With Apple's lightning cable, and USB 3.0, reversible cables are taking off, and I personally think this is very convenient. But we have had better than reversible for a long time, in the form of the headphone jack, which can be inserted in any direction, not just 2 directions. Why isn't a headphone jack shaped connector used for data more often? All I ever see that shape used for is audio and power supplies (I've seen it used once for data, in the iPod shuffle, but thats it).
-
2Contrarian comment: we see more audio on data connectors (USB in particular) than data on audio connectors. At the same time, there exists [something like this](http://www.mouser.com/ProductDetail/FTDI/TTL-232R-5V-AJ/?qs=OMDV80DKjRpyBfF7K%252bVrRw%3D%3D&gclid=CMi4r9q_wr0CFaw-Mgod8FAA7A), which implies that people do use audio jack for serial data. – Nick Alexeev Apr 02 '14 at 19:05
-
Hah, very true. Still though, from a usability standpoint, the round connector shape seems like the clear winner. Why isn't it the other way? – Drew Apr 02 '14 at 19:10
-
21If connectors should indeed be standardized, I'd hope they didn't land on a connector like this, where you short every connection to ground on every insertion and extraction. It's a lousy choice for standardization. – Scott Seidman Apr 02 '14 at 19:37
-
3Also used for the plethora of "card readers" for the iphone and mobiles. Like Paypal Here and Square – David OBrien Apr 02 '14 at 22:52
-
2@NickAlexeev no, audio on usb is still data. Not analog audio. – Passerby Apr 02 '14 at 23:10
-
@DavidOBrien that's still modulated analog audio signals. The original Squares didn't even have encryption on them, they were a straight connection to a mag head. – Passerby Apr 02 '14 at 23:11
-
3CASIO and TI graphics calculators do use an audio-jack-type plug to transfer data. – user253751 Apr 03 '14 at 01:14
-
Did you just call Intel Thunderbolt... "Apple's lightning cable"..? This made me chuckle. – Vercas Apr 03 '14 at 11:06
-
1No... Thunderbolt and Lightning are different things... – Drew Apr 03 '14 at 18:19
-
15Thunderbolt and Lightning, very, very frighting! Galileo! – dext0rb Apr 03 '14 at 19:45
-
@Passerby I've seen D+ and D- multiplexed with analog audio lines. I can't find a reference at the moment, unfortunately. – Nick Alexeev Apr 03 '14 at 23:01
-
1@NickAlexeev some cell phones do this, with a usb mini/micro id pin for a dock mode. It's not very common. – Passerby Apr 03 '14 at 23:03
-
RS232 over headphone jacks was moderately common when RS232 was common - but you can easily do RS232 over 3 wires. – Chris H Apr 04 '14 at 11:32
-
1@dextorb Hahahaha... I was just going to mention the Intel Galileo's serial connector! – gwideman Apr 04 '14 at 12:32
7 Answers
Digital signals are highly susceptible to the noise generated by rotating the plug.
For audio, these noises (cracks) are rarely audible unless they last longer than 50us (simply because of the fact that we're unable to hear frequencies over 20kHz). So, the cracks becomes audible only when the surface of the connector has deteriorated enough that the period of lack of connection is substantially longer.
As a rule of thumb, any connector where there are moving parts while the connection is established, is a terrible idea for high frequency digital data. It might be acceptable for low frequency digital data, as well as power supplies.
Finally, most digital standards require quick detection of the disconnect - even though the above issue could be worked around with proper ECC (Error Correcting Codes), USB assumes that any loss of connectivity for over 2ms is considered a disconnect. (USB 3.0 SS is even more strict).
-
14Ah. This also explains why the iPod shuffle physically restricts the iPod from rotating while transmitting data. – Drew Apr 02 '14 at 19:13
-
I might also add that audio connectors use the shield as a common ground for both left and right. High speed digital connectors like separating the ground from the shield, and even having multiple signal ground paths, up to one ground per signal wire (assuming single-ended, since differential signals are their own return path) – ajs410 Apr 02 '14 at 19:16
-
The point on digital standards - it is probably only defined so short *because* the connectors are of the kind where there won't be long periods of no connectivity once inserted? – deed02392 Apr 03 '14 at 09:02
-
1Not quite - it's always a tradeoff between legitimate disconnect and intermitent loss of connectivity. Otherwise you might end up having a situation where it's unknown what needs to happen - should, for instance, for (Mass) Storage devices - should the program block (hang?) waiting for the device to reappear? Should it return error? Fail silently? Not to mention the exact same tradeoffs that need to be implemented on the device side - which adds complexity to the already complex embedded systems design. – qdot Apr 03 '14 at 18:04
-
The headphone jack was used for video and audio on the Nokia N95, N96 and N97 phones. Possibly others as well (Example product http://www.amazon.co.uk/Nokia-Video-Out-CA-75U-Mobile-classic/dp/B003Z6U89O/ref=sr_1_1?ie=UTF8&qid=1396620321&sr=8-1&keywords=n97+video+cable) – Dreamwalker Apr 04 '14 at 14:04
-
A reason why the round barrel connector with multiple ways may be not preferred is that to insert it, you'll push it through one or two "wrong connections" before it finally comes to rest. Ditto when removing it.
This means that it could short a power supply out (momentarily or indefinitely if not inserted properly). It could also reverse a supply to a chip and this would be an obvious problem.

- 434,556
- 28
- 351
- 777
It's difficult to get more than a few connections in a coaxial form. I have a hard time imagining an equivalent-to-HDMI plug/socket system (19 pins) made coaxially (image from here).

- 376,485
- 21
- 320
- 842
-
2
-
2Sure, as are military circular connectors, Lemo connectors etc. but they're not circular, not coaxial. – Spehro Pefhany Apr 02 '14 at 23:18
-
1DIN connectors are doubly bad: orientation matters (unlike a phono plug), and orientation isn't visibly obvious to the user (unlike USB or Lightning). – Russell Borogove Apr 03 '14 at 03:57
-
A DIN connector could be designed to be somewhat orientation independent with dynamic pin assignment, although I realize that is introducing a ton more complexity. – Drew Apr 03 '14 at 18:22
-
@RussellBorogove pfft. Like orientation is obvious to usb. People are always mistakenly trying to plug them in upsidedown. – Passerby Apr 03 '14 at 23:04
-
20Well, [now you don't have to imagine anymore a jack with 19 pins](http://tinypic.com/r/2lurnyd/8) (; – Francisco Presencia Apr 04 '14 at 00:44
-
1
-
-
@Passerby Relative to DIN, you at least have it narrowed down to two possibilities. – Russell Borogove Apr 04 '14 at 20:56
-
@FranciscoPresencia OMG! That's not a connector, it's a weapon! :-) – LorenzoDonati4Ukraine-OnStrike Jul 18 '15 at 09:24
In fact, theye are. But as explained by qdot, only for low-speed connections. One example that comes to my mind is are the Texax Instruments graphical calculators, like the TI-83+ http://education.ti.com/en/us/products/computer_software/connectivity-software/silver-usb-cable-for-windows-mac/features/features-summary

- 223
- 1
- 7
-
My [iPod Shuffle](http://www.apple.com/ipod-shuffle/) begs to differ. – Cees Timmerman Apr 04 '14 at 15:26
There's another reason which none of the current answers have touched on, but which has a very clear user experience impact. Simply enough, lots of devices use more than a single connector. My MP3 player sitting at the desk as I am typing this has both USB and headphone receptables. The backup HDD at home needs USB and external power. Lots of common devices, particularly those that might benefit from something like this, use two or three connections: one for power and one or more for something else.
Same connector for different purposes has been tried; look at the IBM PS/2 and its (dubious choice of the) use of identical 6-pin Mini-DINs for keyboard and mouse that has caused, I am certain, untold frustration with devices being connected to the wrong port. Now, that's not exactly the type of connector you're asking about, but it presents the same sort of user experience problems. Had IBM used different style connectors for connecting hardware that needed to be connected to the correct port to work (even just different number of pins with one or more unconnected would have prevented the user from accidentally plugging one into the socket for the other), subsequently color coding the mouse and keyboard connectors would not have been needed.
Standard headphone style jacks also momentarily make connection with wrong cables during insertion or removal. This is largely a non-issue for headphones, but it can be a very serious concern if one of those pairs carry voltage at the ability of delivering significant current. While that could conceivably be worked around at the protocol level using something like a Power Good signal, that's another layer of complexity added for at best marginal gain.
Such connectors are also only available in a few different sizes (2.5 mm, 3.5 mm, 6.3 mm, and I think that's about it as far as consumer-appropriate sizes are concerned; I'm not sure if there are other sizes for specialist applications). While this does not necessarily have to be a showstopper, it is a limiting factor particularly if you want to use more than one of them on a single device. Of course, USB too is only available in a few select sizes, but very rarely is more than one USB input needed.

- 222
- 3
- 13
-
The Macintosh SE and Macintosh II had two ports on the back which could independently accept keyboards and pointing devices (one could if desired use two keyboard or two pointing devices). IBM tried to imitate that with the PS2, but their connector had separate non-sharable pins for keyboard and mouse data; if one plugged in two keyboards or two mice, nothing would work. To prevent confusion from that, later machines only ran keyboard-only signals only to one connector, and mouse-only signals only to the other. – supercat Apr 07 '14 at 16:01
-
@supercat Actually, that's my point; identical connectors physically, but completely different from the point of view of hardware compatibility. In today's world USB is ubiquitous for that purpose, and that works because USB was designed with such sharing in mind, the same way the Apple input device bus was. – user Apr 08 '14 at 09:55
-
1I don't think many people know that on early machines the keyboard and mouse connectors were actually interchangeable; the connectors were the same because it didn't matter which plug went into which jack. Further, the design would have allowed the keyboard to have 6-wire jacks on both sides (wired in parallel); run one cord from computer to either side of the keyboard, and plug the mouse into the other. – supercat Apr 08 '14 at 12:18
They are used, but just in a product that doesn't see a lot of daylight. Check it out:
https://support.apple.com/en-us/HT201893
The reasons they aren't prolific have been answered above, but it is interesting to note that it HAS been done, even if it isn't done often.
One additional area to consider is the ability to effectively route the conductors within the connector, with a coaxial connection like a headphone jack the center connector is surrounded by a insulator which is surrounded by another conductor, and then another insulator and another conductor meaning for a simple 3 point connection they had to layer 5 times in a circular fashion effectively making the connector (1+(n*2))width of the conductors/insulators.
wouldn't this make anything larger than 3-4 conductors impractical? also i can't even count the number of headphone cords i have gone through due to their lack of good protection from bumps/wiggles. even the connectors soldered to most boards are prone to disconnect.

- 1
-
This is not necessarily a problem with the type of connector in general, however. You can route separately to the different conductive surfaces, which themselves can be circular. It's not so much a matter of how must be done as it is a matter of how it is currently done. – user Apr 08 '14 at 09:59