- Flow phenomenon: Water boiling? Faucet dripping?
- Visualization technique: Add dye? See light distorted by air/water surface?
- Lighting: Continuous? Strobe? Sheet?
- Visual acquisition: Still? Video? Stereo? Time lapse? High speed?
- Post processing: Creating the final output. Editing: at least cropping the image and setting contrast.
Now that we’ve covered the first two choices in making a flow visual, and it’s time to think about Choice 3: lighting. As we’ve seen, depending on both the phenomenon and the vis technique, lighting can make or break the result, mostly through what the light is pointed at — the flow itself or the background. Our choice for acquiring or recording the visual will also interact with the lighting choice, but in this case the duration of the light — continuous or flash — will be the key decision. If you are recording video, or if your flow is slow, you’ll want continuous light. If your flow is moving fast, you’ll want to ‘freeze’ it by illuminating it for a very short time with a flash or strobe light or use a short shutter time on your camera; more on that later.
On this page we’ll go over a range of light producing technologies and list their pros and cons. While you have to set your lighting before creating your visualization, you don’t have to read this page now – you can come back to it when you need it.
The metric for light intensity is unfortunately complicated, but here’s a simplified version. The base unit is the candela, roughly equivalent to the visible light of an actual candle . If the candle emits light spherically, and you capture about 1/12 of that (one steradian), you get one lumen (lm) . Thanks to a European Union mandate in 2010, the specification for lighting going forward will be in lumens, so it’s a good idea to get yourself calibrated to that. OK, one lumen is pretty weak; the whole candle puts out 12.6 lumens. How bright is the sun, for comparison? Light we get from the sun is 98 000 lux (lumens per square meter) , so compare that to concentrating our one lumen from a candle onto a square meter. An old 60 watt incandescent light bulb, which emits 800 lumens total, provides a more useful guide compared to the 12.6 lm from a candle. In contrast, a new 40 watt (in terms of power consumption) LED can put out 5000 lumens .
Lights are also specified by their size, directionality, color and duration. Sizes can range from a millimeter or two for an LED, to a ‘troffer’, a 2 foot by 2 or 4 foot length panel, originally designed for fluorescent tubes and now available with LEDs. Point source lights are used in spotlights, providing unidirectional or focused light with hard-edged shadows, and in specialty systems (see schlieren). Larger light sources are usually diffused, so the light is emitted from each location on the surface in all directions, resulting in soft-edged shadows. A typical photo studio light system will have one to three lights that are small flash units but are diffused by umbrellas or ‘soft boxes’. The main light is called the key light, the second is a fill light, set to lift the shadows a bit, and the optional third light is usually placed behind the subject for backlighting . On-camera lights such as LEDs on phone cameras, or flash units integrated on other cameras are rarely used professionally. Any lighting device that is close to the camera lens will result in visible glare from glass surfaces (the ‘giant white hole’) and the contours of the subject will be rendered invisible due to the flat front lighting. A small white card (or a sheet of paper) can be used to bounce the light up to the ceiling so that the light will fall from an off-axis direction.
The color of the light is specified by a temperature which is related to its black body spectrum. We’ll get into details later, but for now know that a 3000 K source will give a ‘warm’ light: white with yellows and reds. A 5000 K source, although nominally hotter, will give a white light that feels cooler, with more blue in it.
The duration of light you need has a big impact on what lighting technology to choose, so next we’ll list the technologies by continuous vs flash categories.
Going from oldest technology to newest, we start with the sun. It provides nearly parallel (collimated) light; the diameter of the visible disk accounts for the slightly soft-edged shadows it casts.
Incandescent Light Bulbs
The original ‘Edison’ light bulb is a very straightforward technology. The first version was invented in 1802 , but wasn’t made practical until 1879 . An electric current, either 110V AC or a few volts DC (in the case of old flashlights), is passed through a tungsten wire. The current heats the wire to white hot, and it emits photons over a broad range of frequencies, from the visible down into the infrared. (More on this later.) It is not efficient, since a large fraction of the photons are in the infrared range, not the visible. Over time, the tungsten element will literally burn out; it will oxidize or slowly sublimate. Ordinary household bulbs are made with a partial vacuum inside (providing a nice ‘pop’ when crushed) to reduce the availability of oxygen and prolong the life of the bulb. Halogen bulbs in contrast are filled with an inert gas plus some halogen at high pressure. The halogen makes the tungsten re-deposit on the element. This allows halogen bulbs to run at higher intensities, hence their use in automobile headlights. The high pressure requires quartz to replace the glass shell, and the bulbs are sometimes called quartz bulbs. Quartz, although stronger and more heat tolerant, has two drawbacks: it becomes very fragile when hot, and any contaminant on the surface will concentrate heat there, leading to damage to the quartz and bulb failure. Overall, incandescent bulbs have several advantages: they are inexpensive; they generate true continuous, non-flickering light; and they are broad spectrum. On the negative side, they generate a lot of heat; they are inefficient, requiring high voltages and currents; they are fragile; and they are becoming increasingly rare.
In a fluorescent tube there is a small amount of mercury which is vaporized and energized by electrodes at each end of the tube. The mercury gives off high-energy UV photons. These photons hit a phosphor coating on the inside of the tube. The coating absorbs the high-energy photons, and gives off lower-energy, visible photons. The color of the light depends on the components of the phosphor, giving ‘warm’ or ‘cool’ light. Fluorescents have the advantage of being much more efficient than incandescent lights, and they are inexpensive. Their disadvantages include giving off light at specific wavelengths (usually pink and green), rather than true broadband light. They are generally not very intense, so the light is not concentrated, and makes for poor spot lights. They are hazardous if broken, due to the mercury, and have to be disposed of properly. They also flicker on and off at 120 Hz. This is not usually noticed by humans, but it can make for an interesting strobe effect, as shown in Figure 1. Finally, if you are working with the long fluorescent tubes (T8, 4 ft long), you’ll be irritated by getting them installed in the fixture (troffer) correctly. It’s so fussy!
The availability of white and multicolored LED — light emitting diode — lights in the past few years has dramatically changed lighting equipment. The technology is still changing rapidly, making decisions about what equipment to invest in very difficult, as light panels that were hundreds of dollars a few years ago are now a quarter of the price.
They are available in a wide range of form factors, from fairy lights on a string to large uniform light panels. An individual LED is generally quite small, roughly a millimeter or two. A ‘white’ LED may be a package of red, green and blue LEDs combined. When controlled individually, you can dial the resulting color across the rainbow. Other white LEDs use phosphor to convert some of the blue light to red and green; this type produces a better quality of white light . A large panel is an grid of individual elements combined with a light diffusing sheet. The light intensity of a single element has been increasing, making specifically-designed LEDs appropriate for spot lighting, comparable to incandescents but without the heat load. They are very efficient and long-lasting compared to both incandescents and fluorescents, and they are comparable in price over time.
LEDs do have some disadvantages. They generate some heat, although much less than incandescents; high-intensity bulbs require cooling fins. They are long-lasting, but do eventually fail. They provide stable, continuous light, but only when driven with a DC voltage source. Many inexpensive LED drivers provide poor DC, resulting in LED flicker, usually at 120 Hz.
Students often use the LED flashlight on their phone as a light source for flow vis, which are currently around 50 lumens, according to a Google search. This can work well for small experiments that change slowly in time, but much better lighting systems are available that will result in higher quality visuals.
Continuous wave (CW) lasers differ from other lighting because the light is at a single, pure color, and the light is collimated — that is, comes out as a narrow beam with nearly parallel sides a few millimeters wide. This is great for flow vis applications because a cylindrical lens can be used to turn the beam into a flat sheet of light. Glass stirring sticks actually can make fine cylindrical lenses for this purpose; they have very short focal lengths and reasonable optical quality. The light sheet can slice through a flow, as was seen in Figure 4 in Overview 1. A rotating mirror can sweep a beam, creating the illusion of sheets and other shapes, as in this desktop toy XXXX, but the flow may change during a sweep, so illuminating the whole flow simultaneously is preferred.
Lasers used in flow vis are typically either solid-state (a special type of diode, or LED) or are a gas type. Currently available diode lasers emit light at one specific frequency or color (see Photons, Wavelength and Color), while gas lasers can emit a few colors simultaneously. Gas lasers can produce a very high-quality beam — perfectly round, highly coherent and collimated — while diode lasers emit beams with a rectangular cross section, making focus difficult. The real reason to get a diode laser is that a one-watt green diode laser can be bought for $80 and runs on a 3.7V battery, while a one-watt argon-ion gas laser is $20,000 new, requires a water cooling system, and 220V three-phase power.
In any case, please keep in mind that any laser of more than a few milliwatts can cause serious eye damage, including blindness. A one-watt laser can burn skin and ignite furniture. However, because the light is only in a single color, powers of at least one watt are needed for flow vis applications. Unfortunately, all of the inexpensive diode laser ‘flashlights’ in this power range are on/off only. When setting up a laser for flow visualization illumination, all initial work should be at very low power, for the safety of the experimenter, and then dialed up for the visualization. The driver circuit for a variable power laser diode is not complex and can be put together for less than $100, with another $100 for mounting the laser and optics , but I haven’t seen these offered as a system for a reasonable price. There are a range of techniques that must be employed to use a laser safely; be sure to get trained before attempting to use one. I am not comfortable publishing them here, but I’m happy to train students in person.
</a id=”flash”>The terminology here is not well standardized. ‘Flash’ generally refers to a single short burst of light, used to make a single still image, but ‘strobe’ can refer to a single burst, or a sequence of bursts, like a strobe party light, or the flashing lights on emergency vehicles. Any continuous light source can be chopped into short durations using a rotating disc with holes or by pulsing the power to the light source, but the maximum intensity is won’t be improved. Instead, a true flash or strobe is designed to be much brighter than a continuous light source. Early (~1860) flash technology was electrically ignited magnesium ribbon or powder. In 1930, the flashbulb came along; basically an Edison-type bulb with oxygen inside that burned brief and bright, for a single use only . The modern commercially available ‘electronic’ flash has been around since the late 1950’s although Ernst Mach (yes, as in Mach number) used an air-gap flash in 1888 to look at shock waves (Figure 2).
Strobes and Speedlights
The most common type of strobe (or Speedlight, if you are a Nikon enthusiast) is an ‘electronic’ flash: a xenon flash tube (Figure 3). On-camera integrated flash units and standalone studio strobes are of this type. A charged capacitor dumps current into a small quartz tube filled with xenon gas. The high current ionizes the xenon, and the resulting spectrum is a broadband white light, centered on green . This provides a bright, short burst of light. Unfortunately, brightness and duration are related: the brighter the flash, the longer it lasts. Typical durations are 1 to 10 μsec (1/100,00 to 1/10,000 seconds). Strobe duration specifications are often given as T.5, the time it takes the intensity to decay to 1/2 the maximum, although the tail of the curve (Figure 4) can last well beyond that, causing ghostly motion blur in an image. A more useful specification for flow visualization is thus T.1.
Photographic strobes require significant battery power (4 to 6 AA batteries) and take several seconds to recharge. In contrast, a stroboscope, used for studying rotating machinery, flashes continuously. LED stroboscopes provide low-intensity light bursts at frequencies up to 1600 Hz, and xenon stroboscopes up to 600 Hz.
Strobe units need to be triggered to fire when the shutter on your camera is fully open, i.e., in sync. Triggering can be done either optically (the strobe will fire when it sees the flash on your camera go off), or wirelessly using a trigger device that fits on the hot shoe of your camera. Controlling the intensity can be done in a variety of ways — using a sensor on your camera, on the strobe, or manually — but be forewarned — it will likely be difficult, buried in menus and jargon. Once you figure out what works with a given equipment combination, write it down. Information on how to trigger based on something happening in your experiment is another approach you may need. There is useful information on high-speed triggering at HiVis.com.
LEDs, being solid-state devices, can be turned on and off quickly, and in quick succession. Theoretically, anyways. How fast and how bright depends on the details of the LEDs and circuits involved. The flash on a cell phone is typically of this type because they are small and don’t require a lot of battery power, but they are weak, so not appropriate for anything but phone cameras. I haven’t been able to find any data on typical flash duration or intensities. In any case, there are DIY hackers creating LED strobes, and LED strobes are commonly used on emergency vehicles, but no stand-alone LED strobes for photography are yet on the market.
The ultimate in short-duration flash is the pulsed laser; their flashes are so short that you can freeze any flow. Durations as short as femtoseconds (10-15, a millionth of a billionth of a second) are possible, but are likely to be low power and in the infrared, so not useful for flow vis. Picosecond lasers (10-12; seconds) are available for less than $200, in blue or red wavelengths, sold for tattoo removal at home, but again are likely to be low power, probably around 7 mJ . Since the flash is so short and is spread over many square centimeters in a flow, relatively powerful pulses are needed for flow vis, on the order of 100 mJ per pulse. Such powers are available in nanosecond (10-9 sec) lasers for micromachining and PIV applications, and again for tattoo removal! Yay for the consumer market! $600 will get you a 10-ns, 532-nm (green) pulse at up to 2000 mJ . I don’t have one, but I am tempted. AND THE SAME SAFETY REQUIREMENTS APPLY, DOUBLED! Pulsed lasers are even trickier to use safely for flow vis.