💾 Archived View for soviet.circumlunar.space › zwatotem › diff › connectors.gmi captured on 2023-07-10 at 13:44:15. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2023-01-29)

-=-=-=-=-=-=-

# Does the perfect cable connector standard exist?

I asked myself the question from the title, after seeing some cables on the internet, that added magnetic capability to USB-C by consisting of two de-facto endings, one of which is a regular USB-C connector, and the other which bridges it magnetically to the rest of the cable. I thought then, that USB-C standard turned out to not be enough if people are making mods like this.

So should USB-C have been magnetic?

Not always magnetic connector is the best option. Sometimes you have to opt for higher connection rigidity, in which case holding the contacts together magnetically wouldn't supply enough force. On the other side of this spectrum are specialized cables which are ment for long term, reliable connections. These usually have some kind of latching mechanism, which makes sure, that direct pulling of the cable doesn't break the connection.

What else does matter?

This brought to my mind some kind of connector, which would supply all of the important qualities that you would like to have. I thought about what we like about different kinds of connectors, from USB through RJ45 to XLR. I came up with the following list:

As you can see there is a lot to consider. Let's break it all down and see what we can do.

Designing the perfect connector

I looked at all the aforementioned qualities and tried to come up with a connector, that would have all of them. Unfortunatelly it soon became clear that there are many pairs of variables that need a compromise in a physical sense. Let's take a look at them.

The safety-directionality problem

So I started with the size. Let's say we want something really small and handy - the size of USB-C perhaps. I can send low power through such a connector, but some devices requre higher power. The obvious solution to this is to increase the voltage, but that poses possible problems with arcing (if not upon plugging, then upon unplugging) and safety. For voltage protection the standard practise is to have pins exposed only on one side of the cable. That through blocks the much liked feature of USB-C of cable biderectionality. For shorter cables it's not usually a big problem but if you ever carelessly uncoiled XLR cable across the theater only to learn that you did it the wrong way you know how it fires back. XLR connectors however have one useful feature: they allow for easy extension (daisy-chaining). With that the struggle is increasing: we would need some system that hides the contacts on both sides in such a way that they are inaccessible to the human hand, but still able to contact after plugging, and all this while maintaining the lack of connector gender.

I think this is possible, we'll see later, how.

The shape problem

Let's start fresh and think of the other variable: connector shape. Typically this is single orientation plug (like RJ45), double sided plug (most power plugs, USB-C) or round plug (like audio jack or various barrel plugs). Of course for the ease of use we would want the latter one: it allows to plug the connector first try without even watching and with a bit of guidance in the form of gradient of the device wall we can hit a distant point from the actual socket and still make it. Additionally rotationally symmetric plug can be adjusted afterwards, which is useful in the case of an angled cable termination. If we select a conical-shape plug, like an audio jack (imagine buffed combo jack with even more connections) the safety problem comes back - all the contacts are exposed. The other solution like barrel plug can hide the contacts, but the architecture of concentric circles (for multiple wire connections) would quickly make the plug either very big, or very detailed and britle, requiring high precision to plug. Still - either way we go, we lose the previously discussed biderectionality.

The optical dillema

Fiber optics is thus far the fastest known way to send information (both in therms of throughput and latency (not source-verified!)) so surely we'd want at least partly include a glass fiber in our cable, right? Well, not so fast. It is well known, that glass breaks, and that is no different for glass fibers. Even if they can bend, their bend radus is limited, and copper - while certainly not perfect - is still better than glass in that regard. "Why not allow for optional optics then?" you may ask. I don't wan't optionals in a standard. It usually ends badly, as you end up having to remember all the optional features of a given standard and be extra careful when buying a cable or a device, despite the plugs literally matching. At that point it is better to make a separation and use two cables, so I'd root for that.

With that I already answed my question, but let's keep digging.

The strength problem

About how the cable and a socket hold together.

You can say: "There is no perfect solution for everyone. You either want it firm or you want it light". For me it's not a lost feat. The first contact can be made with magnets, and that would be enough for you in most situations, and pulling the charging cable would not be a threat for your laptop. Then you could push the plugs together to make the connection thighter (like HDMI or USB, which hold really nice and firm after inserting, but still can be pulled with some force). Finally you could close an optional latch to make the connection proof to any accidental pull, making it viable for long term connections.

What to make out of all this?

I will pass on mentioning about things like plug angle or water resistance. Although I was thinking about them, I don't want to confuse you even further so here is, what I came up with in the end:

Think of the simplest way of making two identical shapes that plug together, but don't worry about electrical connections for now; take it purely geometrical. Two flat planes? That's probably not gonna work, so something else. What about a 2⨉2 checkerboard, where a black square represents an outwards "pole" and white square is inwards "hole"? That is an easy one.

Now let's enlong it to 2⨉n where n is the number of pins we would like in that connector, probably about 10-20. Now we have a line and we can wrap that line around into a circle. The inner layer of the circle is the connection layer. On the clockwise and counter-clockwise sides of each plastic pole put a contact pad ranging to a little more than half the hight. This ensures the connector is protected on the top. The outer wall of poles is purely unconductive. They just create a wall to protect the contacts on the sides. Now widen the inner poles ridially towards the center, leaving only a small hole inside. At this point I should make a notice that a whole thing should have about 0.5 to 1 cm diameter, so no finger should reach the contacts and therefore I consider safety as more or less handled.

Now put some magnets inside, and you've got yourself a snap-on solution. Want a better hold? Push the connectors toghether and they are friction-fit. Finally for the latching mechanism - put a free spinning helix spiral in the hole in the middle. You can block the movement of this spiral with a ring on each of the connectors' grip, so you can control each one separately. In particular an internal port on the device should have this helix stationary by default, so that you only need to rotate the cable end. If you latch two cables you should naturally lock both of them.

Conclusion

Is this connector perfect? By no means.

All in all, I'm proud of this design. I think it could handle at least the challenges that USB-C is faced with and invalidate all the hacks that try to make it magnetic or rotational as well as push into the markets that wouldn't adopt USB thus far like analog audio equipment, home appliances power. It could also be useful in places where several cables carry different kinds of data in parallel cables, that could use some kind of wire delegation protocol to give a pair of cables for each separate connection.

One more thing

One additional thought I had, when people started complaining about the removal of headphone jacks in phones: USB is just as good of a connector for headphones as audio jack it's just that the phones should have 2 of them, and everything would go smooth and dandy. Either this or chain a second socket straight after the headphones'/charger plug. Make the USBus a literal bus, how it's supposed to be.

Subscribe for more content :)