Chris' Projects

With this page I would like to explain a little more in detail projects that I've been working on. A brief description can be found in my resume.

Circuit and Physical Design Engineer

TOSHIBA AMERICA ELECTRONIC COMPONENTS, Inc.

August 97 - September 99:

The department I was working in was the Embedded Microprocessor section which then designed an embedded MIPS-based RISC processor, which will be the processor of SONY's next-generation Playstation.
Initially, I started doing HSPICE simulations for critical paths. I also simulated long interconnects to determine the optimum distance between repeaters. Besides the simulation I used Star-RC to extract the parasitics capacitances. (Star-RC is probably the tool that I have filed the most bug reports for although it's not the one that I'm using most. ;-) Besides that, I write many scripts in Perl and Perl/Tk, mostly for generating or processing command files for our place&route tool (Aquarius).
Currently, we are working towards tape-out and now I'm mostly working on clock design, which means clock skew reduction. In our processor we are applying an H-tree in the upper levels and a grid- or mesh structure for the lower level tree. I wrote a bundle of scripts to balance the buffers that drive the mesh. Depending on the load distribution in the DSPF file buffers will be placed. A by-product of this work is a viewing tool for DSPF files. It appears that there is no commercial product on the market. With this tool I can examine DSPF files very efficiently and find problems much earlier.
After the tape-out I plan to continue the evaluation of a library characterization tool called 'libchar' by Library Technologies, Inc. I want to thank Mehmet here in this place for his great support.

Master's Thesis: Design of MMU and BTB

July 96 - March 96:

"Contributions to the Design of MMU (Memory Management Unit) and BTB (Branch Target Buffer) for an Embedded RISC Processor"

These two macroblocks are part of the Emotion Engine processor (see above).
The invention of the BTB is not a very recent one, but until now only a few processors make use of such a unit that can predict up to 90% of all branches correctly and reduce the number of pipeline stalls.
Two novel CAM cell types were developed for faster operation and reduced power consumption. An innovative control scheme could avoid the disadvantages of a Twin-CAM.

Internship: Optimization of Scan Registers

October 95 - March 96:

To reduce the time necessary for manufacturing tests of VLSI-circuits scan registers can be installed on the chip. This additional circuitry becomes redundant after the test but not only the throughput during the test increases, also the fault coverage can be much better. With scan registers many chip internal nodes become directly controllable and observable.
Novel D-flip-flops and multiplexers in CMOS technology were examined and optimized to increase the speed of these scan registers.

Remote-Boot-PROM

March 95 - April 95: (Besides my studies)

The Dualzentrum Dresden is providing network solutions for networks with large numbers of servers and workstations. They also produce software, i.e. for the remote boot procedure that has been defined by IBM. Computers load their operating system from a special server by using the remote boot service. This allows the workstation to operate without any local drives. In addition, and this is even more important, the configuration of such workstations can easily be changed by changing the configuration table, located on the remote boot server. For this purpose a remote boot PROM has to be installed on the network adapter of the workstations. The program of this PROM is run at the start-up of the machine.
A remote-boot-PROM (or RPL-ROM, remote program load) in a LAN-Manager environment was implemented for an SMC network adapter. The program is written in i8086-code and covers the functions of the media access control layer in the ISO-reference-model.

BAM-to-Ethernet Interface

January 95 - June 95: (Besides my studies)

A BS2000 host is usually connected to its terminals by a BAM-interface which is a proprietary SIEMENS standard.
With this interface the cable length is limited 2000m and a server with high costs for maintenance is necessary. To overcome these disadvantages the old cables were replaced by a future-oriented structured wiring system, based on the Ethernet standard.
A new device was required to transform the proprietary BAM protocol into the open Ethernet protocol. This is done by a small IBM-compatible PC (PC104) with a network adapter. In addition, a BAM-interface, based on the Z8530 serial communications controller, was developed. The software is written in C, using LOGICS libraries. Here is detailed information about the project in German.

Internship: Test Software in C and BASIC

March 94 - April 94:

The ZIROX GmbH is a company that produces gas sensors for oxygen, hydrogen, carbon-oxides, lambda and others.
For manufacturing tests of these sensors and to determine their characteristics a test program was developed that displays several parameters while the sensor temperature periodically changes.

Internship: Test devices for automobiles

August 92 - September 92:

The installation of VOLKSWAGEN's power door locks is a little tricky because of the sensors that indicate a locked door. A device for indicating the states of all four sensors and for controlling the system was designed and produced. Although this device was not very sophisticated, it is very helpful and still in use!
Another issue of this practical was to learn the basics of programming robotics and memory programmed controls.

Prior to my studies I was working in electronics and computer circles.
I assembled a clock that would ring a bell in a school at certain times, which was in use from 1988 to 1997. This clock was designed with standard TTL-ICs such as 74193, 7447, 7474 and simple gates. The ring times were stored in a 2716 EPROM.
Another project was a CRT interface for a Z80 microprocessor, also in standard TTL-ICs with a 2716 as character ROM. This was meant to be used with an LC-80 learn-computer for a TV-game. But the whole thing was not needed anymore after the wall crashed in 1989 and everybody could set out and buy an Atari console. Also, is became less interesting as soon as the Z80 could write its characters on the screen. ;-)
During the last three school years I was teaching BASIC to students in a computer circle once a week.

If you have any further questions, please send me an email.

Back to Chris' Home Page