning all dust off and re-spraying with an excellent optical black paint that was recommended to me by an industry insider.
The plate is currently drying in the conservatory and I shall await the hardening of the paint before re-fitting it. Sadly I cannot re-assemble the optical block until that small plate is in place. It is buried deep within the optical block
All work has stopped as a result.
The camera case and all component parts will be cleaned with compressed air to remove all traces of the paint that was floating around inside the camera.
Well that is the story so far. I shall re-assemble the camera and then see what happens when power is applied. An RS232 link will be used initially to talk to the camera as it has less chance of issues than using Firewire. Once I am sure that the camera is working correctly, I will try accessing it via Firewire and video streaming etc.
I hope this little story has been of some interest. I am not detailing the camera model or OEM as I may need more assistance from them, and they may not appreciate me showing a complete tear down of their product. Maybe later
Watch this space
Fraser