SI
SI
discoversearch

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: Mani1 who wrote (54722)4/8/1999 12:06:00 AM
From: Process Boy  Read Replies (1) | Respond to of 1576239
 
Mani - RE: AMD "mask" problem; I can only speculate, but...

The AMD "mask thing" may have been addressed with an early metal layer (interconnect) redesign, or tapeout. This is approximately half way through the fab process, or a little beyond.

However, really serious design problems may take more than one mask taping out, maybe from an earlier mask layer(s).

While production material is waiting for these fixes, it sits and backfills behind the gating operation. Hence, a "bubble". Fab operations people HATE BUBBLES, as it is much more efficient to run the line in a more linear fashion. The best way to run a line is something like "Just in Time", or, when you finish one run, another just happens to arrive, with little or no queue, or backfill. A bubble traveling through the line over time is a major headache.

Mask layers are a part of the process from start to finish. Essentially, the wafers get a layer of material and then the respective layer is "patterned" with a mask. This process is repeated ~20-25 times (for microprocessor designs these days) and takes (with exceptions) several weeks. Chips are "layered" in this fashion, hence "mask layer".

Mani, this is a VERY CURSORY explanation, and I admit I'm not taking the time right now to go into more detail. If any of the engineers / fab knowledgeable people want to jump in here, feel free.

PB