In this article, we propose a double-sided electromagnetic bandgap (DS-EBG) structure for glass interposers (GIs) with low substrate loss to suppress power/ground noise. For the first time, we ...validated wideband power/ground noise suppression in the GI using the proposed DS-EBG structure based on dispersion analysis and experimental verification. We experimentally verified that the proposed DS-EBG structure achieved the power/ground noise suppression (below −40 dB) between 2.5 and 8.9 GHz in the GI. Derived stopband edges, <inline-formula> <tex-math notation="LaTeX">f_{L} </tex-math></inline-formula> and <inline-formula> <tex-math notation="LaTeX">f_{U} </tex-math></inline-formula> based on the dispersion analysis, and 3-D electromagnetic (EM) simulation showed a good correlation with measurements. The effectiveness of the proposed DS-EBG structure on the power/ground noise suppression is verified by analyzing noise propagation in the power distribution network and coupling to the GI channel. Using the 3-D EM simulation, we verified that the proposed DS-EBG structure suppressed the power/ground noise coupling and improved the eye diagram of the GI channel. Finally, we propose a design methodology to broaden the isolation bandgap or miniaturize the dimensions based on the dispersion analysis.
In this paper, for the first time, we designed and analyzed channels between a graphic processing unit and memory in a silicon interposer for a 3-D stacked high bandwidth memory (HBM). We thoroughly ...analyzed and verified the electrical characteristics of the silicon interposer considering various design parameters, such as the channel width and space, redistribution layer via, and under bump metallurgy pads. In particular, we also considered the meshed ground planes used for the proposed transmission lines, which are microstrip and strip lines. Signal integrity (SI) of the proposed channels in the silicon interposer was successfully analyzed and verified using a full 3-D electromagnetic solver and circuit simulations. Based on the extracted lumped circuit resistance, inductance, conductance and capacitance parameters, we thoroughly analyzed the channel characteristics and identified the parameters that dominantly affect SI in relation to each frequency range. From the analyzed insertion loss and far end crosstalk, we verified SI of the silicon interposer by eye-diagram simulations in terms of eye-height voltage and timing jitter in the time domain. In the worst case, the eye-height voltage and timing jitter of the proposed microstrip lines are 0.911 V and 36.8 ps, respectively, with 72 mV of signal coupling. The eye-height voltage and timing jitter of the proposed strip line are 0.887 V and 42.1 ps with 34 mV of single couplings. We show that the proposed channels of the silicon interposer can successfully transfer data at a 2-Gb/s data rate. Finally, we propose concepts and solutions for the next-generation HBM interface with higher data rates up to 8 Gb/s.
This paper, for the first time, proposes a novel stochastic model-based eye-diagram estimation method for 8B/10B and transition-minimized differential signaling (TMDS)-encoded high-speed channels. A ...stochastic model describes a behavior of an encoder with respect to probability. The previous eye-diagram estimation methods are based on an assumption that each bit has the same probability for 1 s and 0 s. However, the assumption limits to estimate an accurate eye-diagram for encoded high-speed channels. We first propose and apply the stochastic model for two types of 8B/10B encodings: 8B/10B and TMDS. For verification, we design the 8B/10B and TMDS encoder within MATLAB. The transient simulation for the 8B/10B encoded channels requires 9700 and 6600 s, respectively. However, the proposed method only requires 23 s in both cases. Furthermore, in the bit-error rate, the transient simulation provides the bathtub curve up to 10 -2 due to processing time and computing resources. In contrast, the proposed method with the stochastic model provides the bathtub curve up to 10 -8 . In conclusion, this paper successfully proposes and verifies the stochastic model-based eye-diagram estimation method for 8B/10B-encoded high-speed channels.
In this paper, we propose glass interposer electromagnetic bandgap (EBG) structure to efficiently suppress power/ground noise coupling. We designed, fabricated, measured, and analyzed a glass ...interposer EBG structure for the first time. Glass interposer EBG structure test vehicles were fabricated using a thin-glass substrate, low-loss polymer layers, and periodic metal patches with through glass vias (TGVs) in glass interposer power distribution network. Using the dispersion characteristics, we thoroughly analyzed and derived f L and f U of the glass interposer EBG structure. We experimentally verified that the proposed glass interposer EBG structure achieved power/ground noise suppression (below -40 dB) between f L of 5.8 GHz and f U of 9.6 GHz. Derived f L and f U based on dispersion analysis, full three-dimensional electromagnetic (3-D-EM) simulation and measurement achieved good correlation. In the glass interposer EBG structure, tapered structure of the TGV and thickness of the low-loss polymer used for metal-layers lamination affected the noise suppression bandgap significantly. The effectiveness of the proposed glass interposer EBG structure on suppression of the power/ground noise propagation and coupling to high-speed TGV channel was verified with 3-D-EM simulation. As a result, the proposed glass interposer EBG structure successfully and efficiently suppressed the power/ground noise propagation and improved eye-diagram of the high-speed TGV channel.
In this paper, we measured and analyzed glass interposer power distribution network (PDN) resonance effects on a high-speed through glass via (TGV) channel for the first time. To verify the glass ...interposer PDN resonance effects on the TGV channel, glass interposer test vehicles were fabricated. With these test vehicles, glass interposer PDN impedance, channel loss, far-end crosstalk, and eye diagram are measured. Based on these measurements, glass interposer PDN resonance effects on the signal integrity of the high-speed TGV channel are analyzed. Due to low loss of the glass substrate, sharp high PDN impedance peaks are generated at the resonance frequencies. High PDN impedance peaks at the PDN resonance frequencies, which affect return current of the TGV channel, increase channel loss, crosstalks, and PDN noise coupling in the frequency domain and degrade eye diagram in the time domain. To suppress these glass interposer PDN resonance effects, a ground shielded-TGV scheme is proposed. The proposed ground shielded-TGV scheme includes two ground TGVs 200 μm away from the signal TGV considering the design rules and includes package ground underneath the glass interposer. Effectiveness of the suggested grounding scheme on the resonance effects suppression is verified with three-dimensional electromagnetic simulation. The proposed shielded-TGV design successfully suppressed the glass interposer PDN resonance effects that results in the suppression of insertion loss, shielding of the crosstalk, and improvement of the eye diagram of the high-speed TGV channel.
In this paper, we propose glass-interposer (GI) electromagnetic bandgap (EBG) structure with defected ground plane (DGP) for efficient and broadband suppression of power/ground noise coupling. We ...designed, fabricated, measured, and analyzed a GI-EBG structure with DGP for the first time. The proposed GI-EBG structure with DGP is thoroughly analyzed using the dispersion characteristics and estimated stopband edges, f L and f U . We experimentally verified that the proposed GI-EBG structure with DGP achieved power/ground noise isolation bandgap (below -30 dB) between f L of 5.7 GHz and f U of 11 GHz. Estimation of f L and f U using dispersion analysis, full 3-D electromagnetic (EM) simulation results, and measurement results achieved good correlation. Effectiveness of the proposed GI-EBG structure with DGP on suppression of the power/ground noise coupling to high-speed through glass via (TGV) channel is verified with 3-D EM simulation. As a result, the proposed EBG structure successfully and efficiently suppressed the power/ground noise coupling and improved the eye diagram of the TGV channel. Lastly, we embedded thin alumina film in the proposed EBG structure and achieved even broader power/ground noise suppression between 2.1 and 14.7 GHz.
A 3-D stacked high bandwidth memory (HBM) becomes a promising solution to satisfy the memory bandwidth for the processor. Due to its unique memory architecture that consists of tremendous number of ...input/output (I/O), it is inevitable to employee Silicon based interposer. Therefore, power distribution network (PDN) design and analysis of HBM interposer becomes one of the important step to guarantee the performance of an entire memory interface. Since the back end of line (BEOL) process technology of a semiconductor industry is applied for HBM interposer, the control of a metal density and the management of wafer warpage are required. Therefore, we designed and analyzed meshed and grid type of PDN for HBM interposer because of the limit of a metal density. In addition, we also designed and analyzed PDN both a single- and double-sided interposer. Because, a double-sided interposer has an advantage of a warpage management compared to a single-sided interposer. For the suppression of simultaneous switching noise (SSN), PDN impedance with a decoupling capacitor scheme must be properly analyzed. In this paper, a single- and double-sided HBM interposer is designed with the five layers and six layers respectively to analyze PDN impedance including though-Silicon-via (TSV). PDN impedance of HBM interposer is simulated and analyzed in the frequency range from 100 MHz to 20 GHz. Based on the designed HBM interposer, we shows the great potential of HBM interposer in terms of the reduction of PDN impedance to suppress SSN with a metal-insulator-metal (MIM) decoupling capacitor.
In this paper, we design and analyze the receiver channel of a glass interposer for a dual band Wi-Fi front end module (FEM). In RF systems, a RF sensitivity is the most important specification for a ...system performance and reliability. To guarantee a target RF sensitivity of the RF system, it is important to maintain a 50 Ohm impedance matching for RF channels. The 50 Ohm impedance matching of various transmission line structures is completely conducted for each layer of the glass interposer in a low band (2.4GHz - 2.5GHz) and high band (4.9GHz - 5.85GHz) respectively. Channel types for each layer are determined considering design rules and constraints. Moreover, we locate ground TGVs near signal TGVs as close as possible to suppress a ground-ground cavity resonance effect in the low band and high band. In order to test a suppression of the ground-ground cavity resonance, simulations that compare pitches between a ground TGV and signal TGV are conducted. Design considerations of receiver channels were analyzed and characterized by simulation results of channel insertion losses.
In this paper, eye-diagrams of High-Bandwidth Memory (HBM) interposer channel with crosstalk reduction schemes on 2.5D / 3D IC are estimated and analyzed. As data rate increases and metal-to-metal ...space decreases to achieve higher system bandwidth, crosstalk effects degrade the signal integrity. Therefore, estimation and reduction of the crosstalk effects are essential on HBM interposer channel. In order to estimate crosstalk effects in short time with high accuracy, PDA-based estimation method is proposed. In addition to the proposed method, wide space and guard trace with ground vias structures are suggested and compared based on the estimated eye-diagrams. With the estimated eye-diagrams, voltage fluctuation on DC levels due to the crosstalk effects can be analyzed. Worst and statistical eye-diagrams of interposer channels including crosstalk effects are estimated. Since the proposed method needs only output and crosstalk response of the channel, the proposed method can be applied to multiple wide I/O channels.
We previously reported the mean 6-year outcomes of total hip arthroplasty (THA) using fourth-generation ceramic-on-ceramic (CoC) articulations. With extended observations, this study aimed to report ...the mean 12-year outcomes focusing on bearing-related complications of CoC THA.
We reviewed primary CoC THAs performed between May 2009 and April 2012 at a single institution. Of the 831 hips (739 patients), 716 hips (635 patients) that satisfied a minimum 10-year follow-up were included. The mean age at operation was 54 years (range, 16 to 83). The occurrence of bearing-specific complications, including ceramic fractures and noise generation, was evaluated. The mean follow-up duration was 12 years (range, 10 to 14).
A total of 2 (0.3%) ceramic liner fractures occurred, as previously reported. No additional ceramic fractures were observed during the extended follow-up period. However, the cumulative incidence of audible noise increased from 6.4 (48 of 749) to 8.2% (59 of 716). Of the 59 hips with noise, 26 (44.1%) developed noise within 1 year, whereas 13 (22.0%) presented with noise more than 5 years after THA. The noise was described as clicking in 35 hips (4.9%) and as squeaking in 24 hips (3.4%). No revision surgeries were performed for noise, except in 2 cases of concomitant liner fractures. The implant survivorship free of any revision was 98.1% at 12 years.
No additional ceramic fractures were identified in this extension study at a minimum follow-up of 10 years. However, the prevalence of articular noise has increased from 6.4 to 8.2% since the previous report. Possible late-onset noise should be considered when performing CoC THA in younger patients.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP