Polyimide-based device-embedded substrates significantly increase the density and performance of electronic circuits by using a chip-stack structure with a thin-film insulation material. Typical thicknesses of our substrates are 0.2 mm for a 4-layer circuit embedded with one chip and 0.4 mm for an 8-layer circuit with 2 stacked chips. These substrates are produced using a co-lamination process, which is suitable for wiring within a high-layer-count circuit board and multi-chip embedding.

For this study, a test vehicle, 0.9 mm in thickness and containing a 3-chip stack, was produced and tested to ensure reliability. The reliability of the test vehicle was verified under the conditions including a three-cycle reflow test after soaking at 30°C 60%Rh for 192 hours (JEDEC MSL3), a HAST at 130°C 85%Rh for 336 hours, and a thermal cycling test -40°C to 125°C for 500 cycles.

We have developed a highly-miniaturized Si photonic transceiver module named "Optical I/O Core" for use in on-board and co-packaged optics. In an optical I/O structure called "optical pin", a laser diode and a driver IC are mounted on a 5 mm × 5 mm photonic integrated circuit. The data rate is 25 Gbps and the transmission media is a multimode fiber. High-temperature operation up to 85°C and multimode fiber links to alleviate alignment tolerance make the optical I/O core usable for various applications. In 2017, AIO Core Co., Ltd. was established to put the developed optical I/O core to practical use. A mass-production line is currently under development.

This paper considers augmenting training images composed of defect and background images in the training phase when a Region-Based Convolutional Neural Network (R-CNN) is applied to an appearance inspection. This approach is applied to the appearance inspection of plastic pieces produced by a press work and the obtained results are reported. In the proposed image composition, firstly, some typical defective patterns are cut from the actual images and then, these defective patterns are pasted to the actual background images by changing the size, rotation, colors, converting them to black-and-white, etc. In the experiments, 81 defective patterns are cut from 50 actual images and pasted on several background images producing 500 unique images by applying the image modifications. R-CNN is trained on these 500 images and then applied to the detection of 53 defective patterns in 40 images. The resulting detection ratio and hit ratio are 81% and 86%, respectively indicating that the suggested approach is promising for practical use with some performance improvements.

In the development of LED lighting and power electronics products, the high output characteristics needed to enhance their performance leads to a temperature rise of the semiconductor elements. Although it is necessary to attach heat sinks to semiconductor chips in order to prevent the chip temperature from rising excessively, it is also desirable for heat sinks to be as light as possible in order to reduce costs. Therefore, we propose a design method for cooling components to reduce their weight while keeping the chip temperature low based on applying a regression formula of the temperature-containing variables related to the size of the components. This paper demonstrates the effectiveness of this design method by comparing measurement data from LED lamps and the predicted value derived from this method.

This paper shows a design method for an Antenna-filter using a multi-layered suspended microstrip line (ML-SMSL) structure. The equivalent circuits using J-inverters, which are able to compensate for the elements of unwanted electromagnetic couplings, are introduced when the design parameters for the Antenna-filter with the ML-SMSL structure are calculated. Although the measured reflection characteristics for the fabricated Antenna-filter are a little different from those for the electromagnetic (EM) simulated one, the characteristics of radiation gain for the fabricated Antenna-filter are very close to those for the EM simulated one.