Mems switch pdf

A computer program was developed at Texas Instruments that accepts all of the constituent components of 5 , calculates both and the chromaticity coordinates Using this tool, different filter passbands were iteratively for the system primary colors, and plots the coordinates generated and analyzed to select the filter set yielding the on the CIE color diagram.

The program also will calcu- best combination of system white point, saturated primary late filter passbands given desired CIE color coordinates. Because most of the energy of the lamp is concentrated into a series of narrow spectral spikes as opposed to a Fig. Achieving a good system white point that is significantly E. Grayscale different from the natural white point of the lamp requires As discussed previously, the DMD is a binary device the sacrifice of energy in one or more of the red, green, or and is incapable of directly modulating the reflected light blue spectral bands, compromising the overall efficiency of intensity in an analog fashion, as is done with most LCD- the color wheel.

The natural white point of the lamp used based light modulators. Instead, binary weighted pulsewidth in this projector tends to be somewhat green. Achieving a modulation is used. In the DDE projector, each of the more natural white in the DDE required sacrifice of some primary colors has 8 bits of grayscale depth, allowing energy in the green passband, lowering. In terms of discrete color rendition, the for the system overlaid on top of the lamp spectrum.

The projector can produce discrete colors. The Keeping in mind that the DDE uses time-sequential color, triangle connecting the three primary color coordinates for a 60 Hz video refresh rate, the projector must display all represents the color gamut of the system. The area within bits of three colors in With equal color display the triangle is the set of all colors that can be generated periods and 8 bits per color of grayscale resolution, the by the projector.

Combining this value with the value for This value is comfortably above the DMD minimum the system, the other loss mechanisms, and the raw lamp switch and settling time of 15 s; however, it is below output in 3 results in a nominal on-screen output of the normal minimum displayable bit time, which is approxi- lumens.

This value equates to a system-level efficiency mately 30 s. This minimum period is explained as follows. First, we be remembered that these are three-panel systems. The must wait for the mirrors to settle, which generally takes 16 time-sequential color technique significantly reduces the s.

The contents of the memory cells then must be cleared luminous efficiency, but the high efficiency of the DMD to turn the pixels off; this process, called a clear, takes itself permits a competitive single-chip system solution that 12 s. Only after both of these events are complete can would not have been possible with LCD technology.

As the a reset cycle 1. Reduction of eye integration errors through the use of bit-splitting techniques. DDE system electronics block diagram. Under normal operation, following the application of One effective technique that Texas Instruments developed the reset pulse, the bias would be reapplied to complete the for mitigating this effect is called bit splitting.

As and spread out across the color field display time, as mentioned previously, the flat-state reflected light is ideally shown in Fig. This technique minimizes the integration out of the projection-lens pupil, resulting in dark mirrors. General Layout image. As mentioned above, the projector accepts 24 bits of parallel digital video data 8 bits per integration of the correct grayscale value, which results color , along with a pixel clock and various timing control in momentarily visible flashes.

This phenomenon can be signals. These data are subjected to several preprocessing especially prominent along spatial grayscale boundaries in operations and reformatted prior to storage in a memory the image where major bit transitions occur.

For example, buffer, where they awaits transfer into the DMD. If all of the displayed B. Video Processing bits that compose code 01 are displayed at The first functional block shown in Fig. YUV formatting separates the luminance and chrominance information and allows the chrominance information to be subsampled. This subsampling reduces the overall bandwidth requirements and has minimal impact on image quality, as the eye is much less spatially sensitive to chrominance information than it is to luminance infor- mation.

This task is easily accomplished by multiplying the YUV data by a 3 3 conversion matrix, as shown at the bottom of the page. In addition to the CSC process, a number of picture control functions are implemented in this processing block as well. All of the coefficients in the above matrices are maintained in the DDE control processor and multiplied Fig.

Eight-bit versus bit de-gamma quantization error comparison. Note that the CSC processing block is easily bypassed by programming the quantization errors. Most video signals have a nonlinear intensity In DDE, while 12 bits of grayscale resolution are brought mapping scheme applied to them to compensate for the out of the de-gamma section, only 8 bits of true resolution nonlinear transfer function of CRT monitors. Because the are displayed on the DMD, owing to the cost implications DMD is a linear device, this encoding scheme must be of displaying all 12 bits.

Higher effective grayscale depth backed out of the data prior to display. The gamma correc- is achieved by using an error diffusion filter. The four tion function imposed on most National Television Systems extra bits of grayscale information, representing the error Committee video signals is a power function that looks like between the displayable value and the de-gamma output the following represented for red data only : value, are added in a weighted fashion to the surrounding pixel values.

The weighting has the effect of smoothing for out grayscale gradients, yielding a contour-free image at for normal viewing distances [11].

The first equation limits the slope of the curve for low- intensity levels. To apply de-gamma, the equations are simply C. Data Formatting and Display Following the error diffusion process, the RGB video for data are passed to the data formatter section, where the for incoming data undergo bit-plane separation.

This process reorganizes the video data into a series of 24 To implement this function in hardware, a RAM-based 1 arrays, with each array representing one of the 8 bits of lookup table LUT is provided that maps input data having grayscale information from each color.

This process can be 8 bits of resolution to output data having 12 bits of easily conceptualized by examining Fig. A series of 24 resolution. The 24 registers are then unloaded into their respective bit Fig. In the formatter, two such register arrays exist bit output quantization. The bit curve lies nicely on so that one array can load while the other array dumps its top of the ideal curve, while the 8-bit curve exhibits large contents into the memory buffers.

Hue Contrl. Picture of complete DDE. DDE data formatter block diagram. The memory buffers are also arranged in a ping-pong fashion, so that one buffer may be loaded with incoming video while the other buffer is read out to the DMD. As shown in Fig. Several buffers into the formatter hardware and is remapped onto very compact, single-chip engine designs have created an a pin data bus.

At the other end of the ap- clock cycles. This architectures are producing images of unprecedented uni- fast load time allows tremendous leeway in optimizing the formity, color fidelity, brightness, and stability [12]. Recently, TI intro- pleasing image.

The loading and time weighting of the bit- duced a new XGA class device mirrors plane data are controlled by a complex series of sequencers based also on the 17 m pitch mirror. Several architectural that generate the memory and DMD addressing and access changes to the mirror structure itself have been incorporated protocols at the proper time in relation to the incoming into production parts that substantially enhance device video vertical sync timing.

Its nominal shrinking it [13]. The gap between the mirrors also was performance characteristics are summarized in Table 2. Other improvements to the mirror formance characteristics of this single-chip engine permit it structure and to device packaging are also under way.

At the stream of new ideas for improvement at both the its introduction, the DDE was among the first SVGA- device and system levels, as well as for new applications, class utility projection engines available in the market. To continues at a rapid pace.

He joined Texas paper. Instruments in , where his initial work was to develop charge-coupled-device CCD image sensors. In [2] R. SPIE Europto, vol. MEMS array of fast digital light switches integrated on a silicon address —, SPIE, vol. Pettitt, A. DeLong, and A. Robert E. Meier received the B. University, Dallas, TX, in Prior to his , , pp. Robertson and J. Fisher, in Television Engineering the B. Benson, Ed.

He is a member of the Group Technical Staff [10] B. Critchley, P. Blaxtan, B. Eckersley, R. Gale, and M. Markandey, T. Clatanoff, and G. DMD pixel mechanics simulation tool, which he has used extensively [12] L. SPIE Mr. Engineers and the International Society for Optical Engineering.

Florence, R. Miller, and T. Michael R. Pennsylvania, Philadelphia, in He has been a reliability Plano, TX. Singh, T. Asutkar, S. In 2nd international conference on advances in electrical, electronics, information, communication and bio - informatics AEEICB , 27 — 28 Feb.

Sahu, A. Saffari, H. Micro and Nanosystems, 9, 80— Article Google Scholar. In The 25 th annual international conference on mechanical engineering, 2 — 4 May pp. In The 16th international conference of iranian aerospace society, Tehran, Iran. Angira, M. Microsystem Technologies, 21, — Lobontiu, N. Compliant mechanisms: design of flexure hinges. Book Google Scholar. Tang, M.

Analog Integrated Circuits and Signal Processing, 40, — Moseley, R. In 19th IEEE international conference on micro electro mechanical systems, pp. Pirmoradi, E. Design and simulation of a novel electro-thermally actuated lateral RF MEMS latching switch for low power applications. He, X. In Proceedings of 5th global symposium on millimeter - waves, 27 — 30 May pp. Li-Feng, W. Journal of Micromechanics and Microengineering , 25 6 , Article Id Bhatasana, P.

Need an account? Click here to sign up. Download Free PDF. Dale Ostergaard. A short summary of this paper. Coupling between electrostatic forces mechanical systems MEMS leads to complex geometrical and mechanical forces can be characterized by mapping the configurations of MEMS [1], which requires new efficient capacitance [11] as a function of the motion of the device.

Lumped models Completely modeling the fully coupled system, TRANS [2] and semi-analytical approach [3] are no longer converts electrostatic energy into mechanical energy and applicable for the devices, where fringing electrostatic visa versa as well as store electrostatic energy. Input for the element the more accurate treatment of electro-mechanical systems consists of a capacitance-stroke relationship that can be by FE [7], boundary element and hybrid [8] methods: derived from electrostatic field solutions.

The element can sequential physics coupling [7], reduced order fully lumped characterize up to three independent translation degrees of Figure 1 [9], distributed mechanical with one 1D freedom at any point to simulate 3-D coupling. Thus, the transducer [2] or distributed mechanical with multi-D electrostatic mesh is removed from the problem domain and transducer [10].

The new experienced convergence issues when applied to the distributed transducer, TRANS element, can be used difficult hysteric pull-in and release analysis [10].

We with both lumped and solid mechanical models Figure 4. The principle of virtual work lays in the basis of the element formulation coupled with the electrostatic degrees of freedom. The energy change is associated with the change of potential distribution in the system, which produces structural reaction forces.

In a bending electrode problem, like an optical switch, obviously, a single TRANS element can not be applied. Fortunately, when the gap is small and fringing is not significant, the capacitance between deforming electrodes can be practically modeled reasonably well by several k capacitors connected parallel.

The capacitance matrix, C , depends on the element However, a new treatment is necessary for more geometry.

The nodal electrostatic reaction force, f, can be general geometries. The new strongly-coupled field transducer du benefits from both methods. This is numerically difficult problem because of the transducer elements as well as transducers and mechanical bifurcation of static equilibrium, which may lead to stability elements balance each other. The mesh, including the air issues. The tip vertical displacement vs.

This procedure current impulse load [13]. The results agree very well with is highly non-linear and huge displacements are allowed for [13] and Runge — Kutta time integration see Figure 8. As interaction. The new element is compatible with structural opposed to EMTGEN, the air region is not restricted to the solid and lumped FE models and thus can be effectively gap, thus, fringing fields are accurately simulated. Contact used for analysis of complex distributed and ROM electro features can be modeled by contact elements, for example mechanical systems.