FLIR Thermal Camera for teardown - my latest patient, the E2

[Fraser]

For all thermal camera fans out there, I have just aquirred a new patient in need of repair.
It is a poorly FLIR E2 that will need a tear down to determine what is ailing it 

This is just the introductory post and will be followed by a photographic record of its disassembly. I hope the journey will be interesting to the readership.

Some background on the camera.

The original FLIR Ex and Exx series were produced in the early 2000's and are not to be confused with the latest Ex and EXX cameras. How confusing of FLIR to use the same family name 

These cameras were intended to be a very portable, yet capable, handheld thermal camera that could be used like a kind of thermal 'torch'. Hence its shape. I tested one of these when they were first released and performance was pretty impressive at the time.
The resolution is 160x120 with a thermal sensitivity of 0.12C and 30 frames per second. The removable standard lens is 24 degrees, but other lenses may be fitted. The lens material is Germanium and it is large in diameter.

The menu structure is simple and all important functions are present. Namely, Auto and Manual modes with setting of centre temperature, span, emissivity, measurement areas etc.

Connectivity options are either PC or direct composite video output. Power comes from a Li-Ion 7.2V 1.8Ah removable battery that is located in the handle. These batteries are available new for £20 on ebay.

The 'patient' that I have just received is an E2 manufactured in 2002. It was being used by an electrician and this was the target market of the series i.e. preventative maintenance tasks. My camera has stood up to a long life in service very well, so build quality appears good.

My cameras fault appeared to be 'failure to boot' in the auction description. On receipt and initial testing it is clear that the camera is booting just fine with no errors shown. The FFC shutter is activating as expected and with normal regularity BUT no thermal image is appearing on the screen. Almost as if the lens cap was fitted. A key symptom is the fact that the spot temperature reading is around -4 degrees ! Something is very wrong there ! I have also noted what appears to be fluid ingress witness marks between the LCD panel and its plastic protector. The LCD is working fine though. Overzealous cleaning with screen cleaner maybe ? That may be nothing to do with the fault but we shall see.

I will be investigating the cause of these symptoms and will have to take the camera apart to do so. As such I thought I would start this topic so that I can add pictures and updates as I delve into this little camera. It is certainly worth the effort to repair it.

I hope this will be an interesting and enjoyable thread for the readership 

Pictures of the patient attached and the cameras brochure.

Aurora

[Fraser]

In case anyone is wondering why I have not torn into the casing as soon as the unit arrived....

I always follow good practice and check all functionality and symptoms before opening the case.

That is to say, check all menu options, look for obvious symptoms during these checks and then do a factory reset so that camera is running a known baseline configuration. Re test camera to see if symptoms have changed.

During the testing I noted the following:

1. Measured spot and area temperatures are always around -4C
2. Setting span to its widest range and bringing centre temperature up to include ambient and 36C resulted in no thermal image data when my hand was waved in front of the camera with  and without lens fitted. The camera is effectively blind.
3. The camera information menu showed serial numbers and information for all modules fitted EXCEPT the lens. From experience, cameras with removable lenses take data from the lens which includes FOV and serial number. That appears to be missing in this case. Germanium lens element Temperature data is also provided by the lens and this is important to the cameras internal correction algorithms. Lens contact gently cleaned but this issue remains.

I am suspicious of the very low -4C measurement but the camera is performing a FFC event without complaint. The absence of lens data is also of interest. I sjall be looking at the three main thermal data providers in the camera....namely, lens temperature, chassis temperatureand the microbolometer outputs. The microbolometer contains both image elements and die temperature monitors. The microbolometer in this camera is circa 2002 and so is likely a conventional (very expensive) hermetically sealed, gold plated, module. They are normally very robust so I do not expect microbolometer failure in this case.

These cameras contain built in diagnostics so I am surprised taht no error messages occur at boot.

I still need to carry out tests on all I/O ports to see what they are producing by way of image data but that will have to wait as duty calls elsewhere.

I shall investigate this as and when time and my energy levels permit   

For anyone interested, the auction I bought the camera from is here:

http://www.ebay.co.uk/itm/331542226167?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Not a great description and the pictures with flash make the screen look awful but it is, in fact, not scratched at all. Just liquid ingress witness marks 

Aurora

[Fraser]

I have had a most enjoyable afternoon disassembling and inspecting my new patient.

How many of you have bought a piece of equipment and had the nagging feeling that someone has been in it before you ? I had that feeling with this camera and I was correct.
Thankfully no harm has been done but it is obvious that someone was looking for the easy fix etc. More on that later.

This E2 came to me with the comment that it belonged to the sellers father who is an electrician by trade. He had apparently intended to repair it himself but lost interest long ago. Hence its was sold as part of a clear-out. ...... that pair of words send a shiver down my spine ...... my wife uses words like that when she see our garage contents 

Anyway, I opened the camera knowing that some form of liquid has got between the LCD and its protective window. I could also see strands of cotton wool on the inside of the screen so someone had to have been in her before me. The worst case scenario was that the E2 had gone for a swim at some point in time. That would have been bad news, very bad news 

With some trepidation I removed all of the screws that appeared to be associated with the case. This camera does not have any concealed screws (unlike the new FLIR TG165) 

The case separates into three parts with the display and keyboard panel lifting off of the top. One screw fixing point was damaged in the display/keyboard panel, possibly the result of a fall impact cracking the plastic around the knurled brass insert.

I was trained in forensic examination of electronics so enjoy analysing a 'crime scene' to ascertain what happened to a piece of equipment. This Camera is going to be subjected to a full inspection as I know there are fluid contamination issues that need to be addressed inside. The good news for the readership is that I have to strip the camera down to its component parts for inspection and lots of photos will ensue.

I continued with the disassembly of the case and at each stage of a PCB removal I inspected the PCB and surrounding area for signs of fluid ingress and residue. I found the following:

1. The rear of the LCD protection screen has been exposed to fluid of type unknown and then wiped with cotton wool. Poorly executed cleaning has left residue on the screen.
2. The LCD surround is tin plated steel and is showing evidence of corrosion.
3. The main PCB is fitted with an aluminium heat-sink and pink thermal transfer pads that cover two large IC's on the board. a layer of fluid was found between the thermal pads and the IC tops. This can sometimes be fluid coming from the pad as occurs with oil/silicone impregnated types. The fluid does not evaporate in warm temperatures so it is unlikely to be water. There is no fluid visible around the pins on the PCB below the heatsink.
4. The microbolometer diaphragm and chassis temperature monitoring housing has evidence of a sticky residue between it and the microbolometer. The shape of the residue is circular yet the microbolometer is a rectangular shape. I have no explanation for this disparity yet. 
6. The micobolometer Germanium window was contaminated with a dried on layer of unknown origin. The layer was easily removed using IPA. There remains some minor marking on the window but this may be fine line corrosion which can occur on Germanium parts. They are not serious enough to effect a 160x120 pixel thermal camera image though.
7. The microbolometer has pins arranged around its periphery. The microbolometer hermetic housing is steel with gold plated pins egressing via glass vacuum seals. On two sides of the microbolometer there is evidence of fluid damage in the form of corrosion of the steel adjacent to the pins (rust has formed). There would also appear to be rust contamination of the microbolometer inter pin spacing. This could seriously impair the data on the affected pins.
8. The display/keyboard case is fitted with a metal finger gasket. This gasket is showing clear signs of corrosion and finger damage due to attempted cleaning using a inappropriate methodology.
9. There is evidence of fluid accumulation around the lower screw mounting of the microbolometer housing where it fits into the lens mount and FFC shutter assembly
10. There is fluid contamination residue on the laser protective window which is located directly above the microbolometer. This causes some distortion and spreading of the laser beam.

Well that is the bad news...... now for some good bits.......

1. The case shows no obvious physical damage due to a drop impact (except the brass bush retainer inside the display section.
2. There is no evidence of fluid presence or accumulation in the lower case halves.
3. There is no evidence of corrosion on any PCB or component pins (except on the microbolometer itself)
4. There is no evidence of fluid ingress through the lens mount and the matt black coating inside is pristine. (fluid would 'watermark' this for sure)
5. There is no evidence of component leakage on the PCB's and no fluid residue visible on them.

Some readers may already be thinking this camera a lost cause ? Well the good news is that I have managed to get it to produce a thermal image, albeit a noisy one. The noise may be being caused by the contamination and rusting of the area around the microbolometer pins The contamination is of an unknown origin and I shall be asking the sellers father for any information that he might have on the incident that lead to such.

This is turning out to be an interesting little project. I have proven that the camera and microbolometer are still capable of producing a thermal image so it is worth the effort to restore the unit to its full capabilities. Nothing I have seen today makes me think this camera is beyond my help.

It Lives 

Pictures to follow

Aurora

[Fraser]

OK here we go....internal pictures time 

[Fraser]

More.....

[Fraser]

and more.......

[Fraser]

and more......

[tautech]

Very enjoyable commentary in reply #2. 

Good luck with your latest patient.

[Fraser]

As I previously stated I noted that the temperature measurement was totally out of wack and that the lens data was missing. The lens data symptom was a very important clue as I knew the lens passes data and temperature to the main board. Without this data a thermal camera will behave very strangely indeed. My first checks were focussed on the lens contacts, PCB and its associated mating pogo stick contacts on the camera. All was well with these. Some quick continuity tests on the ribbon cable that connected the lens contacts to the camera display board. ...... aha no continuity on several lines. The ribbon cable was not correctly seated in its snap lock ribbon socket. It looked OK , but wasn't. refitting the cable correctly cured the temperature reading issue and the lens data now appeared in the camera information page.

The thermal image appeared and responded to my hand in front of the lens and then to my cat sitting on the floor. Focus worked fine and the temperature readings seemed reasonable. The image had a lot of line noise and a poor flat field however so I knew that other issues are causing problems in the camera.

It is my opinion that this camera has suffered an 'event' that caused degradation of the image. Someone, likely the electrician owner, opened the camera  in the hope of finding an easy fix or to investigate how bad matters were inside. Having given up on the repair I believe that person reassembled the camera poorly. This lead to case seals being incorrectly positioned, ribbon cables being pinched and the incorrect insertion of the lens communications ribbon in its display PCB socket. The camera then failed to provide an image so the amateur repairer likely thought they had broken something, and gave up.

Then the camera comes to me for some TLC 

I still need to clean away all corrosion and wash the PCB's in IPA to ensure no contamination remains t cause future issues. The microbolometer will receive lots of TLC to sanitise its case and remove all rust from areas where they may cause issues. The rusted areas will then be carefully painted with a rust stabiliser to prevent any worsening of the corrosion. Once I am sure all is clean, I will retest the camera to see if the symptoms remain.

Note the OEM modifications in the pictures. They are very obvious but well executed. this is a first generation version of this camera as it is dated 2002, which is the year of first release. Hence the modifications.

Once I have the camera working correctly I will carefully set about replacing the lithium button cell whilst maintaining power to its terminals. I doubt it protects the calibration data but I would rather not find out that it does the hard way !

The camera will receive a full restoration of its case as well.

I have a brand new battery on its way to me as well so I will not need to rely upon the somewhat decrepit original batteries that have already been rebuilt once in their lives.

Aurora

[Fraser]

As some members know, I have a high resolution X-Ray capability if needed. In this case the patient did not need an X-Ray, but I have another 'helper' on hand if needed...... the CAT scanner, aka Oliver, or Ollie for short. Picture attached.
When not scanning what I am doing, he sleeps 

Aurora

[mikeselectricstuff]

I wonder why they have a TCXO in there

[Fraser]

@Mike,

I wondered about that as well. My higher spec PM series cameras use normal SMD crystal oscillator blocks and multi clock generators. Even the RTC is a normal 32kHz crystal.

A TCXO seems overkill 

I'll see if I can work out what it serves.

Aurora

[Fraser]

Well I have heard back from the seller. He has yet to speak with his father but it would appear that the camera suddenly stopped working one day. His father is not the original purchaser of the unit (no great surprise there) so it may have suffered 'abuse' in a prior life.

Sadly I was also advised that the poor camera has spent the past 2 years (since failure) in a damp workshop. This may explain the corrosion issues on ferrous parts.

I have re-inspected all of the PCB's this morning and can confirm that there is absolutely no fluid contamination on them or the component leads. No corrosion evident on the leads either. It is possible that I am witnessing condensation corrosion on the microbolometer. This is a lesson to all thermal camera owners....keep your cameras warm and dry when in storage.

I revisited the fluid that was found on the heat sinking pad and tops of the IC's. I can confirm that it is some sort of oil or silicone fluid that the thermal pad contains. Another smaller thermal pad in the lens mount is exhibiting the same sort of 'sweating'. 

It is looking more likely that this camera has suffered a drop event that dislodged parts within. The main large heat-sink has definitely moved since manufacture and is slightly bent. The broken brass insert retainer on the display casing also points to a drop event.

When this happened, I may never know but I suspect someone went inside to 'fix' it. If they did, they made a poor job of re-assembly evidenced by poor seal positioning, trapped flex cables and poor cleaning hygiene on the LCD display protector. It is quite plausible that the thermal sensors cable was not correctly fitted at this time of this 'fix' and eventually lost connectivity with the connector contacts, causing the 'no image' fault. 

I have removed the temperature sensor cable and it has more than one set of pin contact marks on it so it has been poorly positioned in its life. I will detail the thermal monitoring of the camera in a separate message. It uses three DS1624S temperature sensors in the optical path. Quite comprehensive monitoring but this was not a cheap camera when new.

Aurora

[Fraser]

Just remembered..... there is some very bad news regarding the microbolometer in this camera....... it has slight distortion above the knee on its pins down two sides. That is to say, the microbolometer capsule is suspended above the PCB on its pins alone and it is no longer parallel to the PCB. This is classic drop event damage and can be terminal for the microbolometer.

The event plays out as follows...... the camera is dropped from approx 1m height and lands on its rear end. The kinetic energy stored in the relatively heavy microbolometer capsule means it tries to continue moving in the direction of gravity. The thin gold plated supporting leads absorb the energy but in the process bend at the weakest point, namely the horizontal area above the knee. The result is a microbolometer capsule that is no longer parallel to the pcb or perpendicular to the optical path. The other potential areas of concern in such an event are damage to the glass vacuum seals caused by the torque exerted by the leads, and the potential for separation of the thermoelectric cooler from the microbolometer die. Both of these failures would render the microbolometer scrap as it is not DIY repairable 

When I had the camera running, the image was noisy and there was a definite thermal cool sot in the middle of the image. This does not bode well for the microbolometer. Time will tell whether this camera has a future.

Aurora

[Fraser]

Whilst sat looking at the patient considering the work I need to do on the microbolometer, I had a thought about this thread and where it is going. It will likely meander all over the place whilst I work on this camera, and that was my intention really.... a commentary on the investigation and hopefully repair of a piece of equipment that is of interest to many fellow techies these days. It is also a piece of equipment for which no official technical information may be distributed to the public (ITAR regulations).

I am considering staring another thread entitled "Anatomy of a thermal camera" and using this E2 as the subject of the thread. The alternative would be to use one of my many PM5xx series thermal cameras as the subject under investigation. I already have good knowledge of that design. It is somewhat different to the more recent 'pistol grip' designs though. Such a thread may help others to better understand the technology and what lives inside the case of these clever bits of kit. I would have to stick to high level descriptions of the design and no schematics as any deep technical detail (even that learnt from reverse engineering) will place me in breach of ITAR.  I could find myself in hot water if some official in the USA takes issue with what I have written.

Anyway, its just a thought and may come to nothing. At least if the camera is scrap due to microbolometer damage, it will have served a useful purpose. We will see how the repair progresses in this thread.

Expect me to flit around the camera in this thread as I am popping in and out of this repair as and when time and energy permits. You are witnessing my 'quick and dirty' approach to a repair at the moment ! Not my finest work but hey, needs must and I am having a ball 

And for anyone wondering.... yes I am doing the teardown and analysis work on a lap tray in our conservatory and not in my garden based lab. Why ? Well M.E limits what I can do and I can pop in and out of this little project easily if it is sat near me in my nice warm and bright conservatory. Not all work has to happen in a 'shed'  Thankfully my wife is being very tolerant of my 'playing' at the moment.

Aurora

[bookaboo]

I do enjoy a good troubleshooting thread 

[Fraser]

Good,

I'll try to keep it interesting and fun 

Aurora

[Fraser]

OK, those who have been reading this thread up to this point will be aware that the FLIR E2 camera was not producing a thermal image but otherwise appeared to be functioning normally. Important symptoms were the incorrect spot temperature of approx -4C and the lack of any lens information in the camera information page.

For those unaware, FLIR cameras normally have an information page in the camera menu that displays the serial and revision number of key elements within the camera, including the lens. The lens data packet also identifies the lens type to the camera for calibration purposes.

The fact that the lens data was missing immediately told me that camera was not seeing the lens. I am also aware from previous work on FLIR professional cameras that they install a thermal sensor inside the removable lenses to provide Germanium element temperature as this is required for temperature compensation purposes. Germanium lens transmission decreases as its physical temperature increases.

The removable lens communicates with the camera body via four contacts that mate with sprung loaded 'pogo stick' contact that protrude from the lens mount. Contamination can prevent correct contact performance and so cleaning may sometimes be necessary. This was not a solution for my E2.

Upon opening the E2 for the inspection stage I focussed on the signal path between the lens mount contacts and the PCB within. A quick continuity check on the 4 conductors proved that three were not connected to the PCB socket contacts. Further inspection showed that the ribbon cable was not correctly seated in this socket and so had been the cause of the lens data loss.

It does not end there however..... The 4 wire communications ribbon detailed above also carries thermal sensor data from three sensors in the optical path. DS1624S thermal sensors with built in flash memory are employed to monitor the temperature as follows:

DS1624S IC from front back ->

1st = Temperature and identity of lens
2nd = Temperature of the FFC shutter
3rd = Temperature of the Microbolometer capsule and front Germanium window.

All three sets of temperature data were therefore absent when the camera was in fault. I am surprised that the camera did not display an error. Instead it tried to operate without a temperature reference from the optical path and did not produce an image. The lack of the required data appears to have totally befudled the image production process. Not that surprising.

Pictures of the temperature sensors attached.

The flex pcb fits either side of the diaphragm and the curved FR4 pcb is attached to the lens rear.

Aurora

[Fraser]

I'm back again 

Next steps on the repair......

1. Carefully clean the microbolometer to remove all loose ferrous material from capsule and between the pins.
2. Attempt to lift the microbolometer capsule away from its PCB to straighten the distorted pins and correct the error introduced in the face angle.
3. Straighten the bent aluminium heat-sink plate.
4. Temporarily reassemble camera and check operation.
5. Disassemble camera and carefully clean all parts and PCB's to remove any accumulated film on their surface.
6. If camera is still displaying a noisy image, continue investigation 

The noise in the image that I have referred to is not pixel noise but multiple random horizontal lines. These may be caused by the corrosion around the microbolometer pins or there could be a microbolometer related noisy DC-DC converter in the unit. We shall see.

Aurora

[Fraser]

Aaaaah! major brain fade 

The distorted pins on the microbolometer will NOT cause an error in the face angle of the front window relative to the optical path. Why? because the face of the microbolometer is the mounting point to the diaphragm and lens mounts. The micobolometer PCB floats on the leads.

This changes the drop event scenario. If such an event occurred, then the bent microbolometer pins may have been caused by either the kinetic energy of the PCB or by something impacting the PCB from the rear (another module) The impact point would also have been the opposite to that previously considered. Namely towards the front of the unit.

Of course the microbolometer may have had distorted pins since manufacture  It will be interesting to see if any other evidence of a drop comes to light.

Oh dear, time to take a rest, the old brain gets a bit befudled sometimes 

Aurora

[Fraser]

Pictures of the microbolometer. Note the two diagonally opposed mounting points on the front.
Take a closer look and you will see the gap between the PCB and rear of the capsule is not constant. You may also see the bent leads above the knee, though it is not that clear in these pictures.

Aurora

[SeanB]

So it was dropped, and of course hit lens first. Wonder if that would have popped the microbolometer loose on the cooler, and as the front cover is opaque it would be difficult to see unless you use another high resolution camera as a close up to see inside.

[Fraser]

Looks like a potential drop casualty. I have worked on several thermal cameras that have been dropped. A common hazard in industry. Always use a wrist or neck lanyard ! It can get enxpensive if you don't and a drop happens.

I have a superb 'top of the line' NEC AVIO thermal camera that was dropped on its lens. The result was a wrecked lens filter riing (thankfully a replaceable part), a slightly bent lens chassis (aluminium so repairable) and the battery carrier chassis (aluminium) driving into the rear of the microbolometer PCB.
I have got that camera working fine now but just have not had the time to sort out the bent metalwork yet. It will be a lovely camera when re-assembled though. It cost me £700 as a project and they sell used for around £8000 - £10.000  Well worth the effort and why I am taking my time repairing it properly.

Back to the E2. Its weight is at the front so it would naturally fall on its lens. However I can find no evidence of impact on the lens or lens mount    The microbolometer may have suffered damage, I will have to see once I have the unit back together for further testing. Brains a bit too tired to do that right now so watch this space   

I maybe able to X-Ray the microbolometer depending upon the thickness of the steel capsule walls. If the microbolometer die had moved forward away from the thermoelectric cooler I would have expected damage to it but my knowledge of the internal construction is limited to old HP documents and not this type of FLIR staring array.

Aurora

[SeanB]

Carpeted floor will show no damage, but will still have enough force applied to the innards.

[Fraser]

For anyone interested in how I will be cleaning the microbolometer capsule area that sits between and behind the leads ....here is my secret ........

TePe interdental brushes and a syringe full of IPA to flush away detritus 

I'll be using the special ESD safe TePe brushes designed for static sensitive teeth of course 

TePe brushes come in many sizes from tiny to humungous ! Very useful to have in the workshop. They cost around £5 for a pack of 10. Available at all good grocery stores in the UK and from your dentist 

I also use other dental tools, piks etc as they are perfect for SMT working 

Aurora