Welcome!
To access the archive contents (presentation slides, videos, multimedia, etc.) sign in using your LinkedIn account by clicking the LinkedIn button to the right. Full Conference Attendees and Subscribers have access to all content now. FREE guests, once signed in, have access to the presentations after June 1, 2015.
------ This message is only displayed when you are not signed in. ------
2015 Tutorial Barnes
Heidi Barnes
Senior Applications Engineer
Keysight Technologies
Abstract
The ideal transparent socket would make us millions if we could actually find it, but in the real world we cannot simply time travel between two points and ignore what is in the middle. At DC the focus is on the contact resistance, but at the high frequencies of multi-gigabit interconnects one can be plagued by additional dielectric loss, reflections, and complicated multi-mode resonances. This means that if your socket is not working as required at 8, 16, 28 or 40 Gbps then that classic voltmeter sitting on the test bench is not going to do the job. Signal Integrity expert, Dr. Eric Bogatin, emphasizes that a good engineering practice is to always have models or simulations to predict the outcome of a measurement. Often there is a large void between the too simple V=IR calculation and the too complicated full 3D-EM simulation with Maxwell’s equations. Engineers confronted with these two options will typically do nothing on the simulation side and then just point to the data sheet.
A better option is to explore the power of simple transmission line theory and network analysis with scattering parameters (S-Parameters). Simple deconstructed transmission line models can be used in simulations to quickly evaluate the impact of resistance, dielectric loss, reflections, and even resonances. The results of the simulations provide valuable insights into how transparent a socket is for your application. Using hands-on computer labs with both frequency domain and time domain simulations attendees at the workshop will be able to test out their ability to debug a failed test socket measurement and get a socket design that is transparent for an 8.4 Gbps PCIe application example. These simple deconstructed transmission line models also improve the effectiveness of full 3D-EM simulations. Knowing what to expect from the EM simulation insures better setup of the stimulus ports, and effective use of simplification trade-offs.
Another way to make a transparent socket is to mathematically remove its effects from the measurement. This can be a simple calibration of removing a static IR drop at DC, but at high frequencies one must calibrate out both the “attenuation” drop and the interaction of reflected signals. To make things even more complicated, one must also keep track of the phase relationship between the current and voltages at high frequencies. Simple transmission line theory and network analysis scattering parameters (S-parameters) do just that, and with a bit of matrix math they can remove the effects of the socket fixture from a measurement. The signal integrity world calls this fixture de-embedding. Using hands-on computer labs and actual test equipment hardware attendees will learn how to measure socket fixture S-parameters using a two-tier Short/Open/Load/Through (SOLT) with 2-port probing or 1-port open calibration techniques. The socket fixture S-parameters can then be used in the time domain to remove the effects of the socket fixture from the measurement making the socket transparent. This sounds too good to be true, and so again one can turn to simulation to see where it can go wrong and get a practical understanding of how and when to implement fixture de-embedding.
The goal of this half-day tutorial is to give the attendees a toolbox of both simulation and measurement signal integrity techniques for characterizing a socket and ways to make it transparent for high frequency multi-gigabit applications.
Biography
Ms. Heidi Barnes is a Senior Application Engineer for High Speed Digital applications in the EEsof EDA Group of Keysight Technologies, a spin-off of Agilent Technologies. Past experience includes over 6 years in signal integrity for ATE test fixtures for Verigy, an Advantest Group, and 6 years in RF/Microwave microcircuit packaging for Agilent Technologies. She rejoined Agilent Technologies in 2012, and holds a Bachelor of Science degree in electrical engineering from the California Institute of Technology.
Part One
Sorry, you do not NOT have access to this file.
Part Two
Sorry, you do not NOT have access to this file.
Presentation Download
Return to the 2015 BiTS Workshop Archive index page