Photoflash or Lighting Unit for High Speed Cinematography

Chris, Sun Oct 02 2016, 10:46PM

Hello all. First a little background on this project which I have been kicking around for a few years. I have a Hycam II (a old high speed film camera popular until the mid 90s or whenever digital high speed took over) and have been pursuing methods of high frame rate captures mainly for ballistics study. These cameras are quite inexpensive these days, and I even have an extra one for parts. Digital high speed cameras like the Phantom Flex remain quite expensive for amateur pursuits and so film cameras remain the most affordable way for an amateur to step into the world of high speed cinema at least when talking over 1000 pps, and certainly over 10,000 pps. Film and development costs would quickly make digital the cheaper and better option for a professional, but that is not where I am and this project will focus on the Hycam II.

A bit about the camera: This is a 16mm rotating prism camera capable of speeds from approximately 500 to 11,000 nominal frames per second. It takes a C-mount lens. The internal optics give an inherent F-stop value of about 3.8, and so any larger aperture is not effective. The camera takes 400-450 ft spools. The time/footage of film required to reach steady speed depends on the frame rate - at 2000 pps it is almost instantaneous where at 11,000 pps about 200ft of film have elapsed before speed is achieved, and the camera continues to overshoot a bit in speed. There is a 1/4 frame prism/shutter kit which I have which allows frame rates up to 44,000 pps. The 1/4 frame kit exposes frames the same width as a standard 16mm frame but 1/4 the height. With either the full frame or 1/4 frame kit the rotating shutter is open about 1/2.5 of the time, for an exposure time of 9us at 44,000 pps. A 1/2 frame kit was also made but I do not have one unfortunately. The camera has a manual start switch and a remote start function which is activated by a set of dry contacts. There is a shutter pulse output which produces a 15V square wave pulse at each actual frame, although as you get to 40,000 pps it rounds off to look more like a low amplitude sine wave as the circuitry was originally designed for only 11,000 pps. There is also an event output which switches an SPDT relay after a selectable footage of film, based on actual elapsed frames.

These cameras were used extensively in aerospace and for testing of jet engines. Oftentimes at high frame rates copious amounts (as in several MW worth) of halogen incandescent lighting was used (incandescent over other stuff to avoid flicker), and the cameras were simply switched on and allowed to run through the film. Other subjects apparently produced enough light of their own, as in close-up videos of the Space Shuttle Main Engine, or nuclear bomb tests. Lighting will be the biggest challenge and limitation of this camera for amateur use, hence this thread.

Due to the scarceness of 16mm 2R 0.4000 pitch film, I have made a few add-ons to increase the efficiency of film use. I have rigged a mechanical shutter from an old manual film camera onto the original camera lens. This is triggered by a solenoid so that upon demand from an electronic timing circuit (an Arduino for now) this secondary shutter is opened for a mechanically set amount of time (1/15th to 1/500th of a second). It can also be set to remain open as long as the solenoid is active. To make best use of this feature, I have added a black foil piece under the film gate to block just over 1/2 of the width of the frame so that the exposed frame is the width of a standard 8mm frame and half the height. This way the film can simply be flipped over, like in the old Double 8mm format, and the film run several times back and fourth in this way while advancing the timing of the secondary shutter to record different events. I have also found that shutting the camera down at approximately half-way through the spool reduces (but not eliminates) end-of-film shatter at high speeds (saves film, but doesn't mean you don't need to vacuum out the camera after each high speed run). However shutting down the camera much earlier than half-spool results in a catastrophic backlash and crumples all the film on the entire spool. Once the film is developed, I would need to either have it converted to digital format (which would be done in 16mm frames and require post-processing by me) or play it back using the Hycam on low speed as a projector, which I have done by illuminating through the viewfinder with a high powered LED. The camera can run as slowly as 20 pps but speed regulation is poor at such low speeds.

A majority of the film I have in stock for this camera is Kodak High Speed Instrumentation Film, SO-078. I also have some High Speed Range Film SO-0118 and some Ektachrome 2253. The latter film is about ISO 400, while the Instrumentation Film is nominally 400 but can be push processed to 800 or even 1600 with reduced clarity. The SO-0118 is harder to find good information on as it was developed just for military range testing, but the internet seems to think it is Ektachrome color reversal film (and there are examples of it developed with E6 chemistry) and about ISO 200. All the film stock is expired, but kept in the freezer, and should perform decently based on videos and photos I've seen using the same type and era of stock. However I do not expect it to perform as if new.

Now on to the real subject matter. In normal daylight I expect to be able to use ISO 400 film at 2000 to 4000 pps depending on the light transmission of my setup which I am not sure how to find other than still frame testing. For higher speeds, there are two main categories of artificial lighting for the camera which I am open to. It will require further discussion to see which is best and also most feasible. I could do continuous lighting - triggered initially and allowed to run for the duration of the event, or a synchronized strobe triggered by each shutter pulse via a comparator and possibly a delay. It seems that the strobe should allow much sharper pictures, with minimal motion blur if fast enough. Standard photo flash units and strobes can not produce nearly brief enough pulses for the higher frame rates, especially for clearly capturing bullets and such objects traveling in the thousands of ft per second. Continuous lighting will require copious amounts of power. In either category, I am also unclear about the ultimate safety of prolonged flashes used at a distance in daylight, a subject of paramount importance. For a continuous light source, I would expect to need at least 2.6Mlux over the subject area, which could be 4-8 square feet, thus approximately 1Mlumen would be needed. A strobe would average out to the same but have many times higher peak illuminance.

Originally I had wanted to use a PFN built with large capacitors (I have up to seven 51uF 15kV capacitors) producing a square pulse up to 100ms in length through an ultra-long folded arc path seeded with a wire which would be looped back and fourth over a rectangular reflector panel. The seed wire might be coated with sulfur to increase the luminous efficacy such as in a sulfur lamp. Aside from the extreme bulk of this setup and the large inductors it would require, I am entirely unclear of the ohmic characteristics of the long arc because it would be running well below the arc transition current. However if I could determine the correct arc length and get it working properly, I might expect it to achieve some 40 lumens per watt and thus be able to produce the 100ms pulse with just 4-5kJ. The alternative is a short arc (or several; I have somewhere in the neighborhood of 30-60 11.5" parabolic reflector dishes I could use) running on a lower voltage PFN. However I am somewhat limited by the capacitors I have for such a thing, having barely if even enough to do the work at 175V output, which would still be a bit too high for a short arc.

For continuous lighting schemes I have also considered a pyrotechnic material, like flash powder but probably epoxy bonded into a shaped pellet similar to solid rocket propellant. The luminous output and duration would be programmed into the shape and it would sit at the center of a linear parabolic reflector. However I have had no luck finding anything about the luminous output of pyrotechnic compositions, and it would be totally experimental and quite difficult ultimately to get right. A modification on the pyrotechnic idea is an oxy-metal burning torch type device at the center of a parabolic dish (or several). These would use a solenoid valve, could be throttled, and may or may not need an arc ignitor depending on fuel. However this would require some sort of metal bearing fuel - perhaps triethylaluminum (which I am not sure would burn bright enough) or NaK alloy (which I am not sure would burn white enough, being influenced by sodium and potassium spectral lines). None of these fuels are very nice things to synthesize or work with, being highly pyrophoric and all around nasty. I have also considered more traditional fuels (oxy-propane) with magnesium powder injected into the stream by a venturi, however like TEA I am not sure if it would be bright enough. Molten magnesium would be an optimal fuel, but the difficulties of keeping the piping and valve work hot would be great. Molten potassium is also an option, having a much weaker color than sodium alloy, but still a distinct color.

I have also considered a solar concentrator, using several flat reflectors (about 16 to achieve 2.6Mlux) each large enough to project onto the subject area, and each doing so. However these would be difficult to set up and would only work for a very limited time frame without solar tracking, and even with would be limited to the brightest days and times of the year. Compared to other options though it might be pretty nice.

A synchronized strobe would provide better results, but would be difficult. I have recently learned about a photographic strobe using overdriven LEDs which is fast enough, however it is expensive and I don't think it is bright enough nor can repeat quickly enough. I might be able to make a version that is, but there is not much information out there on extreme overdriving of LEDs. An arc method would require extremely fast quenching. If I were to use a pulsed short arc in a moving gas medium, I believe I would want the gas stream traveling quickly enough to move from the focus to exit out the back of the parabolic dish (about 2.3 inches) within the tolerable duration of the flash. I hope I am not entirely correct on this assumption. It would require a device essentially resembling a tiny liquid fueled rocket engine, which would exhaust over the electrodes and out the back of the reflector dish. Even at the gas velocities achieved with an oxy-hydrogen rocket (about 2400m/s) this would result in lingering plasma for about 24us before totally leaving the dish, although it would have moved far out of the focus by then. A more feasible propellant for me would be oxy-ethane. 1us effective pulse width would be more desirable and it is hard to say whether the plasma would have moved sufficiently far from the focus by an amount of time on that order. The pulse generation itself would be far less of a problem than the extreme quenching - I could simply recharge a smaller capacitor from larger ones and then discharge and repeat using thyratrons or high voltage MOSFET arrays. The final discharge probably would be accomplished with a pulse thytratron while the recharge would best use a device capable of interrupting current flow like MOSFETs or pulse triodes (the latter of which I do not currently own). Several pulse capacitors and discharge units could be staggered to accommodate limited charging times and switch duty cycle.

Any other technologies to do this, or input on any of these ideas, let's hear it.