hist78
Well-known member
U.S. chip fab construction is among the slowest in the world -- a complex web of regulations is to blame according to study
It's kind of a mess.
www.tomshardware.com
U.S. chip fab construction is among the slowest in the world -- a complex web of regulations is to blame according to study
- Thread starter hist78
- Start date
www.tomshardware.com
Some of the important standards that influence fab engineering decisions include SEMI S2, National Fire Protection Association, Factory Mutual 4910, National Electrical Code, ISO, IEC and ASTM standards, BSI standards, CE Mark, REACH, RoHS, China RoHS, and the list goes on.
My impression is, the Federal/State/Local split is complex, when viewed from the outside, but not really from the inside; there is a division of responsibilities that is clear to all players. The real slowdowns occur when one of these proliferating global standards comes into conflict with existing assumptions embedded in the US system. The US regulatory code does not "play nice" with the web of global standards that everyone else plays with.
The US system also has rigidly defined roles and responsibilities that translate down to the fab operation level. Certain local codes require certified inspectors and plumbers and electricians which dictates union tradepeople (in some but not all cases), which in turn dictates external bidding, which in turn requires audits and credit checks, which in turn...you get the idea. Evolution. Ultimately, jobs get done safely and predictably, but much more expensively.
My impression is, the Federal/State/Local split is complex, when viewed from the outside, but not really from the inside; there is a division of responsibilities that is clear to all players. The real slowdowns occur when one of these proliferating global standards comes into conflict with existing assumptions embedded in the US system. The US regulatory code does not "play nice" with the web of global standards that everyone else plays with.
The US system also has rigidly defined roles and responsibilities that translate down to the fab operation level. Certain local codes require certified inspectors and plumbers and electricians which dictates union tradepeople (in some but not all cases), which in turn dictates external bidding, which in turn requires audits and credit checks, which in turn...you get the idea. Evolution. Ultimately, jobs get done safely and predictably, but much more expensively.
I didn't really propose a solution, did I? It's not changing the regulations or eliminating regulations. If you were to change something it would make the problem worse; it would create uncertainty. Possibly cause a crash. US regulations aren't really the issue in my opinion. A fab is a "global product"--while located in one place, it has to work in a country of origin (the US, Korea, or Taiwan) and one or two other countries, duplicating the original. Things as basic as the fab shape and number of stories change in translating from one country to another.
So what is needed is greater knowledge of localization strategies in engineering education. You see this a bit with power supplies--they work with multiple plugs and cycles and voltage ranges. Localization engineering is mostly ad hoc, doesn't exist as a discipline currently. As a branch of the major disciplines, we need more engineers who are fluent in the global standards, the unique gaps between foreign and domestic, and how to bridge. No silver bullet solutions here, sadly.
It would help if the US would work on harmonization with global standards. We have our own unique units, our own unique building codes, and that's fine, as I said above, don't change anything. But there is room for some gradual, well planned out transitions and opening up to the global standards. I expect the more highly engineered products like EVs built in Mexico will usher in the global standards through the southern door. So this will happen whether we like it or not, and politicians are acting like it's the end the world already. Metric units are coming! Oh no!
Imperial units are just one of the things we need to modernize and harmonize. But its a slow, slow process. We should recognize our global competitiveness depends on pulling our heads out of the sand.
So what is needed is greater knowledge of localization strategies in engineering education. You see this a bit with power supplies--they work with multiple plugs and cycles and voltage ranges. Localization engineering is mostly ad hoc, doesn't exist as a discipline currently. As a branch of the major disciplines, we need more engineers who are fluent in the global standards, the unique gaps between foreign and domestic, and how to bridge. No silver bullet solutions here, sadly.
It would help if the US would work on harmonization with global standards. We have our own unique units, our own unique building codes, and that's fine, as I said above, don't change anything. But there is room for some gradual, well planned out transitions and opening up to the global standards. I expect the more highly engineered products like EVs built in Mexico will usher in the global standards through the southern door. So this will happen whether we like it or not, and politicians are acting like it's the end the world already. Metric units are coming! Oh no!
Imperial units are just one of the things we need to modernize and harmonize. But its a slow, slow process. We should recognize our global competitiveness depends on pulling our heads out of the sand.
Mils for PCB rules. Just don't get me started... :-(