Agenda – Center for Electromagnetic Compatibility
Industrial Advisory Board Meeting
Rolla, MO
Morning
8:00 amRegistration and continental breakfast
8:30 amWelcome remarks and introductions
Poster Presentation 1
8:45 am1 min. presentations for 20 Posters(P26~P45)
Radio Frequency Interference 1 (C. Hwang)
9:15 amBroadband measurement using a high sensitivity tunable probe(P1)– X. Yan (Missouri S&T)
9:45 amPhysics based dipole moment source reconstruction for RFI analysis on a practical cellphone(P2)–Q. Huang (Missouri S&T)
Signal Integrity (J. Fan)
10:15 am Statistical eye-diagram estimation for HBM channel including ISI, crosstalk, and power noise(P3)–J. Cho (Missouri S&T)
10:45 amBreak
11:45 amDifferential integrated crosstalk noise reduction among multiple differential BGA and via pairs by DoE method(P4)–M. Ouyang (Missouri S&T)
IAB Meeting 1
11:45 amRadiated Two Stage Method from Theory to Practical Application –Yihong Qi
12:15 pmLunch at Hypoint
Afternoon
Shared Challenge Project Reports (J. Fan)
1:15 pm Power supply induced jitter for complex circuit (P5)–H. Kim (Missouri S&T)
1:45 pm A pre-layout power integrity methodology for achieving a Ztarget(P6)–J. Drewniak (Missouri S&T)
2:15 pm Managing and sustaining information and data storage – a repository for the CEMC(P7)–B. Archambeault (Missouri S&T)
2:45 pm Break
Poster Presentation 2
3:15 pm1 min. presentations for 19 Posters(P46~P64)
NSF Presentation
3:45 pmNSF perspective –Denis Gray
Electrostatic Discharge (D. Pommerenke)
4:15 pmESD susceptibility test synchronized with CPU activity(P8) – O. Hoseini Izadi (Missouri S&T)
4:45 pm Identification of the root cause of ESD-induced soft failure using a mixture of full-wave simulation and measurement (P9) –A. Hosseinbeig (Missouri S&T)
5:15 pmOn-die sensor to determine the length of a transient event(P10) – A. Patnaik (Missouri S&T)
5:45 pmPoster and Demo Session, HyPoint
(Drinks and snacks provided)
No formal dinner on Day 1
Agenda – Center for Electromagnetic Compatibility
Industrial Advisory Board Meeting
Rolla, MO
Morning
7:30 amContinental Breakfast
High-Speed MeasurementCharacterization (J. Fan)
8:00 amSFD fixture auto corrections on de-embedding and material extraction(P11) – B. Chen (Missouri S&T)
8:30 amImproved “Root-Omega” method for transmission-line based material property extraction for multilayer PCBs (P12) – S. Jin (Missouri S&T)
Electromagnetic Interference (D. Pommerenke)
9:00 am Signal integrity and EMI evaluation of flyover QSFP connector for 56+ Gbps applications (P13) – A. Talebzadeh (Missouri S&T)
9:30amCharacterization of cable bundle by dual current clamp method for EMI(P14) – S. Walunj (Missouri S&T)
10:00 amBreak
EMC in Bio Applications and EM Modeling (J. Chen)
10:30 am Uncertainty analysis for antenna system using stochastic analysis(P15) – Z.Wang (UH)
11:00amEstimation of the RF induced heating near medical implants under 3T shimming condition (P16) – Y. Wang (UH)
IAB Meeting 2
11:30 amIAB business: project voting, next meetings planning, strategic planning committee
12:00 pmLunch at Hypoint
12:45 pm Group Picture
Afternoon
Power Integrity (J. Drewniak)
1:00 pmSystem-level voltage decomposition in PDN design (P17) – C. Huang (Missouri S&T)
1:30 pmAccurate chip modeling method for Power/Signal Integrity as well as EMI – Y. Lee (Ansys)
2:00 pmFast decap assignment algorithm for optimization of power distribution networks (P18) –K. Koo (Missouri S&T)
2:30 pmZ-directed components (ZDC) technology for power integrity applications (P19) –Y. Cao (Missouri S&T)
3:00 pm Break
IAB Meeting 3
3:30 pmNew IAB member presentations – ASUS & Lexmark
4:00 pmCenter status report
Financial state of the center
5:00 pmPoster and Demo Session, HyPoint
(Drinks and snacks provided)
6:30 pmDinner: Hasselmann Alumni House, 1100 N. Pine Street, Rolla, MO.
Agenda – Center for Electromagnetic Compatibility
Industrial Advisory Board Meeting
Rolla, MO
Morning
7:30 amContinental Breakfast
System-Level EMC (V. Khilkevich)
8:00 amEfficient method for prediction of immunity of Multiwire cable bundles (P20) – D. Nozadze, R. He, and K.Ghosh (Missouri S&T)
8:30 amOptimizing measurement SNR for weak near-field scanning applications (P21)– L. Guan (Missouri S&T)
High-Speed MeasurementCharacterization 2 (J. Fan)
9:00 amApplication of the hurray surface roughness model to transmission lines with slightly dispersive dielectrics (P22) –H. Gao (Missouri S&T)
9:30 amDielectric dissipation factor (DF) extraction based on differential measurements and a 2-D cross-sectional analysis (P23) – S. Yong(Missouri S&T)
10:00 amBreak
NSF FRP Project (V. Khilkevich)
10:30 amSuper-resolution ESM using a spatial filter (P24) – R. He (Missouri S&T)
Radiation Physics (J. Drewniak)
11:00 amCharacterizing EMI and radiation principle analysis for PCB/high speed connectors(P25) – Y. Cao (Missouri S&T)
IAB Meeting 4
11:30 amLife form review, IAB new business
12:30pmAdjourn/Pizza Lunch
Poster Abstract
P1.Broadband measurement using a high sensitivity tunable probe
–Xin Yan (S&T)
High-sensitive near-field probes are desirable for radio-frequency (RF) interference studies, and always apply in various fields. When designing the resonant near-field probes, the resonant frequency need to be considered in different fields of application. In most cases, to cover more frequency range, many probes need to build which have different resonant frequencies. Besides, to measure broadband near-field, several resonant probes need to be used. In this project, an easy way to change resonant frequency is demonstrated based on a magnetic-field probe by varactor diodes. The measured results show good agreements with the simulation. An automatic broadband measurement is presented, showing the broadband and high sensitivity characters of the tunable probe.
P2.Physics based dipole moment source reconstruction for RFI analysis on a practical cellphone
–Qiaolei Huang (S&T)
A practical cellphone will be studied for this RFI study. To understand the radiation physics on the cellphone, phase resolved near field scanning are performed. Equivalent dipole moment will be built to mimic the radiation pattern. This equivalent dipoles are obtained from current distribution analysis, so they have clear physic meaning.
P3.Statistical eye-diagram estimation for HBM channel including ISI, crosstalk, and power Noise
– Jonghyun Cho (S&T)
Statistical link analysis and link budget calculation is the essential part for current high-speed system design. Because of the difficulty of low bit-error rate (BER) simulation at SPICE, several methods were proposed to simulate low BER in a short time. Single-bit response (SBR) method is the most basic method; it uses SBR to construct the arbitrary waveform under the assumption of linear and time-invariant (LTI) system. However it cannot be applied to real system because of nonlinearity of I/O drivers. There exists a method to include these nonlinearity of real system, which is called as multiple-edge response (MER) method. There’s a few paper to explain MER method, but no paper covers statistical analysis for the real system simulation including ISI, X-talk, and power noise. In this presentation, we apply MER method on simple, but realistic HBM channel and also shows calculation burden of MER as the number of channels increase or MER order increases. Additionally, conceptual methodology is proposed to apply MER on real HBM channel with more than a thousand I/Os.
P4.Differential integrated crosstalk noise reduction among multiple differential BGA and via pairs by DoE method
– Muqi Ouyang (S&T)
The Integrated crosstalk noise (ICN) has been widely used for channel crosstalk evaluation in the IEEE 802.3ba standard. In this study, the differential ICN analysis is performed on the package to PCB transitions, including multiple differential BGAs and vias. To estimate the impact on ICN with different input factors such as pin map patterns, physical dimensions and shielding, the full factorial design in the design of experiment (DoE) statistical methods is introduced and verified. In addition, several guide lines for optimizing the design are proposed to maintain the tradeoff among the differential ICN, spacing of designing and manufacturing cost based on the DoE analyzing results.
P5.Power supply induced jitter for complex circuit
– Heegon Kim (S&T)
In this study, an analytical model of power supply induced jitter (PSIJ) for an inverter chain is proposed. Based on the linearized model of the MOSFET I-V curves, closed-form equations of the continuous time-interval error (CTIE) at an inverter chain output induced by power supply voltage fluctuation are derived. First, PSIJ of a single inverter is modeled at rising and falling edges. Because the loading capacitance is dependent to the output state, the PSIJ at each edge need to be separately modeled. In order to obtain the PSIJ of an inverter chain output, the proposed CTIE models at each rising and falling edges are alternately accumulated. To validate the proposed model, the simulated PSIJ by using the proposed analytical model and SPICE simulation are compared. Finally, several design guidance to reduce PSIJ of an inverter chain is discussed based on proposed analytical model.
P8.ESD susceptibility test synchronized with CPU activity
– Omid Hoseini Izadi (S&T)
Likelihood of soft failures occurrence can be dependent or independent of DUT load (activity). A reset caused by an ESD energy coupled to the DUT reset circuitry is independent of the DUT load. However, if the ESD energy is coupled to the power distribution network (PDN), the failures is load dependent. Experiments on CPU of a BeagleBone Black board and main chip of a smartphone have shown that the number of pulses required to cause a soft failure is reduced by a factor of ~5 to 10 if the injections are synchronized to the DUT activity.
P9.Identification of the root cause of ESD-induced soft failure using a mixture of full-wave simulation and measurement
– Ahmad Hosseinbeig (S&T)
A method is developed to predict the soft failure caused by ESD in electronic devices. In order to model the board, a numerical model is created in CST Studio Suite. The active parts are substituted by resistors and capacitors and a simulation is performed to determine the voltages and currents at pins. The pins that have excessive voltages and/or currents are flagged as potential soft failure inducing pins. Then, the model will be verified against measurements for soft failures.
P10.On-die sensor to determine the length of a transient event
–Abhishek Patnaik (S&T)
Testing and debugging of electrostatic discharge (ESD) or electrical fast transient (EFT) issues in modern electronic systems can be challenging. Previous circuits designed at S&T were built to detect the presence of a transient event and determine its peak level. These circuits were built to reside in each I/O pad, so the location of an event could be determined precisely. An extension to these circuits has been designed to measure the length of a transient event. Simulations demonstrate the potential of the concept. Test structures have been designed in a 0.35 um technology to further evaluate the technique.
P11.SFD fixture auto corrections on de-embedding and material extraction
–Bichen Chen (S&T)
The manufacturing variations in the test fixtures will bring errors into the SFD de-embedding procedure. As consequences, such errors are propagating into the electrical performance quantifications of device-under-tests (DUTs), material property extraction, and etc. The fixture error correction algorithm is employed whenever there is the variations between the fixtures of 2X Thru and Total to reduce the de-embedding errors as well as to improve the accuracy of Dk and Df extractions.
P12.Improved “Root-Omega” method for transmission-line based material property extraction for multilayer PCBs
– Shuai Jin (S&T)
Electrical properties of dielectric substrate are critical in designing high-speed products in terms of signal and power integrity. It is important to accurately characterize the dielectric properties to avoid overestimating or underestimating in the design. This paper proposes the improved “Root-Omega” method for extracting dielectric properties from fabricated multilayer printed circuit boards. Based on the electrical properties of fabricated transmission lines, the improved “Root-Omega” method applied to cases with smooth and rough conductors is validated using simulations. Error sensitivity analysis is performed to demonstrate the potential errors in the original “Root-Omega” procedure and the error sensitivity is significantly reduced by the proposed improvements
P13.Signal integrity and EMI evaluation of flyover QSFP connector for 56+ Gbps applications
–Atieh Talebzadeh (S&T)
In this paper, EMI aspects of the flyover Quad Small Form Factor Pluggable (flyover QSFP) connector, as a new approach, for 56+ Gbps applications are discussed and compared with the standard QSFP. This technique can overcome the limitation of large insertion loss of standard surface mount QSFP to PCB Microstrip traces. In flyover QSFP, substituting 7.5 cm PCB traces with the 15 cm twinax cables, at least 7 dB improvement in insertion loss can be attained at 40 GHz. To study the EMI performances of flyover QSFP in contrast to standard QSFP, two test vehicles were designed, fabricated, and tested. The total radiated power (TRP) for both QSFP connectors are obtained up to 40 GHz. Based on the results, it is shown that flyover QSFP interconnector exhibits a better SI performance due to the lower loss in comparison to the standard QSFP, and from the TRP results, it can be implied that if the same power delivers in both connectors, then they present a similar EMI performance at most of the frequencies in the single-ended setup.
P14.Characterization of cable bundle by dual current clamp method for EMI
–Sameer Walunj (S&T)
The common mode loop impedance can be measured using the dual current clamp method. This loop impedance includes the sum of the impedance looking toward the source or load, but does not reveal the impedance of the source or load themselves. The source and load impedances can be determined, however, using multiple measurements of loop impedance and using transmission line theory. Our work focuses on extraction of common mode source and load impedances of a well-defined transmission line system which then can be extended to any general system.
P15.Uncertainty analysis for antenna system using stochastic analysis
–Zhichao Wang (UH)
In this talk, we will discuss an automatic uncertainty analysis system embedded in the CST platform for various electromagnetic simulations. Using the system developed here, efficient and accurate stochastic analysis can be achieved. Examples of a patch antenna design will be presented.
P16.Estimation of the RF induced heating near medical implants under 3T shimming condition
–Yu Wang (UH)
RF induced heating near medical device is great concern especially for MRI 3T systems. Due to the RF shimming procedure, the RF induced heating cannot be estimated very accurately. Currently, only quadrature operating mode can be approved by the FDA regarding the MRI conditional labeling for most medical implants. In this talk, a procedure on how to perform accurate estimation of the RF induced heating under RF shimming is proposed. Both modeling and experiment results will be presented.
P17.System-level voltage decomposition in PDN design
–Chenxi Huang (S&T)
Although the target impedance is a commonly used criterion in most PDN designs, it may not be efficient because it’s usually rather pessimistic. Herein a time domain voltage ripple decomposition approach is proposed to avoid overdesign as well as provide design guidance to PI engineers. Based on a physics-based circuit model for PDN and a switching current generator including both high frequency switching and low frequency power gating, total voltage ripple can be divided into several components. Each component will have a one-to-one correspondence to the real PDN geometry. Thus design curves can also be derived, which can guide PI engineers when making design decisions.
P18.Fast decap assignment algorithm for optimization of power distribution networks
–Kyoungchoul Koo (S&T)
In this work, a very fast capacitor assignment algorithm capable of finding an optimized decoupling solution scheme that meets a predefined target impedance for a given power distribution network (PDN) is presented. The algorithm provides an optimal combination of decap values and decap locations while minimizing the number of decoupling components needed to meet RL-type target impedances. The algorithm only takes a few seconds to find a decoupling solution even when there are a large number of models in the capacitor’s libraries.
P19.Z-directed components (ZDC) technology for power integrityapplications
–Ying Cao (S&T)
An important aspect of ensuring the power integrity of a power distribution network (PDN) design is determining the layout, value, package size, and number of decoupling capacitors. To solve the limitation of the vertical connection of the surface-mounting type capacitor to the power net area fill, Lexmark has proposed a new concept of capacitor denoted the Z-Directed Component (ZDC) capacitor. The ZDC capacitors can be integrated within the PCB substrate directly below the package ball and eliminate the IC to decoupling capacitor horizontal distance across the power-net area fill. Further, the inductance associated with the vertical current paths on the vias from the IC package, and the SMT decoupling capacitors are dramatically reduced or eliminated.Since only the inductance of the ZDC capacitor itself remains, there is the possibility of using far fewer capacitors to achieve a high-frequency target impedance specification.In addition, for typical high-density designs that are common today, the elimination of vias to numerous SMT decoupling capacitors frees up constraints on routing channels for signals.
P20.Efficient method for prediction of immunity of multiwire cable bundles
–David Nozadze, Ruijie He and Kaustav Ghosh (S&T)
Radio frequency immunitystandards require to determine the differential voltage across shielded wires of cable harnesses placed over the metal plate. The presented work proposesthe equivalent circuit model andthe methodology for prediction of the worst case envelope for differential modesignals induced in cable harnesses placed over a ground plane. The induced signals are calculated using the transmission line theory avoiding the need for any 3D calculations. The results are validatedby experimentalmeasurements.
P21.Optimizing measurement SNR forweak near-field scanningapplications
–Li Guan (S&T)
Conventional near-field scanning techniques often employ a general setup such as: broadband near-field probe output connected to a chain of amplifiers through a coaxial cable to a spectrum analyzer. In this paper, we investigated how the signal to noise ratio is influenced by the coaxial connection between the probe output and the first amplifier, types of probes, cooling the probes with liquid nitrogen and the amplifier’s noise figure. Eliminating cabling between probe and first amplifier, and using a low noise amplifiers helped increase signal-to-noise ratio by ~10dB. Further, liquid nitrogen is used to cool down a tunable resonant probe. This increases quality factor of the resonance and improves sensitivity. Thus, SNR is further improved by 10-12dB compared to a similar broadband setup.