Today I took delivery of a very nice FLIR thermal camera that was designed for both Industrial and R&D applications.
The camera came to me as a "spares or repair" project so I was expecting it to not work when tested. I was actually pleasantly surprised by how much of it was working when I connected it to my laptop using Firewire and accessed the built-in web server. The web server permits camera configuration and provides camera specific data. I then used FLIR Researcher Version 2.8 to obtain an image from the camera. This is where a fault revealed itself. The produced image looked very odd indeed and something was clearly wrong with the FFC system. I tested the composite video output and found no video present. This was a previously declared fault so no surprise. I already had a good idea why no video was present, but more on that later.
So going back to the strange imagery that the camera was producing when sending its data to FLIR Researcher...... there was a decent enough initial image, but it was when the FFC event occurred that the weird stuff happened. Whatever was in the field of view disappeared and did not return after the FFC event ! If the camera was moved, the scene reappeared with a ghost adjacent to it. I was using a hot cup of water as a test target so the effect was obvious. Those who repair thermal cameras will know exactly what is happening in this case. The FFC event is occurring and the FFC flag solenoid is operating, but the flag is not sitting in front of the microbolometer. The camera takes the thermal scene as its flat field reference.... hence why it disappears from view. If the camera is moved, the crazy out of wack flat field offsets appear as ghost images in the scene. Once another FFC event occurs, the ghost images disappear, but so does the scene, as once again the scene is taken as the flat field reference. This is classic FFC flag failure territory but I did not know the actual cause yet. I removed the lens to check the FFC flag presence and operation. I found the lollipop shaped FFC flag "swinging in the wind". It was clearly no longer rigidly attached to the FFC solenoid. Such a failure is thankfully a lot simpler to diagnose and repair than something involving the FFC solenoid driver or the solenoid itself. Phew, an easy repair was on the cards. Unfortunately, I also knew that I would have to completely dismantle the camera in order to gain access to the ffC flag and its mount on the solenoid
Not a big deal though as I still needed to investigate the failure of the composite video output and that would involve removal of the PCB stack anyway.
I carried out a complete teardown of the A20 in order to expose the problem with the FFC flag mount. I discovered that the one small screw that secures the FFC flag to the FFC solenoid was very loose and had allowed the FFC flag to uncouple from a locating pin that kept it in the correct position. I found no evidence of any thread locking fluid on that screw. The remedy was to apply thread-lock to the small screws threads, reassemble the FFC flag onto the solenoid and then apply a small spot of laquer to the screw head for good measure. It will not be shaking loose again any time soon. this failure surprised me as the camera has very low hours on it and so a very low FFC event count. The internal web server provides the cameras operation statistics....
Total Runtime: 9 days : 20 hours : 20minutes and 42 seconds.
Cold Starts: 180
FFC Shutter Operations: 3323
This camera has seen very little use in its life ! I can only assume that the FFC flag securing screw was not correctly tightened at the time of manufacture or thread lock was forgotten (other screws have thread lock applied).
The cameras front section was reassembled and the investigation moved onto the absence of video at the composite video output BNC socket. In these cameras the composite video output comes from an ADV7173 video encoder IC that feeds the signal to the BNC socket via nothing more than a small filter comprising a series inductor and capacitor to ground. There is also a 75 Ohm resistor from the ADV7173 output pin to ground. The signal is DC coupled and this is an important point to note. Any excessive voltage applied to the video output socket of the camera is passed straight to the ADV7173 DAC-A output pin. If the voltage and current are high enough, the ADV7173 suffers damage to its DAC-A output
This is not as uncommon an event as you might think !
Upon visual inspection I immediately recognised the sad evidence of an excessive voltage and current having been applied to the video output BNC connector. .... there was a neat hole blown in the top of the 75 Ohm 0603 SMD resistor. Now this resistor sits across the DAC-A output and ground so whatever killed the resistor also got to the ADV7173 and caused damage there as well. The resistor would have needed a decent voltage to blow a hole in it so this poor A20 camera certainly suffered an unpleasant event on its video output. What that event was, we will never know, but it was not simple ESD, that is for sure. The only solution to this damage is to fit a new ADV7173 and 75 Ohm output resistor. The ADV7173 can still be purchased, but it is not inexpensive. Prices range from £14 to over £40 ! I was about to order some from Mouser at around £14 each but then found some old stock in the UK at Silicon-Ark for £4 each
Once they arrive, I will fit the new parts, but in the mean time I could reassemble the camera for testing of the FFC flag function.
Once reassembled I used Researcher to display the image of a hot coffee cup and was pleased to see that the camera now performed normally with no weird FFC ghosts in the system !
I will attach the initial images captured when the FFC system was faulty and an image of the cup after the repair.
For those unaware of the FLIR A20 and A40 series of cameras, these are very versatile units that were intended to provide good quality thermography for use in Industry, monitoring equipment, or for use in an R&D lab. The A series provide not only fully Radiometric thermal imaging, but also various analogue and digital I/O functions. The camera can monitor an external temperature LM35 sensor whilst also providing various alarm outputs if the thermal scene suffers a temperature excursion. The cameras can effectively be a preventative maintenance and failure alerting system in Industrial environs. In the R&D lab the cameras may be mounted on a table stand and used with FLIR Researcher software to carry out serious thermography on a target object or scene. Coupled to the PC and Researcher software they become powerful tools for those needing accurate thermal data.
OK, enough of my waffle, time to go to the pictures.......
Enjoy
Fraser