Template talk:Semiconductor manufacturing processes

Template contents, format, and purpose
This template is not intended as a location for content. It is a navigational aide that lets a reader go to the article that describes the listed process. The template is intended to be right-aligned and vertically oriented in articles that include it, so we want it to be narrow. Please do not add extra information to the entries that will make the template wider. Instead, add such information to the articles. If you wish to add an entry, create the article first, even if it is a stub. If you disagree with this approach, please discuss your concerns here. -Arch dude (talk) 23:59, 25 February 2010 (UTC)

Dates
The dates in this template should match the main date given in the associate article. We need to do this because putting references in the template is unreasonable, but any assertion of fact needs to be references somewhere. If you feel that this in incorrect, please discuss it here. -Arch dude (talk) 15:09, 2 April 2010 (UTC)


 * Hm.. 2022 for the 11nm does really not make sense to me and i work in this field, i also do not see this reflected in the sources. To make a start.


 * Where are these dates based on. When the first prototypes has been successfully produced or when mass production has started?


 * There are tons of WP:RS that speak about 2015 or 2016. Why should we base this "estimation" only on the one source that mention 2022 and ignore all the other sources? IQinn (talk) 15:35, 2 April 2010 (UTC)
 * You are correct: we should NOT base the estimate only on the sources that are currently in the 16 nm and 11 nm articles. Therefore, Please edit those two articles to include your sources, and edit the first paragraph (the "lede") of those two articles to give your new dates. After you have done that, then come back here and adjust this template. I feel that we need to do it this way because that will expose your research to the appropriate Wikipedia review process, such as it is, while simple changes to this template do not provide any easy mechanism to comment on your sources. Oh, and also: Thanks! We really do need people like you to help in the areas they know about. If you need help with formatting your references, or any other editing task, please leave a message on my talk page. -Arch dude (talk) 18:35, 2 April 2010 (UTC)


 * 10 nm is definitely 2015 or 2016 latest. 118.160.101.72 (talk) 12:17, 7 October 2013 (UTC)

Half-nodes
In I tried to show the half-shrinks, but these were later removed, because "the template becomes too wide". Should we show them somehow differently, or the current half-node link at the bottom is sufficient?

In any case doubling the number of list lines doesn't seem a good solution, especially since the half-shrinks are closely associated with the respective node (e.g. 65nm/55nm, etc.), that's why I made it by slightly longer lines instead. Ianteraf (talk) 06:25, 31 August 2011 (UTC)

Reversion from Conversion to "sidebar"
An editor made a good-faith attempt to convert this to a sidebar. the result was about twice as wide and affected the appearance of several of the pages using this template. I made several attempts to fix this by adding parameters from the sidebar and navbox documentation, but failed. I would be happy to work with that editor or others to convert to using sidebar if we can get past the width problem.

There are two additional problems, but those are easily fixed: Thanks. -Arch dude (talk) 00:23, 11 November 2011 (UTC)
 * the change also moved the documentation below the template. That is the usual position for the documentation, but I have been working with this template for a long time, and in my experience we get a lot of bad edits here due to folks taht do not read the documentation. Please discuss.
 * Sidebar by default adds the template navbar (v.d.e) to simplify editing by readers of the article. We have had the navbar on this template twice in the past, and in both cases it led to lots of bad edits. fortunately, the sidebar template has a parameter to turn navbar off, so this is fixable if we can figure out how to remove the whitespace from sidebar. Again before re-adding the navbar, please discuss.


 * I've had another go, this time preserving the size and hiding the navbar. Chris Cunningham (user:thumperward) (talk) 09:15, 11 November 2011 (UTC)
 * Thanks, Chris. That looks great. I will not put the documentation in front of the template again unless we see some more bad edits. -Arch dude (talk) 19:00, 11 November 2011 (UTC)

I added the navbar back minus the edit link (using 1. Feel free to revert if it still causes problems. 50.53.15.51 (talk) 19:36, 8 May 2012 (UTC)

10nm in 2017
From what I've read, in 2016 TSMC will only do risk production - that doesn't count, right? As I understand it, risk production means that they are tweaking the production process. Regular production will supposedly start early 2017, and Intel even later in 2nd half of 2017. Source (German): http://www.heise.de/newsticker/meldung/TSMC-will-Intel-bei-10-nm-ueberholen-2757231.html --91.45.159.136 (talk) 16:39, 13 August 2015 (UTC)

2013 ITRS technology summary includes target dates for 3.5nm, 2.5nm and 1.8nm
I'm loth to edit this template myself but Summary technology trend targets gives predictions for 4 more nodes. - Should we now start 3.5 nanometer etc ? - Rod57 (talk) 20:32, 2 January 2016 (UTC)

Changed 10nm to 2017 to match with the 10nm article
The 10nm article says: "as of 2016, 10 nm devices are still under commercial development. Commercial release is expected to commence in 2017 by Intel.[1]" — Preceding unsigned comment added by 83.165.147.196 (talk) 07:35, 7 August 2016 (UTC)

Time to add 3.5, 2.5 and 1.5nm?
Now these are not official by any means, but are talked about in the industry and are widely discussed and debated on the Internet. Transistor Options Beyond 3nm. I think a ten year foresight towards 1nm is appropriate. Beatitudinem (talk) 06:28, 23 December 2018 (UTC)
 * This is a navigational template to help users get to the appropriate articles. You are free to add these to the template as soon as there are Wikipedia articles on these subjects, but not before. Among other reasons, the template becomes messy and intrusive if you add references to it, but every fact needs to be supported by a citation, so you put the citations in the associated article. You are free to add those articles to Wikipedia if (as you assert) there are reliable sources (WP:RS) for them. -Arch dude (talk) 17:32, 24 December 2018 (UTC)
 * Agreed, and I would love to see more articles about the smaller transistor standards, but I don't have more than average knowledge about the subject, so I'd rather wait for users with more knowledge to create them. Beatitudinem (talk) 08:43, 26 December 2018 (UTC)
 * You don't need specialist knowledge, and in fact it may be detrimental, since specialists sometimes make incorrect assumptions about the readership's knowledge. You just need to find reliable sources, read them, cite them, and summarize them in your own words. A specialist might be able to more quickly evaluate the quality of the sources, but that's all. If you have concerns about the sources quality, then come back here and ask. -Arch dude (talk) 02:33, 27 December 2018 (UTC)

2017 IRDS summary with target dates for 3nm, 2.1nm, 1.5nm and 1.0nm
From IRDS, successor of ITRS. 2017 Summary provides predictions for 4 more node range labels. — Preceding unsigned comment added by 135.0.49.55 (talk) 16:58, 12 March 2019 (UTC)

Added 3nm
(as per comment above) Article 3 nanometer exists. usage "3nm" in common use by manufactures/journalists. ITRS also uses term - eg see :


 * https://irds.ieee.org/images/files/pdf/2017/2017IRDS_ES.pdf Table ES2, p.18 ; "INTERNATIONAL ROADMAP FOR DEVICES AND SYSTEMS 2017 EDITION - EXECUTIVE SUMMARY"

or


 * https://irds.ieee.org/images/files/pdf/2017/2017IRDS_MM.pdf Multiple uses in tables - "INTERNATIONAL ROADMAP FOR DEVICES AND SYSTEMS 2017 EDITION - MORE MOORE"

Xoool (talk)

Indicating the discrepancy between process names and gate lengths
It's fairly hard to understand what the process names actually mean (both absolutely and in relation to one another) since they've become decoupled from the physical characteristics of the actual MOSFETs. This source says:

"[I]n 1997 with the “0.25 micron/250 nm” node [Intel] started over-achieving with an actual Lg of 200 nm – 20% better than the name would imply. This “sandbagging” continued through the next 12 years, with one node (130nm) having Lg of only 70nm – almost a 2x buffer. Then, in 2011, Intel jumped over to the other side of the ledger, ushering in what we might call the “overstating decade” with the “22nm” node sporting an Lg of 26 nm. Since then, things have continued to slide further in that direction, with the current “10nm” node measuring in with an  Lg of 18 nm – almost 2x on the other side of the “named” dimension."

I suggest we mark the 'understated' process names in bold, and the 'exaggerated' process names in italics, with a footnote to explain the formatting. Then, someone looking at the template will be able to tell the relationship between the process name and the gate length at a glance. Galagora (talk) 14:41, 28 May 2024 (UTC)