MEMS and Advanced Radar 1. Hello, I am glad to be here today to talk to you about a marriage between RADAR and MicroElectroMechanical devices (MEMs). MEMs are important because they are built using low cost integrated circuit processing technology. In recent years, advanced radar has come to mean solid state electronically scanned Arrays -- Active ESAs. I believe there is a cheaper, lighter alternative for some applications. I am going to talk about LARGE ESAs which use MEMs components. They will bring down the COST, WEIGHT, AND POWER CONSUMPTION of these radars. While retaining many of the high-end capabilities including beam agility. This technology will open new windows of radar application. And MEM devices will open up new options for waveforms. I will start with a review of active ESA's. And then I will compare passive and active arrays, and I will introduce a new idea: the space fed MEM-tenna. And we will have a look to see what a MEM filter might do for multi-mode radar. 2. Pictured here is the solid state active electronically scanned array (ESA) for the nose of a modern fighter. It operates in X-band, and scans electronically. It provides control of the amplitude and phase at every element. These elements are typically spaced at a half wavelength. So there are a LOT of them when the wavelength is one inch. This array is about 1 ft by 1 ft. EACH element is backed by a solid state T/R module, which is why it is called an ACTIVE ESA. 3. A typical T/R module contains a solid state transmitter, receiver, phase shifter, and gain control. The T/R module brings the transmitter and receiver to the array face. This cuts transmit power loss and receive G/T, relative to passive array designs. The power density achievable by such arrays is hard to match in applications which have limited space available. The current availability of these T/R modules is much due to earlier DARPA efforts in the MIMIC and MAFET programs. A variety of programs (GBR, F-22, JSF) have struggled (and with some success) to bring the solid state T/R modules to a form factor, performance, and cost that allow active ESAs to be brought into production. Such arrays provide an outstanding sensitivity through high power per element, and great flexibility in waveforms. But even in production, they cost $4M to $8M per square meter. They are also heavy, about 700 pounds per square meter in typical modern airborne configurations. And Ground-based variants weigh even more. But the excellent performance and flexibility of an active ESA will NOT be affordable in EVERY application. So let's see what happens when MEM technology is applied to making RF switches. And RF switches are used to make phase shifters. RF switch development started with COTS hardware - the digital micromirror. 4. The digital micromirror device is used in projection TV systems. The DMD contains an array of movable micromirrors that function as an array of optical switches. This suggested an RF switch which uses a DC potential to pull down a membrane such as the I-beam shape in this bowtie switch. 5. A cross-section shows a flexible metal membrane in blue. It could also be in a cantilever configuration. The device acts as either a variable capacitor or RF switch. The electrostatic actuation means that the power required to maintain a switch setting is negligible. In contrast, a diode RF switch normally consumes 2 to 20 mw per diode. For a MEM switch the only power consumption is due to the charging and discharging of the very small capacitance of the actuator. This is a reduction for 2 orders of magnitude in power. So how do they perform? 6. Data from a representative X-band switch shows loss of 0.1 dB at X-band with more than 15 db isolation. 7. Switching times are expected to be about 2 microseconds, with 30-60 volts actuation. Power handling up to 9 Watts at X-band has been reported by Northrop-Grumman, and 4 Billion switch cycles have been completed. For electronic scanning radars these switches may be combined to make a phase shifter. 8. Here six switches used to build 2 bits of a 4-bit phase shifter using microstrip topology. Raytheon constructed this component under DARPA contract on 21 mil high resistivity silicon. The MEM manufacturing technique is applicable to smooth surfaces other than silicon, opening an extensive range of low cost and low loss components. Let's look at the phase shifter results. 9. This shows .6 to .9 dB loss over the entire 6-10 GHz range for the 2 bits. Adding 2 more bits has provided a 4-bit shifter with 1.5 to 1.7 db loss, with very good match. 10. Let's compare the performance of four phase shifter types. Power consumption includes the driver circuitry. With the exception of switching speed, the performance of MEM phase shifters is superior to current technologies in most categories. And while the slow switching speed is a limitation in some High PRF radars, it suitable for medium and low PRF waveforms. So we are ready to examine an ESA which uses MEM phase shifters. I would like to now draw your attention to LARGE arrays. They are of interest when detection requirements demand a large power-aperture product, but power is limited. Or when fire control accuracy demands small beam dimensions, but weight or power, or both, are limited. We will look first at constrained feeds for such arrays, and then look at space fed systems. MEMs have an application to both. 11. Because of volume constraints, ESAs are typically built around a constrained feed manifold. Active ESAs have a single T/R module per element, providing elemental control for low sidelobes and full 2-D scanning. A notional Passive ESA (or PESA) provides passive phase shifters at each element and it uses an order of magnitude fewer T/R modules. The obvious advantage in this configuration is cost, but considerations of prime power consumption have frequently made the Active ESA the selected configuration. The characteristics of MEMs suggest we have a new look. 12. A comparison of active element arrays and sub-arrayed passive ESAs using MEMs can be done by taking into account the losses and power consumption requirements of each alternative. These charts, which I will not discuss in detail, evaluate the total system power required for equal performance. A MEM-based PESA was compared with an active ESA using a 16 element sub-array. The next chart shows the result. 13. This chart examines the total power consumption for different LNA powers. The LNA power here refers to both the LNA and the phase shifter power requirement. To the right are the large arrays. For large constrained feed arrays with power per element of 30 mw or more, AESAs are competitive from a power consumption point of view. But the array cost may be an order of magnitude larger. In weight the systems are likely about the same because of the structure. For larger arrays, with lower power per element, the MEM passive array offers lower power consumption as well as lower cost. And there are further cost and weight savings possible if we re-consider the basic array architecture and replace the constrained feed with a space feed. 14. This chart summarizes the characteristics of the two types of feed. In the space feed, a horn, or series of horns, feeds an array of phase shifters which electronically steer the beam. They form an electronically agile lens. As the schematics make clear the waveguide or microstrip network of the constrained feed makes that it far more expensive than the space feed. Prior to MEMs being available, lens phase shifters were ferrites or diode switched RF circuits. While uncommon in US Systems (Patriot being the prominent exception), space feeds are common in Russian fire control radars. They use ferrite phase shifters and are heavy ground-based systems like the SA-10. But new materials and the use of MEM phase shifters open up the option for lightweight systems. 15. Two obvious potential applications for a lightweight lens are the aerostat and ship radars. Lots of volume and a desire for fire control quality narrow beams. Weight is important for both. DARPA looks forward to hearing your ideas on other applications. Now I would like to show you first a 1-D scanning space fed lens, the RADANT lens, and then a 2-D lens, the MEM-tenna. 16. The Radant Lens is a 1-D scanning lens that is an excellent application for MEMs RF switches. Scanning results from a linear phased gradient introduced in the E-plane at the output. The Conductive septums you see are separate multiple circuit layers that are periodically loaded by RF switches. Currently the switch is a PIN diode. Thirteen copper clad Teflon-glass circuit layers are typically employed for a 4-bit phase shifter. Well suited to assembly automation and using low cost and light materials, this type of lens is suited to large antennas with scan in one dimension. Transmission loss is 0.9 db at X-band. 17. We will examine the impact of lightweight technology by comparing previous RADANT efforts (not shown here), and the results for a small sub-array completed last year, which you see behind me. The Naval Research Lab funded a 3.2 m2 lens for shipboard application which weighed about 275 lbs./m2. This is less than the 700 lb figure I mentioned for active ESA's, but still high. We estimate that if Kapton septums were used, and graphite composite frames, lens weight including support and electronics would be 77 lbs./m2 in a 10 m2 size array. The NRL diode power consumption was about 1500 W/m2. If the diodes are replaced by MEMs, the power consumption will be less than ONE Watt/m2. To confirm these predictions the array shown above, which is 1.2 square feet was built. It scans + 60 degrees and weighed 26 lbs. The lens itself was 98 lb./m2. In summary, a 7-10 fold weight reduction compared to an active array. Cost is estimated to be $125000/m2, more than a factor of 10. In applications needing 2-D scanning, RADANT lens cost doubles, as does loss. But now consider a new concept: the MEM-tenna 18. The "MEM-tenna" concept combines MEMs phase shifters in a 2-D space fed lens. It utilizes multiple feed horns to increase the instantaneous bandwidth. To further reduce weight and cost, the MEM-tenna utilizes an optical projection system, based on the MEM digital mirror. It sends beamsteering commands to each phase shifter, thus eliminating all the control wiring in the array. 19. The DMD transfers signals to the array by rotating the micromirrors to reflect light onto and off of the detector diodes. At each phase shifter a photo detector diode receives and decodes the phase command. Using a 2-bit MEM phase shifter suitable for large arrays, phase shifter loss will be about 1 db at X-band. SPO is developing MEM phase shifters for this antenna, and plans to build 11000 in FY-00, and use them to construct an antenna in FY-01. But I am not done with MEMs radar applications yet. 20. The components shown above incorporate MEM technology into microwave circuit components and allow the possibility of very small low cost microwave filters. The hardware shown was developed for communications equipment by people now in DARPA ATO, and effort to extend the frequency to the microwave range in underway. These components open the way to create T/R modules which will allow simultaneous transmit and receive operation in a radar ESA. This concept has been demonstrated in mechanical scan radars, with great success, but the filter size and cost have always been too great for application in an active ESA. 21. What can I do with simultaneous Transmit and Receive? The Simultaneous Transmit and Receive, or STAR, waveform interleaves several signals, each at a different RF. On receive, a tunable notch filter eliminates the RF currently being transmitted. This allows the transmitter to run continuously, jumping around in frequency. This CW operation increases the transmitter efficiency. For the four channels shown a single radar effectively becomes four. And brings a number of advantages. 22. For starters, the STAR waveform is a nightmare to jam. And, spread over a few GHz bandwidth at X, the waveform reduces target scintillation and multipath. STAR can eliminate range and velocity eclipsing in MPRF radar systems. The frequency diversity during single dwells makes high resolution measurements possible. Interleaved waveforms supporting different modes such as SAR and GMTI mean a true multi-mode radar system is possible. Simultaneous multifrequency multi-angle beams will be possible. Even high capacity AJ data links from the same antenna could be integrated, and truly interleaved air-to-ground and air-to-air modes. The possibilities seem endless. SPO is planning to study and implement STAR technology in a small X-band sub-array in FY00-01. The sub-array could form the basis of Pod radars for such diverse platforms as Global Hawk and business jets. 23. I hope these suggestions for MEM applications excite you and stimulate your imagination, as they have mine. They may well form the backbone of the next step forward in radar. Your good new ideas are always welcome at SPO.