Siltronic AG (formerly Siltronic Corporation) is the world’s third-ranking producer of silicon wafers. The facility where I was employed was the company’s only production site in North America. During my time as an intern at Siltronic I worked in several capacities performing tasks in the field of information management for both Tip Rouse in Management Information Systems (MIS) and Paul McKelvey in Manufacturing Execution Systems (MES).
ADE Status Application (Paul McKelvey, MES) – Development of a text-based application to present at-a-glance manufacturing information
Physical Hardware Inventory (Tip Rouse, MIS) – Comprehensive update of a hardware database with all Windows server hardware
General Assistance (Linda Ahner, Resource Management) – Cell phone recycling, organization of physical files, misc.
In the beginning, my principle task was to bring up to date an excel spreadsheet containing a detailed account of every Windows server in each of the three server rooms spread throughout the various buildings on campus. This required that I actually read the serial number from each “beige box” (or black box as it were) and match it to a unique hostname within the Siltronic DNS (for example “zptlas12”).
In cases where there were discrepancies (i.e. more than one hostname to a serial number, a serial number without a hostname) I would consult Tip and other members of the MIS team to diagnose the problem and adjust the record accordingly.
The reason this work was required was because, as old systems were replaced or reused and shuffled around, the record of where the various servers actually sat would become out-of-date, making it difficult for anybody to put their hands on the actual machine that needed fixed if ever there was a problem.
As time went on, and I worked in this fashion I began to develop a sense for the way the servers were named and managed so that eventually, when this project entered another phase I was in a position to apply what I had learned in the context of an IS solution. In particular, I had some ideas for adding table columns and modifying existing ones. I also wanted to look into changing processes with the aim of further leveraging the database.
The next phase took the form of inputting all the server data I had collected into a hardware database that was part of a larger ERP system designed to give members of the various teams within Siltronic a means of gathering information from various computer systems and applications used in management and operations.
I spent some time interacting with the hardware database as a user and was ready to start looking at the back end and its place in the bigger picture when suddenly Greg Bernards in Shop Floor Automation (SFA), who was in charge of the hardware database and its integration with other systems and who was going to get me started, left the company to find work elsewhere.
Meanwhile Paul McKelvey lost an intern. Given the Circumstances, it was decided that I should go over to Paul’s team and fill the other intern’s shoes.
Before I go on, it is important to describe some background. The manufacture of silicon wafers consists of essentially two parts: Growing, and processing. The silicon for our wafers was grown in monocrystaline ingots and shipped to us from Germany. We then put the ingots through processes such as “grinding”, “slicing”, “final polish”, “epitaxy”, and so on. As part of each of these processes, certain measurements of each wafer are taken by special machines called ADEs. The measurements are then gathered on systems called concentrators and stored in a number of databases.
There existed several legacy concentrators which ran an operating system called VMS (a carry-over from the old VAX systems that used to concentrate ADE data). On these concentrators there was a captive account called “ADESTATUS” which ran a VMS shell script that used a text-based menu system to let users query the concentrators to determine information about various ADES and the wafer lots running on them (wafers traveled through the “fabs’ in lots of various sizes).
The way I used to describe it to laypeople was to say that it was like a text-based library catalog you’d see at a library in the late 80s or early 90s. The only differences were that there was only a handful of books (analogous to the different processes) and that each library patron (a wafer lot) checked out only one book at a time and did so with several books in a particular order (according to its recipe).
The job Paul gave me was to develop a version of the ADESTATUS script which would run on his new Linux concentrators. I was able to do this since I had spent significant time in the months before teaching myself the basics of writing Linux shell scripts (It had been Paul with whom I had first spoken upon receiving my MECOP assignment and he had mistaken me for his own intern, telling me that I could expect to work with Linux and to be prepared).
I set out in late June and by mid September I had completed three of the five main menu options presented in the ADESTATUS script, plus an auxiliary application for the identification and assessment of emerging trends in errors which Paul had tagged on. It should be noted, however that some of the time can be accounted for when one considers the time it takes to become familiar with the concentrator file system and understand the ways data get stored and actions get logged. Furthermore, I spent a fair amount of time documenting and diagramming my efforts.
I kept a detailed daily journal of my internship which is available upon request.
Being that this was my first internship it’s hard to deny that some of my most important lessons came from the simple adjustment from school life to the working lifestyle and from becoming adjusted to work in an office environment. All my previous work experiences had been in direct customer service which has a very different ethic and pace of work. In the technical realm, I would have to say that Linux shell scripting was the biggest skill I learned. Thanks to this development, I am no longer intimidated by a command prompt and I have even gone on to apply this in my personal life.
The benefit I provided Siltronic was aid in their migration from legacy software to Linux, a still-ongoing and complicated process. By working with it firsthand, I gained a better understanding of what this kind of transition entails. On my last day, at my own request, I presented at an IT all-hands meeting with members of several different teams. My presentation covered in more technical detail the purpose and workings of my port of the ADESTATUS script to Linux.
It also bears mention that during the time of my internship I volunteered at a local computer co-op (Freegeek) to gain experience with Linux and with computer hardware.