Thermal camera + CO2 laser?

[ir.ukrm]

Hello, please tell me who has experience of damage to thermal imagers with CO2 laser?

[Fraser]

This document may help

https://www.dataray.com/assets/pdf/wincam-fir_manual_supplement_1011a.pdf

It is definitely possible to damage a Microbolometer with the concentrated energy of a CO2 laser. ND filters can be used but I believe the reflectance type is required as adsorption filters will heat up too much. Much depends upon the power of the CO2 laser being imaged. Some scenarios use a Microbolometer with a lens system but others directly illuminate the Microbolometer window.

Hope this helps a little.

Fraser

[Ultrapurple]

Useful info - thanks. I have often speculated whether I could use a LWIR camera to see a CO₂ laser beam in a similar way that we can see eg a red laser pointer beam path thanks to scattering from dust particles etc. But I've never really thought about using a camera for beam profiling, leastways not in terms of focusing an image onto a sensor array. Instead, I suspect I'd use a high(ish) thermal conductivity target and image the temperature change on that, thereby reducing the energy incident on the sensor by many orders of magnitude. It'd still be 'bright', though.

The quoted damage threshold for the sensor in the DataRay camera is stated as "few hundred mW/cm²". The saturation level is 50mW/cm² and elsewhere it states that damage occurs at about 20 times that level, ie 1W/cm², though personally I'd use the lower limit as a guide. Let's say 250mW/cm² for the sake of argument.

Assuming a 40W CO₂ laser from a desktop cutter and a perfect flat-top beam profile (which it won't be but I'm only estimating here), 40W/(250mWcm²)=160cm², or a beam about 15mm or so diameter. In the real world of course the beam wouldn't be flat-top and I wouldn't want to put anything less than about 25mm diameter anywhere near anything electronic I cared about. In order to make useful measurements (power <50mW/cm²) the beam diameter would need to be nearer 50 to 75mm diameter, still not allowing for its irregular profile.

[ir.ukrm]

I had such a sad but interesting experience yesterday. Two groups of pixels were damaged. It was two beams from two lasers.

[Ultrapurple]

Sorry to see the damage but it's good that the pixels are near the edge if the sensor, rather than right in the middle.

It should be possible to mark them as dead pixels and have the camera 'work around' them, filling I  with a guesstimated value in lieu of real data.

I have no idea whether your camera as an accessible dead pixel marking system but even if it's a return to the manufacturer job it's still going to be much cheaper than replacing the whole shebang.

Damage near the centre of the sensor is much more important because that's usually where you want best image quality. The outer 10% or so of pixels are almost irrelevant in most setups (though more pixels are always nice to have!)

[Fraser]

Another issue with unmarked damaged pixels is that they can really mess up the auto ranging function on some cameras as they can look like very hot or very cold pixels to the ranging algorithm. Even if visually tolerable they really need to be placed in the dead pixel map unless only manual ranging is used.

Fraser

[ir.ukrm]

40 pixels  were damaged. The temperature on these pixels showed degrees by 5-7 more than on the wrong damaged. They were able to hide.

[Fraser]

You may want to research “Microbolometer pixel annealing” as I have read that annealing can recover pixel performance in mild cases of laser ‘damage’. I am not familiar with the process for microbolometers but read of such annealing being used on the CCD imaging FPA’s aboard the Hubble Space Telescope to improve radiation damaged pixel performance. The HST pixels normally operate at approximately -80C but were annealed at +30C for 4 hours. No idea if a similar technique is used on Microbolomters. i.e. exposing the pixels to a ‘flat’ high temperature scene (Such as a heated metal plate with good emissivity) for several hours to heat the pixels evenly across the FPA. Another possible approach may be to remove the Microbolometer FPA and heat it in an oven for several hours ? It may be worth your time to research this and I would be interested in your experience of the process if you decide to try it. Please be careful though, you could potentially make a bad situation worse if you ‘cook’ the FPA at too high a temperature and alter the thermal response of all pixels, invalidating the calibration tables.

The fact that your ‘damaged’ pixels read only 5C higher than those surrounding them appears a positive situation to me. At least the pixels were not destroyed.

Fraser

[Cat]

Highly interesting topic because the same question is bugging me for some time.
The range of stuff I read (unfortunately I can't find the source anymore) was from "Even diffuse backscatter will cut the µBolometer" to "nothing will happen if you don't aim the beam directly in the sensor".
I want to know if the beam of a CO2-Laser is visible with a thermal imager (because of dust like with visible lasers) but I did not dare to power up a 80W tube and look at it with the thermal imager.

I had such a sad but interesting experience yesterday. Two groups of pixels were damaged. It was two beams from two lasers.

Was this from the backscatter of an unmoving beam? What was the target the beam irradiated?

The only thing I could find about using a CO2-Laser and a thermal imager at the same time is this: http://www.rtftechnologies.org/physics/fusor-mark3-microbolometer.htm
Unfortunately only the diverged or expanded backscattered beam was observed. The warning label says 30W max, so the laser is probably in the 10-20W range.

FLIR has some values for the power at the sensor (VOx) in their FAQ (https://flir.custhelp.com/app/answers/detail/a_id/727):
The 100 mW/cm2 continuous input power is without any risk.
You should not exceed 2 W/cm2 continuous input power.

With pulsed lasers in the visible range it's possible to damage a sensor only with backscatter (peak power in the MW-Range, https://www.ilda.com/camera-sensor-damage.htm).

[Vipitis]

As sensor burn in can be healed by just leaving the camera running arcoss very dynamic scene, like moving it for half an hour... it might be worth a shot. Baking your electronics can damage other parts such as BGA soldering etc, be cautious when doing something extreme. Best of luck to recover the sensor or announce those pixels dead and just have them interpolated.

[Max Planck]

An article that will be probably interesting in the context of this topic.

In-band low-power laser dazzle and pixel damage of an uncooled LWIR thermal imager

https://www.researchgate.net/publication/328180319_In-band_low-power_laser_dazzle_and_pixel_damage_of_an_uncooled_LWIR_thermal_imager

Max

[Fraser]

The issue with laser damage, as I understand it anyway, is localised heating. The laser beam can apply a significant amount of energy to a very small area of the Microbolometer die and this can cause very high temperatures in individual pixels. At the extreme end of the scale, the highly efficient heat receptive pixel melts or distorts destroying it. In less extreme cases the pixel heats up to a point where the the pixel is actually changed. The change in the crystal lattice structure of the pixel can change its thermal energy response and so the NUC table can become invalid for that pixel, or pixels. A fresh NUC calibration may be able to compensate for the change in a pixel or the thermal response may fall outside acceptable tolerances and the pixel is identified to the system as Bad and added to the Bad pixel map.

Unless the damage to the Microbolometer pixel is very severe, causing other issues in the Microbolometer like severe outgassing or substrate damage, it is likely that the camera can cope with the loss of the pixels and undergo a new NUC calibration to counter their presence.

Does the effected camera permit user NUC and Dead Pixel table creation ?

Fraser

[Bill W]

CO2 laser profiling was a standard use for the older EEV Pevicon cameras.

As for the dead pixels, if only a few degrees C out that should be fixed by the normal shutter operation. That kind of damage will also usually clear up in a few days running.
As it is not, is this a gain change ?  that is they will look colder on a cold scene and hotter on a hot scene ?

Bill

[ir.ukrm]

Thermal imager Flir T420. Damaged pixels were killed using custom settings. After going through the procedure of killing pixels, the dead pixels were still faintly visible on the screen. The next day, there was no noticeable one damaged pixel. I think the pixels may have been self-stopped.

[Fraser]

Damaging a T420 would make me very sad. I am pleased that the pixels have now been Captured in the NUC and dealt with 

Fraser

[frenky]

Can someone show thermal image of CO2 laser beam?
I plan to make a DIY CO2 laser cutter soon and I  wonder if I could see beam with thermal imager?

[Fraser]

Frenky,

A laser beam is really only actually ‘seen’ when it impacts upon a surface or particles. Hence the use of smoke to reveal visible light laser beams. The light of a visible light laser is normally only visible at the laser exit port and on the surface on which it lands. It is no different with a CO2 laser. The beam itself generates no significant heat or radiated energy whilst passing through the relatively low density atmosphere. If it were passed through a dense particle rich atmosphere some of the beams energy might be visible but remember that a Microbolometer thermal camera is detecting thermal energy acting upon the pixels. A better detection sensor is one that detects photons such as those used in LWIR cryogenic cooled cameras like the old AGEMA 880 LW.  Once the CO2 laser energy impacts upon a non reflective surface, the energy is absorbed by the surface and becomes visible to thermal imaging in the LWIR band. With a Microbolometer type sensor array you are seeing the effect of the photons and associated energy (emitted heat), rather than the photons themselves. Hence why a LWIR Photon detection type thermal imaging system is sometimes a better choice for CO2 laser ‘in flight’ imaging with particle filled smoke as a means to see the beam. If relying upon the heating effect if the energy beam, the surface on which it impacts needs to be of low thermal mass and suitably receptive to the energy wavelength. The heated area will not necessarily show fine detail of the laser beams quality as such detail is easily lost through thermal transmission within the target. Direct ‘on axis’ beam quality imaging is far more effective as the pixels respond quickly to changes in the energy illuminating them and provide fine detail of the beam shape and the energy distribution within it.

As others have stated, the energy emitted from a CO2 cutting laser is to be respected when imaging it is considered. It is all too easy to damage a thermal cameras sensor array, be it Microbolometer or photon based, if too much direct or indirect collimated energy hits the pixels. The risk is highest when working ‘on axis’ with the laser beam directly illuminating the sensor array. ND filters are normally used to help protect the sensor pixels from damage. Even viewing a CO2 cutting laser beam ‘off axis’ carries risk as reflected beams of energy can still damage a sensor array, especially photon based sensors.

Fraser

[frenky]

Thank you for great explanation. 

My simple way of thinking was that my eyes see red color so I can see red laser beam (in dark). And because thermal imagers see IR they could see IR beam.

P.S: I do not intend to point CO2 laser beam straight into my T.E. Q1...

[Dmitryz]

Setup of topic starter Accident:

Two CO2 laser tubes 100W at 60W power  level with 2,5''  lenses and fully enclosed nozzles with beam traps at the 8''.  Thermal camera stay at 1 meter with angle to beam 45 degree. 
like this scheme :

        CO2 tubes           lenses          traps
===============    ()=                 >

===============    ()=                 >
                                              /
                                            /
                             t420 camera look at the gap from nozzle to trap

Two 1 sec impulse.

Scattering from beam trap may be less than 1% of total power with half hemisfere distribution.

[ir.ukrm]

Once I had a camera Flir I7. The matrix was damaged, the reason is not known. See what pictures this camera took -40

[Fraser]

That looks like typical laser beam induced damage.

Fraser

[Ultrapurple]

It looks to me much more like a small lizard that went splat! on the windshield 

... you have checked ...?


All joking apart that does look line a nasty bit of damage. I imagine anyone within earshot at the time may have had their vocabulary increased by a number of classic stressed-engineer words. "Freddled zarking gruntbuggly" is one of my favourite phrases (unless the scale of the disaster is such that it warrants one of those emergency-use-only words such as "Belgium!"

Sorry. I promise I will moderate my language in future.

I did once know the Arabic for "May the seeds of a thousand pomegranates grind remorselessly beneath the dentures of your camel" but it loses a lot of its impact in the translation.

[Gyro]

Could it be direct sun exposure damage? Just a thought. [Edit: I mean ir.ukrm's one]

It's a bit worrying that these cameras are getting 'in harms way' with laser reflections while presumably being hand-held.

[Ultrapurple]

Not wanting to resurrect an old thread as such, just adding a pointer forward in case anyone reds this thread in the future.

I had an unfortunate experience with a Ulis sensor and a CO2 laser, which is documented in this thread. Spoiler alert: the microbolometer spontaneously largely recovered over a period of months.