Questions to ask and review before product and IC
I named these questions Martin Chu's PRE-DESIGN INQUIRY.
by Martin Chu, COPYRIGHTED
Original created August 1 2009
Last Update Oct. 10 2012
1. What are the killer products or services you are targeting now?
2. What will be the killer products or services 6 month, one year or three years from now?
3. What is the life span of that killer product or service?
4. How promptly and reliably that you can design and manufacture that product and make a profit of it!
5. Do you know the strength and capability of your company/team? Do you also fully understand your weakness and limitation?
6. Do you know why some design team take forever to get a design right, while some design teams just take one or two runs to design a real marketable product!
7. Are you sure you are designing to the correct system specification? Also is your design specification really the correct and most up to date specification!
8. Do you verify your design result/layout independently? Are you sure that final layout match with your schematics; your photo masks are accurately DRCed; and using the right PDK version and EDA flow? Or you don't know and/or your design and/or layout teams don't care and you also did not check it! You just blindly trust what you were given from design, layout, mask shop and foundry.
9. Are you using the right technology for this product or design? Why some fabless design houses choose certain foundry over others?
Also, why some IC foundry just larger and better than other companies?
11. Why few companies make an acquisition to other company, later on companies grow and gain competitive edges. However, most other companies just cannot digest the acquired counterparts! Also do you even think and craefully investigate the reasons for the requistion at first?
12. Why some very few design company can design smart phone with excellent communication/video/audio/voice service. While some design companies established at about the same time, only can design very basic and low frequency IC. Also when comparing the compititors in the same field, the profiit from operation is always several tens of percentage lower than their counterparts.
Why certain high performance and complex IC only comes from very
14. Are you using the most suitable SPICE models for your MOS devices, how about bipolars, resistors, capacitors and inductors? Also are you sure that the model your are using predict or trace your simulation and layout?
15. What physics and device behaviors do your SPICE models simulating? Are there something missing? Do you even know it?
16. Are you sure your SPICE model aligned with your devices and/or processing steps and/or your circuit layout?
17. Do you know why your input should be able to handle 32V, but your designs starts to show some current leakage around 26V? Also suppose your are using a highly reliable process, but how come the incorporated products just can not pass some product qualification. Do you know what and how they testing or qualify the foundry processes? Are you sure what they say is what they do? If not then, how to deal with that!
Do you know the reasons why some designs using the same
technology and even no DRC violations at tapeout, and their first
silicon prototype just work
fine and can put into market right away. However, some other designs
just have no-ending
testing and qualification issues. The design just cannot meet some
design/project specification or
even device performance!
19. What final yield are you expecting from your foundry wafer for that technology? Is that the normal Fab yield for that process? How consistent is your chip performance after coming out directly from foundry and after bonding and package?
All these WHY? WHY? WHY? and WHY?
Have you ever given a thought over these!