Cypress Semiconductor

The following report was submitted by Phil Townsend, EE Major

How I got the co-op position

Cypress Semiconductor is a microelectronics company based out of San Jose , California with offices across the United States as well as in India , the Philippines , and Belgium . The company was founded in 1982 by T.J. Rodgers and has expanded to become the current leader in USB device production in the United States . The office in downtown Lexington is part of the CAD (Computer-Aided Design) group and opened in 1996 after Vice President of IT (Information Technology) Alan Hawse expressed a desire to leave the corporate headquarters in San Jose for his hometown of Lexington . The Board of Directors elected to allow Hawse to open an office in Lexington (now located downtown in the newly-renovated Festival Market across from the Lexington Financial Center and Rupp Arena) that has since expanded from fewer than a dozen to over thirty employees. Cypress has a strong relationship with the University and is currently in charge of administering the EE 584 and 555 courses at UK , providing both the software for designing the circuits and the professor for the course, Dr. Elias.

  My experience with Cypress began at the end of April 2006 with a phone call from the office's secretary, Heather Ley, offering a job interview the following week. I received this message on my cell phone while I was at the Co-op office so I was able to tell Donna Hewett about my interview immediately. Hewett advised me not only to read the previous Co-op reports about Cypress that other Co-ops had submitted but also to get in contact with a previous employee of the company, Jeanette Djigenou, to ask for advice on the interviewing process. Hewett provided me with Djigenou's phone number, and when I called Jeanette she told me that I should be prepared for two separate interviews on separate days: a screening interview where I would be asked general questions about my academic experience and career goals and, if Cypress was still interested, a follow-up technical interview that would consist of questions specifically about the various EE courses that I had taken or was currently enrolled in. (Later on I would learn that this second phase is referred to within in the company as the "pack of wolves" interview.)

The Co-op Work

  I spent the next several evenings reviewing the three most relevant courses I had had so far in EE that I would likely be asked about in the interview-EE 280: Digital Logic, EE 281: Logic Design Lab, and EE 360: Intro to Semiconductors. My interview began at about 1 PM in the office's upstairs conference room. Initially I was interviewed by the two supervisors for the site in the anticipated fashion of very general questions about myself such as why I wished to work at Cypress , why I was in Electrical Engineering, and how my previous experience had prepared me for work in the semiconductor industry. However, I soon learned that the company wanted to get new hires in as quickly as possible and as such brought in several other employees to begin the second, technical phase of the interview process immediately. My questions consisted of designing a four-input and output sorting machine given a set of two-input and output sorting machines, an explanation of the recursive algorithm for the factorial function, and drawing a frequency divider using D flip-flops and an inverter (all of which I answered successfully). My interrogators were unable to ask about the four courses most related to what I would be doing at Cypress, EE 461 and 462: Electronics Lecture and Lab, and EE 584 and 599: Beginning and Advanced VLSI, because I hadn't taken those courses yet-thus I was able to avoid many of the more difficult questions they might have asked me. After about an hour in total the interview was complete. I sent a thank you note via email the next day to my interviewers and follow-up asking whether I had gotten the job the following week, the later message resulting in a reply that I had gotten the position. The process of actually starting was held up for several weeks, however, because a formal job offer could not be made until the requisite paperwork was processed and mailed from the corporate headquarters in San Jose. I received an offer letter the third week of May, and on May 30, over a month after my call in RGAN from Heather, I began my job.

My first several days at Cypress involved taking online training courses required for all Cypress employees such as general computer safety and avoiding workplace harassment. Unfortunately, since I had not had the VLSI classes at UK before starting work at Cypress and since VLSI is the core of what the Kentucky Cypress office does, the set of useful tasks that I could perform was extremely limited. My saving grace was my programming background, since I had taken AP CS A and AB in high school as well as CS 216 at UK, giving me a background in C++, UNIX, and Perl (even though I hadn't used UNIX in over a year and had to relearn the majority of it very, very quickly). Thus the majority of my work involved computer programming where, although I may not have had a firm grasp of the circuital concepts at work in a particular process, I could still write computer code that would complete the requested task.

  My first Co-op tour at Cypress served as an introduction to the semiconductor industry and VLSI. Since I had not had any electronics or VLSI courses prior to starting with the company I had to learn the vast majority of the information required to perform daily tasks on the job. Since I was part of the CAD group my projects almost entirely involved computer programming and in several different languages (Ruby, Lisp, Visual Basic, and UNIX). Although the majority of my tasks have involved writing short scripts to help other employees with their work, my largest project, enhancing a program called lvsvcellsim, was a significant modification and was released to the entire company.

Lvsvcellsim is a shortened form of the phrase "layout versus schematic violator cell simulation." All circuits designed at Cypress are represented by, most importantly, a schematic and a layout. The schematic view is an abstract representation of the circuit that Electrical Engineers are familiar with beginning with Circuits I: resistors are represented as squiggles, capacitors as a pair of parallel lines, and so forth. The layout view, on the other hand, represents precisely how one would create the circuit on a silicon wafer, where different materials are represented with differently-colored rectangles in a bird's-eye view of the chip. Layouts and schematics are (at present) created manually in the software by the CAD engineers and must be checked to ensure that they represent the same circuit. Additionally, part of the quality assurance process at Cypress requires simulation of chips with "violator cells" connected to them. (A violator cell produces a strange input for a chip, such as an input voltage at a pin of 8 V rather than the now-standard 1.8, in order to ensure that the chip will not fail catastrophically.) Thus, the program lvsvcellsim is intended to check that the layout and schematic views for a particular violator cells match, connect this violator cell to the chip, and then simulate the chip to ensure that it is not destroyed by an unusual condition.

The enhancements I made involved adding two new simulation programs, Nanosim and Spectre, adding the option to use a different type of schematic view called a simsch, using several different netlist formats (a netlist is a textual representation of the elements and connections in a circuit), and generating an graphical report that is emailed to the user after the simulation process is completed. This project involved learning how the new simulators worked, how lvsvcellsim worked, how to program in Ruby, and how much of a headache porting functional code from one machine to another can be. (My modifications were ultimately released to the entire company through the computer code repository in Mississippi , and much time was spent getting my program to work at all sites, not just Kentucky .) The usefulness of my enhancements became apparent almost immediately as several circuits were shown to have transistors with invalid properties that the old simulator, Eldo, could not detect but that the newly-added ones could.

In addition, I've been assigned several other odd jobs throughout the summer, mostly involving writing Bourne shell scripts and Visual Basic Macros for the other engineers for various purposes. My role has mostly been to complete tasks that the other engineers could complete without much difficulty but that don't have a high enough priority to take care of right away. Thankfully, this has given me extra time to learn VLSI with several trips to the EE 584 website to read the tutorials and lectures. Over time my tasks have begun to require more and more knowledge of VLSI as my knowledge of the subject matter has increased: currently I'm working on automating the process of generating a chip layout given its schematic, requiring a working knowledge of Cypress 's circuit design software and its internal programming language, SKILL (a dialect of Lisp).

  My job at Cypress is not only my first position in my chosen field but also my first full-time job, making it doubly a new situation to become accustomed to. I had few, if any, initial expectations or preconceptions entering my job (after all, that's partly why I decided to Co-op-to find out what the real world's all about!), so the first two months were certainly a period of transition for me. Even though I had taken full course loads throughout college up to starting at Cypress (usually around twenty hours), working forty hours per week took a great deal of getting accustomed to. In an academic setting one has a great deal of liberty in when one works on a task-one might be in classes for three or four hours per day but completing the assigned material can be done at any time, day or night, week or weekend. But in the professional world one has to get tasks done in an orderly manner between 8 and 5, Monday through Friday, with no breaks outside of lunch and no excuses. (Skip a lecture, nothing happens. Skip a workday, lose your job!)

  Yet despite the many challenges I've faced with my new job I've also learned a very great deal. I've had to pick up several programming languages on the spot (Ruby, Lisp, Visual Basic, and relearning UNIX) and have gradually begun to learn the concepts of VLSI that I would have picked up in EE 584 and 599. But beyond the realm of purely technical knowledge I've learned how to effectively interact with my fellow employees and bosses, most importantly summed-up in a discipline called Precise Questioning and Answering. PQ&A is a philosophy of interaction that attempts to make the process of speaking about business issues as efficient as possible by making the questioner (oftentimes your boss) the driving force of the conversation and by keeping the responses of the answerer (oftentimes you) as short and direct as possible. (For example, one would not answer "Have you completed writing your program?" with "Well I ran some simulations this afternoon and some of them haven't quite worked because in the first test case..." with the far more concise "No, I'll be done by the end of tomorrow.")

In addition, I have had to learn how and when to ask for help from other employees, either in person, through email, or over the phone. Learning to ask for help has been perhaps the most difficult lesson for me learn in my time at Cypress because, being a born perfectionist, asking for help goes directly against my personality type. Probably the best maxim at Cypress is "'I don't know' is a step forward, not backward", since the sooner one admits he needs help the sooner he can get that help and move on.