Moby lasers at a concert kill a camera...

[Soulstorm]

White-Light:

My damage is 3 dead pixels.. black.. they simply won't spit anything out. Damage like that video I linked reminds me of LCD screens I replace on broken laptops. Although you could argue that during the video it has a 'bar' as such in the bottom 3rd once the camera dies. Could a beam traveling horizontally be focused down enough via camera optics to effect a small area? That CCD is probably 1-2cm wide at most in that camera.
That video does look weak when it comes to the fans but what if it had just switched to a solid beam when the issue occurred? I doubt it would have captured the frame before it happened.
Just seems quite a big chance for a failure to happen at the exact time as the laser is facing in the same direction. Very hard one to say exactly.

I did notice in the aquamarine footage a few weeks back, that the CCD on the cellphone had received a whack previous - dead pixels/lines on the CCD already from the laser I'd assume.

[White-Light]

Its very hard to tell from video the dangers true. It certainly doesn't look that dangerous.

There are quite a lot of people on here who've experienced camera damage and typically it takes the form of dead pixels. I've never heard of an entire camera being taken out except in the Russian incident and then it was reportedly a 40W pulsed laser so no real surprises there as the pulse would have been many times the 40W rating.

I think its more likely electronic failure therefore although I agree its coincidental it happened during a laser show. However, I'm no expert so stand to be contradicted if anyone can bring forward proof to the contrary.

[mixedgas]

nope, I've lost pixels on cameras with 5 mW of hene, and had a grad student at my former place of employment notice he had burned a spot on a plumbicon type vidicon, (old tube camera for those of you less then 30) so he proceeded to write his initials into the vidicon before he placed the camera on my desk for repair. Also with a hene, although 20 mW at that point. That one did not have a lens, it was a bare detector for a light scattering system.
In fairness to him he hit it with the beam waist from the hene, so it was concentrated.

Depends mainly on the quality of the lens in front of the CCD, and the power and spot size of the laser.

I'm quite certain his ccd was damaged by the laser.

Steve

[Dr Laser]

This CCD killing has always puzzled my. Because I've never had any of my pixels blown or CCD damaged. And yes I had laser shinning directly into the lense of my camera. I think this is a second time Q switching Pulse YAG was used for audience scanning. How many times it has to be told. Pulse YAG = NO AUDIENCE SCANNING!

PS. Forgot to say. The picture is my. Right after this my camera was fine. I still use it and it has no bad pixels at all.

[drlava]

This CCD killing has always puzzled my. Because I've never had any of my pixels blown or CCD damaged. And yes I had laser shinning directly into the lense of my camera. I think this is a second time Q switching Pulse YAG was used for audience scanning. How many times it has to be told. Pulse YAG = NO AUDIENCE SCANNING!

I think you're right. The kind of damage in this video reminds me more of CCD fracture through localized point heating than kiling a few pixels.

[Pangolin]

You guys might want to take a look at what I consider to be a related topic:
http://www.pangolin.com/ubb/showflat...=0&page=0#1692

In fact this may very well be a worst problem, although it's counter-intuitive (both that it's a worst problem, and that it would be a problem at all)...

Bill

[mixedgas]

In my previous life at the university , we measured the nanostructures we made with laser scatter in real time as we made them. Every once in a while someone would slip and get a beam right down the lens, and if it was a cheap cmos CCD, like in some of the cell phone cams, you could get damage with class IIIA. If it was traditional process silicon, it was considerably more robust.

We used bare silicon CCDs with the windows removed (Spiricon Beam Profilers) up to about 100 mW cW , unfocused. But at much more then a milliwatt , the CCD saturated, and you'd see a whole row of pixels go saturated, then more nearby rows. So we had a attenuator wheel in front of the CCD. Put a lens in front of the ccd and that energy per area just went up, 100-1000x times or more. Since my idea of a lens is a 40-250 mm variable zoom c mount, we could hit a pretty small spot size. You may not get anywhere near that good a lens in phone cam.

So when I say blown pixels, I mean that a pixel has failed to be a sensor or is latched on, that does not mean it was vaporized, it means that the next time the camera was powered up, the damage remained. It could be as simple as oversaturating a transistor or silicon resistor on the CCD and popping it from too much current flow,, a analogy might be blowing a eprom data cell during programming, it need not be a hole blasted in the back of the CCD. Either way its a warning sign that light is too bright in the room for human eyes.

I had one student aim a thermal imager into a furnace, despite warnings, he did it anyways, or more precisely he answered his cell while imaging the warm up of the oven to map dead heater elements and got distracted. That oven was a 1x1x1 foot cube and it got orange hot within a few minutes. 5 pixels that were white for the rest of the life of the camera.
7,000$ camera. And on 320 x 180 array, 5 blown pixels is annoying.

Yet on many instances, we had a 300 Watt Cermax xenon arc 3 foot away from a Sony 3CCD broadcast grade betacam (30,000$ camera)
and never saw any damage in 5 years of using the setup. Cermax are to normal arc lamps what a metal ceramic disk ion tube is to a hene laser. It is a ceramic tube with a built in parabolic or ellipse reflector , with a .020" arc at the reflector's focal point. The Cermax was a definate eye hazard if you were within 2 foot of the lamp, yet the camera got blasted for years with no problems. (Note, during operations the light from the setup missed the camera, but it got nailed during alignment)

If you see a pattern here, it was my job to obtain some of the world's highest power density light sources, when you measuring a feature 300 nm across floating in space in real time, you need some serious optical power. And stuff around the wavelength of light or smaller does not reflect like a mirror, it is a isotropic radiator and scatters light through nearly all angles. So you need every photon you can get. In 7.5 years of doing that, we probably popped 4 cameras to optical damage. And it was by bosses attitude that all optical setups were aligned by aiming the light source point blank into the cameras, before installing the masks that limited the cameras to see only the scattered light.

So I'm willing to speculate that it is the fine focusability of the laser light that nails the cameras.

Steve Roberts

[LesioQ]

Here's my DIY CCD damage done using 5W avg. q-switched 532nm (scanned beam).
Piotr.K

[heroic]

Aaaactually... lemme just go and get my cellphone designer hat... ah, that's better.

The cheap CMOS image sensors and the equally-but-not-quite-as-cheap CCDs they use in cellphone cameras have different failure modes.

Firstly, remember that these cameras are really, really, *really* small. The image sensor in most cellphone cameras is only an eighth of an inch across, and that may have as many as eight megapixels on it. Note further that the fixed-focus ones are usually focussed at infinity, since with the tiny aperture on those things it's pretty pointless to do anything else. Light that comes from infinity is collimated. This means that the lens is going to focus however much of the beam that gets through the mask onto a single pixel. That pixel is small. That pixel is really, really freaking small. Each pixel is much smaller than a micron, and you may be putting a fair few millijoules into it. I would be very surprised if a human eye were as vulnerable as one of these tiny sensors. 5 mW of laser light can easily kill pixels on these sensors in milliseconds- far quicker than blinking.

In the semiconductor industry where I used to work, we use a few milliwatts focussed like that to trim the resistors on ICs.

Now, those failure modes. CCDs work like a bucket brigade. There's a buried row of conductors which are covered with an insulating layer. Charge is pushed along the row, and then when it gets to the end, it's read out by a readout amplifier. If you make a spot too hot, or just inject a bunch of charge carriers into the dielectric layer, the charge coupling stops working and you get a row dropout. (You can see this when you point a camcorder at something really freaking bright.) If it gets a little hotter, you can damage the dielectric layer and it will destroy that row. Depending on the design of the CCD, this may stop the entire array working even though only a single pixel is actually damaged.

CMOS sensors are a little different in that they're actually DRAM chips. (In fact, you can actually use CMOS DRAM chips as camera sensors if you pop the tops off them carefully enough.) A DRAM consists of a bunch of tiny capacitors. You charge them up and then some time later you can read them back again. The charge gradually leaks out (the capacitors are *very* tiny, after all) so they have to be refreshed. Interestingly, one of the things that makes the charge leak out faster is if light happens to be falling on the capacitors. This is why DRAM packages are opaque The way these sensors work, then, is to charge up all the capacitors, expose them to light, and then read them out again some time later. The ones that have discharged are the white pixels.

The advantage to this is that you can fab them on almost standard DRAM lines (or, more commonly, on DRAM lines that are too outdated for making modern RAM). This makes them super cheap and that makes them super popular. However, all this superness comes at a price: while CCDs actually consist of only a few layers, there's a bunch of extra Stuff underneath the light sensitive parts of a CMOS sensor. The upshot of this is that if you use a laser to melt a little bit of it, you can short out the charging circuitry that resets all those cells back to their "charged" state. This is generally permanent and causes all sorts of funky badness. Again, those cells are really, really, *really* small, and the gate dielectric in the transistors that drive them is only about five atoms thick (no exaggeration- it's typically only 3 nm thick) so it doesn't take a lot of heat to screw them up really badly.

These chips don't consume a lot of power, so their packages generally don't include any heatsinking, and moreover, the thermal conductivity of silicon (which is pretty similar to glass) is horrible, so if you shine a laser on them they get REALLY hot in one REALLY small spot REALLY fast. Human tissue has a lot more specific heat capacity and conducts heat a lot better, not to mention that most of us don't have particularly perfect optical systems anyway.