Talk:Static random-access memory

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Hardware (Rated C-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by Computer hardware task force (marked as High-importance).
 

Error in Diagram?[edit]

http://en.wikipedia.org/wiki/File:SRAM_Cell_(6_Transistors).svg has two transistors labeled "M5" and none labeled "M3". It seems like the lower-right M5 should be M3. —Preceding unsigned comment added by 12.193.211.18 (talk) 17:11, 14 July 2009 (UTC)

Anomaly in reading paragraph[edit]

Could someone explain why opposite becomes opposfhhghghghghtryrtyrryytytyrtyyeyte? Editing the page didn't show this error Quote: Reading

Assume that the content of the memory is a 1, stored at Q. The read cycle is started by precharging both the bit lines to a logical 1, then asserting the word line WL, enabling both the access transistors. The second step occurs when the values stored in Q and Q are transferred to the bit lines by leaving BL at its precharged value and discharging BL through M1 and M5 to a logical 0. On the BL side, the transistors M4 and M6 pull the bit line toward VDD, a logical 1. If the content of the memory was a 0, the opposfhhghghghghtryrtyrryytytyrtyyeyte would happen and BL would be pulled toward 1 and BL toward 0. —Preceding unsigned comment added by 85.211.82.37 (talk) 12:14, 10 September 2007 (UTC)

address lines[edit]

if there are 'n' address lines and 'm' data line then max size of RAM that can be used is 2^n *m sadf —Preceding unsigned comment added by 14.99.81.238 (talk) 20:19, 11 April 2011 (UTC)

hobby purpose[edit]

the PIC datasheets do not use this term. however, they include small amounts of SRAM onchip and can access serial RAM easily. i do not see a point why a "hobbyist" should stay away from dynamic ram.

now, SRAM is ten times faster than DRAM and keeps the information. DRAM needs to be refreshed. the statement is objective technical documentation should not ... contain objectionAkidd_dublin 2005 0817

I think you meant to say 'technical documentation should not be SUBJECTIVE.' The statement that DRAM needs to be refreshed while SRAM does not is totally accurate and is not subjective. Saying SRAM is faster than DRAM is also accurate. However, it is a bad idea to put a factor on the speed difference (I don't think the article does this) since it will differ somewhat in implementations. This article really needs some cleanup, hopefully I can get around to it sometime this week -- uberpenguin 04:39, 2005 Feb 8 (UTC)

there is a trend away from parallel concepts, see USB. serial ram: [1] ...some unrelated discussion data

  • I have now added asection about SRAM for track buffer purpose.
  • I have re-written the section "hobby purpose", extended and re-named it.
    I have not removed previous formulations (they are included).
  • Feel free to improve the section if "need be". alex 13:18, 1 May 2006 (UTC)

Make the title lowercase?[edit]

According to the Wikipedia policy on article naming, the title of this article should probably be in lowercase (static random access memory, like dynamic random access memory). Would there be any objections to moving this article to a lowercase title? The move would preserve the edit history and automatically create a redirect from the old title to the new title. Aapo Laitinen 22:11, 13 November 2005 (UTC)

It should read "SRAM" (or "S-RAM"). What about "RAM_(static)", "Random_access_memory_(static)"? (the previous discussion data was unrelated and has been removed). probably "SRAM(static random access memory)" ships around the wikipedia policy. alex 12:43, 1 May 2006 (UTC)

Description of operation[edit]

I like the direction some of the recent edits have taken, but I think they are unnecessarily complicated for most purposes. My suggestion would be to use the cross-coupled inverters / T-gates model of an SRAM cell to describe it. This is, I think, much easier conceptually than immediately introducing a schematic of an actual MOSFET SRAM cell. For someone just trying to grasp the concept, it might be less daunting to use a slightly higher level functional model and then at the end speak about real implementation and why the two extra p-channel FETs introduced by the inverter/T-gate model are usually left out of production. -- uberpenguin 23:28, 9 January 2006 (UTC)

I agree, I was also a bit worried that it would turn out to be too complex. I might have overdone it a bit :-) . Help in simplifying and explaining the concept better would be much appreciated. My initial idea would be to simplify to something like this:
A write operation is started by setting the bit lines to the values that are to be stored. Then the word line is set to 1 and the new data is latched into the memory. and
A read operation is initialized by pre-charging both bit lines to a 1. The next step sets the word line to 1 which puts the contents of the memory onto the bit lines.
I do however think that a little bit of more detail than that would be nice, but it might be better put in the end, like you suggest. Abelsson 08:08, 10 January 2006 (UTC)

recent edits[edit]

SRAM in amateur and research applications[edit]

SRAM has advantage in design complexity. No DRAM refresh cycles, address and data bus not multiplexed. The operation principe requires three controls:

Chip enable (CE) causes the chip to listen to the bus. In systems containing multiple SRAM chips, a demultiplexer such as TTL 74LS154 converts upper bits of the address (or other address controls) into chip-enable signals. Write enable (WE) causes data bus data to be stored in the memory cells. Output enable (OE, often /OE) decouples the unit from the data bus while the address is selected, or if the unit is logically inactive. This is called three-state logic. Modern implementations may contain burst logic, which requires additional controls.

  • Write operation
    • The chip-enable is set (if not set already).
    • The address is put on the address bus.
    • The data is put on the data bus.
    • The write-enable controls data storage. This is requires different consideration for static logic, or triggered(clocked) logic. Typically this step requires a specific amount of [nanoseconds.
  • Read operation
    • The chip-enable is set (if not already).
    • The address is put on the address bus.
    • The output-enable is set, and the data is put to the internal output buffer, then the data on the bus gets logically valid.

The logic analyzer diagram's provided by the manufacturer's datasheet contain timing details. Address inputs need to stay constant while the RAM chip transfers data to an output buffer, or if none is present, as long as the data output is enabled.


I do not mind the recent edits. I have consulted a technical manual before writing this. I have tried to put it to easy language. If i insists on this version, it is called POV (point of view). However, i do not see this (above) explanation wrong. Probably it does not belong here. Like superstition, indeeds this does not scroll superstitions, but it gives (wrong) explanation, which is very abstract.

  • Someone professional should edit this article. I have RAM chips at home, and i do not mean the one's in the personal computer.
  • What's wrong with this section? This is not of interest for unrelated people. They should read a short overview. As of today, it reads very complicated to me "should not confuse SRAM with..." - Akidd dublin (abandoned 5/2006) 18:36, 12 May 2006 (UTC)
Hi! Your English is a bit hard to follow, but I removed most of the above section because it is specific to certain old sram chips, it's not information that is general to SRAM. Hobbyist usage does warrant a mention in the article, but discrete SRAM circuits of the type described in the section above aren't very common nowadays. Henrik 19:06, 12 May 2006 (UTC)
I reverted your edits because your English is poor and I could not understand what you were trying to say. I have no doubt that you meant well and that you know your subject; Wikipedia simply demands excellent grammar and spelling in order to be comprehensible. As other have suggested, you may want to try editing the Wikipedia for your native language, as your contributions will certainly be valuable there. And incidentally, I don't think changing user accounts is going to solve your problems here - only changing the quality of your edits will do that. Aguerriero (talk) 20:10, 12 May 2006 (UTC)
That's allrighty. I find it time-consuming to understand technical manuals. I tried to expand a section (hobby purpose). This is now accredited: "poor grammar". Anyway, it does not belong to the article: No "how-to" explanation, it is mentioned in the policies. I do now understand it is an encyclopedia for the public, not for someone looking for instructions "how to use SRAM technically". My mistake. Yy-bo 13:19, 17 May 2006 (UTC)

good english[edit]

The power consumption of SRAM varies widely depending on clock speed.
Fast SRAM is much more power-hungry than DRAM, and 
some ICs can consume many watts at full speed. 
Slow SRAM, such as the battery-powered "CMOS" RAM on PC motherboards,
can have a very low power consumption, in the region of a microwatt when sitting idle.

1. whatever this is, it is not called "intelspeak" (or motospeak). on one side you criticize my english, on the other side you employ weak language constructs, which would never make it into any official documentation.

2. i believe my section understandable for people who have to do with real SRAM. others can just skip it. if it does not belong here (but wikibooks), it is pointless to criticize it for the quality level of its usage of english language. —Preceding unsigned comment added by 84.203.112.173 (talkcontribs)


It seems that for READING and WRITING "1" - the same values are to be set to the bit lines: BL = 1. Please calrify.

Real reference manuals (programming, electronics) are impossible to understand for "Joe Public" (american slang). No pun intended, but it is specific information. Usually there are introduction sections, which are easily understandable. I do not know if it can be demanded, to translate information towards secondary school english (of course grammatically correct, and using sentences of present age).

By the way Joe_Public exists... User:Yy-bo 14:51, 23 August 2006 (UTC)

Where is it used???[edit]

If you are online in the next 10mins from now (5:10pm EST Australia), please answer this question - Where is it used, eg. in small portable computers?

Thanks alot, 'Ritty5 07:16, 17 July 2006 (UTC)'.

Power consumption vs. DRAM[edit]

The article compares SRAM vs DRAM power consumption at the component level. This is misleading because SRAM and DRAM chips have very different capacities and transfer rates. Active power should be normalized with respect to transfer rate. Idle power should be normalized wrt capacity. —Ryan 04:54, 23 March 2007 (UTC)

access time[edit]

it is stated that SRAM is fast: how many ns does a read take, how many ns does a write takes ? why is DRAM so much slower ?

The specific timing will depend on the process, layout, and manufacturer, so there is little that can be said in general. DRAM is slower because of the larger capacitance of the trenched capacitor as opposed to the capacitance on the gates of a few MOSFETs in SRAM. -- mattb @ 2007-04-05T16:18Z

Pseudo-static[edit]

What, no mention of pseudo-static? (I tried to use one once, however it would not work so I had to use a conventional 62256 SRAM. Wish I knew why.) —Preceding unsigned comment added by 24.7.178.142 (talk) 17:39, 30 May 2008 (UTC)

<6T SRAM[edit]

The article says:

Memory cells that use fewer than 6 transistors are possible — but such 3T or 1T cells are DRAM, not SRAM (even 1T-SRAM).

This seems to suggest that the lower limit for SRAM is 6 transistors, but that is not true, you can make a working 5T SRAM by removing one of the access transistors (I know it has been done), and/or I think you can replace one or two of the transistors in the latch with resistors (I think this has been done). Or you can make a different design from scratch (pretty sure this has been done).

But I know very little about how this stuff, so I came here in the hope of some info about <6T SRAM. Is there anybody that knows about this stuff who could add something about alternative SRAM designs, how they work, advantages/disadvantages etc? I can't add anything myself, as I said, I know very little about the subject, I'm not fluent in english, and I'm a total wiki newbie, I get nervous just adding anything to the comment section... Thowllly (talk) 09:10, 18 February 2009 (UTC)

An old standby SRAM design uses 4T and 2R. There have also been various 4T designs, some less "static" than others; the so-called "loadless" or "sort of static" ones are pretty interesting, in my opinion. Dicklyon (talk) 04:30, 19 February 2009 (UTC)

Trademarks[edit]

ZBT is a Micron trademark. NBT, a non-trademarked term, means the same thing.

SyncBurst is also a Micron trademark. The terms "synchronous burst" and simply "burst" commonly mean the same thing.

Gretchenpatti (talk) 20:16, 3 March 2009 (UTC)

Comment and question[edit]

1. "At present appeared the advanced scheme [7] with disconnected by signal record by feedback, which does not require the transistor of the load and it's accordingly saved from high consumption of the energy when writing."

Something is wrong with this sentence.

2. "static RAM used at a somewhat slower pace, such as in applications with moderately clocked microprocessors, draw very little"

What would be the range of "moderate frequency"? The above statement is a little too generic.

ICE77 (talk) 00:54, 4 July 2011 (UTC)

I'd remove the weird uninterpretable bit. Static RAM consumes power proportional to clock rate, so moderating the rate moderates the power by about the same factor. Dicklyon (talk) 01:32, 4 July 2011 (UTC)

Look out for possible copyright violations in this article[edit]

This article has been found to be edited by students of the Wikipedia:India Education Program project as part of their (still ongoing) course-work. Unfortunately, many of the edits in this program so far have been identified as plain copy-jobs from books and online resources and therefore had to be reverted. See the India Education Program talk page for details. In order to maintain the WP standards and policies, let's all have a careful eye on this and other related articles to ensure that no copyrighted material remains in here. --Matthiaspaul (talk) 14:20, 30 October 2011 (UTC)

History[edit]

A history section would be useful.   — C M B J   06:43, 1 March 2013 (UTC)

"but it is still volatile in the conventional sense that data is eventually lost when the memory is not powered."[edit]

"but it is still volatile in the conventional sense that data is eventually lost when the memory is not powered."

The sentence appears in the article summary. Would someone first tell me why volatile is bolded (just curious, dunno if it's wrong). Wgfcrafty (talk) 02:48, 16 June 2014 (UTC) Wgfcrafty (talk) 02:48, 16 June 2014 (UTC)