Viewing CO2 laser with LWIR camera - DON'T!!!

[Ultrapurple]

I finally have a definitive answer to the long-asked question

Can you see a 10.6µm laser beam with an uncooled LWIR camera?

The answer is emphatically yes - but in the name of everything you hold dear, please don't try it (unless you have a camera that you don't mind turning into a paperweight).

I have a 40W CO2 laser in a laser cutter. I set it to lowest power and 'blipped' it whilst viewing the beam area with a Seek Reveal Pro. It all lit up brightly, so I turned to one of my better cameras that can record thermal video. The camera was a couple of feet from the beam path and well off-axis. I gave the laser two test blips again - and discovered, to my absolute horror, that I'd burned out several pixels on the sensor.

Counting the damage later showed that something like 65 pixels have been destroyed. The only good news is that they're not in the centre of the image, though the image below is misleading. It's a crop from the centre-bottom of the sensor; the top two-thirds or so of the sensor remains fully functional.

One very expensive lesson learned.

[zrq]

From my understanding of microbolometers, I think they will work with any wavelength that can pass the lens (Ge) and windows (probably Si), and they should work with the emission from a CO2 laser.

Ref: https://refractiveindex.info/?shelf=main&book=Ge&page=Amotchkina

[Berni]

Yep id say a microbolometer should be able to see pretty much anything that can make it trough the lens material.

So looking up the transmission graph for Germanium:

Yep 10um IR does indeed go trough it rather well.

Combining that with such a powerful 40W laser i can see how the sensor might have had a bad time. Looking at a 40W visible laser spot on the wall with your own eyes would indeed be painfully and dangerously bright. Does show the importance of using eye protection when working with lasers.

[Hydron]

The passband is as much about the anti-reflection coatings on the lenses/windows as it is the material itself. The graph in the post above shows only approx 50% max transmission (I believe it's low because of the very high refractive index mismatch vs air/vacuum, much greater than you deal with in visible light optics) - this wouldn't work for a multi-element lens so the coatings are absolutely needed to bump the transmission up to the >90% region. Coatings are likely to be optimised for just one of the IR bands - don't assume that a LWIR lens will work at MWIR or vice versa, even if they are both germanium.

Ultrapurple, as for the dead pixels, are you able to access any factory setup modes on the camera? If so is there a dead pixel hiding function? Could make it at least behave better if it doesn't use those pixels for AGC any more (won't fix the blind spot, hiding works best on the very small clusters or single pixels expected to be dead as part of normal process yield).

[Fraser]

Oh Ultrapurple , I am feeling so bad for you. I intended to write to you about profiling laser beams with a thermal camera and got distracted. I am so sorry 

Testing laser output beam quality is an area in which SWIR, MWIR and LWIR cameras are commonly used. There are a lot of precautions needed however. The concentrated photon energy In a laser beam of significant power rating will absolutely fry a sensitive thermal sensor in the pass-band. The energy in the beam is such that even a reflection or multi path reflection can still contain enough energy to cause distress to a sensitive sensor pixel or the human eye ! This is why I am horrified when I see eBay sellers offering ‘open site’ engraving laser setups. Used incorrectly, the user is potentially exposing themselves and others in the vicinity to eye damage.

With regard to powerful lasers and the use of thermal imaging systems, if the laser is in the pass and of the optics and filtration it will pass through with low attenuation and fry the pixels. Microbolometer pixels are designed to be super efficient at collecting heat energy and deliberately have poor heat sinking capabilities. They are like lambs to the slaughter where powerful laser sourced photon energy is concerned. Cooled MWIR and LWIR cameras often use photon detectors and these are very easily fried by exposure to excessive energy, even from a Halogen lamp !

It should also be noted that AR pass-band characteristics are not adequate protection against out of band high energy lasers. The laser will burn its way through the AR coating as AR coatings are not perfect and enough energy will pass to cause harm, or localised heating will burn away the fragile AR layers. We must remember how efficient a laser can be at etching and cutting material that absorbs its energy. Any absorption will lead to heating.

Just as with measuring RF energy, Current or Voltage. It is sometimes wise to place plenty of attenuation in the measurement path and Then gradually reduce it to ensure the level is adequate, but harmless. This ensures that the sensing device is not Unknowingly exposed to it’s damage threshold during the test.

I feel so sorry for you Ultrapurple..... I have been where you are myself..... carrying out an experiment and only realising the risk to equipment after harm has been done to it. It makes you feel sick  I truly hope it was not one of your better cameras that suffered damage.

Fraser

[Fraser]

Ultrapurple,

I forgot to say...... thank you for sharing this accident with the forum. Some lesser persons might not have wanted to detail it to others ! You may have saved others from suffering the same damage to their cameras 

Fraser

[Fraser]

Ultrapurple,

Remember this thread ?

https://www.eevblog.com/forum/thermal-imaging/did-someone-know-what-happen-with-this-drone-camera/msg2966286/#msg2966286

[Ultrapurple]

@zrq - yes, the theory has always said that a LWIR microbolometer should be able to see a 10.6µm laser but there are precious few examples on the net of it being done. I posted the message mostly as a cautionary tale because I don't want anyone else to suffer a similar loss.

@Hydron - yes, there is a dead pixel routine but it doesn't want to work. Some of the dead pixels have stuck black but the ones surrounding are white; the dead pixel routine thinks I am viewing a scene with thermal contrast.

Fortunately I can access the raw dead pixel map file, so once I've worked out how it's set up I should be able to adjust it by hand. It'll take time (something in precious short supply) but I'm optimistic that I can improve matters. Although it would theoretically be possible to replace the sensor and recalibrate the hardware for dead pixels, gain and nonlinearity, I don't have access to the tools that set the various bias voltages. The specific sensor in the camera is a particularly 'hot' one (ie sensitive), which I think may have been handpicked by the manufacturer and 'tuned' for best results. I estimate the NETD at or fractionally below 30mK, though I recognise this is at the expense of dynamic range and other aspects of its performance. 

If all else fails I can try the dead pixel repair procedure documented on EEVblog a while ago.

I remain very cross with myself for letting this happen, though the initial test I did with the Seek suggested it should be OK. Live and learn...

@Fraser - thank you for your sympathy. Yes it was one of my better cameras  And yes, I do remember that thread. The sensor in my camera is sun-safe (I have inadvertently proved that a couple of times) but not laser-speckle-safe.

At least I took the precaution of wearing my glass spectacles, which are opaque to LWIR. I'm upset about losing a few microbolometer pixels: I can't imagine how upset I'd be if I'd lost part of my optic nerve.

[Ultrapurple]

On the subject of laser damage (which I realise is off-topic), here's what happens when you shine 5W or so of visible light onto a grape. I suspect the effect on an eyeball would be similar. (The second image is the better of the two as it was done when the cut was much deeper).

[Fraser]

Hmmmm, so the dead pixel routine is confused...... I am thinking out loud here but could one of my Mikron large area a Black bodies bring the whole scene up to ‘white’ so that the black pixels are captured, and then down to ‘Black’ so that the white pixels are captured ? The Dead pixel routine may not be additive in terms of building the dead pixel map and that would be a problem with this technique. However it might generate two dead pixel map files that you can use to manually update the dead pixel map with the required pixels ? Let me know if I can help.

Fraser

[Ultrapurple]

@Fraser - That's a possibility (and a kind offer) but I think for now I'll try the manual approach when I've got some free time. It's easy enough for me to obtain a .PNG or TIFF image that identifies exactly which pixels are damaged so it may just be a matter of using a hex editor and changing some 00h to FFh.

[Fraser]

For those experimenting with 10.6um lasers.....

http://repairfaq.cis.upenn.edu/Misc/laserco2.htm

Fraser

[Fraser]

Remember this numpty ? (I could not find the video on YouTube)

https://v.youku.com/pad_show/id_XMTM4OTQ0ODMwOA==.html?&source=

Fraser

[MadTux]

Quite interesting, that such damage is done by just pointing the camera at scattered 10.6µm CO2 laser beam.
So a CO2 laser is quite effective in blinding thermal cameras. Wouldn't have expected that.

Wonder if that works for terrorists and other people that don't like being tracked on thermal camera by police/millitary. Or IR missile seekers. A small CO2 laser with beam expander optics should easily get that power at a few kms and kill the IR camera, especially in dry air conditions.

[Ultrapurple]

@MadTux - that's an interesting point. You probably wouldn't need anything like precision aiming; even a (say) 40W CO2 laser with a beam expanded to a 30 degree or more cone would probably be bright enough to give a microbolometer something to think about. 

Before anyone rushes out to try this, I should point out that aiming lasers of any kind at aircraft, vehicles or people is a very bad idea and in must jurisdictions is punishable by a hefty fine and/or imprisonment. But I guess if your cargo is a hundred tons of narcotics you've already realised that getting caught would probably spoil your day.

@Fraser - looking at the end of that video when the chap was shining the laser into the sky reminds me of a night many years ago when I had a 250mW or so green DPSS laser that I aimed skywards during a several-day amateur radio exhibition. As you'd expect, it formed a bright pillar of light reaching to the heavens. What I didn't expect was someone asking, in all seriousness, if I would let them string an aerial wire from the far end 

[Fraser]

Any camera can be blinded with enough photon energy at the optimal wavelength. No surprises there. Yes reflections may seem relatively ‘safe’ but just shine a powerful LED torch at a reflective surface and see how much energy is reflected towards your eyes. Now consider a collimated beam and the energy it concentrates into a small area that, if reflected, can become a concentrated reflected beam of energy. It all comes down to the energy density of the reflected beam after the effects of 1st surface absorption, dispersion and reflection at a particular angle towards a secondary impact surface.

Yes powerful blinding lasers have been used in Military systems. They were even developed to blind attacking aircraft pilots but that was deemed inhuman under the terms of the Geneva Convention so we returned to the standard options of punching holes in metal and flesh or burning the pilot alive as the kinder options 

Weapons systems are developed and countermeasures follow. That is standard military R&D life. The Weapon designer then counters the countermeasure and the weapon becomes effective until a new countermeasure is created. And so it goes on. I am not willing to discuss the specifics of thermal guidance Or observation system countermeasures for obvious reasons ! Safe to say, the modern systems are very smart in how they cope with the targets countermeasures, but they are not flawless.

Fraser

[Ultrapurple]

I have a feeling there may be a few notch filters lurking in the more expensive optical systems.

[Fraser]

Ultrapurple.... your story reminds me of an April Fool joke in a Radio Magazine where there was a review of a new laser based aerial/antenna that formed a column of photons that were then fed with the transmitters RF output. The photons were said to form a virtual antenna that could be driven just like an end fed monopole  


Fraser

[chris_leyson]

I've worked with Synrad and/or Coherent slab lasers in the past with output powers of 200W or 400W. Output power is controlled by PWM so at very low power levels the peak output power is still quite high.

[Ultrapurple]

@Fraser - there have been some weird and wonderful antennas tried over the years, including

• Coupling into a tree trunk
• Coupling into a spout of sea-water
• Using the conductivity of white wine for a VHF dipole

[bap2703]

The positive part is that you only damaged pixels and not biological tissues...
Seriously be careful guys, the maximum allowed exposure for continuous lasers is more in the mW range!

About pixel damage I found that :
https://www.researchgate.net/publication/328180319_In-band_low-power_laser_dazzle_and_pixel_damage_of_an_uncooled_LWIR_thermal_imager

[Ultrapurple]

@bap2703 - thanks for that. I had seen the paper before and at one time we thought that I'd purchased the actual FLIR SC-660 camera used in the experiment, but that turned out not to be the case.

[Fraser]

Ultrapurple,

Did the Seek Reveal Pro camera microbolometer survive unscathed ? If so, I am wondering why it was less susceptible to overload damage 

Fraser

[Ultrapurple]

Yes, the Seek had no problem, but it was used from a very slightly different position.

Had I had any suggestion of a problem with the SEEK I wouldn't have moved on to the expensive camera.

[Bill W]

I finally have a definitive answer to the long-asked question

Can you see a 10.6µm laser beam with an uncooled LWIR camera?

The answer is emphatically yes - but in the name of everything you hold dear, please don't try it.

'normal' people use LWIR sensors to check the profile of an expanded beam.
The differences being viewing an expanded beam and the use of the sensor without a lens.  This was one of the main applications for early LWIR tube cameras.

See final page of
http://www.fire-tics.co.uk/datasheets/P8092_datasheet.pdf

My regards to your sadly departed pixels

Bill