optical lightning detector

It's a nikon d3s, so it's the league of fastest cameras out there.

I'm under the impression the return strike or whatever it's called is the brightest one, so there is no loss having a small amount of lag.

I never though about using the moon for a flare test, but streets lights are for me the the "test".

Wide angle lense seem to suffer from more flare where you see weird artifacts because of internal reflections from what I've seen, and a filter, even a good clear super ultra multi coated whatever ones compounds this on the WA lenses. Bright daylight from clouds tends to make telephoto lenses lose contrast as they seem to pick up a "glow" or sorts with no defined shape- it's just that the shadows seem too bright.

Anyway, for the lightning project the lense choice would be a manual focus prime.

This all started when I finally came across a breakout type cable that let's you control a nikon SLR (film or digital) with 3 banana plugs that essentially give you access to the shutter button, but electronically.

Hw many mpix does it have? I have the Nikon D700 and a 10-24 wide angle, if I remember.

Are you ready to make this gizmo? I can give you product #'s for the Phtotodiode amplifier...You will have to be realistic about light gathering power of PD - What Phil Hobbs calls "photon budget." As I see it, the photon problem is to detect flashes on the horizon which don't really give much illumination. IE, if the lightning is a few miles away, is it giving you enough photons to detect?

I am skeptical - but as a crude test, I'd want to be able to hold up a sheet of paper and see the flash on it by eye. I think some close lightning will easily do that, but as you know, intensity falls off as inverse square.

Detecting lightning say 5 miles away is a challenge. I'd consider a light- gathering mirror or lens. Like a magnifying glass. Harbor Freight has magnifying desk lights with lenses about 5 inch dia. You can buy the lens from them as a replacement for $10. I would consider putting the photodiode in the focal point of that - probably about 12 inches. I believe that will give you light gathering increase, but the lightning is very narrow, so the PD may miss it completely. In other words, you may narrow your focus (field of view) onto a certain part of the sky, but then you may miss it completely. Or you may use a wide angle lens, but then the PD won't see the lightning because it is too small to cover the sky image.

I would consider the following modification of the lens: at the focal point, or just in front of it, put a reflective cone narrowing down to the size of the PD

- about 4 mm dia. A flashlight reflector with PD at it's FP? The point is to throw away image quality and get into photon gathering mode...

Phil Hobbs may have some ingenious solution here. I'd suggest a imaging solution with SW pattern detection, but that would be too slow... I'm not having an "Aha!" moment over this. I'm pretty sure this is a problem where our eyes can do better than simple technology. What our eyes are doing is to parallel process all the pixels in the retina...The 192 core Nvidia chip that Google is using in it's tablet technology is the closest thing I know of...

record video for 20-30 seconds at a time, and wait for a input from your flash detector via serial or USB. If there was a flash, save or download the video. If not, start recording again.

time lapse

happens, review

I

It is really not all that hard to filter out the HPS street lighting. They have filters specifically for that. Do not confuse them with LPS filters though.

?-)

The return strike is the big deal, but what you want to do is detect the stepped leader. The stepped leader can be flicking for 10 to 30ms before the return strike hits. Ideally you would open the shutter within 10ms of detecting the stepped leader. If you researching this a bit further, this has been done using external shutters in front of "stock" cameras in bulb mode.

There is much research on the stepped leader since disturbing the stepped leader is a means of preventing lightning.

This paper covers the spectrum of the stepped leader if you want to enhance your set up.

The paper you should really have is "Spectrum of the Lightning Stepped Leader" by Richard E. Orville. (Journal of Geophysical Research, Vol 73, No

22, November 15, 1968.)

If you read research papers on lightning, there are a small number of names that keep cropping up. Uman or course, but Orville is another.

There is a video by Tom A. Warner online that uses a high speed CCD camera. THe last Adobe flash update screwed up my linux desktop machine, and many of Tom Warner's videos don't play. When you find it, it will not be eye candy. It is scientific imaging. I have a still from it and it read "Fri Jul 10

2009" on the lower third. You can watch the stepped leader build.

There is something odd about the shutter lag testing on the D3s. For instance, it is 47ms in the manual focus mode.

The reason this looks odd is the time is nearly the same as the "button halfway down mode" of 43ms.

I would think your first project would be to measure the lag time yourself. You want to measure the time with manual focus and the mirror locked. On the Canon remotes, you can make it look like the camera button is halfway down, so I assume Nikon can do the same thing. That gets around the wake up time. However Canon has a limit on how long you can sit with the mirror locked. It isn't much, like maybe a minute or two. There are hacks around that since the "Magic Lantern" crowd have reversed engineered EOS.

For the shutter lag test, I'm guessing you would have two rows of LEDs that the camera would photograph. One row counts in a thermometer code. The other acts as an accumulator, also in thermometer mode.

For this scheme to work, you would need a lens in front of the photo diode. The effective field of view of the lens as it projects on the diode should match that of the camera. This is to say, the photo diode seems what the camera see.

But the optics don't have to be very good. You can use a cheesy C-mount zoom lens. A busted C-mount CCD surveillance camera will provide the body. That is, you just gut the surveillance camera electronics, just taking advantage of the light tight box and the mechanics to mount the lens.

Note Nikon has a motion sensing mode. It may be possible that you just leave the camera in motion sensing mode and the lightning flash will trigger it. But you will be catching the later flashes since the live view isn't particularly fast.

If you do a patent search, there are lightning detection schemes using pulse transformers rather than AC coupling the photodiode. Remember, you need to reject all the ambient light and just catch the flash.

An AM radio detuned to a open channel will do a nice job of detecting lightning. This is basic the principal of the aircraft lighting warning systems such as "Stormscope". Stormscope patents are worth looking at.

If I had to detect this optically I'd use a small area, fast response photodiode and look for the AC coupled spike.

Google "personal lightning detector" Steve

And the Devil in the Dishwasher here is that thing Dr. Hobbs calls "photon budget." No matter what your electronics do, you have to deliver a certain amount of photons into the front end. A lens magnifies the photons by the ratio of its area to the area of the PD - ideally. BUT - The lightning occupies a small part of the sky image that the detector (and the camera) "sees." So yes, the lens will magnify the photons from the lightning, but the lightning photons are a small part of the "sky image" brought to a focus by your lens. And the more you want to narrow the field of view down, the less surveillance you can do of the overall sky.

You see, we take for granted what our eye-brain easily does. Our system can zoom down onto a few retina pixels where we see a lightning flash in the midst of a large darker sky.

I tried to help matters by putting a concentrator at the focal point. Another trick is to use an array of say 4 photodiodes and detect a differential between them, in an effort to catch the sliver of light.

If you look at the engineering reality of a PD detector, what you get is simplicity and speed. The area of a bwp34 PD, is about 3x3 mm. Compare that to the pixel size of a CCD - say 5x5 microns. 10 mm2 / 25 microns2 abt=

10exp6/25 = 40,000X. That 40K ratio is the rub. The PD wants its area lit up with photons, but a CCD pixel has an SNR 40,000X better right out of the gate.

And I believe this approximate figuring becomes operational when the lightning is farther away than several miles. Just as you will have trouble detecting meteors with a PD, so you will have the same problem here.

There are certain things that simple optics cannot do, and violating the limits of the photon budget is one of them.

Probably a radio detector is inherently more sensitive than an optical detector, due to its narrow band-width. I question whether it's fast enough to catch same-strike photography, but lightning does have recurring strikes. The lightning detector can turn on the D3s time-lapse photography feature and start capturing.

Here's the reality of professional digital photography: Out of 500 photos you take, you get maybe one good/great shot. Then you erase the 499 and call it a day. Why do the pros shoot like this? Because there are many factors for a good shot, many fleeting. My best photo I ever took was a child in the wreckage of Haiti. The photo was there only a few seconds, and, with the superior auto-focus and handling of the Nikon, I grabbed it. This would have been impossible with a Canon.

With Cydrome and the D3s, I still think you have to take this philosophy - take massive numbers of photos and erase 99.9%. Set your camera up in a place that will be striking, such as the DC fountain area, etc.

It strikes me that there are lots of flashes that are not lightning, and lots of radio impulses that are not lightning either, but the coincidence of optical and radio may be pretty reliable. Others had mentioned that the leaders are dimmer than the return stroke, and I'd guess that a coincidence detector tuned to microsecond pulses may work, and one can set the thresholds lower without being swamped with false alarms.

Joe Gwinn

But the point is you don't want to detect lightning in the overall sky. Rather, you want to detect lighting in a manner that replicates the field of view of the DSLR. That is why I suggest using an old C-mount surveillance type camera body and hack it to mount the photodiode in the focal plane. Based on the dimensions of the photodiode, you can pick a lens such that it sees what the camera sees.

Allow for multiple exposures. The strike will be far brighter than the surrounding landscape anyway.

ton

ain

he

a)

ut

of

it

how about a photo diode mounted on the viewfinder ?

-Lasse

I suggest you rethink the RF detection method. Although most of the energy is in the 10-100KHz range, some detectors run at much higher frequencies to prevent long range RF falsing.

LIGHTNING DETECTION METHODS AND METEOROLOGICAL APPLICATIONS

Lightning detector schematics:

Lightning detector chip (500 KHz):

Lightning Radar:

Wikipedia article on Lightning Detection:

Lightning detectors:

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

I beg to differ (as usual). Two problems.

The lightning flash tends to be behind the buildings in the foreground. Then lightning hits at a distance, all the building are backlit resulting in a black building. You can see the effect on many single frame lightning photos, where the photographer spent much time with Photoshop trying to increase the contrast on the buildings. This is where HDR (high dynamic range) photography really works. Instead of a single image, a series of images is taken and added together. Buildings being to look like buildings and small streamers in the lightning strike begin to appear. The trick is to trigger early, and shoot as many frames as possible. When to stop shooting is a problem.

Another problem is the field of view. There is no way you can predict where in the sky the bolt is going to hit. It's highly unlikely to start directly in front of the camera. More commonly, as in cloud to cloud strikes, the bolt starts to the side and move horizontally across the viewing area. If you trigger on the viewing area, you'll miss some strikes or get a late start where a partial bolt appears on the image. RF detection will pickup bolts behind the camera, which are rather wasted, digital storage is cheap, and easily accommodated with the erase function.

Half full disclosure: In about 1976, I helped a friend in Florida do some lightning photography. There's very little lightning in California, so I had no easy way to test my trigger circuits. So, I fired up an old Tesla coil and immediately destroyed most everything in range. Hint: Do not run tests indoors, near hi-fi's, TV's, or other electro sensitive devices. I later discovered that any small spark will work. Anyway, after 6 months, much fine tuning, and 3 major revisions, it mostly worked. I ended up with two loops, tuned to two different frequencies in the beacon band (about 300 KHz) feeding an AND gates. If there was noise on both inputs, it's a lightning hit. If there was noise on only one input, it's the Navy or a local navigation beacon. Separating the two loops as much as possible reduced falsing from local sparking (auto ignition noise).

Oh-oh, not good: "First impressions of the AS3935 lightning detector"

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

He gives a few clues how he does HDR under his "Helpful H taken of the Pittsburgh skyline (the first one in the post). I clicked my remote, the shutter opened and BOOM! It hit.

asked for a more perfect view of it either. So, he's using a wired or wireless remote shutter trigger, and relying on blind luck and his reaction time to obtain usable photos.

Incidentally, "a huge bolt straight down, spidering up" is not quite correct. Lightning bolts start on the ground, and go up. Does lightning strike from the sky down, or the ground up? The answer is both. Cloud-to-ground lightning comes from the sky down, but the part you see comes from the ground up...

Spoiler: Average duration of a lightning bolt is about 50 microseconds. The optical trigger will need to be fairly fast in order to catch that image with a 1/1000 (1 msec) shutter speed. About the only way to do that is to anticipate the strike. I couldn't find a design for a negative time delay circuit, so that's unlikely. What seemed to be happening with my RF trigger was that the lightning was arriving in a series of 2 to 5 strikes. The camera would miss the first few strikes, but pickup on the later ones. I was scratching my cranium for quite a while trying to determine why there were so many blank photos.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

No viewfinder when you lock the mirror. You lock the mirror to reduce the lag.

Actually you frame the shot, then hope for lightning. That is, if you are trying to do serious photography rather than just lightning photography.

Can you predict where lightning will appear? Well probably not in Chicago, but no problem in the desert. If there is virga, that is where the lightning will happen.

Photos of virga don't do it justice. When you are in the desert, the virga is very easy to spot.

Go back and look at your link. That is text lifted from Uman's book. The full stroke will be 100ms on up.

Time duration of flash 0.2 sec