Originally posted by mattdm Paul, the Pentax flash manuals quote the T.5 flash duration. This is an industry-standard measurement with some practical use in comparing flashes, but not so much in terms of actual use. The flash pulse is a curve — on, curve up, peak, curve down, off. The T.5 duration is the length of time for which the curve is over 50% of the peak power. There's quite a lot of light left before and after that.

Another standard number is the T.1 number (given by Metz in their manuals). This is the time the curve is above 10% of peak, and is more representative of the actual impact on the exposure. Normally, this is about 3× the T.5 time, but that's just a rule of thumb.

Pentax USA won't answer my queries about the T.1 flash times. I Pentax Japan just directs me back to Pentax USA. (I suspect Pentax USA *can't* answer, actually — I don't think anyone at Pentax USA knows anything technical about their flash system beyond what can be found in the manual and whatever secret repair manual they have.)

Maybe one of these days I'm going to have to find someone with a digital oscilloscope hooked to the right sensors and measure myself.

Matt

I have been considering the same approach to measure duration. One thing to consider, however, and I am thinking about this as I write, so forgive me if some of it seems disjointed, is that the flash pulse, or more appropriately either the current through the tube or the voltage applied to the tube is going to be a exponential. I will write the equation using spreadsheet formula notations, to make it clear

Value = exp(-t)

the value will be what ever dictates power in the flash during discharge, probably the current from the capacitor.

time constant t will be the discharge ov the capacitor through a current limiting resistor

As a result, if you have the time to half value, you should be able to predict the time to 0.1

There are a lot of assumptions here, that will cause errors such as the voltage across the tube as a function of current, which may cause additional errors, but in a first order, you should be close.

If we start with the guess that the time to 0.1 is 3 times the time to 0.5 then this would result ina duration of 1/400, or 2.5mS that is a substantial period in time, and would only allow a duration for each shutter curtain, leading and trailing of 1.25mS for the sweep. Now, this is something that could be measured. i.e. how fast a curtain travels across the sensor, using a repeating pattern on a scope, and photographing it.