Radar Stealth

http://www.airplanedesign.info/52-radar-stealth.htm

Radar Stealth

And now we get to what people really mean when they say “stealth”. This is the new technology that makes the latest combat aircraft look really funny. Why do they look that way? What is it that those shapes are trying to do?
To understand that, you need to first understand how radar works. In theory, radar works like visible light: A source shines it onto an object, the “illuminated” object reflects it towards a sensor, the sensor picks up this reflected light and can thus locate and identify the object. There is one big difference, though: With visible light, we are used to having many sources all over the place. Either it’s the sun and the sky and the things they illuminate, or it’s a bunch of light bulbs in at least a few spots in the room. With radar, however, there are no other sources, just the one by the sensor. This would be equivalent to wondering around a football field in a moonless night trying to with only a flashlight in your hand. Say I tell you that, somewhere on this field, there is a black pole sticking up a few feet from the ground, with a model airplane at the tip. How do you find it? You’ll probably sweep the flashlight beam slowly around you until you see it illuminate something that is not the ground.

If the model airplane is black, it will be harder for you to see it.
If the model airplane is shiny, it will reflect light from the flashlight… But it might reflect all your light away from you, and all you will see on it is the reflection of darkness, which is indistinguishable from the darkness behind the airplane, unless it reflects your light towards you and you catch the glint of the reflection of your flashlight on the shiny model.
Thus, to be stealthy to radar, an airplane can do two things: Absorb the radar reflection (be “black”), and reflect it away from the sender (be “shiny” and shaped in a way that does not reflect radar in every direction).
There is a third thing, but it’s a little trickier to understand. The airplane surface can reflect the radar signal twice, such that each reflection is half a wavelength out of tune with the other, so the reflections cancel out. That’s the basic idea behind noise-canceling headphones: They hear the sound outside and play for you the same sounds but half a wavelength off, thus canceling much of the ambient noise. They can do this and play music at the same time, or they can just play the external noise half a wavelength off if you just want some peace and quiet.
Of these three goals (reflecting radar away from the sender, absorbing the radar energy, and reflecting the radar waves in a way that cancels them out), reflecting radar away from the sender is by far the most important.
It is the thickness of your airplane’s skin, and the relationship between the skin and the internal components, that will determine whether you’ll reflect radar in such a way that it will cancel itself off.
It is the materials on your airplane skin that will either absorb the radar energy, reflect it, or let it pass through.
It is the shape of your airplane that will determine whether the radar gets reflected back to the sender or away from the sender.
Denys Overholser, a Lockheed mathematician and electrical engineer who had the brilliant idea of using a highly swept, wedge-like, faceted design for stealth (which eventually became the F-117), once famously said that there are four elements that are important in reducing the radar reflection of an airplane: “shape, shape, shape, and materials”. So you can guess which one of these really matters.
Before we go into how the skin shape, materials and thickness ought to be chosen, let me explain just a little bit about how radar reflections are described. In other words, how do you measure how “stealthy” a stealth plane really is?
Well, different shapes reflect radar in different ways when in different orientations. An airplane may reflect a lot of radar at you when seen from the side, but almost none when seen from the back, for example. So a small airplane seen from the side might return about as much radar as a large airplane seen from the back. Radar engineers and operators measure the radar return of a given airplane from a given angle in terms of its Radar Cross Section, or RCS. This is measured as an area, correspondent to the cross-sectional area of a sphere that returns the same amount of radar. So, if a 747 seen from the back has a Radar Cross Section of ten square meters, then this means that it returns as much radar as a sphere with a cross-sectional area of ten meters. (This sphere would be about 3.5 meters across). However, from the side, the 747 may have an RCS of 100 square meters (equivalent to a sphere about 11.3 meters across). Seen from below, it might have an RCS of a few hundred square meters. I just made all these numbers up, but you get the idea (and they’re probably in the ballpark for a 747’s RCS).
So the challenge is to reduce the RCS from as many angles as possible, especially from the front. Reducing RCS from the nose-on angle is by far the most important, as that is the angle your enemy will see you from while you are on your way to bomb them.

The problem with reducing RCS is that the detection range varies with the fourth root of the RCS. This means that if a given radar can detect an airplane from 100 miles away, then reducing the RCS by half would mean this airplane can be detected by this radar from 84 miles away. You’d have to reduce the RCS by a factor of sixteen, down to 6.25% of what it was, so that this radar could detect you from “only” 50 miles away – not really much of an improvement as far as evading the radar, given you had to reduce your RCS by a HUGE amount. You’d have to reduce your RCS to less than one percent of normal before you can fly with relative impunity from most radar systems. Many radar systems have a range of a few hundred miles. But to destroy them, you must get to within a few miles of them. To make this possible, RCS must be reduced by at least a factor of 10,000 – equivalent from taking an airplane the size of a large combat jet but having it return as little radar as a pigeon would.
Like I said, there are three ways of going about reducing RCS: absorbing radar (skin material), reflecting radar waves that cancel themselves out (skin thickness, “depth” of internal components), and reflecting radar away from its origin (aircraft shape). Of these, the last is by far the most important. So let’s start with what is the least important, and today the least commonly used:

Radar-Absorbent Materials (RAM)

The idea of using RAM to evade radar detection dates almost as far back as the first widespread military use of radar, naturally. During World War 2, England developed a wide and effective radar network to protect itself from German ships and air attacks. Towards the end of the war, aircraft (British, German, and American) started carrying radar to find enemy ships and other aircraft. The Nazis figured out that, if a material absorbs radar the same way that black things absorb visible light, then an airplane covered in this material might be able to slip through British radar.
Whether or not a material absorbs radiation of a certain wavelength has to do with the energy levels of the electrons in the atoms that make up that material’s molecules, as well as with the masses and structures of the atoms that make up the molecules. By finding a material whose molecules can vibrate in frequencies similar to those of radar waves, and/or whose electrons can absorb quantities of energy similar to those carried by photons of radar radiation, there is a good chance this material would absorb radar. Carbon products were found to absorb radar well. In addition, radar waves create small magnetic fields as they hit iron, so many small bits of iron could create magnetic fields in such a way as to absorb most of the radar energy. It turns out that small round particles coated with carbonyl ferrite (“iron balls”) are the best absorbers. They should be embedded in a dielectric material (usually a plastic like neoprene, possibly other materials) which slows down the radar waves and gives them more time to be absorbed.
The first use of RAM was to coat German submarine periscopes with the stuff. It is unknown how effective it was – the material covered what was already a very small target, one that was surrounded by waves a large fraction of the time, and thus almost impossible to detect with radar to begin with. The Horten brothers, designers of flying wings who were working on a twin-jet flying-wing bomber for the Nazis, suggested the use of this material on their bomber. The builders had developed a glue-like RAM made of adhesive, sawdust, and carbon, and the skin of the flying wing was to have been made of two layers of plywood sheet with the RAM glue sandwiched in-between. Only the first of three prototypes was finished by the time the war ended, and it did not have this RAM skin, although the second and third prototypes would have. These kinds of materials were later tested on Canberra bombers, but the RCS reduction was not substantial, so attempts at RAM use were eventually dropped in Europe.

When the US started investigating the possibility of building aircraft with radar-stealthy shapes in the 1970s, development of RAM was again picked up. The F-117, the first airplane designed with stealth as a primary design objective, was entirely covered with tiles made of a soft dielectric plastic (like neoprene), and these tiles had, embedded in them, round particles covered with carbonyl ferrite (“iron balls”). Any maintenance or damage to the aircraft had to result in the repair (patching) of this soft plastic tile surface. Dielectric putties, tapes, and adhesive sheets (with the “iron balls” embedded in them) were developed to make this process easier. Eventually, a dielectric paint was developed that had the “iron balls” suspended in it, and this made the process even easier, although the solvent for this paint is extremely toxic.

More modern stealthy airplanes rely almost entirely on their shape for stealth. As you will see below, the shape of the airplane skin can be designed such that radar is almost always reflected away from the emitter. These more modern aircraft use RAM only in places where radar is necessarily reflected in many directions, such as edges (like the lip of the engine air intakes, the perimeter of the wing, etc), joints/cracks (like where control surfaces, access panels, landing gear doors, the canopy edge, or bomb bay doors, meet their fixed neighboring surfaces), and other non-smooth features. RAM can now be sprayed on, making the maintenance of RAM surfaces even easier.

Lastly, it should be noted that some of the radar energy does penetrate the aircraft skin into its internal structure. Most of it is reflected or absorbed, but not all – the remainder goes bouncing around inside the airplane, and eventually comes back out. Metal does tend to reflect radar very well. Because of this, structures made of composite materials (many of which are carbon-based or can have carbonyl ferrite easily embedded in them) are widely used in stealth aircraft, rather than the more usual titanium, steel, nickel, or aluminum. Metal components – such as the engines and any other necessarily-metal parts – must be surrounded by materials that absorb radar or reflect it away, or must be at the right depth under the skin so as to use wave cancellation to cancel out any outgoing radar reflections. Which brings us to…

Wave Cancellation

Given that some of the radar waves will penetrate the skin of the aircraft and possibly bounce back towards the radar emitter, how do you minimize the impact from these waves? Well, you try to have them bounce from an internal structure (like a second skin inside the first skin) that is one-fourth of a radar wavelength inside in the airplane. In other words, you set up a second skin, which is one-fourth of a radar wavelength under the surface of the airplane. When radar bounces off this internal skin and leaves the airplane, it meets up with the radar reflection of the real surface. But the path of the radar wave that reflected internally is half a wavelength longer than the path of the wave that bounced off the surface (one quarter of a wavelength in, one quarter of a wavelength out), which means they are now approximately out of phase, and largely cancel each other out. That’s the theory, anyways.

The first attempt at radar stealth in the US used this approach. When the U-2 spyplanes were being flown over Russia in the late 1950s, it was known that the planes were being detected, and that they were not shot down only because they flew too high for a missile to hit them (a decreasing safety margin, one which eventually disappeared altogether). It was a matter of time before the U-2 was shot down – Lockheed knew this since before the U-2 was even operational, and the Blackbird was designed to fix this (it replaced the U-2 after less than 10 years). But until the Blackbird was finished, the engineers at Lockheed thought they could keep their U-2s from being detected by having them create a secondary radar reflection half a wavelength off the primary one. Under the top secret Project Rainbow, a U-2 was covered by a grid of thin but firm wires, held a quarter of a typical radar wavelength away from the skin of the airplane (and from each other) by non-conductive poles (originally bamboo, eventually fiberglass).
These U-2s came to be known as “dirty birds”. (An airplane in a “dirty” configuration has the landing gear, flaps, bay doors, weapons, or other things extended, adding to the drag). These wires were definitely a huge dent in the U-2’s stellar low-drag aerodynamics. And, it turns out, they did not make the U-2s stealthy: When the Soviet Union protested about the U-2 flights in early 1958, their report contained precise data about the “dirty birds” flight paths. At that point, the Rainbow apparatus was removed and never used again.
When Lockheed started investigating RCS reduction, they developed a RAM coating made of a dielectric plastic with carbonyl-ferrite-covered particles embedded in it. The figured out that the dielectric plastic slowed down the radar waves. While radar waves have wavelengths ranging from around one inch to a few feet while going through the air, they were slowed down in the material. This means the longer waves were squeezed into a wavelength of a few inches, while the shorter ones were a small fraction of an inch long. The thin RAM coating would mostly absorb the shorter waves and would reflect them in such a way that they would cancel out, as its thickness is about one quarter the wavelength of the shorter waves. But what about the longer, more penetrative waves? Under the RAM, engineers placed a thick layer of fiberglass honeycomb, which could be made to have its surface be less dense and its deeper layers becoming progressively denser. This meant that the medium waves would bounce somewhere along the middle of the honeycomb layer, and come out such that they had traveled the extra half-wavelength necessary for cancellation. The longer, more penetrative waves would travel deeper before bouncing, and thus would also travel approximately the extra (longer) half-wavelength. Engineers likened the RAM-on-honeycomb set-up to a multi-channel stereo system: You have a tweeter, which just takes care of the high-frequency (short wavelength) stuff, and you have a woofer, that is best equipped to handle lower frequencies (longer waves). Individually each of them misses a lot of the waves, but together they can handle a wide variety of wave lengths. Typically, longer-range radars use the lower frequencies (longer, more penetrative waves), while smaller radars on fighter planes use medium waves, and missile radars use the higher-frequency, short-wavelength waves. RAM alone cannot protect from longer-range radar waves, so the ingenious graduated-density fiberglass honeycomb layer absorbed some of it and cancelled out what it reflected. It acted as an internal skin, almost always one quarter of a wavelength below the surface. This multi-layered skin is another important feature of stealth aircraft design.

One of the main developments in radar technology since the advent of stealth is the ability to detect aircraft using longer and longer wavelengths. These longer wavelengths, made up of lower-energy photons, are not absorbed much by RAM and are not cancelled out by the quarter-wavelength effect because they are so long. This means only the shape of an airplane, not its materials, can guarantee that is it hard to detect by modern radars.
These longer wavelengths start overlapping the range of wavelengths used in radio communications and other uses, therefore much noise is encountered. Interference with radio transmissions is undesirable for the radar operators (since it means more noise amidst which an aircraft has to be picked out) and for the radio operators and radio receivers (since the radar pulses become incorporated into the signals the radio people are trying to generate and to demodulate / listen to). However, modern computer algorithms do a good job of picking out an aircraft from the noise, so long-wave radars are being used in many countries’ defense networks nowadays.

Reflection

This is the most important challenge of all, the one that really allows stealth aircraft to be stealthy and that causes them to be shaped so strangely.
So far we have been talking about absorbing radar waves or canceling them out. But if you can reflect most radar waves away from the radar antenna to begin with, then only the rest need to be absorbed or cancelled, minimizing the need for complex multi-layer skins and for RAM. (And, like I just mentioned two paragraphs ago, modern long-wavelength radars emit energy that is not easily absorbed or canceled). However, this goal forces your airplane to be shaped in unconventional ways, some of which help aerodynamically, some of which do not.
Let’s get back to our “looking for something on a football field at night using a flashlight” analogy for radar.
Let’s say that the thing you’re looking for is shiny, rather than dull. That means that a narrow beam of light hitting its surface will bounce off at a certain angle, rather than being scattered all over the place. So as you look at this object in the middle of a dark field, what you see is really what it reflects. If some of the surface is at just the right angle, it will reflect the flashlight beam back at you, so you will see the reflection of the flashlight, and you will see the object. However, even if your flashlight beam hits it, if it only reflects the beam AWAY from you, then all you will see on the surface is the reflection of the darkness, and so you would not be able to pick out the object from its dark surroundings.
That’s the basic idea of making a shape that is stealthy to radar. If, when it is hit by a radar beam, it can reflect all the radar “light” away from the source, then the radar can’t see it (since the radar only sees the radar energy that is reflected back the way it came).
Say, for example, that I have a smooth, cylindrical-shaped piece of shiny metal. Say I shine a light onto it:

The light will be reflected in every direction, including me. In other words, if I shine a light on it, I will always observe that the middle of the cylinder looks very bright (as that is where the light I shine will always be reflected back in my direction).
Now say that I have a smooth, prism-shaped piece of shiny metal. Say I shine a light onto it:

The light will always be reflected away from me:
UNLESS, that is, I happen to be exactly at right-angles to one of the sides:

So, as you can see, something round reflects light in many directions, while something flat usually reflects light in one direction:

This was realized by the Lockheed team that built the F-117: Faceted surfaces reflect radar only in a few directions, and you’d have to be perfectly perpendicular to one of the facets to get the radar reflected right back at you. Since the airplane is always moving, the angle of each facet is always changing from the point of view of someone on the ground, so even if you ARE perpendicular to one facet, this only lasts a moment.
Now, let’s make things a little more complicated. First, radar does not bounce off a surface just like a laser off a mirror. SOME scattering does go on, similar to a surface that is shiny but not so perfectly shiny that you could adjust your hair by looking at your reflection. So beams bounce off mostly at the same angle they came in, give or take:

This means that you don’t have to be PERFECLY perpendicular to a facet in order to get energy reflected back at you. You just have to be kinda close to perpendicular:

The best way to ensure that no energy bounces back to where it came from is to try and have your surfaces at some angle relative to the radar source.
This is impossible to do from all directions: A surface that is angled away from one direction, is inevitably angled TOWARDS some other direction. So which direction is most important? Well, when an airplane flies towards enemy territory, all the enemies start out in FRONT of the airplane, so most of the radar sent its way comes approximately from the FRONT, at least at first. The more you reduce the reflection of radar energy sent from the FRONT, the longer you maintain the element of surprise. Anyone who has been caught speeding on a highway can appreciate the importance of a low frontal RCS (This has led the designers of many airplanes that are not very stealthy, like the F/A-18F, Gripen, and Eurofighter Typhoon, to work to reduce the “nose-on” radar reflection, which gives the most bang for your buck as far as implementing stealth technology goes. For example, much of the frontal radar reflection comes from the fan blades on the front of the engine compressor. Having your intake be curved (serpentine, S-shaped) so as to hide the engine compressor blades goes a long way in reducing frontal RCS).
So it you can have all the surfaces angled away from the front (that is, have the nose and leading edges look like a thin, pointy pyramid), then this is as good as a faceted design can get. The nose is like a wedge that pushes radar energy only a little bit out of the way:

This “wedge nose” idea guided the design of the Lockheed Have Blue prototype, the first aircraft that is stealthy in all modern senses of the word, and which evolved into the F-117. The Have Blue, however, flew very badly because of its extremely low-aspect-ratio and extremely high-sweep-angle wings. It was very draggy and very unstable. When the design was turned into an operational military aircraft, the angle was made not-quite-so-pointy. This may have increased the radar cross-section by a tiny amount, but it lowered the drag and tremendously improved handling characteristics.

Now, what do you do with the back of your airplane? There might be radar coming from there as well, right? You can’t make it pointy like the front, because you would then have a thin, unstable, diamond-shaped airplane, and it would be hard to add engine exhaust nozzles or a tail into that design, let alone ailerons or flaps. (When competing for the Have Blue contract, Northrop did have such a diamond-shaped configuration in their proposal, a design referred to by many as the “hopeless diamond”. As we know, Lockheed got that contract, and their Have Blue prototype eventually evolved into the F-117. Interestingly, Northrop later did fly a stealthy “diamond” during the X-47 program, but a revised version of the X-47 (the X-47B) did add wings to the “diamond”, making it look basically like a smaller and pointier B-2).
So you can’t really have a diamond. The wing needs to have a not-too-swept trailing edge, and the fuselage must stick out behind the wing root so that a tail (and some engine exhaust nozzles) could go all the way in the back:

The problem with this design is that, if any radar comes from behind, it will be perpendicular to one of the edges on the back if the radar comes from one of four directions. These four directions are so spread out, any radar coming from anywhere in the back will be dangerously close to perpendicular to ONE of the four edges.

Lockheed’s clever idea was to have the right wing trailing edge swept at the same angle as the left fuselage trailing edge, and the left wing trailing edge swept at the same angle as the right fuselage trailing edge. This means that there are only two directions between the four trailing edges:

This idea, of lining up edges into parallel groups, minimizing the number of directions from which a radar would be perpendicular to an edge, is one of the key concepts of modern stealth design:

One other thing to note is that no two edges – and no two surfaces – should be at 90 degrees to each other. Any two edges or surfaces that are at 90 degrees will ALWAYS reflect a radar beam RIGHT BACK to where it came. For example, here is a pair of lines at 90 degrees. Notice how a beam can hit the pair from any angle.

The beam bounces off the first line at (180 degrees minus original angle) and it bounces off the second line at 180 degrees minus THAT, which is (180 degrees minus (180 degrees minus original angle)), which is equal to the original angle! (This is why, if two mirrors meet at 90 degrees, you will ALWAYS see a reflection of yourself when you look into that corner, no matter where you stand). So it is important that edges and planes on an aircraft not meet at 90 degrees.
This means that, if an airplane has vertical stabilizers and horizontal stabilizers, any radar energy that hits one of them will bounce off it, then off the other, and then right back to where it came from. Same thing for the internal walls of rectangular air intakes. This means that having diagonal stabilizers and non-rectangular intakes drastically reduces the radar cross section of an airplane:

For the same reasons, the edges of landing gear doors and weapons-bay doors are also lined up with the airplane’s edges: If the door edges reflect radar, at least they should only reflect radar in the direction where the wing edges already reflect radar. Most doors and bumps on the F-117 are hexagonal – basically a rectangle aligned with the direction of the airplane, plus a pointy front and a pointy back to line up with edge directions.

But a bay door cannot always be pointy – there might not be enough room for it to stick forward enough to have a pointy front at the right angle. The solution? The front and back edges can be serrated:

Now let’s go back to the “wedge” idea. Of course only the nose can be a wedge in every dimension (that is, a pyramid), but if the aircraft has chines – that is, if the fuselage sides are a sharp ridge rather than a near-vertical wall – then basically the whole airplane is a “wedge”.

One other way of looking at this is: Most radar comes not from straight above or straight below but from a shallow angle, so it hits the airplane near the sides. So it is the vertical sides that reflect back the radar. Get rid of the vertical sides, leaving only the top and bottom, and this shape will only reflect the rare bits of radar energy coming from the top or the bottom. (Besides, if you’re flying right over a radar site or right under a fighter plane, you’re in trouble already…)

This approach was first tried in the Blackbird. The Blackbird was originally going to have an F104-like fuselage (pointy nose but pretty much round in cross section), plus delta wings – not too different from, say, a B-58. The center of lift was pretty far back, especially at high speeds, so it was difficult to (as stability demands) keep the center of gravity near the front. Canards were considered, and successfully tested in wind tunnels. But then the radar engineers at Lockheed suggested that chines be added to the sides. That way, the sides would not reflect radar back to where it came from. The aerodynamicists at first opposed this – it increases the overall surface area and makes skin friction much higher! But wind-tunnel tests showed that these chines generate vortices over themselves which actually create a lot of lift! All the way at the front, too! So no more need for canards, plus the airplane could turn tight without stalling, and takeoff and landing speeds were reduced as well. (In fact, these low-aspect-ratio lifting surfaces are now prominently used in most fighter jets to increase high-alpha lift, such as the F-16, F/A-18, MiG-29, and Sukhois).

The first pictures of the Blackbird to be publicly released were taken in profile (from the side), so that these chines were not apparent. It looked like just another delta-winged plane, kind of a cross between an F-104 and a B-58, albeit with odd (or, should we say, “retro”) engine placement (engines in the middle of the wings) and big spikes at the inlets (smaller spikes were not unusual in fighters at the time).
Modern stealth UAVs make use of these two shape principles to stay stealthy: chines that make for a wedge-like cross-section, and edges that only reflect radar in a few directions:

Even the noses of the JSF and F-22 have a slight “edge” along the sides, rather than a flat and purely vertical side:


The Sukhoi-32 and -34 “Platypus” has horizontal ridges going back along the sides of the nose and forward fuselage, much like the F-22 and F-35 and X-36, somewhat like the Blackbird and X-45. The Russians claim these ridges reduce the RCS of the Platypus. However, the rectangular engine intakes, the engine’s compressor blades placed not too deep in a non-serpentine intake, plus all the surfaces at right angles (weapon pylons and wings, engine pods and wings, engine pods and belly, horizontal and vertical tail fins), probably make the Platypus’ RCS too great for chines to really make any significant dent in it.
Now, you will notice that many of these modern stealth aircraft have curved surfaces, not facets. You may say, “But we just saw how curved surfaces reflect energy in more directions!!!”. This is true, but it’s all right, for two reasons:
One; If the curved surface is at an angle to the radar source, it will reflect the beam away. Remember the “wedge”. Sure, a slightly curved wedge might reflect the beam in more directions than a straight-sided wedge, but those directions will all be AWAY from the radar source anyways, so you’re ok. As long as the aircraft is, overall, fairly flat (that is, made up of few slightly curved surfaces (ideally just a top and a bottom) that meet at a sharp edge), it will only strongly reflect radar if it comes from straight above or straight below (and if you’re flying right over a radar station or right under a fighter plane, then you’re in trouble already).


Now hopefully you can see why flying wings have such low RCS.
The second reason why curves are ok is that, on modern stealth planes, their radius is rarely constant. They never look like circles when seen from any angle, and no part of the surface is spherical or cylindrical: they always look like squashed ellipses blended together with hyperbolae and parabolae. What that means is, as the orientation of a curvy stealth airplane changes in relation to you as it flies around, the part of the curvy surface that is perpendicular to you RIGHT NOW will have a different radius of curvature than the part that was perpendicular to you one second ago. As the airplane flies, the part of the surface that is perpendicular to you changes. But since each point in the surface has a different local radius of curvature, the parts that are reflecting energy back at you keep changing their radius of curvature as the airplane moves. Why is this important? Because something with a small radius of curvature (something very curvy) reflects less radar back at you than something with a large radius of curvature (something flat).

This means that the amount of radar energy being returned by the airplane keeps fluctuating. So even if a radar IS perpendicular to the surface (like a fighter plane right above it or a radar station on the ground right below it), it will be hard to get a radar lock, or even to tell the airplane from the random static around it. Non-constant radii of curvature ensure that the radar reflection changes a lot as the airplane moves (flies), making it hard to lock onto.

In summary...

... a stealthy shape will reflect radar back to the receiver only if the receiver is in one of a few directions. You can minimize the number of directions by having your airplane made of flat (or flattened) shapes, and by having groups of parallel edges (all of the B-2’s edges line up along only two directions). A wedge-like shape (be it the pointy front of the F-117 or the sharp chines on the Blackbird or X-47) tends to deflect radar light away from where it came from. If nothing else, hiding particularly radar-reflective parts (like the fan in the front of your jet engine) goes a long way. And use curves with non-constant radii, since this will cause your radar return to fluctuate, making it harder for it to stand out from the background noise and for the enemy to get a radar lock.

The Future

What is the future of radar stealth? Two emerging technologies that may become used more widely are plasma stealth and active signal cancellation.
Remember how I said that radar wave cancellation works like noise-reducing headphones? The idea is that a wave of a certain profile, when mixed with a copy of that wave that is inverted or half a wavelength out of phase, is mostly cancelled out. Noise-reduction headphones work by “listening” to the ambient noise with a microphone, and then playing this noise to you but inverted and/or out-of phase, so that the noise it plays cancels out the real noise. Now, stealth aircraft skins are made of multiple layers, so that the radar waves bouncing off the internal layers are half a wavelength out of phase with (that is, an inverted form of) the waves that bounce off the surface. But what if an aircraft could sense the incoming radar wave and figure out what direction it’s coming from, figure out what the aircraft’s radar reflection wave would look like, and then emit an inverted / out-of-phase version of that wave? The first part – detecting the radar wave and where it’s coming from – is easy and already an important part of most combat aircraft’s defenses. The second part – figuring out what your reflection would look like to that radar and from that direction – is trickier but not impossible. The third part – emitting a radar burst that cancels out the original – is much harder, because this burst must stop as soon as the original burst ends, otherwise your plane is sending out radar waves that are not canceling anything out and so can act as a beacon, signaling enemy radars to your airplane’s presence and location. It is also extremely important that the emitted radar-canceling wave matches the naturally-reflected waves very well, otherwise instead of cancellation, you get MORE radar waves coming from your airplane. Because the direction, phase, and timing of the emitted radar waves must be very precise, this system is significantly more complicated than the noise-reduction headphones it mimics. However, there are unconfirmed rumors of such systems being in use on the Rafale (the latest French fighter jet) and on the B-2.
“Plasma stealth” is not as spectacular as it sounds. Basically, it has been found that ionized air does a good job of absorbing radar waves. I actually have personally worked with trying to ionize air flowing around an object, even a wing – in my case, though, this was done for the sake of aerodynamic advantages, not stealth. Still, the fact is, with the right electric field, air flowing over a surface can be ionized easily and with little power. All you need is a high-voltage electric field. (In my case, I used a high-voltage high-frequency alternating field between two conductive strips, one of them exposed on the surface and one of them hidden just under the surface no more than a couple millimeters away). The strong electric field ionizes the air that flows into it – pull electrons one way and atom nuclei the other way. But the field can be set up so that the ions don’t actually hit the electrodes (if the field’s frequency is high enough, for example, the ions just vibrate and don’t really move very far), so you only need really low currents – not much power at all. And you don’t have to ionize air right on the surface of the airplane: You could also have some air from the engine intake or exhaust be fed through a strong electric field somewhere inside the airplane, which would ionize this air, and then pump it out somewhere near the front of the airplane. From there, it would be blown back, bathing the airplane in a shroud of ionized air. Or some completely different ionization system could be used, I don’t know. But the point is, once you have ionized air around your airplane (something that is quite doable), it might be much harder to pick up by radar. Again, there is no evidence that such an idea is operational or that it is even being tested on real aircraft, but that doesn’t mean it’s not, and it might be in the future.