Talk:RS-232/Archive 2
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 | Archive 4 | Archive 5 |
What are the other pins for?
I have a barcode scanner with pin 11 of a DB25 plug connected. But the article does not mention pin 11; and might there be other pins in use of which the article makes no mention?--83.105.33.91 15:23, 17 October 2007 (UTC)
232 or 232C?
I just learned that the standard is sometimes referred to as 232C. An internet search doesn't bring any useful results, and the article as it is now contains two random references to 232C. Can anyone address this? Binba (talk) 23:14, 5 December 2007 (UTC)
- The letter is the revision. Those are mentioned in the History section. --J Clear (talk) 16:01, 19 January 2008 (UTC)
Transmitter Timing
In the timing signals section the article mentions that the DTE can provide a timing signal (TT) for both transmit and receive data. But I believe TT is only in reference to the transmitted data and that receive data is solely timed by the DCE. Since the primary reason to use TT is to overcome problem where the relation between send timing (from the DCE) and transmitted data is variable with the cable length and design of the DTE. There should be no such problem between receive data and receive timing signals as they will traverse the same cable length. In fact using TT for RxD would just reverse the timing issue.
Anyway I think I talked myself into rewriting that paragraph. --J Clear (talk) 16:19, 19 January 2008 (UTC)
Early history?
The earliest date in the article is 1969, when the C revision was issued. Could someone who knows please add dates for the earlier versions? Loganberry (Talk) 19:19, 5 March 2008 (UTC)
Does the standard define the cable length or not?
This page says that it does: EIA standard RS-232-C as of 1969 defines: [...] and cable length
The Serial Cable page says it doesn't: The RS 232 standard states that a compliant port must provide defined signal characteristics for a capacitive load of 2500 pF. This does not correspond to a fixed length of cable since varying cables have different characteristics.
I've googled around and a lot of pages suggest that there is no per-standard defined maximum length (BTW most pages suggest 50 feet @ 20Kbps as a rough approximation) Ndemou (talk) 16:05, 11 March 2008 (UTC)
- I screwed this up back in April 2005 and no-one has noticed it until now. I've fixed it. I'm sure the standard doesn't define a cable length and I'll review it when I get home. I hate to think how many mirror sites now have wrong information due to my error - which will never get fixed, because they only "clone" Wikipedia at erratic intervals. When you see this sort of error WP:Be Bold and fix it. (I like to think that on the whole I've made more good edits than errors.)--Wtshymanski (talk) 19:39, 11 March 2008 (UTC)
- In general old standards often defined cable lengths, and new standards define capacitive loading. So this may be one of the differences between RS232 and RS232C 150.101.166.15 (talk) 08:11, 13 August 2008 (UTC)
- Probably not, and not "in general". Page 8 of revision C recommends short cables, less than 50 feet or 15 metres, but says longer cables are OK up to the 2500 pf total capacitance limit. So, the length is not specified in C as a hard limit but 2500 pF is. I continue to be frustrated that of all the hundreds of thousands of editors on Wikipedia, not one of them admits here to having a copy of the current edition of the standard. --Wtshymanski (talk) 04:15, 15 August 2008 (UTC)
- In general old standards often defined cable lengths, and new standards define capacitive loading. So this may be one of the differences between RS232 and RS232C 150.101.166.15 (talk) 08:11, 13 August 2008 (UTC)
Should there be a nice box below the photo?
Even if RS-232 isn't a 'bus specification', could it benefit from a nice summary box below the photos, as in IEEE-488 and various DVD formats? Also, even approximations of maximal lengths of various cable types is better than none. Geologist (talk) 19:33, 4 June 2008 (UTC)
- I'm not sure what would be in the summary. The IEEE 488 example box does not strike me as particularly informative or authoritative or even accurate (since the 1960's - c'mon, we should be able to reasearch things a little better than that. "Specifications - external yes" - gibberish. Manufacturer Hewlett Packard - I'm sure there are literally millions of devices out there with an IEEE 488 that never saw the inside of an HP plant. Etc., and so on). It's a lot of work to format a box for only one article. --Wtshymanski (talk) 21:40, 4 June 2008 (UTC)
Miscellaneous corrections and clarifications
I've gone through this article and made a few corrections and clarifications along the way. Most should be fairly self explanatory but a few may need a little additional comment.
- To state that USB is higher speed than RS-232 is oversimplifying things a bit. Certainly USB has much greater bandwidth, but also much higher latency. USB is a polled bus and response times under 1-2 ms are not possible. RS-232 is a point to point link where it is easy to add interrupts - response times can be sub-microsecond with the right OS.
- Also comparing it to USB, the higher voltages employed are not in themselves a bad thing. All other things being equal higher voltages give better noise immunity. OK USB uses differential signalling for the same end but that does not make the higher voltages bad. Slew rates may limit the ultimate bandwidth acheivable with higher voltages but that is also not a disadvantage here - RS-232 was never intended to be super-fast and was naturally designed for the speeds that it was intended for.
- Cable length - the 50' thing is a myth and always has been. I've run cables ten times that distance using plain Cat5 cabling and had no trouble driving the terminals at 57.6kbaud. I've seen lines around a mile long with nothing more than a line booster at each end. I've added a relevant cite there.
- The 16550 hadn't even been invented when the original IBM PC was launched, much less used in its design. I've expanded that section to reflect both the original and modern UARTs.
CrispMuncher (talk) 20:02, 29 September 2008 (UTC)
- All good points. I too have run EIA232 over far longer runs than fifty feet. And you get a "special mention" for adding a cite rather than simply quoting personal experience in the article. Jeh (talk) 23:08, 29 September 2008 (UTC)
- "Latency" is a weird thing to claim as an RS 232 advantage - a character time takes at least 50 uS. USB is faster - that's why it's used. The rule-of-thumb is given in RS 232C itself (see Section 3.1 of EIA RS 232 C - you do have a copy of the standard around, don't you? It says " The use of short cables (each less than 50 feet or 15 metres) is recommended; however, longer cables are permissible,provided that the resulting load capacitance (Cl of fig. 2.1), measured at the interface point and including the signal terminator, does not exceed 2500 picofarads. " ), and EIA shows many other standards better suited to long-distance communication. See Horowitz and Hill "Art of electronics", for example. This article is not the history of the IBM PC and so ibm trivia does not belong here. There is no INS 8250 article, likely never will be (I hope), so a link to the 16450 is the best that Wikipedia can do. I don't doubt that someone out there is sending 19200 baud data with RS 232 hooked up to a rusty fence for a couple of miles...but it's not the way it's done in the real world.
- Firstly, information may be transferred in much less time than a character time - consider what happens when you assert CTS for instance. Latency is a very real factor in many links. One particularly severe instance is a 27Cxx EPROM programmer I have here. This uses a rapid dialog between the programmer and the host computer of the form:
- >>> [high address][low address][data]
- <<< [OK]
- >>> [high address][low address][data]
- <<< [OK]
- >>> [high address][low address][data]
- ... and so on.
- That works at 38.4 kbaud and programs at a relatively slow 800 or so bytes/second. Hook it up through a USB↔RS-232 converter, however, and performance collapses to around 100-110 bytes/second. That isn't due to some interaction between the RS-232 and the USB connections or because the RS-232 port is being bit-banged in some non-standard way - it is simply that USB is not capable of two-way communications of the kind of speed that device needs in order to run flat out. To simply state that USB is faster, period, when it is unable to transfer 3.2 kbytes/second in this instance is either grossly simplistic or just plain wrong. This isn't a particularly unusual situation either - many devices only ever transmit in very short bursts.
- Firstly, information may be transferred in much less time than a character time - consider what happens when you assert CTS for instance. Latency is a very real factor in many links. One particularly severe instance is a 27Cxx EPROM programmer I have here. This uses a rapid dialog between the programmer and the host computer of the form:
- "Latency" is a weird thing to claim as an RS 232 advantage - a character time takes at least 50 uS. USB is faster - that's why it's used. The rule-of-thumb is given in RS 232C itself (see Section 3.1 of EIA RS 232 C - you do have a copy of the standard around, don't you? It says " The use of short cables (each less than 50 feet or 15 metres) is recommended; however, longer cables are permissible,provided that the resulting load capacitance (Cl of fig. 2.1), measured at the interface point and including the signal terminator, does not exceed 2500 picofarads. " ), and EIA shows many other standards better suited to long-distance communication. See Horowitz and Hill "Art of electronics", for example. This article is not the history of the IBM PC and so ibm trivia does not belong here. There is no INS 8250 article, likely never will be (I hope), so a link to the 16450 is the best that Wikipedia can do. I don't doubt that someone out there is sending 19200 baud data with RS 232 hooked up to a rusty fence for a couple of miles...but it's not the way it's done in the real world.
- As for cable length, I don't have the standards in front of me - that will have to wait until Monday now. From memory I believe you are correct but maybe placing to much emphasis on a woolier part of the standard - as you yourself state it is the capacitance that is ultimately the limiting factor. I don't regard Cat5 (or even worse, telephone wire) as particularly "special".
- If we don't have a page on the 8250 that is not reason to silently link to 16550 because it implies more equivalence than actually exists. We have a generic UART article to link to instead. However after a little looking I found an 8250 article but in an unlikely place - I've moved it to its present position. I would hope that change at least is uncontroversial so I'll go ahead and make it now. I disagree with your views on coverage of IBM serial ports - in my experience a little coverage acknowledging the PC world helps to ensure that no-one comes along in the future and rewrites the article solely in terms of PC - witness computer bus for instance. However I understand where you are coming from with your thoughts so I'm quite prepared to cede that point. CrispMuncher (talk) 19:56, 4 October 2008 (UTC)
Pinout: 10P10C
10P10C is a type of connector. Could someone please describe the standard according to which this is the 10P10C pinout? Hpa (talk) 22:06, 1 December 2008 (UTC)
- This is not defined by RS 232 and so is not really on-topic for this article. --Wtshymanski (talk) 23:54, 1 December 2008 (UTC)
- A collection of pinout information (not defined in the standard) is not on-topic here; put it in the Wikibook. --Wtshymanski (talk) 14:51, 8 December 2008 (UTC)
Oscilloscope Diagram Slew Rate
The simulated oscilloscope diagram shows a signal with a non-conforming slew rate. A real signal has to cross the transition region in no more than 4% of the signalling unit, with a maximum of 1ms. The diagram looks more like 10 to 15%. (Also, line diagrams should be done in lossless formats.) --David Woolley (talk) 23:57, 30 December 2008 (UTC)
Power Down Detection (RS232 C paragraph 2.5).
There is no mention of the powered down interface detection requirements, which mean that the effective negative going threshold for some signals is actually greater than 0 volts. In fact, the last paragraph of the voltage section implies that zero volts is never defined, when it is actually defined for certain critical control signals.
In practice, this requirement results in all IC implementations of line receivers having a negative going threshold of >0V, and the de facto ability to drive RS232 C receivers with TTL level drivers.
-- David Woolley (talk) 00:03, 31 December 2008 (UTC)
Headline Picture shows Non-Compliant Connector
The headline picture for the article shows a connector that does not comply with the standard, as it has too few pins and does not use the pin numbering specified in the standard. I think this partially results from a popular confusion between the standard and the loose use of RS232 to refer to PC asynchronous interface ports. (Another part of this confusion is to assume that RS232 equates to asynchronous.) -- David Woolley (talk) 00:13, 31 December 2008 (UTC)
- Noted. Please add your recommended changes to the article. Do you have a later edition of the standard text than rev. c, which is now quite out of date? --Wtshymanski (talk) 05:28, 31 December 2008 (UTC)
- There is an inconsistency here. On the one hand, Wtshymanski, you have previously argued against changing the article name to TIA-232-F, on the basis that "What a thing is called is what people call it." And I can see that by the "principle of least astonishment", "RS-232" is in such common use that even a redirect page (RS-232 redirecting to TIA-232-F) would cause too much astonishment.
- But it seems to me that it is also certainly the case that those 9-pin connectors are called "RS-232" when they're on the backs of PCs. (Then again they are also commonly called "DB-9" when they're really DE-9, but that is another topic for another article's talk page.) I think this problem with this picture could have been easily and more usefully fixed by changing its caption. Jeh (talk) 21:15, 31 December 2008 (UTC)
RS-432
What is RS-432? Should that redirect here? 2fort5r (talk) 12:57, 22 August 2009 (UTC)
- 60 seconds with Google says that RS 432 is " Standard for Magnetic Tape Records: Endless-Loop Cartridges for Eight-Track Stereophonic Records at 3.75 in/s (9.53 cm/s)" which pretty obviously answers the second part of that question. Or do you mean RS-423 ? --Wtshymanski (talk) 14:14, 10 September 2009 (UTC)
Copyright
Wikipedia external links ought not to point at web sites that are apparently distributing copies of standards in violation of copyright. --Wtshymanski (talk) 00:55, 1 July 2010 (UTC)
With the exception of open standards. For example, ITU-T Electron18 (talk) 08:19, 1 July 2010 (UTC)
The law limits the copyright when using them in the student's view.Electron18 (talk) 08:23, 1 July 2010 (UTC)
- Which law? Handing a few photocopied sheets to your students is one thing; posting the whole text of the copyrighted standard for the whole world to download is quite another. We'll see if TIA takes any action. --Wtshymanski (talk) 15:03, 1 July 2010 (UTC)
You are absolutely correct placement is a violation of copyright! The question of whether a violation of copyright link to this site?Electron18 (talk) 15:47, 1 July 2010 (UTC)
Weeding pin outs table
Looking at the pin outs table I feel we could do with certain amount of weeding since it has grown a few things that perhaps shouldn't be there. We have some explanatory text in the signal descriptions (OOB control signal etc) that repeats itself is each entry and that would be better moved to explanatory text rather than having such redundancy in the table. The same goes for textual indication of the direction which duplicates the originator column.
Finally, what is DEC MMJ doing in there? That is RS-423. Interoperability by playing fast and loose with the respective standards is one thing but it doesn't belong here, and certainly not without a note that it isn't RS-232. Crispmuncher (talk) 13:27, 7 July 2010 (UTC)
- Blow them all away, put them in serial port - there's only one pinout (two pinouts) given in the standard, the rest is all application depenedent. --Wtshymanski (talk) 14:09, 7 July 2010 (UTC)
- I note you removed the additional pin outs at the time. I've just finished the job I described above - I don't see the point of saying pretty much the same thing in every line, especially when we have discussed the nature of the signals previously in the article. I also snipped a line advocating linking SG and PG - mixing signal and chassis grounds is not just "not essential" but actively bad practice. Crispmuncher (talk) 09:07, 2 May 2011 (UTC)
16550A maximum speed 1.5 Mbps
- bit rates for transmission, although the standard says it is intended for bit rates lower than 20,000 bits per second. Many modern devices support speeds of 115,200 bit/s and above. Devices such as the 16550A UART permit maximum speeds of up to 1.5 Mbps.[1]-96.237.78.13 (talk) 11:22, 20 August 2010 (UTC)
- Possibly true, but such high rates aren't used with RS 232 compatible voltage swings. --Wtshymanski (talk) 13:39, 20 August 2010 (UTC)
Need pinout for common PC connector
Someone should add the pinout for the common 9-pin PC connector. Sbmeirow (talk) 02:04, 9 October 2010 (UTC)
- That's not in the RS 232 standard. Take a look at the tables under Serial port. --Wtshymanski (talk) 02:15, 9 October 2010 (UTC)
- I don't understand why there are 2 articles, RS-232 and Serial Port. Too much of the information is identical. It would be far more intelligent to merge these articles and create a section within RS-232 just for PC and recent implementations. Sbmeirow (talk) 03:30, 9 October 2010 (UTC)
- That's why this article should talk about the standard and leave the specifics of baud rates and alternative pinouts to other articles. Think of this article as supplemental to a more general discussion of serial communication. After all, there's RS 485 and current loop and RS 423 and a whole bunch of other standards, too. --Wtshymanski (talk) 04:07, 9 October 2010 (UTC)
- ok, now I understand :-) Sbmeirow (talk) 04:36, 9 October 2010 (UTC)
- That's why this article should talk about the standard and leave the specifics of baud rates and alternative pinouts to other articles. Think of this article as supplemental to a more general discussion of serial communication. After all, there's RS 485 and current loop and RS 423 and a whole bunch of other standards, too. --Wtshymanski (talk) 04:07, 9 October 2010 (UTC)
- I don't understand why there are 2 articles, RS-232 and Serial Port. Too much of the information is identical. It would be far more intelligent to merge these articles and create a section within RS-232 just for PC and recent implementations. Sbmeirow (talk) 03:30, 9 October 2010 (UTC)
Info box/Disinfobox
I added an info box my rev: http://en.wikipedia.org/w/index.php?title=RS-232&oldid=415181558
Wtshymanski removed it citing Wikipedia:DISINFOBOX. However The disinfobox page is just an essay not a policy and interestingly all the examples of 'bad' infoboxes it gives are of articles that currently contain infoboxes (albeit in improved form) so can hardly be said to represent a general consensus for the removal of infoboxes.
I don't think the box I added wonderfully transforms the article but I certainly don't think it detracts. What I like most about infoboxes is the structure they can add to wiki. The information in them can be more easily electronically processed for example. Also they do give some significant info at a glance, never going to deal with the subtleties that the body of the article can but truth is an infinitely divisible thing and sometimes you just want a broad outline.
Anyway I'm for putting it back in or finding a better infobox or improving the existing one. Any views?
IanOfNorwich (talk) 11:22, 22 February 2011 (UTC)
- For one thing, RS-232 isn't a "computer bus." It's a point-to-point connection, and as such can even be used without computers at either end. Another issue is that the "ad hoc" nature of RS-232 interfaces is not well described by an infobox. (A variety of different connectors are used; sometimes only two pins are used while other times more than nine are used; there are a variety of different flow control methods including "none at all;" etc.) Jeh (talk) 12:27, 22 February 2011 (UTC)
- The box was also wrong - revision C came out in 1969, obviously RS 232 was "invented" earlier than that. Real encyclopedias don't have "infoboxes" and I don't know why we have to reduce a whole article to an illiterate sound-bite. The box adds no useful "structure" and insofar as it contains wrong statements, actually detracts from the article. --Wtshymanski (talk) 14:11, 22 February 2011 (UTC)
- I don't agree that infoboxes are always bad. As you've pointed out in a fair number of places, Wtshymanski, this isn't created anything like the way a "real encyclopedia" is, and so will never be regarded as one (which doesn't mean it isn't worth contributing to...). In some cases - say, articles on the US states, or on countries - an infobox can be a quick way to see a standard set of stats like date of founding, population, area, a cute little picture of the flag, etc., without digging through the article (and due to the "independent" way WP articles are written, this stuff will be in a different place in the text of every article). And though I haven't looked at one in a long time I wouldn't be surprised if "less formal" encyclopedias like the World Book had something very similar on such articles. But the standard date for RS-232C isn't a particularly useful piece of info and there's almost nothing else you can say about RS-232 for which you can't find common exceptions, except that it's point-to-point and it's serial. Your point about "adds no useful 'structure'" is exactly right. This is a great example of how an infobox can be a disinfo box. Jeh (talk) 14:58, 22 February 2011 (UTC)
- Well I'm willing to concede that the info box isn't going back on the basis that it's a poor fit. It was a spur of the moment edit on my part in any case. However, the rational that it contains incorrect information is not a good one for removal but for correction. The idea that 'real' encyclopedias don't contain infoboxes is either wrong or means either that wiki isn't a real encyclopedia (as it does contain infoboxes), so presumably different rules apply!IanOfNorwich (talk) 13:45, 24 February 2011 (UTC)
- I don't agree that infoboxes are always bad. As you've pointed out in a fair number of places, Wtshymanski, this isn't created anything like the way a "real encyclopedia" is, and so will never be regarded as one (which doesn't mean it isn't worth contributing to...). In some cases - say, articles on the US states, or on countries - an infobox can be a quick way to see a standard set of stats like date of founding, population, area, a cute little picture of the flag, etc., without digging through the article (and due to the "independent" way WP articles are written, this stuff will be in a different place in the text of every article). And though I haven't looked at one in a long time I wouldn't be surprised if "less formal" encyclopedias like the World Book had something very similar on such articles. But the standard date for RS-232C isn't a particularly useful piece of info and there's almost nothing else you can say about RS-232 for which you can't find common exceptions, except that it's point-to-point and it's serial. Your point about "adds no useful 'structure'" is exactly right. This is a great example of how an infobox can be a disinfo box. Jeh (talk) 14:58, 22 February 2011 (UTC)
- The box was also wrong - revision C came out in 1969, obviously RS 232 was "invented" earlier than that. Real encyclopedias don't have "infoboxes" and I don't know why we have to reduce a whole article to an illiterate sound-bite. The box adds no useful "structure" and insofar as it contains wrong statements, actually detracts from the article. --Wtshymanski (talk) 14:11, 22 February 2011 (UTC)
Unless a bunch of article are all presenting the same information (arguably, then should be a table in one article), such as chemical elements, there's no point in an infobox. Infoboxes that are "one shot" items used only on one or two articles are useless. Very often boxes are more concerned with pretty boxes than with getting the facts right. It's a rare infobox that is properly designed to contain relevant fields. Lots of other objections to infoboxes but I'd just be parroting the essay - I agree with it strongly. Infoboxes should be used sparingly and only where there size and conspicuity is warranted by the value and accuracy of the information contained. --Wtshymanski (talk) 14:11, 24 February 2011 (UTC)
Merge from Ring Indicator
Suggest merge from Ring Indicator which is not particularly notable by itself and which would be placed in proper context as part of this article. --Wtshymanski (talk) 04:42, 27 November 2011 (UTC)
- Concur. Ring Indicator can be left as a redirect for the few who might happen to look for that name. Jeh (talk) 08:04, 27 November 2011 (UTC)
- This article contains text from the former Ring Indicator article. --Wtshymanski (talk) 00:30, 31 December 2011 (UTC)
- Since there has been objection to loss of material on such merges... what I would suggest regarding these "an article for every signal" things would be to merge substantially their entire content, particularly the stuff on alternate uses, into the Serial port article. Info on all the nonstandard uses of RI and CD or how RI causes the AA LED to blink and so on really doesn't belong here, but reasonably complete information on what the standard says about them certainly does. Jeh (talk) 06:25, 31 December 2011 (UTC)
- That makes sense. Use this article to talk about the standard and what the standard says the pins do, and use "serial port" to collect all the innovative perversions. --Wtshymanski (talk) 06:37, 31 December 2011 (UTC)
- Since there has been objection to loss of material on such merges... what I would suggest regarding these "an article for every signal" things would be to merge substantially their entire content, particularly the stuff on alternate uses, into the Serial port article. Info on all the nonstandard uses of RI and CD or how RI causes the AA LED to blink and so on really doesn't belong here, but reasonably complete information on what the standard says about them certainly does. Jeh (talk) 06:25, 31 December 2011 (UTC)
What the heck does this mean?
"Valid signals are plus or minus 3 to 15 volts; the ±3 V range near zero volts is not a valid RS-232 level" ~~ the least of my issues here is that "±3 V" should probably be "±3 volts" to be consistent, but the first sentence is completely contradictory with the second. No idea what to make of this. BrainSlugs83 (talk) 06:12, 16 September 2012 (UTC)
- Seems clear enough to me, although if it confuses anyone, it needs looking at.
- There are two valid voltage ranges, + & -, representing 1 & 0. One range is from +3V to +15V, the other is from -3V down to -15V. The gap in the middle (like all voltage-based binary systems) is excluded as indeterminable. Voltages between +3V and -3V should not be generated for transmission and cannot be recognised on reception. Andy Dingley (talk) 09:29, 16 September 2012 (UTC)
- I did some copyediting, "A signal between +3 and +15 volts is interpreted as a logical zero. A signal between -3 and -15 volts is interpreted as a logical one. Between -3 and +3 volts is not a valid RS-232 level." -—Kvng 23:06, 19 September 2012 (UTC)
V.24 Circuits
Here on it.wiki I put, I hope, a complete list of V.24 circuits and some RS-232 interfaces. --Abisys (msg) 17:37, 12 gen 2009 (CET) 15:37, 12 January 2009 (UTC) — Preceding unsigned comment added by 69.112.242.91 (talk)
Spitting Image Sketch
While I accept this page is about the technical aspects of the interface I think it might be worth mentioning the effect it had on popular culture in the 80s or 90s. Spitting Image did a sketch about RS232 Interface Leads which can be viewed at http://www.youtube.com/watch?v=CDlj0jBtYmQ. I won't add it myself but think it should be considered.
Jumbles1971 (talk) 01:21, 14 August 2012 (UTC)
- Oppose. Pop culture trivia isn't appropriate for most articles (aside from Pop culture trivia). --Wtshymanski (talk) 13:50, 14 August 2012 (UTC)
- The standard criteria apply: was there a notable real-world impact? Did the subject take notice of the reference? Did secondary sources pick up on the reference? In its present form, none of those criteria were met by this section, and it shouldn't go back in until such point as that's been remedied. Chris Cunningham (user:thumperward) (talk) 10:40, 4 January 2013 (UTC)
RS232 level question
I don't understand why the diagram showing the data bits of 1101 0010 is described as 4b (hex). Looking at the wiki article on hexidecimal, there is a table there showing that 1101 0010 is d2 (hex) which is not an ASCII character (wiki ASCII page). The only way I can get it to make sense is if the data is reversed to 0100 1011 (4B hex). However, the diagram (rs232 page) clearly shows a start and stop bit which surely indicates the start and end of transmission, and therefore the order of the bits? Any clarification would help my understanding. Many thanks, Dave. Trellictower55 (talk) 12:22, 5 March 2012 (UTC)
- Possibly someone has been fiddling with the definitions of 1 and 0 again - there should be an explanation about mark, space, idle, 1 and 0 that explains this. It's not the order of bits, it's the definition of 1s and 0s that is the problem - its unfortunate that the character chosen was so symmetrical. --Wtshymanski (talk) 14:20, 5 March 2012 (UTC)
- I did a sanity check, but a couple of plausible refrences say UARTS send the least-significant bit first, and the levels and order of bits look OK to me. --Wtshymanski (talk) 16:20, 5 March 2012 (UTC)
- RS-232 is most definitely Least Significant Bit first. Hpa (talk) 23:02, 6 March 2012 (UTC)
- Well, most serial ports using RS 232 are LSB first, but the standard itself doesn't dictate which way about the bits are, or even how many bits in a character. --Wtshymanski (talk) 23:25, 6 March 2012 (UTC)
- Or the use of start and stop bits, for that matter. Synchronous signaling is possible and used to be fairly common (e.g. Bell 202 modems). ASCII over RS232 async does always send the LSB first, but that certainly is not only thing ever sent over RS232. I know I've dealt with at least one use of RS232 (probably synchronous rather than async) that went MSB first, but I can't recall which one. Re this issue, the diagram could be annotated more thoroughly. Jeh (talk) 00:15, 7 March 2012 (UTC)
- Per the IBM 2741 article, those machines sent the MSB first. Jeh (talk) 21:00, 4 January 2013 (UTC)
- What does the RS-232 standard say? Whilst not disputing that the IBM 2741 was big-endian, that's just as likely to be because IBM were non-conformant to a standard as it is that the standard was ambidextrous. Andy Dingley (talk) 00:09, 5 January 2013 (UTC)
- As Wtshymanski said above, it doesn't. Nor does it even mention start and stop bits, bits per character, parity bits, recommended specific bit rates, etc. Jeh (talk) 03:48, 5 January 2013 (UTC)
- Not what Wtshymanski said, what does the standard say? Also, AFAIR, it does mention stop bits et al. It doesn't specify how many are used in each instance, but it does define what they are, where they're placed, and the possible number of them that can be used. How does it treat endianness? Is this defined and stated to be the upper protocol layer's choice, or is it also mandated as a single possibility? It could easily have been written either way, so this must be sourced, not merely relying on 25 year old memories. Andy Dingley (talk) 11:38, 5 January 2013 (UTC)
- I guess I was not clear. I meant that the standard says, or rather does not say, what Wtshymanski said it does not say: The published standard doesn't mention bit ordering.
- Re. framing and so on, the only mention of start and stop bits I can find is in the title of another, referred-to standard: EIA/TIA-404-A, Standard for Start-Stop Signal Quality for Non-Synchronous Data Communication Equipment, this being mentioned in the context of distortion tolerances. I can't find any other occurrences of "start" or "stop", let alone "start bit" or "stop bit". In fact it barely even mentions the word "bit", except in the maximum bit/second spec (20,000), and a mention of a "Unit Interval (bit time)" in the context of transition rate. There is quite detailed explanation of the meaning of the control signals, but for the data signals there is (in rather stark contrast) practically nothing, particularly for "non-synchronous" systems. (Nor, of course, is there any mention of how special bit sequences and bit-stuffing are used to provide framing in synchronous systems; that's up to bisync, SDLC, etc.) Nor does it say anything like "this standard does not address details of e.g. character framing for non-synchronous systems".
- If this seems rather cavalier, it might help to recall that this standard originated when an async DCE, like a Bell 103 modem, did not have to know anything about "bytes" at all. It just relayed "marking condition" and "spacing condition" to the modem at the other end and was not even concerned with bit boundaries, let alone bytes. (Which explains how IBM could get away with using 1.5 stop bits on the 2741.) So getting "marking condition" and "spacing condition" between the DTE and the DCE was all this spec had to be concerned with.
- What has likely happened here is that various articles and user manuals and so on have mentioned these details while talking about "RS-232 ports", and it is a reasonable conclusion that the same standard that tells you what voltages to emit on TD (old nomenclature, but YKWIM) on your DTE will result in a mark vs. a space being received by the DCE, should also spec how to manipulate marks vs. spaces to send bytes over an asynchronous link. Reasonable, but mistaken. Jeh (talk) 15:49, 5 January 2013 (UTC)
- Not what Wtshymanski said, what does the standard say? Also, AFAIR, it does mention stop bits et al. It doesn't specify how many are used in each instance, but it does define what they are, where they're placed, and the possible number of them that can be used. How does it treat endianness? Is this defined and stated to be the upper protocol layer's choice, or is it also mandated as a single possibility? It could easily have been written either way, so this must be sourced, not merely relying on 25 year old memories. Andy Dingley (talk) 11:38, 5 January 2013 (UTC)
- As Wtshymanski said above, it doesn't. Nor does it even mention start and stop bits, bits per character, parity bits, recommended specific bit rates, etc. Jeh (talk) 03:48, 5 January 2013 (UTC)
- What does the RS-232 standard say? Whilst not disputing that the IBM 2741 was big-endian, that's just as likely to be because IBM were non-conformant to a standard as it is that the standard was ambidextrous. Andy Dingley (talk) 00:09, 5 January 2013 (UTC)
- Per the IBM 2741 article, those machines sent the MSB first. Jeh (talk) 21:00, 4 January 2013 (UTC)
- Or the use of start and stop bits, for that matter. Synchronous signaling is possible and used to be fairly common (e.g. Bell 202 modems). ASCII over RS232 async does always send the LSB first, but that certainly is not only thing ever sent over RS232. I know I've dealt with at least one use of RS232 (probably synchronous rather than async) that went MSB first, but I can't recall which one. Re this issue, the diagram could be annotated more thoroughly. Jeh (talk) 00:15, 7 March 2012 (UTC)
- Well, most serial ports using RS 232 are LSB first, but the standard itself doesn't dictate which way about the bits are, or even how many bits in a character. --Wtshymanski (talk) 23:25, 6 March 2012 (UTC)
- RS-232 is most definitely Least Significant Bit first. Hpa (talk) 23:02, 6 March 2012 (UTC)
- I did a sanity check, but a couple of plausible refrences say UARTS send the least-significant bit first, and the levels and order of bits look OK to me. --Wtshymanski (talk) 16:20, 5 March 2012 (UTC)
Who most likely navigates here in 2013 and seeking what information? I.e. lab devices such as scientific balances?
IMHO this is the foremost question all of us contributors should keep in mind when working on any Wiki article. I'm clueless about computer terminals, just a lay visitor here. I'm shopping for a balance (a weight scale) for use in a perfumer's lab. I'm considering a model manufactured by Ohaus, a well-known brand. Their literature states that this model can "accommodate both RS232 and USB connectivity." (Btw, that's reason enough to keep, or at least redirect to, this nomenclature). My goal here is to learn whether I need RE232 connectivity; I intend to connect the device to an iMac, hence Apple OS X operating system. The writer stated that this type of connection protocol(?) is "now" mainly used only in scientific equipment because often this equipment is older and expensive to replace. That's nice to know. But a couple more sentences there might have given me a clue as to whether an RS232 connection might be easier for a Mac user to deal with. One might argue that this is stretching the topic a bit far, but if what the writer here says is true, more specifically that RE232 is "now" mainly used for scientific equipment, a common type of visitor here comes in search of further practical information regarding its use in scientific equipment. To my mind at least, this is an example that demonstrates the need for contributors to write not only historical and encyclopedic information but also information which is useful and practical for those individuals most likely to drop by.
Lasty, I can't resist adding:
I guess it's self-evident that I completely disagree with a contributor's statement made above under this talk page section "Name change from RS-232 to EIA-232" (see: User:Wtshymanski 03:26, 17 June 2007 (UTC)): "An encyclopedia should not be so quick to bow to popular usage." What? Says who? I guess an encyclopedia comes across as more "serious" and "official" if it is lofty, pedantic but lacking approachability and practicality? Mykstor (talk) 14:04, 29 December 2012 (UTC)
- This "encyclopedia" is about as serious and lofty as a debate between occupants of adjacent barstools. It's too bad we don't have an article titled "How to interface the Binford model 6100 scale with J. Random Computer", but just wait a while, it will be here eventually. We used to have a guideline that Wikipedia is not a how-to tutorial guide but why should that mean anything? After all, doesn't the Encyclopedia Britannica teach you everything you need to know about cleaning and cooking your Christmas turkey?
- What, exactly, would the article have contained to tell *you* if *you* need RS 232 connectivity? Do you have an RS 232 port on your computer? Is it supported by the computer, software, and the device? Do you care about the speed and power supply available through a USB port?
- You typed "RS 232" after the description in the ad you were reading, and came to this article. That's a really solid argument to rename it "TIA 232F". --Wtshymanski (talk) 19:39, 29 December 2012 (UTC)
- "is to learn whether I need RE232 connectivity", well this article tells you what RS232 is ie an electrical interface definition. RS232 doesn't even define a protocol.. Perhaps the ingress should be more clear on this distinction as a lot of confusion comes from the misuse of "rs232" to mean the electrical definition and the asynchronous transmission (8N1). Perhaps a new section "Usage" besides "Role in modern personal computers" should be written to clarify when this interface makes sense and when it doesn't. Electron9 (talk) 23:33, 31 January 2013 (UTC)
- Don't we already have a Wikibook which is the place for a tutorial? Encyclopedias aren't supposed to give you how-to information. --Wtshymanski (talk) 14:47, 1 February 2013 (UTC)
- This was about what RS-232 is good for, WHY it's of any use at all. In particular in the rs232+async combo. Electron9 (talk) 22:09, 1 February 2013 (UTC)
- As WP editors we have to be careful about making such judgments, unless we can find RS's. Jeh (talk) 01:06, 2 February 2013 (UTC)
- This was about what RS-232 is good for, WHY it's of any use at all. In particular in the rs232+async combo. Electron9 (talk) 22:09, 1 February 2013 (UTC)
- Don't we already have a Wikibook which is the place for a tutorial? Encyclopedias aren't supposed to give you how-to information. --Wtshymanski (talk) 14:47, 1 February 2013 (UTC)
- "is to learn whether I need RE232 connectivity", well this article tells you what RS232 is ie an electrical interface definition. RS232 doesn't even define a protocol.. Perhaps the ingress should be more clear on this distinction as a lot of confusion comes from the misuse of "rs232" to mean the electrical definition and the asynchronous transmission (8N1). Perhaps a new section "Usage" besides "Role in modern personal computers" should be written to clarify when this interface makes sense and when it doesn't. Electron9 (talk) 23:33, 31 January 2013 (UTC)
Categorization to Category:Serial_buses ?
In or out? As the category page has it, it's a category for point to point links (which RS-232 is), as well as multi-drop buses.
There's also an argument to be made that RS-232, as the progenitor of much work on serial comms in general, has adequate relevance to the topic of serial bus comms, just on that basis. Andy Dingley (talk) 10:59, 6 February 2013 (UTC)
- A bus permits connection to multiple devices along its length and has a way to address one of those devices at a time. We can allow in things like PCI-E which have more of a "star" topology, but still, there is addressing - devices connected to a PCI-E hub can be individually addressed. RS232 is a point-to-point link with no possible means of addressing and no other way to select among multiple devices. As for multidrop, multidrop isn't part of TIA-232-F. There was multidrop on links with "RS-232-like signaling" but multidrop is covered in another standard. Jeh (talk) 14:47, 6 February 2013 (UTC)
- Perhaps if the category page defined itself as "This category comprises serial buses, serial communication point-to-point interconnections, and chocolate cake recipes." the problem with the category page would be a little more apparent. A dog still has only four legs, etc. Since great amounts of doomed ingenuity have gone into trying to make RS 232 into a "serial bus", perhaps it (and any other strictly point to point link) should not be in the category. --Wtshymanski (talk) 14:56, 6 February 2013 (UTC)
- I don't see any chocolate cake, but the category does already define itself as "This category comprises serial buses and serial communication point-to-point interconnections." By which definition, RS-232 is included. As other standards like MIDI (which I believe to be daisy-chained point-to-point links, with addressing) are in there as well, it's not such an esoteric question.
- Really the useful question is not "what does the category say?" but "What is a useful scope for making an encyclopedia work?" Then we change use and definition, as needed.
- BTW - there's plenty of prior art for addressing via RS-232. Where there was some part of a protocol permitting fan-out (RTTY being an obvious one), then addressing by in-band broadcast of an address packet was long established as a low-speed addressing mechanism for packet radio etc., and it ran fine over the numerous RS-232 links involved. Mind you, I know nothing of railway Morse. In the UK it wasn't really used – we went for application-specific bell codes instead (and yes, some of those were multi-drop). Andy Dingley (talk) 15:03, 6 February 2013 (UTC)
- Er, but none of that involves RS-232 per se. Sure, an RS-232 link can carry address info that is interpreted by something at the other end, beyond the RS-232 receiver. But to RS-232 itself it's just data. What you are saying here is akin to saying that Ethernet is a file transfer protocol. It isn't. Ethernet has no concept of anything remotely like a file. It can be used as a transport layer by file transfer protocols, of course.
- Regarding "what is a useful scope for an encyclopedia," I think it is useful to not dilute the term "bus" to the point where it has no meaning distinct from any old point-to-point connection. Jeh (talk) 19:47, 6 February 2013 (UTC)
- Perhaps if the category page defined itself as "This category comprises serial buses, serial communication point-to-point interconnections, and chocolate cake recipes." the problem with the category page would be a little more apparent. A dog still has only four legs, etc. Since great amounts of doomed ingenuity have gone into trying to make RS 232 into a "serial bus", perhaps it (and any other strictly point to point link) should not be in the category. --Wtshymanski (talk) 14:56, 6 February 2013 (UTC)
- ... however, it appears that the category is already a disaster in that respect.
- My suggestion: Rename the category to reflect its content. Create two new subcategories within, "serial buses" and "serial point-to-point interconnects". RS-232 can then go in the latter. Jeh (talk) 20:01, 6 February 2013 (UTC)
- This makes more sense than expanding the category to include cake recipes. I'm not sure what the object was in lumping in point-to-point connections with real multidrop serial buses. --Wtshymanski (talk) 22:00, 6 February 2013 (UTC)
Inclusion of DE-9 connector in article
An IP user, 216.240.30.23 made a good faith edit and and added in references to the DE-9 connector along with its pin assignments. Our resident edit warrior Wtshymanski reverted the edit with a pointless comment that it is not in the standard. Another IP editor 86.169.32.152 restored it with the observation that the connector, "... is more than sufficiently widespread to warrant inclusion" (I cannot argue with that). Wtshymanski nevertheless reverted it again with a repetition that it is not in the standard. It can only be assumed that Wtshymanski is attempting to get yet another article edit blocked through his continued edit warring.
Granted, the 9 pin DE-9 connector is not covered by the RS-232 standard and so technically at least, it would seem to not belong here. But this is supposed to be an encyclopedia where users come to find information. One such item of information that such a user might seek is the pinout of this connector. This is particularly likely because in the personal computer world, the connector exists almost to the exclusion of the official DB-25 connector. It is therefore right and proper that the pinout details are included particularly as the likely search key would be "RS-232" which is what the port is officially known as in this conext. The article can, and of course should, note that the connector is not covered by the official standard, and the edit made above correctly did precisely that. The article might also observe that the connector is often (incorrectly) designated as 'DB-9'. I B Wright (talk) 14:55, 5 April 2013 (UTC)
- I agree with keeping the DE-9 info here. Where would the reader find this info about this connector using RS-232 signals and so common in the real world? In my lifetime I must have looked for this information up several dozen times. Too bad WP pops up, now, after spending hundreds of hours to find this info. Technical expertise is belittled by use of ad hominem comments. 174.118.142.187 (talk) 17:27, 5 April 2013 (UTC)
- No, we've got an article for Serial port. This article talks about the standard - the 9 pin connector is not in the standard, and this article points you to a place where you can learn about some of the myriad other ways people have wired serial ports. Not that long ago this article was full of helpful tips on configuring baud rates, port addresses, and IRQs for the IBM PC - *those* aren't in the standard, either. --Wtshymanski (talk)
- Yes I see that now. I didn't read this article that closely and perhaps a stronger ref to it would be in order. oops... we have one right at the top! I noticed it also has a good connection chart. We really don't need multiple ones to maintain, just one good one. Wasn't the DB-9 introduced as a standard EIE-232 connection along with the 8-wire telephone style clear flat connector RJ-45?. Much rack mount equipment comm equipment used those when massive amounts space would be taken like in the front end of SCADA masters. In view of these details I have to change sides with Wtshymanski on this one. It does make sense to keep it out. 174.118.142.187 (talk) 19:27, 5 April 2013 (UTC)
- It's the DE-9, and no, it isn't in the standard (TIA-232-F) that this article addresses. Its pinout is a widely accepted industry convention, at the very least, but it is not in the standard that is the subject of this article. Part of being an "encyclopedia," instead of RS-232 for Boneheads, is being rigorously correct about such matters, rather than adding to the confusion that comes from too many inferences from popular usage. (For the same reason I don't think the article should be called "RS-232" either, but rather "TIA-232-F" with redirects from the other names... but that's another argument that we've already had; WP:COMMONNAME prevailed.) Jeh (talk) 19:45, 5 April 2013 (UTC)
- To IP 174.118.142.187: Glad you came around. For future reference: Sorry but "I found it useful" is not sufficient reason for inclusion. See WP:ILIKEIT. Jeh (talk) 19:47, 5 April 2013 (UTC)
- More to the point, if enough people believe that it is useful to include it in the article then it should be in the article. As already noted, the article does document that it is not officially covered by the standard. No one person has the right of arbitration as to what may or may not be included in any article. Wtshymanski, could you please advise us when you believe this right was confered upon yourself as you seem to believe that you have it as betrayed by your constant edit warring? I B Wright (talk) 11:34, 6 April 2013 (UTC)
- What "enough people believe" is actually completely beside the point. WP does not run by majority rule, but rather by compliance to policies and by arguments that are based on those policies. And per those policies, the MOS, etc., articles should be about the subject as described by the title and the lede. So it doesn't matter how many people "believe" that something not in the standard should be covered here in detail equivalent to that used for material that is in the standard. This article is about RS-232 and its later revisions. TIA-232-F does not even mention a nine-pin connector, let alone describe a pinout therefor, so why should the DE-9 be detailed here? Should we also cover details of asynchronous framing, number of data bits per character, parity bits (or not), typical speeds, etc.? Of course not; those aren't in RS-232 either. We have the Serial port article for all that other information. Jeh (talk) 14:19, 6 April 2013 (UTC)
- Agreeing with Jeh's last statement; if we didn't have another place to cover this I would agree to include it here but, we have a good home for readers to find this material and lots of references to it in this article. Stick to the specific and specialized topic at hand on this one. 174.118.142.187 (talk) 23:56, 6 April 2013 (UTC)
- What "enough people believe" is actually completely beside the point. WP does not run by majority rule, but rather by compliance to policies and by arguments that are based on those policies. And per those policies, the MOS, etc., articles should be about the subject as described by the title and the lede. So it doesn't matter how many people "believe" that something not in the standard should be covered here in detail equivalent to that used for material that is in the standard. This article is about RS-232 and its later revisions. TIA-232-F does not even mention a nine-pin connector, let alone describe a pinout therefor, so why should the DE-9 be detailed here? Should we also cover details of asynchronous framing, number of data bits per character, parity bits (or not), typical speeds, etc.? Of course not; those aren't in RS-232 either. We have the Serial port article for all that other information. Jeh (talk) 14:19, 6 April 2013 (UTC)
- More to the point, if enough people believe that it is useful to include it in the article then it should be in the article. As already noted, the article does document that it is not officially covered by the standard. No one person has the right of arbitration as to what may or may not be included in any article. Wtshymanski, could you please advise us when you believe this right was confered upon yourself as you seem to believe that you have it as betrayed by your constant edit warring? I B Wright (talk) 11:34, 6 April 2013 (UTC)
- To IP 174.118.142.187: Glad you came around. For future reference: Sorry but "I found it useful" is not sufficient reason for inclusion. See WP:ILIKEIT. Jeh (talk) 19:47, 5 April 2013 (UTC)
- It's the DE-9, and no, it isn't in the standard (TIA-232-F) that this article addresses. Its pinout is a widely accepted industry convention, at the very least, but it is not in the standard that is the subject of this article. Part of being an "encyclopedia," instead of RS-232 for Boneheads, is being rigorously correct about such matters, rather than adding to the confusion that comes from too many inferences from popular usage. (For the same reason I don't think the article should be called "RS-232" either, but rather "TIA-232-F" with redirects from the other names... but that's another argument that we've already had; WP:COMMONNAME prevailed.) Jeh (talk) 19:45, 5 April 2013 (UTC)
- Yes I see that now. I didn't read this article that closely and perhaps a stronger ref to it would be in order. oops... we have one right at the top! I noticed it also has a good connection chart. We really don't need multiple ones to maintain, just one good one. Wasn't the DB-9 introduced as a standard EIE-232 connection along with the 8-wire telephone style clear flat connector RJ-45?. Much rack mount equipment comm equipment used those when massive amounts space would be taken like in the front end of SCADA masters. In view of these details I have to change sides with Wtshymanski on this one. It does make sense to keep it out. 174.118.142.187 (talk) 19:27, 5 April 2013 (UTC)
- No, we've got an article for Serial port. This article talks about the standard - the 9 pin connector is not in the standard, and this article points you to a place where you can learn about some of the myriad other ways people have wired serial ports. Not that long ago this article was full of helpful tips on configuring baud rates, port addresses, and IRQs for the IBM PC - *those* aren't in the standard, either. --Wtshymanski (talk)