Talk:Latch (electronics)

From Wikipedia, the free encyclopedia

Contents

[edit] Latches and Flip-Flops

I would like to see an explanation on how the reset input controls the different latches and Flip Fops. Also it would help to see truth tables and charts for the different devices with the relation to the reset line. (Does the reset always return Q to a 0 state and is the reset dependent on a clock signal.)----(User:treese 1/10/07

An asynchronous reset (or set) is not dependent on a clock signal (hence the name), this is mentioned in the introduction to flip-flops. Check the timing-diagram in the datasheets for 7400-series parts (eg. philips/nxp, TI, etc) and you will get a feel for it, as many such parts have both asynchronous and synchronous inputs. /HenkeB 13:32, 17 January 2007 (UTC)

A clearer explanation of the function and use of a latch in the introduction would be really useful. --User:DV 23:18, 1-1-2007

"A clearer explanation"? What do you like to know? The technical function is described (briefly) later in the article. Please give me a hint, or even better, a few concrete questions, and I will try to clarify. /HenkeB 17:20, 4 January 2007 (UTC)

The article should explain the relation between latches and flip-flops, what is the difference? how do they relate to each other?

A flip-flop is a clocked latch. --Yuje 00:09, May 18, 2005 (UTC)

Latches are a bit more complicated than implied in the article. Let me attempt an alternative introduction:

Latches are the simplest binary memory circuits. They have two or more inputs and one output. At least one combination of input values (a so-called input event) causes the output to be 0, at least one input event causes the output to be 1, and at least one input event causes the output to retain the output value which which the previous input event had produced. This latter property is the momorising property of the latch.

Latches with only two inputs are called elementary latches. There are 36 elementary latches of which only 28 are technically feasible. Some of these latches can be realised by devices such as a thyristor, certain pneumatic valves, or certain electric relays. Neither of these devices incorporates a so-called feedback.

All latches can be designed using logic gates and a feedback signal. The feedback (or feedback signal) is the time-delayed output signal used as an additional input (i.e., it is the output signal fed back to the input). To specify and calculate a latch, it is advantageous to use a Karnaugh map or a reduced Karnaugh map. They allow you to design latches that are free of hazards and quick to react to input changes.

The theory of latches is extensively covered in: S.P.Vingron: `Switching Theory. Insight through Predicate Logic.' Springer Verlag 2003. ISBN 3-540-40343-4. --Ving 16:07, 17 October 2005 (UTC)

Latches can be clocked as well. Remember that these things are just circuits, and the inputs to the circuits can be any sort of digital siganl - one could put the clock into the data input, and a data input to the clk - it doesn't change the type of device it is - only its use. A latch is either level-triggered or always transparent, a flip-flop is edge-triggered - the clock has nothing to do with this. Fresheneesz 21:01, 3 November 2006 (UTC)
As [1] states, the difference between a latch and a flip-flop is that a latch doesn't have a clock signal, and a flip-flop does. Yes, you can apply an oscillating signal on a latch's inputs an say "this is a clock". But if the circuit is considered to be a latch, such a signal will be considered not a clock but just a sequence of input states!
Unlike this, flip-flops have a signal that is specially designed to be the clock: it is considered to be not a data but a management input.
Based on these considerations, I suggest to remove all 'clocked' stuff from this article.
Right now, i'm going to write a preface to clarify the above statement. All circuits that can be considered both latches and flip-flops should be present (at least, as links) in both articles and have a note about the ambiguity. — Vano 17:57, 22 November 2006 (UTC)

This oversimplified "distinction" between latches and flip-flops is both completely arbitrary and historically incorrect. The first flip-flops were not clocked (see flip-flop). In fact, they had exacly the same properties as that of a "latch"! A more appropriate categorization could be:

  • Simple bistables (cross coupled inverting elements)
  • Gated bistables (same as above plus some gating logic)
  • Clocked bistables (normally implemented as master-slave devices)
  • Edge-clocked bistables (creates an optimal (short) internal clocking-pulse from a single edge)

Even this is simplified btw. /HenkeB 02:18, 26 November 2006 (UTC)

[edit] Don't Care conditions in the truth table?

Wouldn't using don't cares for the two truth tables make more sense than what's currently there? I mean, I don't want to go ahead, edit them in, and have someone get confused and complain. EDIT: Gah, forgot to sign--Sporkot 03:53, 21 January 2006 (UTC)

[edit] smaller truth table

I added a smaller truth table: I find the bigger one very confusing, while the smaller one (according to me) clearly explains the behaviour of the device and why we use it.

With the small one, the bigger one gets redundant, so I think we might remove it. I didn't do it already because I thought it was better to discuss it. Alessio Damato 22:25, 11 May 2006 (UTC)

I agree the larger one is redunant and is no longer necessary--Rehnn83 14:06, 21 June 2006 (UTC)
I've changed the tables to the way they gave them at lectures (simple and clear). If the terms are uncommon for US (i'm from Russia), you're welcome to change them. (btw we say RS- instead of SR- and trigger instead of latch or flip-flop ;) ) — Vano 21:59, 15 November 2006 (UTC)
Similar in Swedish. We use the term "vippa" for all types of bistable elements. Types are distinguished by prefixes such as clocked, transparent, RS, JK, etc.
/HenkeB 09:11, 26 November 2006 (UTC)
SR? I've never heard them called SR. In fact both this page, and the flip-flop one that links to it had me confused for a few seconds. A search for "RS flip-flop" vs. "SR flip-flip" show RS coming up 5 times as much. A search for "SR latch" vs. "RS latch" shows equal usage. Unless modern EE texts are now calling them SR for a valid reason, I'd like to see RS. If SR is now the standard, I'd like to see an explanation about why RS is no longer used. —The preceding unsigned comment was added by 68.60.59.250 (talk) 10:17, 6 January 2007 (UTC). Ooops, sorry. I'll sign it now. 68.60.59.250 16:33, 11 January 2007 (UTC)

[edit] disambiguations

hi, i'm not sure how to do this, but can one if you make RS Latch (and the other kinds of latches this page mentions) link to this one when searching? Thanks. —The preceding unsigned comment was added by 128.32.77.56 (talk) 02:43, 12 February 2007 (UTC).

[edit] Error(s)

There is an error with the right drawing of the JK flip-flop. Either J and K or Q and -Q have to be swapped (not the labels but the wires) for it to work as given in the truth table. And for the left picture... This does not seem to be correct at all. (15th march)


I agree, this error is correct. The image should removed, or fixed