Antwan said:
Hopefully in 5 years time LED tech will be at an appropriate stage so I can make one for my final uni project :-\
The technology is there now! The salient point is that to take a photo you need more light than you think.
The efficacy of commercial LEDs is around 150 lm/W now. Lab samples are over 200, but its not going to get too much higher than that, because there is a theoretical limit, which for white light is under 300. Doubling the efficacy "does little" for photography - e.g. all it does is half your exposure time or allow you to stop down the aperture by one stop. The only reason for you to wait five years is in case the price comes down.
There is a well-established formula that relates the candela-second rating of a light source to the photographic guide number. That's why I said that you needed about 10 joules of electrical energy to take a photo. Obviously that depends on the speed of the "film", and one advantage of a digital camera is that you can - these days - turn the gain up quite high, i.e. it works well in low-light conditions.
The "figures of merit" you need for taking a photo are therefore the guide number and the camera settings such as aperture and film speed. OK, we know all that, but the point Im making is that guide number is related to the candela-second rating and NOT to the lumen rating of the light source. All this fixation with lumens is slightly wide of the mark
The reasons being a) the candela is a measure of the concentration of the beam (if it is more concentrated, it is brighter of course, for the same lumens), and b) you need to take into account the exposure time. Viz... if your light source is 1,000,000 lumens and its on for only a microsecond, that's great for high-speed photography but that's only the same exposure as a very modest 10 lumens that is on for a tenth of a second. Both exposures are only one lumen-second.
To take a photo you may need, say, 1500 lumen-seconds or, lets say, around 4000 candela-seconds. Obviously these figures are all very vague. Ive taken them from an article Ive written for the next issue of the CREG journal. Its very long and very mathematical, but it ends up deriving the formula for guide number in terms of the BCPS (beam candlepower seconds) figure for a light source and goes on to give a brief example of a 7W LED which, when used with the right reflector, might result in a guide number similar to that of a flashbulb or a 'pretty-good' flashgun (say 40m @ 100ASA), for a 1.4s exposure time. See
http://bcra.org.uk/cregj
To get back to your point about a project: things arent going to change much in five years - apart from a modest increase in efficacy, and a decrease in cost. So there is no reason not to start designing that LED flashgun now! The main design point is to do with exposure time. In the example, above, I said that a 7W LED was used with a 1.4s exposure time (i.e. about 10 joules of electrical energy). If you want to reduce that exposure time, to make a true "flash" gun, you need to bring it down to, say, 28ms, so you would need 50 LEDs. (I dont think over-driving the LEDs by "much" is a good idea - you will drastically shorten their lifetime. Heat is not the only problem - its the current density in the semiconductor that might be more significant a problem).
In some respects, the design is simpler than making an LED lamp because you wont have to think about heat-sinking. Getting rid of getting-on-for 300W would be a problem if it was on a continuous basis, but if you limit yourself to 1 flash every 10 seconds, that's only 30W. In fact, rather than building one large gun, it might be better to use many small ones, all slaved together. This gives you some redundancy and takes me back to a project of many, many years ago, the "redundant array of little flashguns" (RALF).
Ive mentioned guide number. Those of you who use flashguns will know how vague a term this is. Its a "guide" and caves are not the "typical scenes" for which is was derived. What would be really helpful would be if someone could establish the guide number of some typical LED lamps, via some carefully-planned practical tests. For example, Footleg and others have described using LED lamps for cave photography in the pages of the CREG journal but I dont think they actually give a figure for "guide number". Apologies if Im wrong about that. Having some info in the public domain on "guide number of caving lamps" would be good. If you want to write an article for the CREG journal, please contact Rob Gill via the CREG or BCRA web sites.