Vela flash

[Garry Edwards]

That's a good result.
But, in addition to streamlining the circuit to achieve a faster peak, it might be an idea to take note or Richard's suggestions (HoppyUK) to maximise the light output.
It seems to me that lack of power is your biggest problem right now, because lack of power = inadequate depth of field, which makes the image look unsharp, and many people won't understand that it's a depth of field problem and will simply assume that the flash duration is too long...

Higher power, or more efficient harnessing of whatever power you can get, will allow higher f/numbers to be used (and so, to a dramatic extent, will getting the light much closer to the subject) and produce greater depth of field. The alternative to that is to use an even high ISO setting, but this will impact on image quality.

Another approach would be to use a camera with a much smaller sensor (compact camera) to achieve much greater depth of field, but this will limit the ISO setting.

A possible 30% power gain won't actually do very much, you need a more creative approach to gaining greater depth of field - look at some DOF tables to work out which options will be of greatest benefit - this one is a bit basic but simple can be good, at least you can select a variety of sensor sizes via camera model

 

[ascorbic]

I actually worded that wrong. Rather than a 30% increase, I should have said that this was giving around a 30% drop, so quite a big difference there. But yes, I agree. Really we needed a whole day to play around with arrangement, timing and everything. Even a reflector would have helped a lot. Ideally a second Vela flash. But this is the first time we've shot with a real gun, and the first time we've shot on a range so we'll get better at it (and be able to share what we've learned).

Thanks for the DOF tips. We shot this on a 5D mk II, but I have an OM-D E-M5, so a smaller sensor there and acceptable high-ISO performance.

 

[sk66]

So you're getting about 2/3 of a stop less out of this one... it's definitely worth worrying about, but not all that much really. Perhaps a more efficient reflector design (parabolic).
Has the LED lifespan been established yet? It would seem to me that overdriving them this hard with short pulses is a bit of an unknown. I would also suggest that the design allow for easy LED panel replacement. Especially if going to a design where they are all surface mounted to a single board. I would not want to spend ~ $1k for a light that isn't (economically) serviceable.

I find the CRI rating of 70 troubling... partly because the ratings are often (usually) overstated, and because it makes the idea of grouping them in order to get more light somewhat problematic. But in single unit use it's probably not of much concern.

 

[Garry Edwards]

So you're getting about 2/3 of a stop less out of this one... it's definitely worth worrying about, but not all that much really. Perhaps a more efficient reflector design (parabolic).
Has the LED lifespan been established yet? It would seem to me that overdriving them this hard with short pulses is a bit of an unknown. I would also suggest that the design allow for easy LED panel replacement. Especially if going to a design where they are all surface mounted to a single board. I would not want to spend ~ $1k for a light that isn't (economically) serviceable.

I find the CRI rating of 70 troubling... partly because the ratings are often (usually) overstated, and because it makes the idea of grouping them in order to get more light somewhat problematic. But in single unit use it's probably not of much concern.

CRI is always problematic in any discontinuous spectrum light source, but unless I'm missing something obvious here, I don't see how this would be a problem for what is in effect a scientific tool - perfectly accurate colour rendition surely isn't needed?
In what way would the problem be greater with lights grouped together?

 

[sk66]

Mixed color light sources...

I suppose if the only thing you care about is the study of the (frozen) motion it might not matter... but that takes it a bit further away from being a "photographic light source" and further towards being a "scientific tool" (IMO).

 

[Garry Edwards]

In fact, there are no mixed light sources, the shots are taken in total darkness so the only contributed light is from the flash.
In actuality, it is a scientific tool, not a photographic light source per se

It's something that interests me personally although I have no practical application for it - I've spent time in the past photographing bullets in flight and I've always failed to get anything decent with even the fastest conventional flash - this seems to me to be a viable alternative to an air gap flash.

 

[sk66]

In fact, there are no mixed light sources, the shots are taken in total darkness so the only contributed light is from the flash.

Maybe you misunderstood me?
The unit itself is meant to be able to combine with additional units into a larger panel or daisy chained together for greater functionality.

 

[Garry Edwards]

I have some good news about the power though. I've been doing some more testing this week and it turns out I'm getting less power from this prototype than from the previous, non-pretty one. From looking at the waveforms it's clear that this is because the rise time is a lot slower than in the previous version. From more tests it seems this is down to parasitic inductance from the wiring of the luminaire (the bit with the LEDs). The good news is that more tests show we can significantly improve the power by changing the way the LEDs are wired. It looks like it could be 30% more power at least (but no promises yet). I was already working on a new design for the luminaire so make it easier to assemble (right now it's nine individual holders wired together, which is quite fiddly). The new version will have a lot lower inductance, as the LEDs are mounted directly to the board. This means the rise time will be a lot shorter, meaning the total exposure over the period is higher.

Good news on achieving beating your funding target too