Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Technology

USB-C Can Hit 120Gbps With Newly Published USB4 Version 2.0 Spec (techcrunch.com) 79

Posted by BeauHD from the new-and-improved dept.
An anonymous reader quotes a report from ArsTechnica: We've said it before, and we'll say it again: USB-C is confusing. A USB-C port or cable can support a range of speeds, power capabilities, and other features, depending on the specification used. Today, USB-C can support various data transfer rates, from 0.48Gbps (USB 2.0) all the way to 40Gbps (USB4, Thunderbolt 3, and Thunderbolt 4). Things are only about to intensify, as today the USB Implementers Forum (USB-IF) published the USB4 Version 2.0 spec. It adds optional support for 80Gbps bidirectional bandwidth as well as the optional ability to send or receive data at up to 120Gbps. The USB-IF first gave us word of USB4 Version 2.0 in September, saying it would support a data transfer rate of up to 80Gbps in either direction (40Gbps per lane, four lanes total), thanks to a new physical layer architecture (PHY) based on PAM-3 signal encoding. For what it's worth, Intel also demoed Thunderbolt at 80Gbps but hasn't released an official spec yet.

USB4 Version 2.0 offers a nice potential bump over the original USB4 spec, which introduced optional support for 40Gbps operation. You just have to be sure to check the spec sheets to know what sort of performance you're getting. Once USB4 Version 2.0 products come out, you'll be able to hit 80Gbps with USB-C passive cables that currently operate at 40Gbps, but you'll have to buy a new cable if you want a longer, active 80Gbps. In a statement to CNET, USB-IF said they don't expect to see supporting products for the new spec for "at least 12 to 18 months."

The USB Implementers Forum also updated the USB Type-C Cable and Connector and USB Power Delivery specifications today to accommodate USB4 Version 2.0.
This discussion has been archived. No new comments can be posted.

USB-C Can Hit 120Gbps With Newly Published USB4 Version 2.0 Spec More

USB-C Can Hit 120Gbps With Newly Published USB4 Version 2.0 Spec

Comments Filter:

  • Perhaps stop naming cables based on protocols that go over them.

    Perhaps use a classification like they do for networking cables where the cable capability is described like CAT5, CAT6e.

    Let's go with UCCn (Universal Cable Category), eg, UCC1.

    • Re: (Score:2)

      by ceoyoyo ( 59147 )

      USB-C is the connector. It's not confusing at all. The name "USB" in it is useful, because it tells you what it's for. Like "ethernet cable/jack/plug/port." People who insist on confusing it with the USB version are weird. Nobody confused USB A, B, mini and micro with the USB version.

      The USB version naming is confusing as hell, but less so now that they've stopped withe the super duper uper speed shit.

      • Each standard also comes with its own updated USB-C connector standard. Makes sense since youâ(TM)re increasing the power and data rates, pins need to become bigger and better shielded.

        Sucks for those people in the EU that are forced to have a specific version implemented on the legacy non-standard USB-C-like connectors until they get around to voting for a new legacy USB-C connector.

        • What the hell are you talking about? USB-C specifies the physical form-factor of the connector, and there's only one.

        • What you just said is simply false. The USB-C connector has not changed since its introduction, and has no need to change. Cables can, of course, be manufactured to different specs, but the form-factor remains the same, and you can absolutely put a new 240W-capable cable into any USB-C receptacle and it'll fit just fine and negotiate as much power as the host says it can safely deliver. Stop pulling "facts" out of your ass.

          • Re: (Score:2)

            by Shaitan ( 22585 )

            It literally says they changed the standard to accommodate this new spec in the TFS:

            "The USB Implementers Forum also updated the USB Type-C Cable and Connector"

            • You misread it. Read it as a sentence.

              "The USB Implementers Forum also updated the USB Type-C Cable and Connector and USB Power Delivery specifications today..."

              They updated the specifications, not the connector. The title of the document is "USB Type-C® Cable and Connector Specification Release 2.2". There were no changes to the connector design, which you can see if you click the link and look at the actual spec. There's even a redlined version that highlights the changes in red.

              • Re: (Score:2)

                by Shaitan ( 22585 )

                You are technically correct in what you've said. Per these specs the new cable will work in the old receptacle and in the unlikely event manufacturers follow the specification will be marked visibly. But the devil is in what you omitted. Existing cables will not work in all cases. So your bin of USB-C cables will randomly work or not work in various applications.

                Worse they've retroactively changed the standard for some tests like attenuation so that existing cables that pass the existing tests and were sold

                • You're conflating USB4 with USB-C here. Different cables with USB-C connectors have different capabilities. USB4 cables must comply with the new USB4 specs. No USB3.x cables suddenly fell out of spec for that standard.

                  Older cables not supporting newer standards but maintaining connector design is neither unique nor new.

                  Use a CAT5 cable on a CAT6 network and you ain't gonna be seeing CAT6 speeds. Same connector, different cable specs.

                  Use a SATA1 cable on a SATA3 port and you're not gonna see 600Mbit/s. Same

                  • Re: (Score:2)

                    by Shaitan ( 22585 )

                    "You're conflating USB4 with USB-C here."

                    No, I'm not. That document is the USB-C specification which has been modified along with the introduction of USB 4. 3.7.5.1 USB 2.0-only Cable Assemblies, Table 3-30, item D+/D- Pair attenuation changes the requirements for the EIA 364-101 test.

                    "Use a CAT5 cable on a CAT6 network and you ain't gonna be seeing CAT6 speeds. Same connector, different cable specs."

                    Same connector but different cable and the marking is on the jacket of the wire at 1ft intervals. Not to men

                    • Why not? It is a USB-C cable and USB4 still uses USB-C. You should expect that any USB-C cable works where a USB-C cable is required, as distinct from usb-micro or usb-mini. And that is exactly what people are going to expect. They did this crap with HDMI and it has been a nightmare for everyone but retailers who can always con people into thinking they have to buy a new highly marked up cable.

                      Almost* any USB-C cable will work where a USB-C cable is required, just not at the maximum data rate/power delivery that the host/device supports. Also, remember that Micro-USB actually did get a new form factor when Micro-B Superspeed (3.2) came along and it became that weird elongated double-plug. Even then, you could just use a USB2 Micro-B cable, ignoring the added pins, and be limited to USB2 speeds.

                      Finally, the USB-IF has actually got a plan to address the problem you're focussed on; going forward, US

                    • Re: (Score:2)

                      by Shaitan ( 22585 )

                      "Almost* any USB-C cable will work where a USB-C cable is required, just not at the maximum data rate/power delivery that the host/device supports."

                      That meets my definition of not working and is the worst kind of failure... silent. Not delivering the data rate and power requirements the interconnected devices are rated for means random and inconsistent results for the devices/applications you are interconnecting. How does this play out in a consumer scenario?

                      Bob has a bunch of usb-c cables accumulated and b

          • Re: (Score:1)

            by guruevi ( 827432 )

            Off course the USB-C connector has changed, there are updates to its spring mechanism and a number of mechanical directives since the first version was released. Obviously there needed to be updates to solder pad and metal gauge specifications, and shielding/harness as you increase data and power rates. You can't electrically run 240W over a circuit board that has been designed for a specific 100W specification (which the EU mandates).

            And FYI, the EU did not implement USB-C, instead it used a derivative IEC

            • Once again, the form-factor has not changed. The changes you mention do not prevent a previous generation cable being used in a USB4 port, or vice-versa. You just won't get the full speed/power delivery if you use a previous generation cable on a USB4 port.

        • Re: (Score:2)

          by ceoyoyo ( 59147 )

          They changed the specification document. One of the changes was to add this note:

          Note 3: USB4 Gen4 passive cables are electrically equivalent to USB4 Gen3 passive cables. For USB4 Gen4 active cables, refer to Chapter 6 for the USB4 Gen3 and Gen4 electrical requirements.

          The connectors are exactly the same and will stay that way for a long time. The passive cables are the same. The active cable chapter got mostly rewritten, but most people have never come across an active USB cable.

  • Intentional confusion so that they can change the cable ends to be physically incompatible. "Make it easier for the consumer" they'll say. That'll show the EU to mess with the free market!
  • Make it happen. You know you want to.

    Stop with this .0 spec naming and just rate cables class 1 2 3 4 5 up to whatever new whiz-bang speeds or power delivery you need. No one cares about actual throughput numbers at this point. Class 1,2,3,4 for speed and Class a,b,c,d for power. SO cables can be simply class 1A or 3A or 2C etc. KISS.
    • Comment removed based on user account deletion

      • Re: (Score:3)

        by Shaitan ( 22585 )

        That is just the beginning. At some point USB-C will win the franchise wars and all cables will be USB-C. No more HDMI, power cables, mouse/keyboard cables, etc it'll just all be USB-C cables. You'll even slap a USB-C connector on the end of your HAM radio antenna.

      • Three phase power baby! It should be possible to power a mainframe using a USB-C cable...

        Wait are you the chair of the USB-IF?

  • Complicated? (Score:5, Insightful)

    by TechyImmigrant ( 175943 ) on Wednesday January 18, 2023 @08:35PM (#63221366) Homepage Journal

    It isn't complicated if you bother to explain it.

    There's a connector. That has evolved over time with confusing names like A, B and C. A goes at one end, B the other. C goes anywhere.
    There's a protocol, with right hand numbers for the major version confusingly numbered 1.0, 2.0, 3.0 etc. and speed variants within that.
    There's a power protocol (PD). That only has one name, so it's never called out.

    What you get results from the combination of the wire, the connectors and what's at the end points. So naming the wire after things it doesn't control doesn't make sense. Naming the wire after thing it does control (the high speed wires (or not), the PD wires (or not), the connectors (which type?) ) does make sense.

    Just like wired ethernet, where there is one name (IEEE P802.3 / Ethernet) with a myriad of sub types of connectors, wires, L2 protocols, power protocols. We just call it ethernet, but it will negotiate down to the lowest common feature set (just like USB) and if you want to make sure you're getting the features you want, you need to read the feature set (just like USB).

    • Re:Complicated? (Score:4, Informative)

      by AmiMoJo ( 196126 ) on Thursday January 19, 2023 @04:40AM (#63221828) Homepage Journal

      A plugs can be on both ends now. With USB 3.0 we started to get hubs that have a USB A port as the input. I don't know why they did it, it's dumb. The point of A and B was to make sure you didn't plug a dumb power supply into another dumb power supply and blow up your equipment.

      PD got screwed up as well. 100W wasn't enough for some laptops, so manufacturers made their own versions. Lenovo has a 130W USB C charger. I think most of the machines will charge from 100W USB PD as well, but not the more common 65W.

      There is confusion around USB 4 and Thunderbolt compatibility too. In theory USB 4 doesn't need to support Thunderbolt, but Microsoft made it a requirement for Windows so all laptops do support it. I think.

      • I understand there was a desire to support computer to computer communication. There were various other use cases put forward to justify it, but I discussed it with the people owning the standards and it was computer-computer they were thinking about. Whatever it was, it did not catch on.

        PD - my involvement was in the authentication protocol, so I don't have any insights into why they limited to what they did.

        The missing link is the labelling. I want to be able to look at the wire and see markings as text o

    • Re: (Score:2)

      by Shaitan ( 22585 )

      Ideally you have different connectors so that you can tell what it does by looking at it. Yes in many cases the same connector is used for ethernet for different variations but the wire is literally designed on the jacket of the wire so you can still tell what it is by looking at it.

      What good is having a giant tangled box of cables that all look like the same thing but secretly have different capabilities?

    • Re: (Score:2)

      by drhamad ( 868567 )
      It's not confusing if you're generally involved in the industry. But there's no easy way for the average consumer to know wtf is going on. Heck, it's the same problem with HDMI. Historically, there were different connectors for different types, at least over more than a [relatively] small range of difference.

  • How fast will usb4 gen3 x4 2v2.1 be? (Score:5, Funny)

    by iAmWaySmarterThanYou ( 10095012 ) on Wednesday January 18, 2023 @08:35PM (#63221368)

    Because I've got a usb4 gen2 4v2.8 cable and a usb4 gen4 2v2.4 device and can't figure out if that should be faster or slower than usb4 gen3 x4 2v2.1 or if I should switch to usb4.2 gen3.1 v6?

    Anyone know?

    • No one cares, all you need to do is look at the pretty pictures and logos that the USB-IF have created. They realised that people who claim they are smarter than others aren't actually so, and thus dumbed USB down for them.

      • So you looked at how they fucked it up again, right, and didn't just see a post from someone with a trigger name and immediately fire off what you thought was a clever slam. Nope, not you, you knew exactly what was being referred to and totally got it because you are so smart and not easily triggered at all. Oh and brave. You're very brave too.

        • Yeah what was being referred to was an idiot who can't read the number on a logo and realise the biggest number of the two thingies that plug into each other is what will work.

          You're username is completely backwards.

    • Comment removed based on user account deletion

      • Sweet! So it only took 4 versions, or was it 7 or maybe it was Gen x4 3.2 versions to reduce to one cable that is now universal, except all the ways _you_ say it isn't. To say nothing of future versions, features, and so on that will quickly result in even more cables and oh wait that physical plug can't support this other new class of device so we only need this other port for those.

        But it's all really easy now, right?

        You know what easy is? When my mom can go into a store and buy a cable and it will 100%

  • Suggestion for the next spec (Score:5, Interesting)

    by TomR teh Pirate ( 1554037 ) on Wednesday January 18, 2023 @10:15PM (#63221488)
    put QR codes on the devices, codes on the cables, then scan both so a computer program can tell you if the combo will work. I'm only half kidding
    • You insensitive clod, that doesn't work if I have to plug in the cable for the internet to work to look up the QR code! Clearly they need to include a cross reference chart and picture reference on a flash drive built within each end of every cable.

    • Re: (Score:2)

      by kmoser ( 1469707 )
      Or make it so the O/S can query the port and cable to provide *useful* information about the devices attached to it, without making the user look up all this stuff. Something like "Port is type X with a max speed of Y, currently attached cable is type J with a max speed of K, can provide power, etc."

      It's 2023. Why is this so hard?

      • No thanks. I don't want any active components in my USB cables. If I need to have different features on different USB-C cables, I'll personally standardize on a color for each (red for PD cables, yellow for 80Gbps, etc).

        • I have bad news for you, PD compatible USB-C cables already have active components in them. They have chips called emarkers that inform the host about the capabilities of the cable.

          • Re: (Score:2)

            by AmiMoJo ( 196126 )

            It's a shame there is no way to view that data on a computer or mobile device.

            Then again, the manufacturer can always lie about the capabilities.

            • Re: (Score:2)

              by kmoser ( 1469707 )
              The point of queryable cables isn't to ferret out false claims of capabilities but rather to programmatically determine (claimed) capabilities. Even if you could somehow guarantee the claim was correct, a cable could still fail to function as claimed due to mechanical issues, in which case the verified claim was useless in determining capabilities.

              If you want to make it harder for manufacturers to build cables with false capability claims, you could require the cable's firmware to be cryptographically sig

          • Re: (Score:2)

            by Shaitan ( 22585 )

            Interesting. Is that why it is so damn difficult to find usb-c power distribution splitters cables?

            All I want for late Christmas is to power all the usb-c accessories from one usb-c female port on my VR headset without a hodgepodge of adapters or braiding/chaining a 4ft long only splitter I could find on the internet down to size.

    • You only half need to kid. The USB-IF has already prepared a set of logos that define port / cable maximum capabilities. Forget the QR code, just look at the pretty pictures.

    • Nah, we need color codes, as used on resistors.

    • put QR codes on the devices, codes on the cables, then scan both so a computer program can tell you if the combo will work. I'm only half kidding

      And even if they do supposedly work together...you will still spend 20 minutes copying 10 GB of files to that thumbdrive.

  • USB-C is a security nightmare waiting to happen (Score:3)

    by bubblyceiling ( 7940768 ) on Wednesday January 18, 2023 @10:58PM (#63221536)
    I really don't like this putting ALL-THE-THINGS in one port push of USB-C. Ports should be specific to their purpose. HDMI for Video, LAN for Data, USB for generic data transfer

    USB-C gets around this by adding a controller on the port, which can decide how to configure the port. And so it can pass power to power, video to video, etc. But this just leaves another attack avenue. All one needs to do is trick the controller and you can access anything on the machine.

    Even things that are supposed to be off limits, such as - PCIE Lanes, RAM, etc
    • Different ports (and corresponding cables) for different things certainly has advantages (many of which you have listed). The disadvantage is, of course, for smaller devices (phone, tablet, some laptops), the physical area of all of those ports turns into clutter. You can't put an ethernet or HDMI port on a phone because the size of the dimensions of the jack are larger than the thickness of the phone. That's why we had mini-HDMI ports and all kinds of other variations. Coming from that situation, USB-C

      • Re: (Score:2)

        by Shaitan ( 22585 )

        There is a whole world between a phone and datacenter. tablets, laptops, desktops, etc and for all of them space is far less of a premium.

        It is amazing all the things a phone can do these days but important to remember it still sucks at all of them except the one nobody does anymore, make phone calls. And it sucks at all of them for the same reason you mentioned, there is no room for a useful human interface on a phone. I don't think USB-C for everything is the future of the phone interface though, I think

        • Yes and for some of those form factors, regular Ethernet ports (for those who need them), HDMI ports, et cetera make a lot of sense. On the other hand, I find it particularly useful to sit my laptop on my desk, plug in a single USB-C cable attached to a hub and everything is now docked. That wouldn't happen with PS/2 keyboard/mouse interfaces. We used to have proprietary docking stations with proprietary connectors that were intentionally changed with every model to force you to buy a new one. This isn'

          • Re: (Score:2)

            by Shaitan ( 22585 )

            Agreed and that is consistent with your comment. Putting it back into the context of the overall conversation though it stops being a better alternative when all usb-c cables are not equal. Ethernet cables are at least labeled on the plastic sheathe and there are some distinction in the connector.

            We already have some of this problem with ports that are various usb specs... to those of us who understand technology we understand what is happening when a port isn't fast enough or can't deliver enough power but

    • Your point has been relevant for every standard since DMA was created. Unfortunately security should not get in the way of performance progress. If you are inclined to be doing the kind of work where a nefarious actor has the desire to physically access your machine, get a hardend machine.

      USB-C isn't anything new here. Your same post was made about the likes of the Firewire in the 90s.

  • "Not sure, but I think it is a USB C USB 4 Version 2 category A type 3 TB4 120 Spec PD PCIE DP Symmetric 3x3 PAM-3 AM 2.0"

    "No freaking idea"

    "Gotcha"

  • Doesn't that mean it's almost twice as fast as PCIe 5 x16?

  • I am glad we have now technology to understand the language of computer connectivity.

  • That should also be able to charge my car with 42KWh so that will be intersting.

    • Re: (Score:2)

      by Shaitan ( 22585 )

      It'll also replace the attached wire on your headphones. In the future all wires are USB.

      • Re: (Score:2)

        by drhamad ( 868567 )
        A wire? On headphones? What is this, 2015? But more to the point, yes, if you're saying analog wires are replaced by digital, I mean, USB headphones have existed for quite a while - not to mention Lightening. That idea isn't new.
  • I'll wait for USB 4 Version 2.0 Gen 2x4 Rev B to come out next year, thanks.
  • What we have here, is a failure to define HOW things communicate. USB-C is a connector type, nothing more, nothing less. There are standards for the protocols that will use USB-C, and those standards are why so many people are confused. This confusion has gotten to the point where articles show up talking about what the CONNECTOR can do when it is actually the cable between each end of the cable that is the most important thing to understand. Like Cat-5 and Cat-6 cable, it's just a cable, does nothin
  • So these fast data rates seem great, but I keep wondering what kinds of devices they are connecting to that could utilize that speed? I guess maybe a really fast SSD, but I can't imagine needing that kind of bandwidth for a "casual transfer". In server farms, maybe this would be a good idea, but you'd still have to have some kind of disc array that can handle all that speed. Not that I'm opposed to faster.... and of course when one thing in the chain gets faster, it forces the others to all catch up... bu

  • Make it simpler for end users

Slashdot Top Deals

A committee takes root and grows, it flowers, wilts and dies, scattering the seed from which other committees will bloom. -- Parkinson

Close