New for October 2022: Dr. Eli Brookner was a well know lecturer on phased arrays, we gathered some YouTube videos of him on this page.
New for September 2018: here's a page on cylindrical phased arrays
New for December 2017: here's a phased array tip, from Colin!
Click here to go to our page on AESAs
Click here to go to our page on PESAs
Click here to go to our page on T/R modules
Click here to go to out page on time delay units (TDUs)
Click here to go to our page on grating lobes
Click here to go to our page on RMS error calculations
Click here to go to our main page on antennas
Click here to go to our main page on phase shifters
Click here to go to our page on ferroelectric phase shifters
Go to our download page and get the phase array spreadsheet!
Check out the aperture gain rule of thumb at the bottom of this page!
Phased arrays are the opposite of microwave career killers. Much of the material on this page was contributed by Arne Lüker, a friend of Microwaves101! For two excellent primers on phased arrays to our book page and pick up a copy of Stimson's or Skolnik's books.
Applications of phased arrays
Phased array antennas can be electrically steerable, which means the physical antenna can be stationary. This concept can eliminate all the headaches of a gimbal in a radar system. It can keep an antenna locked onto a satellite, when the antenna is mounted on a moving platform. It is what allows a satellite to steer its beam around your continent without having to deal with the "slight problem" associated with trying to point things in space where every movement would require an equal and opposite mass to move in order to keep the satellite stabilized. A phased array receiver can be flush-mounted on the top of a commercial airplane's fuselage so that all of the happy passengers can receive satellite television!
You might be surprised to know that the inventor of the phased array was Karl Ferdinand Braun, sometime around 1905, for long distance radio communication by Marconi and Braun. Here is an example of an early L-band test array developed by Sperry Rand and evaluated by MIT Lincoln Laboratory in the early 1960s. Moving forward in time, the photo next to it is the well-known phased-array antenna for the Patriot missile. What does the acronym "Patriot" stand for? Phased Array TRack to Intercept Of Target. It replaced the "Homing All-the way Killer", or Hawk missile. There's some trivia you won't learn on Wikipedia!!!
Patriot radar, image from Wikipedia.com
So far there are not many consumer applications of phased arrays with the recent exception of some of the newer WiFi routers. This is because they can be quite expensive, due to the need for many microwave phase shifters and their control signals. On top of the phase shifter expense, phased arrays usually need a low noise amplifier at each element for receive operation, and a power amp at each element for a transmit signal. One consumer market that is developing for phase arrays is satellite television for vehicles such as RVs. For a couple thousand dollars, your kids can now watch eight Disney channels, while you tour the painted desert in your Winnebago. Life is good, especially if it appears on a small screen! Of course, the main driver for all developments in consumer technology is pornography, in this case, now you and your date can watch pay-per-view flicks on the Playboy channel from the comfort of your recreational vehicle!
The physics behind phased arrays are such that the antenna is bi-directional, that is, they will achieve the same steerable pattern in transmit as well as receive. In many applications, both transmit and receive systems are needed; the solution to this problem is known as the transmit/receive module (T/R module), which will be the subject for another day.
Phased array antenna properties
The principle of the phased array is to synthesize a specified electric field (phase and amplitude) across an aperture. The resulting beam approximates the Fourier transform of the E-field distribution. The individual antennas are frequently space about a half-wavelength apart. Sparse arrays use much larger element spacing of course but their behavior and utility are probably outside the scope of what you want to learn here.
Adding a phase shift to the signal received or transmitted by each antenna in an array of antennas allows the collective signal of these individual antennas to act as the signal of a single antenna with performance vastly different from the individual antennas in the array. Here is a list illustrating the some of the results of arraying many antennas.
- Power: The signal collected is the combined signal of all the individual antennas and so is stronger.
- Beam Shaping: The antenna pattern of the combined antennas can be much narrower than any of the individual antennas.
- Beam Steering: The direction of the peak sensitivity collective antenna can be altered without mechanically re-positioning the individual antennas. For an array with electronically variable phase shifters, you can switch the beam position as fast as you can switch the phase shifts. Big antennas move rather slowly.
- Reliability: For a single antenna, if the positioning system fails, you can’t point to anything except down the line of sight of the antenna. For the array antenna, if one antenna fails, all the rest continue to function and the collective pattern is modified slightly (called graceful degradation.)
- Weight: For airborne applications the weight of an phased array is less than that of a comparable rapidly-steerable, gimbaled, single antenna.
- Cost: A very large mechanically steered antenna may be replaced with a collection of less expensive smaller antennas without loss of resolution (although a single cryogenic receiver may cost less than a collection of cryogenic receivers) but cost comparisons are difficult without looking at the detailed system requirements.
- Multiple Beams: Utilizing the wide range of control provided by the phase shifters, you can synthesize multiple beam responses if you desire.
- Digital or mixer Option: You can actually do without the analog phase shifters by down-converting to the base band and then filtering and shifting the signal digitally. On the other hand you could shift the phase of the IF or LO signal instead of the RF signal. In either the Digital or IF/LO mode of operation, the complexity and expense of the individual receivers goes up as you have to distribute the LO signal to every antenna. For large arrays I’m not sure that it’s worth it but some automotive radars are frequency scanned.
But as usual you don’t get something for nothing.
- Scan Angle: If the target is low on the horizon, the mechanical antenna tilts over and the entire aperture collects the signal. When you phase up a phased array to point in that direction, the aperture's physical configuration (say parallel to the ground) doesn’t change but the effective collecting area of the aperture deceases. If the target that is on the horizon, the array is edge-on to the target and intercepts very little signal. On-the-other hand the phased array concept allows for the antenna elements to be positioned in a curved surface, like a dome, but, as usual, trade-offs need to be evaluated before proceeding.
- Bandwidth: Most phase shifters are designed to provide a constant phase shift over a band of frequencies. The nasty secret is that we have been approximating the delaying of a signal with the phase shifting of a signal. A delayed signal has a linear phase shift with frequency. For wide instantaneous bandwidth signals and a phased array set for a narrow beam width, the phase shift controlled beam will vary its pointing direction with frequency (squint). This can be enough to steer the beam entirely off the target, resulting in a greatly reduced collected signal. It’s not all bad. You can use this property to scan the beam by scanning the frequency.
The phase shift to line up waves before combining is 2π(d/λ) sinθ where d is the array element spacing, λ is the wavelength, and θ is pointing angle direction. Note that it is based on a frequency dependent parameter - wavelength. If instead you use delay to point the beam, the delay needed is (d/c) sinθ with c being the speed of light (propagation). Note here that the pattern is now not dependent of frequency.
- Mutual Coupling: Another approximation we have been using is that each individual antenna performs the same even when surrounded by other antennas. Unfortunately, this is not the case. The design must be modified to reduce the influence of mutual coupling.
Definitions and acronyms
First let's define a few terms and acronyms (which we'll also put in the Microwaves101 acronym dictionary):
ESA: electronically steered array (as opposed to a mechanically steered array or MSA)
AESA: electronically steered array
CPA: cylindrical phased array
PESA: passive electronically steered array
AOA: angle of arrival, also known as the look angle
ULA: uniform linear array
UCA: uniform circular array
UGA: uniform grid array
TDU: time delay unit
How to Design Phased Array Systems
Our buddies over at Keysight Technologies have uploaded a video on this very topic over on their YouTube page.
This video will provide a discussion on the most important considerations for phased array system design, especially popular for proposed 5G architectures. It begins with the basics of phased array design, then covers four key parameters of phased array architecture. It then shows how the far field pattern affects the overall system performance through an example, and what factors influence that far field pattern as part of the design process.
Dr. Murthy Upmaka from Keysight explains how to design phased array systems
Antennas for phased arrays
Phase shifters are mostly used in phased array antennas (radar systems) but now alsoin some automotive radars. It is well worth it to step a bit back to have a closer look on the antenna aspect.
An antenna should be viewed as a matching network that takes the power from a transmission line (50Ω impedance, for example), and matches it to the free space "impedance" of 377Ω. The most critical parameter is the change of VSWR (voltage standing-wave ratio) with frequency. The pattern usually does not vary much until the start of unacceptable VSWRs (> 2:1). For a given physical antenna geometric size, the actual radiation pattern varies with frequency.
The antenna pattern depicted in Figure 1 is for a dipole. The maximum gain is normalized to the outside of the polar plot and the major divisions correspond to 10 dB change. In this example, the dipole length (in wavelengths) is varied, but the same result can be obtained by changing frequency with a fixed dipole length. From the figure, it can be seen that side lobes start to form at 1.25 and the side lobe actually has more gain than the main beam at 1.5. Since the radiation pattern changes with frequency, the gain also changes.
Figure 1. Frequency effects
Array effects
Figure 2 depicts phase/array effects, which are yet another method for obtaining varied radiation patterns. In the figure, parallel dipoles are viewed from the end. It can be seen that varying the phase of the two transmissions can cause the direction of the radiation pattern to change. This is the concept behind phased array antennas. Instead of having a system mechanically sweeping the direction of the antenna through space, the phase of radiating components is varied electronically, producing a moving pattern with no moving parts. It can also be seen that increasing the number of elements further increases the directivity of the array. In an array, the pattern does vary considerably with frequency due to element spacing (measured in wavelengths) and the frequency sensitivity of the phase shifting networks.
Note: we've had a number of comments on an apparent mistake in this figure. Instead of fixing the figure, we'll tell you what's wrong with it according to Tom:
" I was looking at the section titled "Phased Antenna Arrays" and noticed a possible mistake with the middle drawing in Figure 2. The radiation pattern shown for 1/2 wave spaced antennas fed 90 degrees out of phase is actually the pattern of 1/4 wave spaced antennas fed 90 degrees out of phase.
The critical variable left out of the section on "Phased Array Antennas" was the influence of antenna spacing on the array's pattern. Given a ULA (uniform linear array), in broadside mode, the pattern is always symmetrical (figure 8 shaped) for any element spacing. Spacings at even multiples of 1/4 wavelength are also symmetrical in the endfire direction. Spacings at odd intervals of 1/4 wavelength are asymmetric in endfire mode gradually progressing to be symmetric as the element phasing rotates the beam around to broadside mode."
From another antenna guy Justin:
"He's right. That figure is wrong:
k*d = pi for half-wave spacing
To steer a half-wave spaced array out to end-fire you need a 180 phase shift:
necessary phase shift = k*d sin(theta) = pi*sin(90) = pi = 180 degrees
Those NASA guys don't know squat!"
(The original figure came from NASA...)
Figure 2. Phase/array effects
A linear phased array with equal spaced elements is easiest to analyze and forms the basis for most array designs. Figure 3 schematically illustrates a corporate feed linear array with element spacing d. It is the simplest and is still widely used. By controlling the phase and amplitude of excitation to each element, as depicted, we can control the direction and shape of the beam radiated by the array. The phase excitation, (n), controls the beam pointing angle, 0, in a phased array. To produce a broadside beam, 0=0, requires phase excitation, = 0. Other scan angles require an excitation, (n) = nkd sin (0), for the nth element where k is the wave number (2/). In this manner a linear phased array can radiate a beam in any scan direction, 0, provided the element pattern has sufficient beamwidth. The amplitude excitation, An, can be used to control beam shape and sidelobe levels. Often the amplitude excitation is tapered in a manner similar to that used for aperture antennas to reduce the sidelobe levels. One of the problems that can arise with a phased array is insufficient bandwidth, since the phase shift usually is not obtained through the introduction of additional path length. However, it should be noted that at broadside the corporate feed does have equal path length and would have good bandwidth for this scan angle.
Figure 3. Corporate fed phased array
Diamond versus square lattice
Coming soon!
How to avoid grating lobes
We now have a separate page on grating lobes, located here. With cool pictures!
A grating lobe occurs when you steer too far with a phased array and the main beam reappears on the wrong side. Elements must be spaced properly in order to avoid grating lobes. The equation for maximum spacing is a function of wavelength of operation and maximum look angle:
Thus for a 30 degree look angle, dmax is (2/3)xlambda, while for a 60 degree look angle, dmax is 0.54 lambda.
Calculating antenna gain in a phased array
Gain at broadside in a phased array is both a function of the individual element gain and the number of elements. The aperture gain is calculated by:
Here's a Microwaves101 rule of thumb contributed by Glenn:
The number of elements required in an electronically-scanning phased array antenna can be estimated by the gain it must provide. A 30 dB gain array needs about 1000 elements and a 20 dB gain array needs about 100.
The gain of the individual elements is a function of what radiator is used. This is a case where you don't want the element to have too much gain, because the entire idea behind a phased array is that you want to maximize scan volume; you don't want system gain to rapidly drop off as you move away from broadside due to the element pattern. In practice, most radiators used in phased arrays provide about six dB gain.
So, what happens to gain as you scan off of broadside? The gain drops as cosine of the angle. Thus at 60 degrees you are at 1/2 the gain at broadside, and when you get to endfire condition, gain is down to zero. This is the one problem with phased arrays that might make you want to reconsider a gimballed approach. To get a full 360 degree coverage usually takes four phased arrays, you could do that with a single, rotating antenna.
Time delay units (TDUs)
TDUs are used at the sub-array level in a phase array, to improve how the array performs over frequency. Time delay is required to get all of the phase centers to be approximately equal phase length to the receiver or exciter, otherwise the beam will distort over frequency. Check out TDUs here.
Learn about when to use TDUs versus phase shifters here.