Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Process Boy who wrote (72321)	9/19/1999 12:37:00 PM
From: Dan3	Read Replies (2) \| Respond to of 1572507

Re: what is the tradeoff in process maturity and process variation? Thanks for your reply. From what you wrote, it sounds as though it may not be an option, maybe AMD needs copper at .18 due to characteristics of their design. If it were copper or a redesign, and they were pretty confident that copper was doable, maybe it was the best solution for them where it wouldn't have made sense for an Intel design. Don't know if the above makes any sense at all, I sure don't know how these things are designed and manufactured. :-) Dan

To: Process Boy who wrote (72321)	9/19/1999 5:15:00 PM
From: Dan3	Read Replies (2) \| Respond to of 1572507

by: Chevyman1963 65127 of 65130 Fact.. Before Liu was introduced to contact 2 layers of circuits you did what is called "contact etch" which basically meant you put a barrrier over the first "circuit" layer and etched contact "holes" that were filler with tungsten to contact up to the next"layer of circuit. With 3 layer metal theis was pretty simple and not very involved. With the advent of the 5 layer and later 6 layer metal circuits problems started to surface. Esentially to keep chips small we went up instead of out and in so doing added circuits on top of one another. The problem is that sometimes the bottom layer needs to be connected to the top or upper layers and therefore the li etch has to extend down 3/4 or even 5 layers and must not be covered up by the circuits above it. The li etch is not a direct etch like the contacts were eg: a straight hole straight down. The li etch is an etch that shoots for a specific area the the underlying circuit should be at. Then you lay the next layer onto the chip and the holes are filled. One of the problems that I alluded to in another post was that sometimes the lower contact area gets etched or eaten away before it is filled at the later metal layer laying. That is what happened to the inital runs of k6. When the resist was removed the contact area for the li etch was eaten away by the chemicals that removed the resist. This was not suppose to be possible and took AMD months to track down. The reason it took so long to track down was that all the li etch areas were covered by no contact was made because the lower area was already gone. Remember todays etch is a dry etch and is very very harsh because of the chemicals used along with the plasma etch process. Once the problem was identifies by partitioniong off the processes yields jumped from 3% to 75% within days. OIbviously I am not detailing everything that is done just giving you the gist of what is done. Remember it takes ~255 processes and ~50 days to complete a wafer. Obviously I have tried to simplify and there is a lot more involved. This is all information and those of you that want to believe it can and those that don't .. Go back to the Intel board. I recommend no investments and just share info to those that ask..If I know I may answer ..if I don't I'll tell ya. Did ya see those Cowboys??? Gooo Chevy Gooo AMD