Macro the 1/f rule and flash duration

Started Sep 8, 2008 | Discussions thread
JimH Forum Pro • Posts: 12,911
An interesting subject :)

The tests I did of the 580EX showed that the average brightness of the emitted light is increased as you set the power level higher.

Here are the three relevant captures from my testing:

Above: Flash pulse waveform.

580EX flash, set to High Speed Sync mode, 1/32 Power, camera at 1/1000th shutter speed.

Above: Flash pulse waveform.

580EX flash, set to High Speed Sync mode, 1/16 Power, camera at 1/1000th shutter speed.

Above: Flash pulse waveform.

580EX flash, set to High Speed Sync mode, 1/16 Power, camera at 1/2000th shutter speed.

Note that the actual amplitude (brightness, shown on the vertical scale) is approximately doubled for the 1/16th power tests versus the 1/32nd power test.

The durations of the two emissions are the same for the two that were shot at 1/1000th, and much shorter for the one shot at 1/2000th. Which is to be expected since you don't need to keep the flash firing if the shutter isn't open.

So the implication is that the system adjusts the duration of the "burst" to match the shutter speed.

And it adjusts the duty cycle of the power applied to the tube during the burst to adjust the average brightness of the emission.

What I have not tested is whether the duty cycle is adjusted by altering the pulse widths or the pulse rates (PWM or PRM).

And it might be hard to tell because from the shots that I did make at higher sweep speeds, to try to show the detail of some of these bursts, I don't see very fast rise/fall times for the optical emission. So we might never really be able to know what the shapes of the pulses driving the tube actually are.

Here's a shot showing a better detail of how the light output is modulated during one of these HSS bursts:

And here's another, with an even faster sweep speed:

You can see that the light does not drop to zero, nor rise to a really high level during these bursts. The testing system has a response time of approximately 50 nanoseconds so the waveshape you see in that capture should be very true to the actual light emission.

And it seems to show that the light emission really doesn't have very sharp edges. It's about a 10 microsecond risetime. But what we can't see here is the waveshape of the signal being fed into the flash tube.

I need to ask a coworker of mine who has designed flash tube drivers what he thinks of all of this.

I suppose it'd be interesting to do a whole series of these, starting at 1/1 power and going down to 1/32 power, but with the shutter speed kept constant so we could see if the average brightness continues with the pattern we see comparing the 1/16th power and 1/32 power bursts.

I need to build my photodetector into a more "self contained" unit so I can bring it home and play with it here. As it is, I use a separate power supply and sort of a major kludge when I use it. And it's all at work so about the time I'm really motivated to play with it, I'm home, and then when I'm at work, I've got other things nagging at me or I don't have my camera gear

-- hide signature --

Jim H.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow