Paul Buff
Offline
Account Locked
|
 p.1 #10 · understanding flash duration? | |
howardm4 wrote:
I don't know what the measuring standard/requirement is for strobe units but if a strobe can dump it's energy in, say, 1 second vs. a competing unit that takes 2 seconds, then the instantaneous intensity is greater but the overall output is the same once time is factored in.
a pool of water only has so much water regardless of how fast you filled it.
This is the closest analogy here, but IGBT control brings in a whole new equation.
Starting with the basics, if you take a 100W tungsten lamp (say 2000 lumens) it continually emits 2000 Lumens. Lumens is the actual intensity of the lamp in its raw form and is not altered by beamwidth (IE reflector angle).
If you turn this lamp on for one second, it produces 2000 Lumenseconds of light (the AMOUNT of light. It is the amount of light that determines exposure. . . intensity times duration = LS.
Now if you turn the lamp on for 1/1000 second, the intensity remains the 2000L but the amount is 2000 times 1/1000 = 2 LS. It is a flash of light with a duration of 1/1000 Second. Because you are only using the 100W light for 1/1000 Second, the energy is 1/10WS
So if you are going to get 2000LS from the lamp at 1/1000 Sec you must increase the intensity of the lamp by a factor of 1000 . . . to 2,000,000 Lumens (a 100,000 Watt lamp). The energy is now 100WS.
With Xenon flash, the energy is not emitted continuously, but as a brief flash of very intense light. If we (incorrectly) assume the flash is switched on and off like the tungsten example for 1/1000 second (flash duration) and we want to achieve 2000 Lumenseconds within this time window, the intensity as, once more, 2,000,000 Lumens for 1/1000 seconds. But if we settle for a 1/500 second flash duration, the intensity required drops to 1,000,000 Lumens X 1/500 = 2000LS.
So the first lesson is that, for a given amount of exposure from flash, the intensity of the flash is higher for a short duration flash than for a long duration. The base flash duration is dependent on the design of the tube and the operating voltages. The design tradeoffs are the shorter the duration the higher the instantaneous power and intensity. If the base flash duration is made too fast, the system is stressed, the tube life is shortened and reliability and color temperature can suffer.
Now we have to look at the real world waveforms of a Xenon flash. The light doesn't switch clearly on and off. Instead, the light quickly rises to maximum peak brightness, then trails off following an exponential curve until the tube stops conducting and shuts off. This is where the terms t.5 and t.1 come into play.
The t.5 term defines the length of time it takes for 50% of the tube's light to be emitted. But that leaves the other 50%, which takes much longer to emit and which can cause considerable motion blur. So the term t.1 was created to better define action stopping capability. t.1 defines the time it takes to emit all but 10% of the total energy. But even this 10% induces blur at about 3 f stops below the primary exposure.
t.1 more closely relates motion blur to that obtained by an equivalent mechanical shutter speed. In a conventional flash system, the t.1 time is typically three times as long at the t.5 time. So if a flash is specified as have a "1/1000 flash duration" this is always the t.5 time unless otherwise specified. So one can usually assume the t.i time is around 1/300 . . . this is the number that really tells you the action stopping capability.
Now, on to IGBT control. In an IGBT controlled system such as Einstein or high end Broncolor systems and low power speedlights, when you reduce power you do so by abruptly shutting the tube off, thus eliminating the ever-declining tail end of the normal flash waveform. The more you reduce power, the more quickly you shut the tube off, thereby dramatically shortening the t.1 time . . . using the very high intensity initial flash waveform and discarding the blur-inducing tail.
An excellent discussion of this process can be found at:
http://robgalbraith.com/bins/content_page.asp?cid=7-10044-10303
The specifics of the Einstein design are as follows:
640WS of capacitor energy are dissipated with a base t.5 time of 1/1600 second and t.i time of 1/540Sec. The peak flashtube current is about 2000A from 500Volt capacitors. Thus the peak energy is 1 million watts. Since well designed Xenon systems achieve an efficacy of about 45 Lumens per Watt, the peak Lumens emitted are on the order of 45,000,000 Lumens. This is equivalent to a 2.6 million watt tungsten lamp.
Since 640WS are dissipated at an efficacy of 45 Lumensecond per WS, about 29,000 Lumenseconds are delivered. At Full Power, the action stopping capability is the same as conventional flash units having a 1/540 Second t,1 time (1/1600 t.5). But as power is reduced, the t.1 duration rapidly decreases to as short as 1/10.000 second. By comparison, dropping the power of a conventional flash unit causes the flash duration to get longer -typically rising from 1/540 Second t.1 to about 1/300 at low power settings.
See also discussions at http://www.fredmiranda.com/forum/topic/835412/5#lastmessage
Edited on Jan 07, 2010 at 01:44 AM · View previous versions
|
Note that while the snow was streaking in the 1/600 sec. of flash duration, the train was not, despite moving at the same speed.