Equipment Engineering
I worked for 5 years as an etch equipment engineer in semiconductor manufacturing at Broadcom and Samsung Austin Semiconductor. I have experience with troubleshooting daily mechanical failures and with implementing long term, root-cause solutions. I love studying data trends to identify and resolve problems that have gone unnoticed. In my experience, I have:
Saved $100M+/month by eliminating bottle necks and increasing OEE and throughput. Identified commonalities and root causes with 8Ds and 5WHYS and implemented long term solutions by qualifying new parts, publishing BKMs, writing and automating new processes, and creating SPC interlocks.
Resolved chronic quality issue, resulting in 50% less failures and 3% stoploss reduction. Convinced management in Korea and in Austin with clear data and communication to implement changes to the fab’s automated system to reduce stoploss by 1.75%.
Designed and qualified parts that promoted safer and simplified installations.
Wrote and edited dozens of documents for maintenance, troubleshooting, and OCAPs (out of control action plans). I take pride in writing succinct documents.
Troubleshot equipment issues pertaining to electronics, RF and MW plasmas, robotics, mechanical systems, vacuums, pneumatics, temperature controls, computer communications, and gas delivery systems.
Directed technicians daily by providing action plans and assigning work based on fab priorities. Received the Director’s Award and 5 Spot Awards for my teamwork, self-direction, leadership, and detailed troubleshooting.
Wrote data visualization programs using JMP and Excel VBA.
DESIGN-driven solutions
I liked to make technicians’ jobs as simple as possible. How did I do this? I collaborated with them, listened to their complaints, and brought them data and ideas of how we could fix issues together. Here are a few examples of how I resolved complaints.
“The lifter’s locking pin is so easy to lose!!”
Problem: The lifter’s locking pin was unsecured and at risk of falling during production or maintenance. The pin was easy to lose because it was small and required multiple re-installations when adjusting lifter heights. The issue took priority during a tool-down situation due to a lost pin and 0 replacement stock.
Solution: We established that the new pin should: 1. secure the lifter, 2. eliminate the need to remove the pin when making adjustments, and 3. use a non-conductive material. After days of ideating, I realized we could modify the existing part by tapping a hole in it and fastening a small stopper. I CAD’d and 3D printed a prototype, which worked perfectly, then sent out existing stock for modifications.
“These pads are a pain to remove and install!!”
Problem: Removing and installing pads was a pain because their clips were difficult to insert, easy to break, and even easier to lose. Additionally, installing pads required disassembling and reassembling a complicated shelving system.
Solution: A technician recalled using rubbery mushroom pads on another equipment set. I ordered similar pads that then required no shelf disassemble, were unbreakable, and were easy to insert.
“These chambers power-off and cause more problems when I perform routine maintenance!!”
Problem: Technicians complained that chambers powered-off during routine maintenance. Turning on power requires extra work and can induce electrical problems. To avoid this, technicians bypassed safety interlocks with makeshift jumpers.
Solution: We designed and installed jumper inserts on the equipment and distributed jumpers with large warning tags labeled with each technician’s name to keep them accountable. The SOPS were updated and the technicians received training to ensure that they use proper LOTO when using the jumper. This is a much safer solution than the make-shift one of installing stray wires directly into circuits.
DATA-driven solutions
“The load ports are impossible to recover!!”
Problem: I noticed that loading ports were losing communication to the system. The technicians said ports lost communication when 1 bad lot would land on a port, cause an error, leave, then repeat this issue on the next port. After an hour, all ports would lose communication, no product could load in our step of the assembly line, and we became a bottleneck. The technicians would fix the issue by physically chasing down the bad lot, removing it from production, and re-establishing communication to the load ports.
Solution: I found commonality to one process causing the issue. I spoke to the process’s engineer, identified the root cause, and added a step in between our steps to fix the issue. As a further precaution, I limited that process to one frame (rather than 30) so that we protected the fleet and alarmed much sooner. I also showed the technicians how to stop the lot from loading using software rather than chasing down a lot. This allowed us to process an extra ~$100M worth of production per month by not becoming a bottle-neck.