optical lightning detector

Roll your own... might have to use a photo-tube... that was what I used in my first slave-flash when I was ~15 years old ;-) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

Did these tubes act like a giant light activiate SCR?

I don't recall the details now... it has, after all, been 60 years, but I used it to directly fire a flashbulb. (I didn't have a strobe until a few years later.) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

Digital, camera? Most of them have massive delay between pushing the button and actually taking the picture. There often is a sports mode that reduces the delay. But, I suspect this will be a problem with most digital cameras, they will take a picture of black sky, a large fraction of a second AFTER the lightning.

Jon

With a vidicon-type camera, that could work: the image remains readable for a few microseconds after the light is received. With solid-state, though, there's a cycle of erase/integrate/readout. If you keep the camera in erase mode, the flash is over before you can get to integrate mode.

Best, might be to alternate two cameras, having one always in the integrate phase, and capture the associated readout data only when a flash was detected during the integrate phase. So, you'd be discarding a lot of frames of a video feed.

It's easier with film: just open a shutter, and close it if/when enough incident light is seen. Probably a used point/shoot 35mm camera is the best way to proceed; with a tripod, using an IR remote to start the exposure. Extra points for electric film advance...

the camera is fast, so lag won't be a problem, but figuring out how to adjust everything will be, as I can't summon a lightning storm at any time.

One strategy is to just concoct a trigger like you envisioned, the camera is in "bulb" mode which opens the shutter for say 20 seconds each time. Since lightning often occurs in clusters of strikes, you have a chance at catching the follow-on strikes. 'Pro' phtographers do this manually.

The trigger has to catch a sudden burst of light. How to do? A simple way is just trigger on any bright flash. Take a bwp34 photodiode and lens. connect to a transimpedance op-amp and then to a schmitt trigger circuit to trigger the camera open for 5-30 s. I'm assumimg you have dark condx.

And, in case the lightning flash is weak, I wonder about the following comparator circuit. Assuming the photodiode is put in a transimpedance configuration, a standard way of wiring PD's to minimize capacitance effects.

Now suppose you take a sensitive comparator and put the output signal of the PD amp on BOTH input pins of the comparator. ONE pin of the comparator has a cap to ground to provide a bit of time delay. When a fast changing signal comes through, one pin will lag and the comparator will fire. Slower signals will not fire it, and its schmitt one shot.

I wonder if anyone has more insight into this type of simple circuit - a delay comparator in time. Be interested in hearing about.

The two signals must obviously be isolated from each other -

OH, and bear in mind that if you get 1/500 hits with a digital camera, its OK, since you just erase the 499 and keep the one. This greatly relaxes the precision requirement of your detector in time and space. Of course the more you miss, the longer you have to wait for a good pic. A detector that shoots and misses often may be better.

I thought about this area a bit because I wanted to catch grasshoppers springing from grass into flight. A trick is to use a bright flash even in daylight to catch them. But they are very fast, so I had trouble with manual shooting, as they can hear you coming and spring fast.

Buy one of the consumer digital cameras that has a SDK. Set it to 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.

Steve

The DSLR shutter lag really is significant. We're talking on the order of

100ms, and this is assuming you have a mirror lock capable camera and are not using autofocus.[Prefocus and leave the camera in manual.]

The lightning flash lasts about 100ms

The way these shots are done at night is you just leave the shutter open (B) and stop it down a bit. That gives you a long exposure time. When the flash occurs, close the shutter.

In the daytime, they use a neutral density filter. Same basic technique.

To do what you want to do will require an external camera shutter. Some can fire in under 20ms.

Uman has a few books on lightning. Well worth the money.

Yes, that's clever. An even simpler way is to not use a detector. Do time lapse photography every night, each 20 s. long, and if a thunderstorm happens, review your photos. If not, erase the photos. (total chip erase.)

simple. easy.

With a very fast camera, that might even work for night time photography, however, for day time photography, the change in total light input for a wide angle detector is not that great to make a reliable detection.

Have you tried a video camera (preferably setting it a bit underexposed) and in post processing skip the "uninteresting" frames ? Doing some change detection between frames should help in elimination.

I don't know how he does it here, but he seems to mostly take HDR photographs.

Greg

I'm boring and offtopic, but check out what can be done with a cheap Canon camera and CHDK (Canon Hack Development Kit):

With digital cameras you can do continuous exposures and pick the frames that contain a strike. Or the video+motion detector mode can detect the pre-strike and start the exposure in time.

Previously I got some very nice strike pictures on slide film by doing long exposure after an another and closing the shutter only after a strike or some experimental time when the frame was already exposed by the background.

Shutter delay of cameras is usually too slow to catch a strike even from pre-strike trigger when using the external camera trigger connection.

-- mikko

Well, I think using a camera flash (especially cellphone camera) indoors would be a good lightning simulator. I'm pretty sure you could use a small photocell as a sensor, reverse biased to some modest voltage, and capacitively coupled to a current to voltage (transimpedance) amp. Then, high-pass filter the output of the amp. Final calibration of the trigger threshold would need a storm, but you can probably get it close with a flash.

Jon

the camera has shutter lag of 40ms or less, so it won't be too slow. Lightning really isn't all that fast, I just want an electronic trigger system that only triggers the camera when a flash of light of the correct speed and intensity appears. It's fairly bright in Chicago so long exposures when you're inside a storm mostly pickup reflected light from the city itself, which is an ugly orange color. Yeah one could use blue filter and all that, but it's still very bright.

Cloud to cloud lightning causes lots of lower intensity flashes which I'd like to just tune out so to speak as there's nothing to see there but a flash of diffused light.

I will look for the book though.

this might be ok out in the country, but 20s exposures are going to pick up too might stray light in a major city.

It may be possible to set a super low ISO in the camera and use a super dark ND filter to attentuate the garbage light, but for the hell of it, I want an electronic solution vs. brute force. I've not found a screen cap from an oscilliscope of the light intensity of a lightning flash as measured by a photodiots. This would be interesting to see.

Are you sure that 40ms isnt that "trigger button halfway down" mode? Most cameras are in a sleep mode when just sitting there in manual. The link on this Mk III shows this (model specific of course).

The stepped leader has a number of 1us pulses. That is what you would want to trigger on. There is a very high frame rate video of a lighting strike on the internet, but I can't find it at the moment. Too many false hits on google. The stepped leader is not as bright, but the very short duration steps would make for a good trigger, giving you some advance notice before the return strike hits. The video wasn't on youtube, but hosted on the authors website. I snagged a copy of it with a little hacking, but I need to locate the file.