Talk:PS/2 port

Computing Low‑importance

	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.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing articles
Low	This article has been rated as Low-importance on the project's importance scale.
	This article is supported by Computer hardware task force (assessed as Mid-importance).

Damage

Yes, it really is true -- at least on some older motherboards, unplugging a PS/2 device while powered will occasionally at least semi-permanently break the motherboard, such as blowing a special fuse. -69.87.204.161 01:03, 19 March 2007 (UTC)[reply]

Really? I've done it dozens of times and never ever had a single computer fail because of that. I have also plugged mice in the keyboard plug, and the other way around, and I'd swear they worked, but I'm not sure on that one. — isilanes (talk|contribs) 09:45, 12 March 2008 (UTC)[reply]

As stated, it was primarily on older systems. I had a friend who co-op’d at the school board ~1996 where he did it with an IBM system (which was probably a few years older than that) and burned out the motherboard. On the other hand, the PS/2 connector on my EPoX EP-4B2A2 from 2002 eventually developed a bad connection after a few years and the keyboard would cut in and out, so I ended up having to disconnect and reconnect the keyboard countless times while the system was on, without hurting it. If I was in Windows, what would usually happen is that the typematic rates would get reset and I would have to change them back, otherwise nothing bad happened. Synetech (talk) 19:03, 20 June 2011 (UTC)[reply]

I used to work on PS/2 keyboard and mouse firmware in the 1990s and never saw a damaged motherboard due to hot plugging. All motherboards and interfaces, going right back to the original IBM systems, included protection devices for static discharge into the pins, overcurrent limiting and wrong polarity. Obviously the protection devices could be damaged with purposeful abuse, e.g. a high negative voltage for a long time into the data lines could overload the protection diodes, but normal use would rarely damage the ports electrically. I did see one motherboard with a fusible protection device on the power line which would require replacement on significant over current. None of these issues would be likely to occur through normal hot plugin. However, some BIOS/hardware could lock out the PS/2 interface if bad signal sequences were received (I saw many times while debugging kbd/mouse firmware). This would have either been a hardware shift register wrong state issue or bad BIOS code or a BIOS protection mechanism, ignoring apparently dodgy signals to prevent rubbish input for the user. Maybe this is what is being referred to? However this was never a hardware issue and caused no damage, only requiring reboot of the host system. Lkingscott (talk) 06:59, 29 September 2022 (UTC)[reply]

Spelling

I corrected some spelling errors. jpoke89 (talk) 11:49, 22 November 2007 (UTC)[reply]

Mouse in keyboard port and vice versa

Care should be taken to avoid connecting a keyboard to a mouse port or vice versa.[citation needed]

That's BS. Electrically they're the same. They signal the same. On all modern motherboards both ports even go to the same chip. The data format is different, but you won't harm or break anything by plugging the wrong thing into the wrong port. —Preceding unsigned comment added by 72.29.167.236 (talk) 19:22, 20 September 2008 (UTC)[reply]

Interestingly the 1st reference on this subject appears to be a comment, rather than a reference... PrimalBlueWolf (talk) 10:59, 25 March 2020 (UTC)[reply]

Pre PC/AT Connector?

What was the connector used in the original IBM PC? Family Guy Guy (talk) 04:34, 19 November 2008 (UTC)[reply]

The IBM XT used a DIN connector - that used a different set of Scan codes. See this if you are interested in some technical details. Megapixie (talk) 04:43, 19 November 2008 (UTC)[reply]

PC keyboards use the same 5-pin DIN connector as AT keyboards and the pinout is the same as well (although not all ATs supported the RESET signal any more). However, the data going over the interface is different, PC keyboards use 9 bit frames, whereas AT keyboards use 11 bits. Also, PC keyboards send scan codes in what later became known as scan code set 1 (the only set supported by these keyboards), whereas AT keyboards use scan code set 2 by default (and MF keyboard introduced yet another set 3). For compatibility purposes, the keyboard controller on the mainboard translates these interface scancodes into codes according to set 1 for IRQ 1 (INT 9). There are very few exceptions (including some "obscure" IBM machines) where the scancodes are not translated by the keyboard controller ("pass-through mode"). In these cases, AT keyboards will have to be forced into scan code set 1 or the INT 9 interrupt handler of the keyboard driver will have to translate the codes itself. Scan code set 3 was meant to be used by PS/2 machines, but it is rarely used. While all MF keyboards supported all three scan code sets in the early 1990s, later low-quality keyboards were "optimized" to only support scan code set 2.--Matthiaspaul (talk) 01:26, 23 January 2012 (UTC)[reply]

PS/2-mice and keyboards plugged into USB ports (and vice versa)

The article should contain a section in which the compatibility of PS/2-to-USB and vice versa is discuessed. Simple adapters exist both for plugging PS/2-hardware into USB sockets, as well as for plugging USB-hardware into PS/2 sockets. -- Alexey Topol (talk) 16:31, 27 March 2010 (UTC)[reply]

Yes, a section about adapters is appropriate. I bought an inexpensive adapter to connect a Logitech M-S48 PS/2 mouse and a Microsoft PS/2 Intellimouse to USB sockets. Neither mouse worked with the adapter. After the Logitech mouse was connected through the adapter for about 10 minutes, the Y axis photodetector overheated enough to melt the window and melt three spokes of the encoder wheel. The MS mouse survived a shorter connection time. Only an adapter from a reputable supplier should be used. An inexpensive nameless adapter can ruin a mouse. Conceivably, it can even cause a fire. Regards, PeterEasthope (talk) 21:10, 26 April 2013 (UTC)[reply]

In case anyone is interested, this is the link to the faulty adapter. Hopefully it is removed from the market soon.

http://www.ebay.ca/itm/300492713758?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649

Regards, ... PeterEasthope (talk) 02:40, 27 April 2013 (UTC)[reply]

This sort of thing really isn't WP's job. We're not a product review or advice site. And everything written in WP must be verifiable by reference to reliable sources, regardless of editors' personal knowledge. If you can find an RS that says something like "some PS/2 to USB adapters have damaged the mice plugged into them", that could be included with the cite of the RS. If you can find a review site that approves of one adapter or disapproves of another, a link thereto might find a place in the "External links" section. But I'm afraid your own report of the faulty adapter is, by WP standards, pure original research and therefore not acceptable for inclusion. Jeh (talk) 04:49, 27 April 2013 (UTC)[reply]

"...almost all desktop computers still have PS/2 ports...."

Really? I've not bought a retail PC for years but almost no mobo's for self build had PS/2 ports when I was self building in '08. Anyone else think this is a bit out dated and that now almost all KB and mouse are usb as well as PS/2 being largely dropped from new pc's? —Preceding unsigned comment added by 77.197.60.204 (talk) 12:38, 25 August 2010 (UTC)[reply]

No, not really. I just looked through like 10-15 high-end Socket 1366 motherboards and not one was lacking a PS/2 port. It seems to be the same story with cheaper motherboards also. Retail PC's are another matter though and may leave them out for lack of space on the I/O panel. So "almost all desktop computers" is definitely an exaggeration but PS/2 (at least for keyboards) is alive and kicking and as it seems is now partly geared toward enthusiast users who are aware of the advantages of PS/2 that exist over the USB and who also might not want to give up old PS/2 keyboards they have owned for a long time, and still kept for compatibility reasons as well. There is the problem of USB keyboards sometimes not being usable at boot which in those cases makes PS/2 absolutely necessary. 83.226.206.82 (talk) 00:16, 5 October 2010 (UTC)[reply]

That, and who wants to waste a perfectly usable USB port when there's a PS/2 port available? USB ports are in high demand on my system, and I'm sure plenty of people agree. 207.144.204.209 (talk) 22:27, 7 January 2011 (UTC)[reply]

newegg PS/2 stats
	T O T A L	# W I T H O U T	% W I T H
	T O T A L	PS/2
AMD	148	3	98	95
Intel	317	23	93	95

I went through every AMD and Intel mobo listing at newegg.com the other day before I removed the (unreferenced) blurb about PS/2 ports "frequently" not being included (results at left).
I didn't bother with the "Motherboard / CPU / VGA Combo" and "Server Motherboards" categories, but 100% of the "Top Sellers" in them have PS/2 ports.
As for retail PCs, I bought a compact tower Acer from BestBuy last year that has seven USB ports and individual PS/2 ports for keyboard and mouse (as well as many other ports). ¦ Reisio (talk) 23:29, 7 January 2011 (UTC)[reply]

I've undid the change partly because it removes a lot of other useful information, and partly because it *is* legacy: PC 2001 (and probably even PC99) says "legacy devices such as PS/2-compatible mouse devices" and "legacy devices such as PS/2-compatible keyboard devices" and refers the reader to the Legacy Plug and Play Guidelines. ⇌Elektron 00:23, 31 January 2011 (UTC)[reply]

Some more research: PC 98 and 99 "prefer" USB to PS/2; only PC 2001 and the Legacy PnP Guidelines (dated before PC98) call PS/2 "legacy". I'm not aware of major advantages of PS/2 over USB except price, and that PS/2 communication is (by default) device-initiated, so you potentially get lower power consumption and latency (but the maximum PS/2 mouse "sample rate" is apparently 200 Hz, so it only really helps button-clicks and keypresses). The USB-keyboard-in-BIOS problem is not an issue if PS/2-less motherboards implement working USB support; I've only noticed problems on boards with PS/2 ports (e.g. the DG33BU does not support keyboards attached via hubs, and many keyboards have integrated hubs; the GA-D525TUD occasionally fails to reset USB at (re)boot, but there might be a BIOS fix; some boards let you disable USB keyboard support which might happen on a CMOS reset). Of course, anecdotally, a friend bought a PC and later noticed that it didn't specify any PS/2 ports, so he bought a converter (the computer arrived with PS/2 ports). ⇌Elektron 01:06, 31 January 2011 (UTC)[reply]

1: The survey described above is a couple of years old, and furthermore is original research.

2: Motherboards sold by places like Newegg are one thing. Motherboards installed in preassembled PCs from Dell, HP, etc., are another. You can't make any judgments like "the majority of motherboards have PS/2 ports" without considerng them. A quick glance at the machines on display at Fry's last week told me that most of the mass market PCs they were offering had no PS/2 ports. The same is also true of laptops: very few laptops have PS/2 ports any more. My belief is that Newegg sells mostly to hobbyists and gamers who need (or think they need) PS/2 ports to support a beloved old Northgate or Model M keyboard (or for the latency advantage in gaming)... but that of course is WP:OR.

3: regarding the "cn" tag: WP:V says that "Even if you're sure something is true, it must be verifiable before you can add it." When someone writes "the question is how to cite it", that is not reason to remove the "cn" tag, it is reason to keep it in place.

3A: The link that was there previously, a broken link to a product description for one motherboard model, was most certainly not sufficient to back up the claim.

3B: WP:V also says "Attribute all quotations and any material challenged or likely to be challenged to a reliable, published source using an inline citation." Well, I'm challenging this material, so a citation is indeed needed.

3C: And finally, WP:V says "Any material lacking a reliable source directly supporting it may be removed." There wasn't and still isn't a WP:RS for this claim. So any editor would be within their rights to remove the claim outright. A cn tag is a completely reasonable interim step. If anyone wants this claim to stay on the page they need to step up and WP:PROVEIT. Jeh (talk) 07:52, 21 January 2013 (UTC)[reply]

You’d rather have information that’s inaccurate by omission, I suppose? Don’t care. Think that violates WP:VERIFY? The omission does also, because I can (and have) verified that the text without the information is inaccurate; and if that weren’t enough I could always show you WP:IGNORE or WP:CONSENSUS (no, you against me [and all these other editors] does not mean you have a consensus, in case you were wondering; in fact it makes a consensus against your position).
A simple matter of wording. If you want to change the wording from “motherboards” to “motherboards for individual sale” or the like, be my guest. The burden of proof is, however, on you as well as I; only my end has already been proven.
See 1.
1. I didn’t add it, and have taken no action against its removal.
2. See 1.
3. See 1.

¦ Reisio (talk) 17:45, 21 January 2013 (UTC)[reply]

1: As a matter of fact, yes, omission is preferable to an unreferenced, challenged claim. That is WP policy. "The standard for inclusion is verifiability, not truth." If it can't be verified then it doesn't belong here, no matter how strongly you "know" it's true or how important you think it is to the article. And besides... you might be wrong.

1A: Your claim that an omission violates WP:V, just because of your WP:OR, is absurd. WP:V says that all material here must be verifiable. It in no way says that anything that can be verified must be included. Material verified only by WP:OR? That's the absurd part.

2: I'm not making such a change because there is no support for that assertion either. Your WP:OR, a survey of one vendor's offerings to a particular market segment, is not sufficient and is WP:OR besides.

3: Consensus on WP is about deciding between possible alternative content, not whether or not to follow WP:V. Consensus cannot override WP:V nor any other core policy. Nor can WP:IGNORE.

4: Why do you object so strenuously to a CN tag? Jeh (talk) 07:25, 23 January 2013 (UTC)[reply]

Almost, consensus is general agreement, and it can be on any matter at all. There is no trump order amongst policies that I’m aware of, though it can safely be said that WP:IGNORE can trump any other policy (unless consensus is against it, obviously, because that’s a numbers game).
Because as already explained (and clearly evident to the majority), this issue is indisputable. If the tag didn’t mean some yahoo would come by later and remove the assertion thinking it was utter nonsense, I wouldn’t care; but since that does happen, and without it the article is inaccurate, I do.

¦ Reisio (talk) 19:20, 28 January 2013 (UTC)[reply]

"Indisputable"? I'm sorry but WP does not recognize your or any editor as having the power to declare a matter under dispute as "indisputable." I'm disputing it, based on considerable experience with PCs sold as complete PCs. I'm talking in particular about the models that are installed on the high tens of millions of corporate desktops in the US alone. PS/2 ports are like hen's teeth in that environment and these machines vastly outsell motherboards sold as components or high-end PCs sold to gamers.

Re. consensus, it is most decidedly not a "numbers game." WP:CONSENSUS is not based on voting, but on logical arguments:

Consensus is determined by the quality of the arguments given on the various sides of an issue, as viewed through the lens of Wikipedia policy.

"Most" vs "Some"

Would you accept a wording change to something like "PS/2 connectors are still present on some current PCs and component motherboards"? I believe that that claim is sufficiently noncontroversial that a citation is not needed. I would also ask that you not delete my add regarding availability of new PS/2 compatible mice and keyboards. No, it is not proven, but that is what a CN tag is for. Jeh (talk) 02:21, 6 February 2013 (UTC)[reply]

I have already twice suggested you could merely change the wording rather than utterly remove the text, but you seemed to have ignored that. I would accept a wording change to something like what I’ve already mentioned, but not the use of the phrase “present on some current PCs”, because as has already been established, it is clear the majority of (at least individually sold) motherboards, and not merely “some”, have PS/2 ports. Your addition expressly contradicts that which we have already been discussing, and I’m not sure tagging something {{citation needed}} while simultaneously adding it is not disingenuous; nevermind that you would apparently promote one assertion even though you have no source (and even less of what you call original research) while opposing another that is contradictory. ¦ Reisio (talk) 03:33, 6 February 2013 (UTC)[reply]

Far from ignoring your suggested alternate wording, I have replied twice that there is no WP:RS evident for your modified claim either. So I would insist on a CN tag there as well.

"Some" still allows "most" as a possibility. It doesn't rule out "most".

Please note that I am not trying to "utterly remove the text." I just want a CN tag on it. The fact that that claim seems "clear" to you does not establish verifiability as far as WP is concerned. All you have "established" is that your own two-year old survey of one web retailer that caters mostly to the enthusiast market supports the claim. But that result is not representative of all PCs, nor even of all individually sold motherboards, and it is OR besides. It's not something that can be cited, as it is WP:OR. Therefore, a citation to a WP:RS is still needed. And that's all the CN tag means. Jeh (talk) 08:36, 6 February 2013 (UTC)[reply]

“"Some" still…”

Then it may as well be the more accurate “most”.

“Please note…”

I’ve replied to this sentiment at the DNR discussion. You may be truly content to repeat yourself in not one but two different locations, but I have better things to do with my time (believe it or not :p).

¦ Reisio (talk) 01:03, 7 February 2013 (UTC)[reply]

But there is a citable source for "some", whereas there is none for "most". Jeh (talk) 21:54, 7 February 2013 (UTC)[reply]

CN vs. "doubtful" tags

CN means "it may be that this claim is true, but no reliable, published sources were given to verify it and the information is not considered common knowledge". That is a direct quote from the template description. It goes on to say that the template should be used "when there is a general question of the verifiability of a statement, or when an editor believes that a reference verifying the statement should be provided."

This is why I insist on a CN tag for your "most." You see, the CN tag does not mean "I claim it's not true." It does not even mean "I'm doubtful about this." (There is a "doubtful" tag for that.) I'm saying it's not self-evident the way "the sky is blue" is, nor common knowledge like "the moon orbits the Earth" is, so by WP:V it needs to be cited. And the survey you did two years ago, while providing sufficient motivation for the inclusion of the claim, is not something that we can use as a citation nor sufficient to remove the CN tag.

If I thought the "most motherboards" claim wasn't true I'd have used a "doubtful" tag instead. That validity of the claim is not what this dispute is about. The fact that it lacks a citation is. Jeh (talk) 08:36, 6 February 2013 (UTC)[reply]

You seem to be so caught up in bureaucracy that you ignore the meaning of your own words. Reliable. Published. Verifiable. These words have meanings, and had them long before random internet denizens on Wikipedia wrote up articles on them. The information I have provided is reliable, it is published, and it is verifiable (although I can only assume you have not taken the time to verify it). I won’t again go into how WP:V does not trump other policies, but I will say I don’t see why you bother taking a position no dictionary editor would. It may not be self-evident the way claiming “the sky is blue” is, but it is self-evident all the same, and if you don’t see that, you have just been ignoring me this entire time. My survey might not be enough for you, but it is for me. :) As far as I can tell, it is not your clear distinction between {{citation needed}} and {{doubtful}}, but your stubborn inability to accept reality over bureaucracy that motivates you. The validity is the only thing that matters. It is valid, and I wish you’d accept that; and that despite the myriad of Wikipedia bylaws, there is not an appropriate paper to push in every single situation. ¦ Reisio (talk) 01:14, 7 February 2013 (UTC)[reply]

No, I'm not ignoring you. I'm trying to get you to see that you can't simply ignore WP policy. The "self-evidentiary" nature of your claim is not what is under discussion. What is relevant to inclusion on Wikipedia is whether it is verifiable.

I'm afraid that words like "reliable source", "published", and "verifiable" do indeed have specific meanings in the context of WP policy. And regardless of how they are used elsewhere, when writing for WP and defending your work here, you must use WP's meanings. Otherwise your position here may be indefensible.

In particular, you write "The validity is the only thing that matters." I tell you three times: WP policy disagrees with you. WP:V states that all material added to WP must be verifiable, according to WP's definition of "verifiable", from what WP considers "reliable, published sources." Otherwise it can't be here, no matter how true it is, no matter how true you "know" it is, no matter how satisfied you are from your own research. That's WP:V in a nutshell. And there is no provision I can find that would allow any other policy to override that.

I really think you need to read all of WP:V, and as well, the essay "verifiability, not truth") for some of the rationale behind it.

And then you need not to say "but, I can ignore all that, because I know this claim is true."

Similarly the criteria for "reliable" generally excludes self-published sources (like online catalogs); see WP:RS.

However! I think I found an "out". WP:RS does allow that "Self-published or questionable sources may be used as sources of information about themselves" (emph. added). So a statement along the lines of "At least one PS/2 port can be found on most of the component motherboards offered by one on-line vendor", with a link to Newegg's "motherboards" listings, seems to me to be acceptable. (It's too bad that PS/2 ports are not among their selection criteria for motherboards, as they are for keyboards and mice.) A much stronger statement would be to do this with three vendors and say "several vendors", leaving their names to the refs. Jeh (talk) 21:44, 7 February 2013 (UTC)[reply]

PS/2-compatible peripherals

Since CN does not mean "I don't think this is true," adding and simultaneously tagging something as CN is not at all disingenuous. Adding new text to WP and adding a "doubtful" tag at the same time would have been disingenuous. But CN and "doubtful" do not mean the same thing.

Regarding the claim itself: I can provide for the "most currently sold keyboards and mice are not PS/2 capable" claim the exact same level of OR that you provided for the "most PCs have PS/2 ports" claim... only mine is from today, not two years ago. If you do the product searches at Newegg you'll find that PS/2-capable (including both PS/2 only and USB+PS/2) keyboards and mice are in the distinct minority. For keyboards, only 177 products out of 810 total are listed as PS/2-capable; for mice, only 27 out of 634.

However, even though that result supports the claim, it's not cite-able, any more than your survey is. A CN tag is therefore still needed. But the claim is now supported at least as well as the "most motherboards" claim, so can I expect that you will not delete it again? If not you'll be quite inconsistent. Jeh (talk) 08:36, 6 February 2013 (UTC)[reply]

“Since CN…”

It would be amusing to see how long an editor including {{citation needed}} with every single addition to any article would go on unimpaired. :p

“Regarding the claim…”

Then by all means reinsert that claim. See how easy it is to accept a believable claim?

“However, even…”

Well then by your own protocols you shouldn’t have inserted it, and therefore wouldn’t have cared that I removed it.

¦ Reisio (talk) 01:25, 7 February 2013 (UTC)[reply]

As I have explained, my own protocols do allow for adding such material with a CN tag. My OR can establish believability, plausibility, likelihood, etc. But not verifiability. Jeh (talk) 20:31, 7 February 2013 (UTC)[reply]

Color code

IMHO this wire color table should be removed completely. There are no standards or conventions for these things in the kinds of places where cheap commodity consumer goods (like computer keyboards) are made. This is just confusing to people who come here and need that kind of information most. The very presence of multiple combinations should be a clue that this isn't really any kind of reliable information. At best, it is a collection of random anecdotes. If people want to know which color is which signal, they should be instructed to verify the pin assignments on the connector. That is the ONLY place that is guaranteed to be correct. Even the silk screen nomenclature on the PC board internal to the keyboard has been known to be incorrect. Rcrowley7 (talk) 20:07, 21 September 2010 (UTC)[reply]

I agree. Be bold. ⇌Elektron 23:55, 30 January 2011 (UTC)[reply]

PS/2 & USB - different voltages?

Both use 5V DC, how can there be different voltage through an adapter? Zac67 (talk) 19:35, 26 May 2011 (UTC)[reply]

On a motherboard I have next to me the 5v of both connect to the same point. The ground is the main system ground.--Tagno25 (talk) 04:41, 1 October 2011 (UTC)[reply]

Both use 5V for power supply (although the original IBM PC apparently used 6V instead for its DIN keyboard interface), but the voltage levels for the data lines are different. The PS/2 interface uses 5V CMOS logic levels, whereas the USB data lines work at a much lower voltage level (however, they are 5V safe as well). PS/2-USB-adapters do not contain any electronics, they are passive adapters using a "tricky" wiring allowing the keyboard to detect if it is connected to a PS/2 port or not. If not, the controller in the keyboard will switch protocols (if the keyboard's firmware has been programmed to support both protocols), thereby giving the illusion of the adapter translating between PS/2 and USB protocols. Of course, this does not work, if the keyboard supports USB only. The PS/2 data lines are pulled up to 5V on the mainboard, but I don't see how this could become a problem because the USB data lines are 5V-safe as per the USB specification, and the keyboard must be designed to withstand this voltage on the data lines. --Matthiaspaul (talk) 00:32, 23 January 2012 (UTC)[reply]

Thanks for clarifying, I was 'huh'-ing over this one. It could be made clearer in the article - I'll go do this in a min :) ☭ cmn ☭ ( ❝❞ /✍ ) 15:43, 8 April 2012 (UTC)[reply]

Hot-plugging note

While hot-plugging is not supported and the system may not detect the new (or even old) device, it can usually be accommodated by pressing the reset button, causing the motherboard to POST and pick up the device on hard-reboot. Synetech (talk) 19:03, 20 June 2011 (UTC)[reply]

Actually, no. (And what would be the advantage of your method, anyway, if you still have to shut down the operating system?) The problem with the PS/2 port is that it is not hot-plug safe by design. Removing or inserting devices while the system is powered up (no matter if the operating system has been booted or not) can cause physical damage to the keyboard controller or the chipset of the motherboard. Originally, there was no protection for this, you just were not allowed to do it (the same applies to the parallel port, BTW). This should no longer be a real problem in recent motherboards with PS/2 ports, but originally it was - and quite a few people lost their mainboards this way. Hardware manufacturers have tried to make the PS/2 port much more fail-safe by changing the circuitry over the years, but it is still not a garanteed property and the plug itself is electro-mechanically not designed for hot-plugging. --Matthiaspaul (talk) 00:09, 23 January 2012 (UTC)[reply]

It's not so much the circuitry, but the fact modern PS/2 connectors have a 'make-first break-last' ground shell, which alleviates ungrounded current spikes, which is what destroyed many a mobo in the 90s. With that said, 99.99% of BIOSes check PS/2 status on input, and set registers appropriately (both host and device side), not touching them again until reboot. It's a holdover from the olden days. Modern PS/2 plugs can be hotplugged perfectly safely (from an electromechanical standpoint), just many manafacturers choose not to implement it because it's a legacy port and not worth the cost/effort tradeoff. ☭ cmn ☭ ( ❝❞ /✍ ) 15:53, 8 April 2012 (UTC)[reply]

Hmmm? Make-first break-last ground connections are completely standard and even expected in physical interfaces designed for hot-plugging. Look at the Serial ATA article, for example. Anyway, my understanding is that it is (or rather was) the circuitry. In the original PS/2 and clones that used the same or similar chips, damage would most commonly happen not because the ground shell was guaranteed to contact when nothing else did, but because there was no guarantee on the "make" sequence of the pins. The lengths of the pins weren't staggered, so which "made" first on connection was a matter of happenstance. There was about a 50% chance that data pins would be connected before power, with the unhappy result that the keyboard would, for a moment, draw power through the data pins. The chip on the mobo couldn't handle that... poof. Jeh (talk) 17:43, 8 April 2012 (UTC)[reply]

The problem with hot-plugging isn't about power being drawn through data lines - as long as these have short-circuit protection nothing much can go wrong there (presuming the connector isn't plugged shifted or twisted). The point is, if the ground connection has not been made yet while you connect data lines you'll have the potential equalization voltage and current present on the data pins which can easily kill circuitry. Connecting power before everything else is not essential but a good idea (so you can have hardware initialization before mating and avoid garbage on the signals). Furthermore, the line protocol should be designed so that garbage data is detected and discarded while mating.

On the PS/2 connector, prior ground contact isn't granted but looking at the construction it is highly likely. Early power contact is nowhere near granted and the protocol isn't designed for hot-plugging either (which you could possibly neglect for keyboard or mouse data). Zac67 (talk) 09:03, 9 April 2012 (UTC)[reply]

Sorry but the problem was of power drawn through data lines, because the chips didn't have short-circuit protection. And connecting power after ground, but before data lines, would have addressed that issue, so it was a lot more important than simply avoiding data glitches. Jeh (talk) 06:20, 14 April 2012 (UTC)[reply]

Combo-port

I’ve noticed that newer motherboards tend to have a single combined (half-green, half-purple) PS/2 port instead of separate ones. I’ve read conflicting information on what exactly they are for. Some people say that they make you choose either a PS/2 keyboard or a PS/2 mouse and have to use a USB one of the other, while (more) people say that you can use a PS/2 splitter (like the kind that laptop users sometimes use) to plug both devices into the dual-port. Motherboard manuals seem to vague or even lacking in their discussion of the combo-port.

It would be good if someone with a definitive answer could add it to the article. Synetech (talk) 19:03, 20 June 2011 (UTC)[reply]

Splitters work by connecting one device as normal and connecting the other to the two spare pins (not sure which is which offhand). It would be down to the individual motherboard manufacturers whether they actually implement those two extra pins. So you aren't going to get a definitive answer only a "it works/doesn't work with board x" answerPlugwash (talk) 04:15, 16 January 2012 (UTC)[reply]

Microsoft driver

There's note, that windows input driver v.8 no longer supports PS/2, but this topic Input and HID - Architecture and Driver Support shows the opposite. — Preceding unsigned comment added by Bigbes (talk • contribs) 10:56, 31 December 2011 (UTC)[reply]

Interrupts

I seem to remember that the legacy I/O devices take up valuable IRQs. A floppy, a COM port, a Keyboard socket and a mouse socket will take one interrupt each. Several devices can be attached to one USB bus and only take up one interrupt, although they may not perform as fast eg when using a keyboard for gaming. — Preceding unsigned comment added by 92.26.151.41 (talk) 23:23, 22 January 2012 (UTC)[reply]

Since the full support of PCIbus, APICs, etc., IRQs are no longer "valuable" anyway. Shared interrupts are expected and fully supported on PCI and PCI Express. Jeh (talk) 02:47, 30 January 2013 (UTC)[reply]

PS/2 power connector?

I have a Deltaco external hard disk case, and its DC connector is just similar to PS/2, I tested it and a mouse connector fits perfectly. This article has nothing to say about power cable usage. 85.217.42.90 (talk) 17:09, 8 September 2013 (UTC)[reply]

And it should not. What you're talking about is the physical connector type, which is a six-pin Mini-DIN connector. This article describes one use of such a connector. Similarly we have an article on D-subminiature connectors, and we have articles on Serial port and RS-232, but the latter two do not mention the many other ways 9- and 25-pin D-sub connectors are used. (Not to my recollection - if they do, that should be fixed too.) Jeh (talk) 17:44, 8 September 2013 (UTC)[reply]

Sorry, I thought this was the main article. Didn't realize PS/2 is one of many. 85.217.42.90 (talk) 00:16, 9 September 2013 (UTC)[reply]

No problem at all, you pointed out something that IMO should be fixed by renaming this article. Btw I too have a few external HD enclosures with the six-pin mini-DIN power connectors... And I have a few others with a connector that looks identical... but look close and you see there are four thick pins instead of six slender ones. Far better for a power connector, and much more rugged, but nonstandard. Jeh (talk) 17:04, 9 September 2013 (UTC)[reply]

Proposed move to "PS/2 port"

Anent the above, this article is not so much about the physical connector (which, as the IP above pointed out, is used for other things, even in this exact configuration, the standard 6-pin type). It is about the signals and protocols and pinout that appear on the connector in this particular usage. Note that the articles on Serial port and RS-232 are not called "Serial connector" or "RS-232 connector".

Accordingly, I think this article would be better named "PS/2 port". I so propose. Please discuss. Jeh (talk) 17:44, 8 September 2013 (UTC)[reply]

You might be right on this. 85.217.42.90 (talk) 00:17, 9 September 2013 (UTC)[reply]

I call one of these things a “PS/2 port”, and so am in favor of such a change. TOOGAM (talk) 17:56, 16 October 2013 (UTC)[reply]

Agree. Rename and add either a disambig or a redirect -- Akb4 (talk) 04:35, 5 December 2013 (UTC)[reply]

The Mini-DIN connector is the physical connector. A PS/2 connector is simply a Mini-DIN connector using specific signals and protocols. So I think the terms PS/2 connector and PS/2 port are just synonyms that mean the same thing, as are Serial connector and Serial port. – Wbm1058 (talk) 14:13, 15 September 2014 (UTC)[reply]

Well then, you should be happy, as there is already a redirect from PS/2 connector to PS/2 port, just as there is for Serial connector to Serial port. Jeh (talk) 22:59, 15 September 2014 (UTC)[reply]

usb over ps/2?

The article notes that a number of peripherals were shipped with a chipset smart enough to detect whether it was plugged into a ps/2 port or a usb port and could auto-switch between the two, and that such devices generally had usb connectors on them, plus a passive usb to ps/2 adaptor. Is the reverse ever true? Are there devices that have an auto-detecting chipset and a ps/2 plug on them and a passive connector with a ps/2 socket and a usb plug? And are these passive connectors, either usb-plug-becomes-ps/2 plug or ps/2-plug-becomes-usb-plug (if those exist) interchangeable, or is there more than one possible pinout? -- Akb4 (talk) 04:35, 5 December 2013 (UTC)[reply]

USB 6-key limit

I read this as hearsay all the time over the internet and I am perplexed because it simply can't be a hard rule. The USB keyboard I use myself, Microsoft Sidewinder X4, manages to get around this mythical limit somehow, and it isn't simply due to their special keyboard software because it retains its ability to recognize like 15 keys at once in Linux as well.

Furthermore the "source" used to back this statement up is a link to another wiki.174.45.212.205 (talk) 01:36, 1 February 2014 (UTC)[reply]

non-PC use of PS/2 connector

In the "port availability" section:

> PS/2 mouse and keyboard connectors have also been used in non-IBM PC-compatible computer systems,

This reads as (non-IBM) (PC-compatible computer systems), when in fact it is trying to say (non-(IBM PC)-compatible) (computer systems).

96.50.85.160 (talk) 20:18, 8 April 2014 (UTC)[reply]

WP:SOFIXIT. Since you spotted the problem, you can probably come up with a wording that fixes it. Jeh (talk) 20:38, 8 April 2014 (UTC)[reply]

External links modified

Hello fellow Wikipedians,

I have just modified one external link on PS/2 port. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

Added archive https://web.archive.org/web/20140811112340/http://www.usb.org/developers/devclass_docs/HID1_11.pdf to http://www.usb.org/developers/devclass_docs/HID1_11.pdf

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Y An editor has reviewed this edit and fixed any errors that were found.

If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 11:32, 5 December 2017 (UTC)[reply]

Does the host generate the clock while in host-to-device transmission?

In the Communication Protocol section, it says "To send a byte of data back to the keyboard, the computer pulls Clock low, waits briefly, then toggles it with a clock signal generated by the computer, while outputting a frame of bits on the Data line, one bit per Clock pulse, (...). The computer releases the Clock line when it is done.".

Does this imply that the host generates the clock while in host-to-device transmission?

Other sources [1][2][3][4] say that the device is always responsible for generating the clock cycle. When the host wants to transmit, it first pulls the CLOCK low to inhibit the transmission from the device, then it pulls DATA low, releases CLOCK and waits for the device to start generating the clock pulse.

[1] https://www.avrfreaks.net/sites/default/files/PS2%20Keyboard.pdf

[2] http://www.networktechinc.com/ps2-prots.html

[3] http://www.mcamafia.de/pdf/ibm_hitrc07.pdf

[4] https://web.archive.org/web/20180830070412/http://retired.beyondlogic.org/keyboard/keybrd.htm

Obs.: References [3] and [4] are included as references for this article.

I don't know if it is the case, but maybe the host is responsible for generating the clock on an older PS/2 protocol that I am not aware of? If not then the text should be corrected. — Preceding unsigned comment added by 2A00:1398:4:3601:D37B:883D:9C3E:8857 (talk) 21:38, 28 December 2020 (UTC)[reply]

I completely agree, I've rewritten the text. AFAIK, both AT and PS/2 have always had the device generate the clock. In any case, this is the usual mode, and if there are exceptions they should be called out explicitly. By the way, while I love Beyond Logic, that source ([4]) seems to have a mistake as it says

This bit will be read into the keyboard on the next falling edge, after which you can place the next bit of data.

whereas the other three sources all agree that on host-to-device transmission, data is clocked in on rising edges instead. Digital Brains (talk) 11:31, 17 January 2021 (UTC)[reply]

Add 2022 date when referring to continued availability

When I came across this article I was surprised by the statement that motherboards still commonly include a PS2 port. I checked and as of 2022 this is correct, however as this is date relevant and the ports will most likely disappear or become less common at some point in the future then the statement will then no longer be true. Therefore I have added a date when this is known to be true. This should be updated in the future if the interface continues to be fitted to motherboards. Lkingscott (talk) 07:32, 29 September 2022 (UTC)[reply]