Canon Speedlite 580EX Flash Durations: How long the flash is on for powers settings down to 1/128.

Electronic flashes vary power output by changing the length of time the flash is on. Knowing the duration of the flash is very useful for estimating what duration is needed for stopping the action when the velocity of the object to be frozen is known. It's also necessary when the flash duration and blur are used to measure the speed.

Since these are both things I do from time to time I was understandably disappointed to discover that the manual for the Canon Speedlite 580EX flash did not include the flash durations for the various power settings. I sent Canon two emails requesting this information but while they responded to the emails they said they couldn't find the information. I then called Canon and was told, "that information is not available," and the technical representative refused to dig any deeper for it. (This surprised me because just a week earlier I called them with a lens question and when the tech rep. couldn't answer it she walked me through two levels of engineers until we found someone who could. It appears the quality of Canon's technical assistance is dependent on the mood of the person answering the call.)

It is a well known axiom that,"Hell hath no fury like an engineer scorned." Frustrated by Canon's inability or refusal to provide this information I determined to figure it out myself.

The process turned out to be simple. I looked around the house and found the fastest moving thing I could lay my hands on: a rotary saw. Knowing the blade radius and rotational speed I calculated that the teeth were traveling at 1658 inches per second. By taking pictures of the teeth at various flash power settings and measuring the length of blur I was able to approximate the duration of the flash.

The tooth gap captured at 1/128th power.

The lowest power setting I could go was 1/16 before the blur closed the tooth gap and made measurements impossible. Fortunately, the data from tests at 1/16, 1/32, 1/64, and 1/128 power were enough to get a good idea of the flash durations. Four shots were taken at each power setting, the blur lengths measured for each shot, and the values statistically averaged to find the estimated flash duration for each power setting. The calculated values for these four settings were:

1/16 power = 1/15,000 second

1/32 power = 1/19,000 second

1/64 power = 1/31,000 second

1/128 power = 1/35,000 second

Inaccuracies in measuring the blur and speed variations of the blade indicates that these values are accurate to plus-or-minus 25 percent. So, the flash duration for the 1/128 power setting is somewhere between 1/26,000 and 1/44,000 second. (I admit this isn't very precise but it's a lot better than I got from Canon.)

As any engineer I love graphs so here's one I drew up for the Speedlite:

The black line represents what the flash duration should be if the unit was constant in it's efficiency. At full power, power setting 1.0, the manual states that the flash is 1/833 second. If the efficiency remained constant then the duration at 1/2 power would be 1/1667 second and so on to the 1/128 power setting, which would have a duration of only 1/107,000 second. However, in the real world as the flash duration for xenon flash lamps becomes very small their efficiency decreases. To compensate for this loss of light the flash is programmed to remain on long enough so that the output is what it should be.

The red line in the graph represents my data.

Return to the mainpage and browse 90 other subjects including several other pages related to cameras, lenses and photography.