Measuring flash duration?

V

vaizki

Suspended / Banned
Messages
1,048
Name
Jukka
Edit My Images
No
Any tips on an easy way to figure out the flash duration of speedlights and studio flash? I don't need exact figures, just an easy way to find the limits of the gear.

Had a play with my 1D today, doing extreme flash speeds with a YN-560 (which is a fully manual "dumb" flash).. I shot some test shots at around 1/4000 and it was pretty clear that the flash (at full power) was taking longer than that to output. The exposure changed every time I changed the shutter speed.

I think somewhere around 1/1250 was where I started getting consistent results (only the background was affected by the shutter speed). The YN-560 specs state 1/200s which is veeeeeery looooooong for a flash duration, they must have measured the "complete capacitor dump" time or something else. :gag:

So any tips how could I easily experiment to find a "safe" shutter speed for the different flashes? I was thinking maybe of shooting frames of a white wall with a large aperture so that it's JUST BARELY blown out on 1/200s, then increasing the shutter speed gradually until it's not.. shoot me down! :bonk:

Note: The 1D uses an electronic shutter and is spec'd for a sync speed of 1/500 but I can easily do the full 1/16000 without having any artifacts (except extreme loss of light over the full frame).. :thumbs:
 
The only way to do it accurately is to use an oscilliscope connected to a light dependant resistor connected in series with a 10K resistor.

But you can get a good idea simply by photographing a subject that's moving at a known speed, with the image taking up a known about of space on the frame - something like a wheel with markings on the rim.

Bear in mind that the flash duration times typically quoted with studio flash are the t.5 time, which is the time it takes for HALF of the flash to discharge. When Hotshoe flashes are fired at less than full power the flash actually fires at full power but then a controller quenches the flash, shutting it off early and very abruptly, it doesn't 'tail off' in the way that a studio flash or a hotshoe flash set at full power tails off.
 
Thanks Garry, I found someone on Google who had used a circular saw to measure flash duration :clap: Might consider something similar..

I'll have to scrounge up an oscilloscope later just for the fun of it :)

Of course this is mostly an academic exercise for me, as my main flash has HSS and I don't take the studio lights out to kill the sun.. but it's giving me some ideas already.
 
I've seen that circular saw test (though can't find it now!). I recall he was using it to test the flash duration of Canon guns, because nowhere does Canon state what they are at anything less than full power.

From memeory, didn't he get a 580EX down to something like 1/35,000sec at 1/128th power? And I seem to recall that at higher outputs than 1/128th and 1/64th, the reduction in flash duration was pretty linear.

I'm not sure I believe the YN's quoted time, for the reasons already stated, and given the vagaries involved in t.5 etc, I bet that if you said the YN was 1/1000sec on full power, 1/2000sec at half power, 1/4000sec at quarter power and so on, that the figures you'd get that way would actually be at least as close to actual as you could realistically measure by DIY means (without an oscilloscope).

Have a look here, at NK Guy's site, and check some of the links he posted. They're at the bottom of this page http://photonotes.org/articles/eos-flash/index3.html#links I've seen some oscilloscope traces of all sorts of things, I'm pretty sure from one of those links, including high speed sync in action and other interesting stuff. Sorry I can't be more specific, but if you spend a little time searching through that lot I think you'll find the answers you're after without having to resort to the circular saw!
 
You will need an attachment for the scope to keep the trace on the screen. Have you ever tried to visualise a trace after it has been and gone. I made a simple tool by having a circular card with a cross on it and mounted to the fastest motor I have. The aim was to take a flash shot of the card. You then measure the blur of the cross and caculate the time. This was a no go as the cross showed a perfect cross with no blur, even at normal flash times.
 
classcams said:
You will need an attachment for the scope to keep the trace on the screen. Have you ever tried to visualise a trace after it has been and gone. I made a simple tool by having a circular card with a cross on it and mounted to the fastest motor I have. The aim was to take a flash shot of the card. You then measure the blur of the cross and caculate the time. This was a no go as the cross showed a perfect cross with no blur, even at normal flash times.
Click to expand...

You need something spinning very fast with a large radius. 1/20,000 sec is enough to freeze the wings of an insect.
 
You must log in or register to reply here.
Back
Top