Talk:Logic family

Definitions

Definition: A group of compatible ICs with the same logic levels and supply voltages for performing various logic functions have been fabricated using a specific circuit configuration which is referred to as a logic family.

Types:

1. Bipolar Logic Families

 i) Saturated
   a) Resistor-Transistor Logic(RTL)
   b) Direct-Coupled transistor logic (DCTL)
   c) Integrated-injection logic (I2L)
   d) Diode-transistor logic (DTL)
   e) High-Threshold Logic(HTL)
   f) Transistor-Transistor logic (TTL)
 ii) Non-Saturated 
   a) Schotty TTL
   b) Emitter-Coupled logic (ECL)

2. Unipolar logic families

 i)   PMOS
 ii)  NMOS
 iii) CMOS               --Krishnavedala 08:10, May 15, 2005 (UTC)

This seems like a very old list of logic families. Perhaps there should also be discussion of:

1. Static Logic

 a) Pulsed Static CMOS
 b) Differential Cascode Voltage Switch (DCVS)
 c) Cascode Non-Threshold Logic (CNTL)
 d) Pass Gate/Transmission Gate Logic
 e) Complementary Pass Gate Logic (CPL)
 f) Push-Pull Logic
 g) Output Prediction Logic (OPL)

2. Dynamic Logic

 a) Domino
 b) Footless Domino
 c) NORA/Zipper Logic
 d) Multiple-Output Domino
 e) Compound Domino
 f) Dual-Rail Domino
 g) Self-Resetting Domino
 f) Sample-Set Differential Logic
 g) Limited Switch Dynamic Logic

Bisaksen 02:20, 24 February 2006 (UTC)

Yes, but there's two kinds of logic families here. One is available as packaged building-block circuits. The other kind are really families of design techniques used within integrated circuits. Your expertise is needed here - give us an outline! --Wtshymanski 00:51, 21 October 2006 (UTC)

Thank you, Bisaksen. I've copied the above list of "static logic" and "dynamic logic" design styles to the article.

--DavidCary (talk) 04:04, 27 August 2013 (UTC)

Non-CMOS families rarely used anymore

I agree with Bisaksen, this article is very out-of-date. CMOS has dominated the industry for the last 20 years but you can't tell that from this page. Digital design is 100% CMOS these-days, with Bipolar famiies only used for analog. TTL dip packaged logic is rarely found anymore in mass market electronics, only used in low volume stuff. Dyl 14:48, 6 August 2006 (UTC)

You know what to do. Write up this oberservation and put it in the summary page; though there's still a lot of TTL etc., being sold each year. A reference would be good - do you have an industry article that says no-one is really designing in the bipolar families any more? And mass market electronics is never representative of the state-of-the-art - if you found an article that says Agilent Technologies hasn't bought any TTL for years, that would be highly interesting and relevant. And provided this is properly noted, I think it's entirely appropriate that an encyclopedia article talk about all the historically important logic families, not just what's used today. After all, the article on Germany talks about a lot more than what happened in 2006 so far. --Wtshymanski 15:42, 6 August 2006 (UTC)

Well, if you look at any PC motherboard or crack open a cellphone, you'll see very little dip packages these days. These items are both mass market AND cutting-edge. Also, in the 1990s, mainframe computers switched to CMOS (away from bipolar ECL) due to lower cooling costs and increasing CMOS frequencies. No, I don't real numbers, but it's relatively obvious as the biggest semiconductor companies these days aren't making their money on dip packages. Yes, an unsubstantiated claim but I don't have the time to do the research. Dyl 16:54, 7 August 2006 (UTC)

If DIP packages were so obsolete as you claim, why would they still be made? If no one was buying them, electronics distributors wouldn't sell them and then the companies that produce them wouldn't bother making them either. However since this isn't the case, I'd say DIP packaging must still be in use today. OracleGuy01 08:49, 3 November 2007 (UTC)

I think the classification that I made was THE classification at that time. I accept that the classification is out of date. But as OLD IS GOLD, we just cant omit the previous work. Just put it in a HISTOR section? --Electron Kid 04:20, 21 October 2006 (UTC)

NMOS and PMOS are hardly discussed anymore, yet why does RTL and DTL even deserve mention? As I recall, NMOS was the big thing before CMOS came along. Most of this old technology will NEVER return for a number of reasons, so only deserves passing interest. You can't even buy most of this stuff anymore. --71.245.164.83 (talk) 02:13, 21 February 2011 (UTC)

The student of CMOS would do well to understand its ancestors. And why are you replying to 2006 comment now? Better to start a new section if you have a point to make. Dicklyon (talk) 02:49, 21 February 2011 (UTC)

Threshold Logic

There seems to be too much detail in the TL section, with too many references cited. Shouldn't such a discussion be provided in a new page for the topic? The summary alone can be provided here, I think. —Preceding unsigned comment added by Vivek22 (talk • contribs) 10:03, 26 February 2008 (UTC)

I've moved the whole works here...it doesn't seem to be a family of ICs, and was out of place where it was. --Wtshymanski (talk) 19:29, 29 March 2008 (UTC)

Threshold logic is a means of switching states build on the concept of a threshold. The concept of a threshold is aligned to a level, e.g. below the level (threshold) may infers a logic zero (‘0’) and above a logic one (‘1’). The initial ideas behind threshold Logic (TL) [1-5] date back to the 1950s. Akers et al. [6] developed TL circuits for recognition tasks. In the 1970s a press release relating to TL appeared [7]. In 1974 [7] TL circuits were combined with Boolean processors and a pattern store in distributed processing elements that make up a pattern recognition system. An extensive survey of TL was carried out by Beiu et al. [8], which categorises the field from the 1950s to the present day.

[1] R. McNaughton, "Unate truth functions," Stanford University, Applied mathematics and statistics Lab, Tech. Report. No.4, reprinted in IRE Trans. on Elec. Comp., Vol. EC-10, pp. 1-6, 1961. October 1957.
[2] M. C. Paull and E. J. McCluskey, "Boolean functions realizable with single threshold devices," Proc. IRE, vol. 48, pp. 1335-1337, 1960.
[3] S. Muroga, I. Toda, and S. Takasu, "Theory of majority decision elements," J. Franklin Inst., vol. 271, pp. 376-418, May, 1961.
[4] P. M. Lewis and C. L. Coates, "Linearly separable switching functions," J. Franklin Inst., vol. 272, pp. 360-410, November, 1961.
[5] R. Winder, "Threshold Logic," Princeton University, N.J. (Ph.D. dissertation), also pub. as Air Force Cambridge Research Lab. Bedford, Mass., Scientific Rept. No. 1 (AFCRL-62-318) May 1962.
[6] S. B. Akers and B. H. Rutter, "The use of threshold logic in character recognition," Proceedings of the IEEE, pp. 931-938, August, 1964.
[7] D. Grover, "A Vision of perfection," in Computing Europe, 1974, pp. 10-11.
[8] V. Beiu, J. M. Quintana, and M. J. Avedillo, "VLSI implementations of threshold logic-a comprehensive survey," IEEE Transactions on Neural Networks, vol. 14, pp. 1217-1243, 2003

Year introduced

Is it true that integrated logic circuits were first introduced in 1962? And that "the IC version of RTL" and "MECL" were each introduced in 1962? (Currently one place in the article says "1962", but another place says "1963"). Is there a better name for "the IC version of RTL", to distinguish it from the much earlier "RTL built out of discrete components"? --68.0.124.33 (talk) 18:38, 23 April 2008 (UTC)

What is true? What does "introduced" mean? A handful of $1000 chips built for a missile or boxcarloads of gates at 25 cents each made for VAXes and similar products? Check out the Computer Museum reference, and Horowitz and Hill's table. --Wtshymanski (talk) 22:20, 23 April 2008 (UTC)

Logic level

The logic level article is quite stubby and could be replaced by a dictionary definition. It is linked from few other articles. The issue of different logic levels only arises when interfacing between families and so I think "logically" belongs here. Explaining logic levels in a free-standing article needs so much context that I think it's better to explain it in a more comprehensive article, rather than splitting out a paragraph into its own article. --Wtshymanski (talk) 13:40, 21 August 2008 (UTC)

I agree. There seem to be 3 closely-related things here that are difficult to explain in a stand-alone article without also mentioning the other two:

• Sets of packaged building-block circuits designed to be directly wired to each other -- described in the logic family article. (Also, the voltage level translation required when connecting a chip from one set to a chip from an incompatible set.)
• The "design styles" used in CPUs and ASICs (and used internally in the above building-block circuits) -- briefly touched on in Dynamic logic (digital electronics), static core, and low-power electronics. And,
• The voltages (or in some cases, the currents) on the wire connecting the output of one gate and the input of another gate, whether both are on the same chip or are on separate chips -- briefly touched on in logic level -- such as so-called "TTL levels" ("TTL-compatible logic levels") mentioned in articles such as MAX232, Differential TTL, Digital signal, etc.

I suggest we temporarily merge logic level into logic family. Also, perhaps some of the content of the above-mentioned articles should also be merged into logic family. I hope that a better way of splitting this content into 2 or more articles (WP:CONSPLIT) will then be more obvious. --DavidCary (talk) 04:15, 27 August 2013 (UTC)

• Oppose. The notion of logic level (mapping an analog quantity into logic values) and logic family (circuit structures to implement logical operations) are reasonably separate ideas. Packaging is an independent issue. Design styles involve engineering choices that impact the choice of logic family, but that seems to encourage confusion. Many designs use two or more logic families to reap the benefits of a hybrid. A merge now only to split later does not seem appropriate. Glrx (talk) 18:04, 28 August 2013 (UTC)

TTL and EV6

Shouldn't there be some mention of EV6 vs TTL or just a external link? I am not an electrical engineer, but... --Ramu50 (talk) 21:46, 10 September 2008 (UTC)

Destroying information

To User:Wtshymanski that reverted edit destroying usefull information on 14:48, 19 January 2010. I wrote this note in your talk page, but you emptied your page, so I put here the note:
Next time try to be constructive rather that simply reverting usefull edits. Do you know there is a discussion page for the article? For example, the complete list of families that I added have red links, but can be converted to a simple list without links. Good work. More bad you removed the BiCMOS from the table that were not red links. To be usefull to the community try to edit more and revert less--Efa (talk) 09:51, 21 January 2010 (UTC)--Efa (talk) 17:02, 24 January 2010 (UTC)

Corrections and comments

This article is pretty lousy. I reorganized and fixed some of it but I am still not happy about the contents. I noticed things that don't look good and I list them below.

1. I reorganized the list to respect the chronology of the technologies. I also linked versions of technologies to the main technologies.

2. RTL: RTL section says 1962. However, the table at the bottom says 1963.

3. DTL: "Diode logic goes back as far as ENIAC". I don't know if it should be DL or DTL.

4. TTL: TTL section says 1963. However, the table at the bottom says 1964.

5. CMOS: CMOS section says 1968. However, the table at the bottom says 1970.

6. I moved "The logic level problem" and "Other families" under CMOS. I also moved "Improved versions" under BiCMOS. I did this for consistency.

The chronology of the technologies is very unclear. Somebody should take the initiative and find out when each technology was truly invented or placed on the market for the public.

ICE77 (talk) 02:23, 15 February 2011 (UTC)

Diode logic in ENIAC. WP:SOFIXIT applies. --Wtshymanski (talk) 15:36, 15 February 2011 (UTC)

And don't worry so much about consistency between sections. Worry about consistency with cited sources instead. Dicklyon (talk) 03:59, 18 February 2011 (UTC)

DTL Not ENIAC

DTL was not used in ENIAC. ENIAC had no transistors. It used vacuum tubes. DTL stands for Diode Transistor Logic. ENIAC used Diode Logic with vacuum tubes. This should be corrected.Thingmaker (talk) 10:29, 13 August 2014 (UTC)

Diode-tube logic? Crystal diodes, anyway, instead of thermionic diodes. --Wtshymanski (talk) 18:33, 13 August 2014 (UTC)

I am not certain what you are saying. In DTL the T denotes Transistor. I have no problem claiming ENIAC used Diode Logic and that they worked with Vacuum Tubes like DTL works with Transistors. Literally the T might denote tube but I believe DTL is commonly accepted as meaning Transistor. At the very least it would be confusing or misleading to say DTL was used in ENIAC. If you allow DTL to include Vacuum Tubes then all the logic circuits using T could mean Tube and the T becomes meaningless. Could it mean Thyristor? That is an amplifier also. I doubt if the term DTL was used in the 1940’s. I know I never heard of it then. Maybe we should coin a new term DVTL. Honestly I am not trying to be combative. I believe this is important and we have an obligation to the readers to make it right. Let’s work together to get it right, whatever that is. If the T in TTL, DTL, and RTL means Vacuum Tube OR Transistor then we need a new notation. Maybe A for amplifier like DAL or TAL. I am only a very junior member of the Wikipedia club but I am a very senior engineer. If what I suggest is wrong then you or someone should correct my work. To make certain you noticed I did change the article.Thingmaker (talk) 19:43, 13 August 2014 (UTC)

I guess you can ignore my comments above. I just saw you changed my edit. I have no problem with what you did. I do think it was worth saying that ENIAC came before the discovery of transistors. My concern for the original article was not about saying vacuum tube diodes did not qualify for DTL. My concern was only for T meaning Tube in DTL. I would like to see the "ENIAC before transistors" added but I will let you decide that. I think it says a lot to say that ENIAC existed that early.Thingmaker (talk) 20:08, 13 August 2014 (UTC)

http://archive.computerhistory.org/resources/text/Knuth_Don_X4100/PDF_index/k-8-pdf/k-8-r5367-1-ENIAC-circuits.pdf page 758 where it describes implementing the boolean functions. (Also interesting is the 5.7V filament voltage....) Glrx (talk) 20:19, 13 August 2014 (UTC)

Silicon on insulator

Can we reliably classify chips based on whether they built on insulator or not? Any sources on that? AXONOV (talk) ⚑ 22:46, 2 December 2021 (UTC)

Do any sources call "silicon on insulator" a "logic family" ? --Wtshymanski (talk) 23:40, 8 December 2021 (UTC)