Search results

The objective of this paper is to demonstrate a method for producing embedded horizontal micro‐channels using a commercial line‐scan stereolithography (SL) system. To demonstrate that the method is repeatable, reproducible and capable of producing accurate horizontal micro‐channels, a statistical design of experiments was performed.

Demonstration of the technique was performed using a 3D Systems Viper si2^TM SL system and DSM Somos^® WaterShed^TM resin with polytetrafluoroethylene (PTFE)‐coated wire having diameters of 31.6 and 57.2 μm. By embedding the wire and building around the insert, the down‐facing surfaces were supported during fabrication enabling accurate fabrication of embedded micro‐channel geometries. The fabrication method involved first building an open micro‐channel, interrupting the SL process and inserting the wire, and then capping over the wire with multiple layers. After fabrication, the part with the inserted micro‐wire was post‐cured to harden any uncured resin around the wire. The micro‐channel was produced by simply pulling the wire out of the part. Scanning electron microscope images were used to examine and measure the geometries of the fabricated micro‐channels, and characterization through a statistical analysis was accomplished to show that the process was capable of producing accurate horizontal micro‐channels.

The measured data showed that the micro‐wires were successfully removed from the channels, leaving high quality micro‐channels, where the mean measured diameters for each wire were 2.65 and 2.18 μm smaller than the measured wire diameters (31.6 and 57.2 μm). Based on the statistical results, it is suggested that the method described in this work can rapidly produce repeatable and reproducible circular, embedded, and accurate micro‐channels.

The method developed in the current work was demonstrated on simple straight channels and a statistical study was used to show that the process is capable of repeatedly and reproducibly producing accurate micro‐channels with circular cross‐section; however, future studies are required to extend these procedures to more realistic and complicated geometries that may include non‐straight channel paths and non‐circular cross‐sectional geometries. The process can be used for micro‐channel fabrication with not only circular cross‐sectional geometries as shown here but potentially with a wide range of additional cross‐sectional geometries that can be fabricated into a PTFE‐coated micro‐wire.

This work demonstrates a process using commercial line‐scan SL and embedding a PTFE‐coated micro‐wire that is subsequently removed for producing repeatable and reproducible horizontal embedded micro‐channels of circular cross‐sectional geometries.

The purpose of this paper is to investigate the use of embedded Shape Memory Alloy (SMA) nitinol wire for the enhancement of vibration and damping characteristics of filament-wound fiber-reinforced plastic composite hollow shafts.

The plain Glass Fiber-Reinforced Plastic (GFRP) and plain Carbon Fiber-Reinforced Plastic (CFRP) hollow shafts were manufactured by filament winding technique. Experimental modal analysis was conducted for plain hollow shafts of C1045 steel, GFRP and CFRP by subjecting them to flexural vibrations as per ASTM standard C747, with both ends clamped (C-C) end condition to investigate their vibration and damping behavior in terms of first natural frequency, damping time and damping ratio. Nitinol wires pre-stressed at various pre-strains (2, 4 and 6 per cent) were embedded with CFRP hollow shafts following same manufacturing technique, and similar experimental modal analysis was carried out by activating nitinol wires. The first natural frequencies of all the shaft materials were also predicted theoretically and compared with experimental measurements.

Among the three materials C1045 steel, plain GFRP and plain CFRP, the vibration and damping behavior were found to be the best for plain CFRP. Hence, CFRP shafts were considered for further improvement by embedding nitinol wires at pre-stressed condition. For CFRP shafts embedded with nitinol wires, the damping time decreased; and damping ratio and first natural frequency increased with increase in percentage of pre-strain. In comparison with plain CFRP, 7 per cent increase in first natural frequency and 100 per cent increase in damping ratio were observed for nitinol embedded CFRP shafts with 6 per cent pre-strain. Theoretical predictions of the first natural frequencies agree well with the experimental results for all the shaft materials.

The effect of nitinol on vibration and damping characteristics of filament wound hollow CFRP composite shafts with different pre-strains has not been studied extensively by the previous researchers. This paper addresses the effect of embedded nitinol wires pre-stressed at three varied pre-strains, that is, 2, 4 and 6 per cent on the vibration and damping characteristics of composite hollow CFRP shafts manufactured by filament winding technique.

Increasing speeds combined with the level of integration that can be obtained with advanced IC technology has dramatically changed the interconnection requirements for high performance electronic systems. With much of today's circuitry being implemented in custom silicon, IC technology has allowed both a dramatic reduction in size and a tremendous increase in performance. However, in terms of the interconnection problem, the by‐product of advanced IC technology is a new generation of IC s that often require several hundred I/OS, exhibit rise times of 150 ps to 300 ps, and dissipate several watts per device. As demanding requirements are placed upon circuit boards, the complexity of the design task increases dramatically, since a working solution must simultaneously address interconnection density, signal integrity and thermal performance. This paper examines embedded discrete wiring technology as a high density solution that meets the requirements necessary for transporting high speed digital signals.

Flex‐rigid circuits have been used for many years, primarily by the military, as a method to reduce the size and increase the reliability of electronic systems. However, in today's emerging designs where high speed ASICs are often the dominant components, flex‐rigid circuit assemblies are now an attractive solution for providing high density transmission line interconnects from board to board. Much of today's circuitry is being committed to ASIC designs to increase both circuit density and speed. Following this path, designers are faced with the task of interconnecting high lead count SMT packages often with as many as 300 to 500 leads per device, each dissipating several watts. At these power densities conductive cooling through the circuit board is often a necessity, dictating the use of either metal cores or heat exchangers. To make efficient use of the core and minimise weight, designs generally require SMT packages to be mounted on both sides of the core with electrical communication from side to side. However, as more exotic core materials (carbon fibre matrix, beryllium, etc.) and liquid cooled heat exchangers are used, electrical communication through the core has become difficult, if not impossible, in some cases. Instead, high density flex‐rigid assemblies are used to partition the circuit, allowing the board to ‘fold’ over the core. This results in hundreds of signal lines that must cross the flex, obeying the electrical design rules dictated by the rigid sections to maintain impedance values and crosstalk margins. This paper focuses on recent work at AIT, producing high density flex‐rigid circuits using embedded discrete wiring technology to meet the above requirements. Using 0.0025 in. diameter polyimide insulated wire, as many as 100 lines per linear inch can pass over the flex region on a single layer. This generally results in a single flex layer where all wires can be referenced to a continuous ground plane from board to board. Controlled impedance is easily maintained due to the uniform wire geometry, and high frequency attenuation is significantly lower than on equivalent etch circuit designs due to the smooth surface finish on the wire. In addition, the high interconnection density offered by this technique reduces the overall thickness of the rigid sections, thereby minimising the thermal resistance to the core.

In future generations, electronic systems will rely extensively on advanced IC technology to achieve higher performance levels. However, with limits placed on the level of integration that can be obtained on a single IC, a need still exists for an interconnection hierarchy to provide the necessary density transform between system components. A recent addition to many high performance interconnection structures has been the Multichip Module. By eliminating the conventional IC package, MCMs have dramatically reduced the electrical length between devices, thereby minimising propagation delay, crosstalk, and attenuation. Although MCM techniques will offer many performance advantages, they also present many design challenges at subsequent levels of interconnection. This paper will focus on the requirements of MCM backplanes interconnecting several modules and, as a solution, will present recent work on advanced metal core substrates. MCM substrates provide a tremendous density advantage, however, the interconnection between modules is still a formidable task. Modules often have I/O densities of 300 to 500 leads per square inch and typically dissipate 10 to 50 watts per square inch. In addition, with sub‐nanosecond rise times, the distance between modules is often sufficient for signal paths to be treated as transmission lines. In an effort to meet these requirements, metal core circuits based on copper, copper Invar, and copper molybdenum have been fabricated using 0·0025 in. diameter embedded discrete wiring technology. Combined with a Kevlar surface layer suitable for wire bonding and blind laser drilled vias to access the internal wires, this technique offers many benefits. As many as 4 conductors can pass between holes on 0·050 in. centres in a single wiring layer only 0·018 in. thick. With the absence of interstitial vias, additional substrate area can be dedicated to create a sizeable thermal path, essential to conduct the heat from the MCM to an internal metal core. Together, these features have made this an attractive approach for interconnecting multichip modules.

In the never‐ending quest for speed, designers are now turning to digital GaAs integrated circuits both to extend the bandwidth of current designs and in some cases to generate a whole new class of products never before possible. The engineer well versed in high speed ECL design techniques generally understands the problems associated with this transfer to GaAs logic. However, even with the design task well defined, the exact solution for interconnecting devices is often difficult and stresses the capabilities of existing multilayer printed circuit techniques using conventional dielectric materials and processing. This paper examines the design task in detail, and will present recent developments in shielded discrete wiring techniques as a possible solution for GaAs packaging.

The purpose of this paper is to investigate the stress distribution in shape memory alloy (SMA) composite due to phase transformations in the fiber in view of the applied boundary conditions on the matrix.

A consistent homogenization of a SMA wire‐reinforced polymer composite volume element undergoing quasi‐static deformation was performed and SMA wire‐matrix interface behaviour was presented. For the SMA wire, a one‐dimensional phenomenological constitutive model was used. Eshelby's inclusion theory was employed for homogenization. A strain averaging approach was reviewed in which the average strain was substituted back to obtain the expressions for the effective stiffness, the inelastic strain, and the average stresses in the constituent phases. In order to study the stress distribution in SMA composite and constituent phases (fiber and matrix) as a consequence of the SMA wire‐matrix interface effect, interfacial stress model was derived. Interfacial axial and shear stress distribution is characterized for forward and reverse phase transformations. Finally, the thermomechanical behaviours were computed by applying strain energy approach incorporating the interface effects.

The results presented show that due to the difference between the shear modulus of matrix and SMA wire, and because of the strain non‐uniformity at the SMA wire‐matrix interface, shear stress is developed within the matrix under the axial loading of the representative volume element (RVE). The shear stress increases more rapidly as the SMA wire radius is increased but not with increase in the length. However, the axial stress does not increase much with increase in the SMA wire radius and length. Further, the average stress equation of the RVE at the SMA wire‐matrix interface is effectively addressed. The modeling approach is successfully validated extensively for different geometric and volumetric parameters for different loading conditions. It is evident that the interface effect of SMA wire composites is SMA stiffness dominated due to the fact that the geometric parameters do not influence much the stresses as compared to the change in SMA wire stiffness.

The approach is based on modeling the fiber matrix interface effect using homogenization scheme. Further, the strain energy approach is applied to compute the stress‐strain response. This indicates the importance of modeling the SMA wire‐matrix interface effect, and in particular, the energy exchange between the constituent phases. The results have been compared for different geometric parameters as well as volume fractions of the constituent phases under different loading conditions.

AT the request of the Director‐General of the International Labour Office Mr. Petre Lupu, Rumania's Minister of Labour, has described the benefits brought to his country through setting up a Management Development Centre.

The objective of this work was to develop practical experimental techniques for monitoring corrosion in “difficult‐to‐test” conditions such as corrosion under insulation (CUI).

An electrochemically integrated multi‐electrode array namely the wire beam electrode (WBE) method has been used in combination with noise signature analysis for the first time to monitor the penetration of corrosive species under simulated corrosion‐under‐insulation conditions. Corrosion of aluminium exposed under insulation materials such as rock wool, glass wool, cotton wool and tissue paper has been successfully monitored.

A typical potential noise signature of a major potential jump from AA1100 WBE was observed which corresponded to the corrosive species reaching the WBE surface in WBE current distribution map. A good correlation between the galvanic current maps and the corroded surface was also observed.

The preliminary results suggest that the proposed novel electrochemical method is capable of monitoring CUI.

The paper describes a new electro‐optical board technology, based on the discrete wiring principle. Isolated copper wires are embedded in the circuit board to realise the electrical interconnections. Glass optical fibres are embedded to obtain optical interconnections. The technology allows for crossovers and for electrical and optical interconnections on one layer of interconnection. As the technology can be applied on the level of package or multichip module, circuit board and backpanel, it has the ability to offer a complete solution for chip to chip electrical and optical interconnections. The paper will describe the basic manufacturing technology of the boards. The benefits of the technology from a system designer's viewpoint will be addressed. The problem of coupling light in and out of the embedded optical fibres will be discussed and the realisation of a first on‐board optical link via embedded optical fibres will be described.