6 ms·
But not everything can go through the USB-C cable you have on hand. That's the annoying bit with USB-C. We may have (almost) standardized on a single plug/sock
by TeMPOraL 5y ago
But not everything can go through the USB-C cable you have on hand.
That's the annoying bit with USB-C. We may have (almost) standardized on a single plug/socket shape, but we didn't escape the essential complexity - the fact that one type of connection cannot handle all the use cases we'd like it to. We just pushed that complexity into cables. Instead of having to deal with separate data, network and graphics ports, users now have to deal with potentially separate data, network, graphics and charging cables. I'm not convinced this is an improvement, because USB-C cables are a bottom-feeder market that will not hesitate to outright scam the buyer.
At this point I'm not sure it's an improvement. I feel like the optimum point would be a small amount of standards targeting mutually incompatible applications. That, or forcing some specification requirements on USB-C, and standardize some capability labels.
- infogulch 5y agoI hope that this is what USB4 will bring, since iiuc, USB4 is basically the IF's name for Thunderbolt-4-capable USB-C. This was enabled by Intel contributing the TB4 spec to the committee, in a shockingly benevolent move that I guess may have been the greatest internal political feat Intel staff pulled off in the last decade. Edit: Oh and presumably the ports on the Framework are USB4, they just can't say that yet because the certification is still in the works.
- mixmastamyk 5y agoReally? Interesting if true. Is that the case with other new intel gear?
- infogulch 5y agoAh it seems I was slightly off, it's TB3 not TB4, "The USB4 specification is based on the Thunderbolt 3 protocol specification." [0] But it does require: USB-PD, PCIe & DP tunneling, minimum 20 Gbit speed, max 40 Gbit speed. Stated goals to "minimize end-user confusion". I've seen some peripherals and such with it, no laptops yet though. The spec was released in 2019, so considering hardware cycle time we should start to see more devices soon. It's pretty cool that Framework will likely be on the leading edge of that wave. [0]: https://en.wikipedia.org/wiki/USB4 https://en.wikipedia.org/wiki/USB4
- curt15 5y agoThe Dell XPS 13 9310 has two TB4/USB4 ports.
- iSnow 5y agoIs Dell still gimping them to 10Gbit/s like they used to? Considered an 13" XPS for years, but then turned to Apple because of this ridiculous decision.
- Dylan16807 5y agoThey'd better not be. Thunderbolt 4 mandates 32Gbps of data. The site says each port has 4 lanes of pcie 3.0
- jjoonathan 5y ago> PCIe tunneling Hell yeah, USB over PCIe over USB over PCIe here we goooooo!
- hfjrkekdkfj 5y agoI do not want all my USB-C cables to be able to handle 90W. That would make them very thick and expensive. I do not want all my USB-C cables to support the maximum 40 GBps speed (or whatever it is). That would require them to have all the 19 wires and shielding and all and again, would make them expensive and short. And just imagine how much a 90W maximum speed 3 meter cable would cost... I prefer having one power cable, one fast cable and then a bunch of disposable cables for general use cases.
- kragen 5y agoUSB-PD permits 100 W with 20 V at 5 A. If that's carried over just two round copper wires (I don't know whether it is in USB-C) they would need to be 18-gauge or thicker for safety—about 0.94 millimeters. If they're copper, that's about 7.3 grams of copper per meter, 14.6 grams including the return path. Copper is expensive: almost US$7/kg. So a 3-meter 5-amp DC or two-phase cable would weigh 45 grams and contain 15¢ worth of copper. You could drop both the cost and the weight by going to aluminum. If you were designing the system from scratch, you could use 3-phase AC to cut the weight by half again, and use 48 volts to cut the weight by another 58%. I don't have any idea how thick the 19 wires have to be for USB 40Gbps (GBps?) but I imagine the answer is "not nearly that thick".
- ethbr0 5y agoAfaik, signaling frequency is typically cross-talk limited, vs diameter limited. Think more shielding, stricter tolerances over thicker.
- kragen 5y agoIt sounds like you could maybe make substantial progress by just separating the differential pairs (?) by a millimeter or two of dielectric, giving you a ribbon cable, with much lower crosstalk than the round kind. Bonus points if you color the dielectric rainbow colors.
- Dylan16807 5y agoI feel like it's harder than shielding and tolerances, given that longer passive cables don't seem to exist despite the very high prices people are paying for active cables.
- int_19h 5y agoIt's definitely an improvement, because you can still carry the one cable that does it all, and use it for everything, even the things that don't actually require it.