Going from Transistors to Gates to.....IC's?

Question

So I am trying to understand I guess what is probably a basic concept of computers. I was taking a look at this previous question: From Transistor to Gates

And it's accompanying link: http://www.cs.bu.edu/~best/courses/modules/Transistors2Gates/

So let me get this straight, we use Transistors to build logic gates correct? Are these all usually Bipolar Junction Transistors? I also see that we also stuff multiple gates into an IC like so: http://www.kpsec.freeuk.com/components/74series.htm (Btw most of these have like....4-6 logic gates within them....is there such thing as a SINGLE gate you can buy, or would you have to build that single NOT/AND/OR/Whatever gate out of other transistors)

So...like for example, the IC's above....are those built using CMOS technology or TTL? Im a bit confused at what the difference between a TTL built transistor and a CMOS transistor. Like is this a TTL transistor or a CMOS transistor? (it is a biopolar one correct?):

I know this is a lot of questions, but Im just trying to understand how we go from Transistors to Gates to IC's. I mean.....the IC's I linked above (like the 74 series Logic IC's) are built from gates.....which must be built from transistors (CMOS or TTL?) But how come we never see like SINGLE gates being sold.....and if that's so then why do big transistors such as the one above exist if we can fit 6 logic gates in a small IC.

And do FET's come into play in Logic Gates at all....or not so much.

Whew....

edit: ALSO any books/websites that talk about this is a HUGE help as well.

Answer 1

The first comprehensive logic series was the TTL series 74xx. This used BJTs (Bipolar Junction Transistors). Later there came variants like the often used 74LSxx, where the "LS" stands for Low-power Schottky TTL. As the name implies these used less power than the rather power-hungry TTL, and were faster too. At the same time the CMOS 4000 series was developed. The "C" in CMOS stands for Complementary, meaning it's a combination of N-channel and P-channel MOSFETs. Their construction is simpler than TTL and they use far less power. Later standard CMOS developed into HCMOS, "H" for High-speed. Most 74LSxx types have been released as HCMOS in the 74HCxx series, or the 74HCTxx series, which is TTL compatible. Later more variants were developed, like Advanced CMOS (74ACxx).

Microcontrollers are built in HCMOS technology, so they use MOSFETs. AFAIK JFETs aren't used for logic ICs. The transistor you show in the picture is a BJT, which you can tell from the pin designation:

E = Emitter
B = Base
C = Collector

For a MOSFET the pins would be

S = Source
G = Gate
D = Drain

respectively.

Many ICs in the 74HCxx series were originally released in 14 or 16 pin DIL packages, which meant that they would fit four 2-input gates. With miniaturization (SMT) came the demand for smaller packages, even if they contained less gates. Several manufacturers offer single-gate and dual-gate versions of logic gates. For example, NXP has a 74LVC1G00 (single 2-input NAND) and a 74LVC2G00 (dual 2-input NAND) version of the classical 74HC00. 74LVCxx is yet another HCMOS technology. This page lists all NXP logic families.

Answer 2

Where to start...

Logic Gates used in VLSI chips are not constructed from Bipolar Junction Transistors (BJT) but rather use Field Effect Transistors (FET) (Metal Oxide Semiconductor (MOS) to be exact). The topologies used in this realm are commonly composed in Complementary P-type pull-up and N-type pull-down networks. Basically I think these types are used in this context because they can be created by a reliable manufacturing process on a very small scale. Perhaps more important is that in CMOS, what counts is the voltage levels, noise margins and switching speeds. All things that are very important in computing technology.

TTL Gates (ala the 7400 series) are integrated (IC) versions of the 3-terminal BJT's as you show in your picture except not quite. They are integrated into appropriate topologies (including integrated "resistors") such that they logically behave as boolean gates from an input/output voltage perspective. Some of the BJT's used in these gates also tend to have multiple emitters for inputs (this is equivalent to multiple 3-terminal BJTs with their collectors and bases connected together). TTL is logically similar to CMOS technology, but the gates can actually drive and sink an appreciable amount of current (e.g. 20 mA+).

The 3-terminal BJT devices tend to have much higher current carrying capacity than the integrated versions, but I'm sure you can find integrated varieties that are comparable. 100mA is common, and upwards of 1A are available. You can immitate TTL logic topologies with them, but that is rarely, if ever, done in practice.

One last thing I'll add is that the terminals of BJT's are commonly labelled E for emitter, C for collector, and B for base. The terminals of a FET are commonly labeled S for source, D for drain, and G for gate (not to be confused with a "logic gate"). The terminals are sort of analogous, but when employed to implement "logic circuits," the topologies are much differnt for these two technologies.

Answer 3

This question already have two good answers, but it's worth mention that this subject is quite extensive to go deep into all the details without writing too much. I highly recommend you to watch these MIT lectures, it's long but worth every minute:

http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-002-circuits-and-electronics-spring-2007/

Lectures 4 and 5 cover what you asked, and lecture 13 is a good addition.