In order to tinker with LED-based flash lighting, it seemed like a good idea to have some way of measuring flash intensity and also flash duration. Intensity seemed like it might be quite difficult to get results in meaning units without buying an expensive piece of kit that has been calibrated for use for that purpose. So, I decided to settle for measuring duration, and with a bit of jiggery-pokery, to be discussed in another post, relative flash intensity. Thus, the starting point in this blog post, is the measurement of flash duration. That may seem pretty straight forward – take a photosensitive device of some kind and hook it up to a ‘scope and fire a flash at it and hey presto. However, it turns out not to be quite as simple as that. Strobe flashes, if they are to be useful, are generally pretty brief, and it turns out that many devices are simply not fast enough; they may be sensitive enough and show a change in current, voltage or resistance, but they may significantly distort the time-course of the light flash.
The masters of the measurement of all things to do with light are Hamamatsu. They have an excellent series of handbooks on how to measure light intensity and flash duration. One of the best devices for measuring flash duration is the PIN photodiode – a diode with a wide, undoped intrinsic semiconductor slab sandwiched between a p-type and an n-type semiconductor region. For the ins-and-outs of using them, and getting them to respond as fast as possible, see: http://www.hamamatsu.com/resources/pdf/ssd/e02_handbook_si_photodiode.pdf. PIN photodiodes can respond extremely quickly and some devices can work well into the Gigahertz range. Rather than go into the technicalities, with many which I have only just become familiar and thus can claim very little expertise, I show below the circuit recommend for the fastest response times using a BPX65 photodiode, a relatively cheap device (about £5:00). With a 50 ohm load and a 5V bias voltage, the chip has a 12ns rise time – plenty fast enough for measuring the flash durations most photographers are interested in. The circuit employs a PIN photodiode in ‘photoconductive mode’ – in this mode the diode is reverse biased and its capacitance, that would otherwise slow its response, is minimized.
This circuit was mounted in an earthed metal box and the output taken to a BNC socket. On this web site http://imajeenyus.com/electronics/20130110_high_speed_photodiode/, I saw that someone had cleverly incorporated a facility to change the load resistance thus making it possible to change the sensitivity and rise-time. I stole the idea. I used a 100 ohm load resistor – changing this for 50 ohms makes little or no difference to the results.
With the circuit built, I figured it was time to measure the flash duration of some flash guns. I started with my pair of Yongnuo 560II guns. I hooked the circuit up to my Siglent ‘scope, connected a 100 ohm load resistor, and set up the triggering to catch the output from the photodiode.
So, the t=0.5 is about 1460us. Here is the output for the lowest power:
For t=0.5, about 50us. Note the change of scales.
A quick test of the repeat-ability at 1/128th power – the two references flashes (pink) were made with the gun at different distance so it is only the duration that can be compared. Scaling by eye, the durations and rise-times are pretty much identical.
How similar are the two guns?
The two guns are close to identical at 1/64th power and this is true at all powers.
Here is a table of the t=0.5 flash durations. These are pretty close to those measured by others (see for example – http://www.gock.net/2012/01/flash-durations-small-strobes/).
Power Duration (us)
Finally, here is an example of the light output from a 3W LED driven by ~10A (see previous post). The voltage trace has been filtered to reduce noise and that is why it looks a bit ugly. The time course of the light output from the LED is practically identical to the current pulse used to drive it – in this case the pulse was 10us in length.