Hacker News new | past | comments | ask | show | jobs | submit login
USB 3.2 is going to make the current USB branding even worse (arstechnica.com)
249 points by nottorp 27 days ago | hide | past | web | favorite | 153 comments



USB-IF has to be doing this to deliberately confuse end users, right? I get that it's a specification number, not a product name, but they have to know that this will only cause problems for customers trying to find the appropriate cable.

Incredibly, their language usage specifications doc https://www.usb.org/sites/default/files/usb_3_2_language_pro... begins with this:

> USB-IF emphasizes the importance and value of consistent messaging on USB product packaging, marketing materials, and advertising. Inconsistent use of terminology creates confusion in the marketplace, can be misleading to consumers and potentially diminishes USB-IF’s trademark rights.

I simply don't understand USB-IF's motivation to make this so confusing for everyone. Their board consists of Apple, HP, Intel, Microsoft, TI, Renesas, and STMicroelectronics, so it isn't like it's controlled by low end trashy cable manufacturers trying to make a quick buck from confused customers.


If I were being cynical, I'd say their motivation is to get people to throw up their hands and say "I don't understand USB any more! Just give me the latest version of Thunderbolt. I know it supports the highest speeds and highest power, including displays and eGPUs, and all USB devices, and works on any new Mac."


It allows manufacturers (yes, many of the same ones on the USB-IF board) to "upgrade" their products by printing another label. USB-IF doing this repeatedly while being fully aware of the consequences can't have any other explanation.

This is just like the 4G-5G-5Ge debacle. Both being motivated solely by financial gain.


> This is just like the 4G-5G-5Ge debacle. Both being motivated solely by financial gain.

You are forgetting 4G LTE... ^__^;


The motivation is the same for re-branding 4G-LTE as 5G (or 6G). Sell the same thing with a new name so customers are tricked into thinking you're selling something newer and better based on the spec sheet without context.


The 5G thing is the carriers's doing, not the standards body.


It’s like they created Long Term Evolution (no more ‘G’enerations) then immediately realized the marketing issue.


So is there going to be a generation of young people who grow up thinking that all peripherals are USB, and that the naming system is so complicated/nonsensical that it isn’t even worth trying to figure it out?

When I was growing up, we had USB, FireWire, and a couple size/speed variations. It was easy to understand and you could tell what a cable was by looking at it.

For people who grew up in this era, the current situation is super annoying. Perhaps the young people of today will view peripheral standards as some sort of super-obscure language that isn’t intended to be decipherable by laymen?


Yes, and they will be forced to fall back on heuristics like "newer / more expensive / brand specific cables tend to fail less."

Muddying the waters is a diabolically genius way of attacking the generic cable manufacturers.


You kids had it so good. In my day, we had about 15 different peripheral busses that all ran over RS-232C or DB9 cables. You could fry a $10000 printer by plugging in a cable that fit perfectly on both ends.

And we felt damn lucky to have them.


So far all of their moves in regards to USB 3+ and type-c seem to have been to cater to various consumer companies.

If we put it in that context, then this new re-branding has to be done for the same reason - so that the same companies can claim to support a "newer" USB standard (and thus a reason for you to upgrade to the new devices), even though nothing has changed.


They seem to go out of their way to find perverse naming. Apart from 3.0/3.1/3.2 there’s the highspeed/superspeed nonsense. I can never keep any of it straight.


Off topic, but this is how I see your quote btw: https://i.imgur.com/dIFTp4Y.png


When quoting, please prefix lines with > instead of using preformatted text: preformatted text for blocks of prose is horribly unusable, everywhere.


Unfortunately, HN is so bad at publicizing and making this easy. I spent years assuming others knew some magical tool I didn't, before realizing people just do everything manually and stick to the same unwritten style guide.


...it felt pretty obvious to me? ">" is pretty commonly used, in email for instance.


agreed, I just thought I was missing some hidden feature. Like, I have to do the line breaks myself? I guess?


It’s even better to also italicize (using asterisks), e.g.:

> USB-IF emphasizes the importance and value of consistent messaging on USB product packaging, marketing materials, and advertising. Inconsistent use of terminology creates confusion in the marketplace, can be misleading to consumers and potentially diminishes USB-IF’s trademark rights.

Then at a glance there is a distinction between the comment itself vs. quotations.


Apologies, I read the site every day but post so rarely that I forgot that's the consensus way to block quote. Thanks!


Drives me nuts! For everyone on a phone:

> USB-IF emphasizes the importance and value of consistent messaging on USB product packaging, marketing materials, and advertising. Inconsistent use of terminology creates confusion in the marketplace, can be misleading to consumers and potentially diminishes USB-IF’s trademark rights.


Who the heck came up with USB 3.2 Gen 2×2? That has got to be the dumbest idea. No one cares how it works, we just have to go through the utter crap of buying a cable. Does the low end manufactures pull the most weight in this org because that's the only way this makes any sense.


All of this naming doesn't matter anyway, because everybody will use USB-C cables. And, as we all know, these can support (or not) almost any combination of anything (thunderbolt, video, various power combinations, USB in all its incarnations, etc) and there is no way to tell what will work by looking at the cable or socket you are plugging into.

In other words, it's a crapshoot anyway, so who cares? We'll just keep plugging until something sort-of works.

(do I sound crushed by the mind-boggling trans-galactic humongousness of the borderless expanse of human (and especially commitee) stupidity?)


>because everybody will use USB-C cables

That is assuming all USB-C Cable will support all power combinations ( They don't ) All USB 3.2 Speed ( They Don't ), and All Thunderbolt Spec ( They Don't ).

And it is precisely this comment that makes me thick why USB-C has won. Until the problem becomes so wide spread you have to test each cable whether it is working as intended before people will call for changes.


I don't quite understand why this naming is a big issue. USB is now so complicated you have to look at the data sheet if you want to know what you PC or whatever provides. It's not just the speed. It's also how much power can flow, in what directions and whether it does that when the power is off. It's not just USB, but whether it supports pin reassignment to DisplayPort, Thunderbolt and what not. If you care you are going to have to look it up.

The point is you will be able to look it up. With cables you can't. I wish it was mandatory to publish the rated max speed, amps and max number of insertions on every USB-C cable. It's only 3 numbers - but there is no chance of it happening.


For now I am refusing to buy anything that need a USB-C cable


To clarify: At the moment I have no use for the added functionality of USB-C, I understand why it is useful to many. For my needs USB-C only means compatibility problems.


Thunderbolt compatible cables are their own pain in the butt. Apple loves their power only, and I had to order from Monoprice Thunderbolt 3 40Gbps 100W cable, and not get suckered with a 60W or a 20Gbps cable. To save others the trouble, here is the link https://www.monoprice.com/product?p_id=24722 because that seems like the easiest if you have to have a single cable type around.

To think, memory used to be the number soup......


> Who the heck came up with USB 3.2 Gen 2×2? That has got to be the dumbest idea. No one cares how it works

Well, you will if you end up trying to use a part which supports 2×1 with a part which supports 1×2, since they both support the same top speed, but together they only support the lower speed of 1×1.


I cannot tell if you are joking, and that scares me worse than if you are completely serious.


I'm mostly serious, though I really don't know why 1×2 exists and why you'd build a part that supports it but not 2×2 (or, alternatively, that supports it instead of 2×1, which is, but for a name change, an older standard with the same nominal bandwidth), and the 2×2 will also support 2×1.

But maybe there is an electrical engineering reason why it might be practical to do 1×2 but not 2×1 in some device.


Probably it’s for cables, in that a 1x2 cable can (I assume) be longer for cheaper than a 2x1 cable.


1x2 requires lower frequency electrical signals I suspect? That might be preferable in some scenarios. Still insane.


Just wait, Usb 3.2 1x1x1x1 is just around the corner for high power draw devices!


Yes lower frequency. There is a big jump in implementing a 10G PHY vs a 5G PHY


Why not enumerate it based entirely on features, also toss in what power delivery it's rated at in watts and volts.

USB 3 20Gbit PD 100W 20V


USBFG-9000 4K 6G Deluxe


They all support the same voltage, and there are only two meaningful power delivery levels.

So they only have to put the speed class and PD or not PD. It would make things so much better.


What's "3" for?


socket format?


you've got A, B, C, Micro, and Mini connectors all that conform to 3.0 and 3.1 spec...

Alas the standard needs standardised....


And A is always the same socket regardless of version while B is not. It's a mess.


> And A is always the same socket regardless of version

Even this is technically incorrect: USB-3 A-type sockets, though backwards compatible for f/m and m/f connecting with USB-2, have 5 more contacts compared to USB-2 type A sockets.


Returning to this topic, maybe...

USB-C-3 for the new connector format.

USB-C-2 for some (many) cellphones.

USB-Micro-2 (for many phones).

USB-Mini-2 (older phones)

- Edit -

I changed my mind, the connector is more important than the wire version.


> Alas the standard needs standardised....

https://www.xkcd.com/927/


Totally agree. I’d much rather have. Simple branding such as USB A5, USB C10, USB C20 to indicate the maximum speed and connector possible for each iteration.


It makes more sense for high end manufacturers, who want to sell expensive high end cables and make it hard for people to find and buy cheap compatible cables from anyone else.


Actually, I think its the other way. The confusing branding helps any low-end manufacture to do the least amount (cheapest) work to get something that looks like a high-end designation.


I don't think confusion can help them here. It helps to steer consumers towards less scary choice though, i.e. to the brand they trust.

Remember how in the Nokia days manufacturers of mobile phones all made incompatible chargers and data cables. Eventually they were hit with regulations. I think this is similar, but more advanced plot against consumers with plausible deniability.


USB 3.X / USB-C / Thunderbolt 3, all of it together is confusing enough for those technical, add marketing speak to the mix and it's just going to end up with a mish mosh of different words that mean absolutely nothing and tell the end user very little.

Add to the fact that everyone is skimping on the actual USB 3 standards and a large amount of shitty products on the market that don't do what they say they do (due to poor quality control, making wires longer than spec expecting them to work to spec), we have to rely on guys like Benson to crowdsource working and non working products.

https://bensonapprovedcom.wordpress.com/


For real. Why not just call it USB 4.0 and divorce the branding from the technical specification? You can put a thing on the back of the box that says "implements USB specification 3.2.16a-rfc#4619" or whatever.


> Why not just call it USB 4.0 and divorce the branding from the technical specification?

That is exactly what the "SuperSpeed USB" branding is all about. The technical spec is USB 3.whatever and the brand name is " SuperSpeed USB". But neither the press nor vendors seem to be able to stay in the line and actually consistently use only the branding, so here we are


But SuperSpeed was 5Gbps. Then it was SuperSpeed+.

This was the nice aspect of Firewire - once 1394B was out - everyone referred to them as Firewire 400 and Firewire 800. No ambiguity.

Heck even Thunderbolt does this better - 1 and 2 are interchangeable device/cable-wise, and just support higher speeds, but it's just a new 'version' and it just doubles the bandwidth.

With USB you have to read all the fine print to make sure you get what you want/need.


SuperSpeed+ was never supposed to be exposed to consumers:

> NOTE: SuperSpeed Plus, Enhanced SuperSpeed and SuperSpeed+ are defined in the USB specifications however these terms are not intended to be used in product names, messaging, packaging or any other consumer-facing content.


Oh great so it's all just meant to be "SuperSpeed".. "How super?" "Guess".


> USB-IF’s recommended nomenclature for consumers is “SuperSpeed USB” for 5Gbps products, “SuperSpeed USB 10Gbps” for 10Gbps products and “SuperSpeed USB 20Gbps” for 20Gbps products

Yes, guess how super “SuperSpeed USB 20Gbps” is


But the lowest speed doesn't specify the speed - so unless a user happens to have read the USB-IF recommendations for naming, they won't know that there is a difference, if the manufacturers even follow said recommendations


Maybe be USB4 is already taken/planned?


USB 4.0 will actually be USB 1.0 with a new name, same speed.

4.1 will be what was USB 1.1

4.2 will be what was USB 2.0

4.3.0 will be what was 3.0.

4.3.1GXY1 will be what was 3.2 1x1 (which was 3.1 Gen 1 (which was 3.0)

4.3.1GXY2 will be what was 3.2 2x1 (which was 3.1)

4.3.2GXY1 will be what was 3.2 2x2

4.4.0 will be double the speed of 4.3.2GXY1 (what was called 3.2 2x2) but will max out at 25mm (~1") cables and provide no power at all.

4.4.1 will be like 4.4.0 but will expand the length to 30cm (~1')

4.4.1ZORB22 will be like 4.4.1 but also carry up to 5W power for bus powered devices.

Go ahead and laugh, and then realise this isn't really that different than what we have now from USB-IF.


This makes just as much sense as the current situation.


Here is a helpful chart of USB versions and their connectors: https://preview.redd.it/s9dun6a1xmpz.png?width=960&crop=smar...

e.g. USB 3.2 has the same connector as USB-C

edit: I personally care far more about the connector than exact specs of each version


This is unhelpful. The whole point of the article is that both of those middle columns are also USB 3.2, i.e. you have this:

* USB 3.2 Gen 1 5 Gbps (formerly USB 3.1 Gen 1) (formerly USB 3.0)

* USB 3.2 Gen 2 10 Gbps (formerly USB 3.1 Gen 2)

* USB 3.2 Gen 2x2 20 Gbps

Literally everyone that came up with this ridiculous naming scheme and literally everyone who approved it should never be allowed to participate in a working group ever again. This level of incompetency is just absurd. These people cannot be allowed to pollute critical standards with their terrible ideas.


The specification is USB3.2. Gen1, Gen2, Gen2x2 are modes. The USB3.2 spec is more than just the speeds. It can include protocol updates and errata.


End users don’t care about protocol updates and errata.


> USB 3.2 has the same connector as USB-C

USB-C is a connector, so you cannot say "has the same connector as USB-C".

It is the only connector for USB 3.2, and a connector that is compatible with USB 3.1 Gen 2.

Edit: I meant it is the only connector for USB 3.2 "Gen 2x2", and compatible with 3.1 Gen 2 (now called 3.2 Gen 2)


> It is the only connector for USB 3.2, and a connector that is compatible with USB 3.1 Gen 2.

From what I've been able to find, not even that is reliably true. Looks like the older connectors can be labeled USB 3.2 as long as they use the "x1" scheme in the label and are capable of reaching the speed of that scheme.


There is no USB 3.1 Gen 2. There is only USB 3.2 Gen 2. ;D


> edit: I personally care far more about the connector than exact specs of each version

hopefully usb-c connectors are tougher then micro-usb's


In my experience, the device side is much tougher. I've had 2 cables fail but even after about 3000 operations (that's about estimated 5 operations per day for 2 years in my case), the device connector still feels new with a fresh cable.


As it should be. I’d hate if the device side wore out


People confused with the branding don't realize that USB is by manufacturers for manufacturers, so anything that lets them claim on the box they support USB 420.69 while doing the least amount of work and spend is a-OK.

This is what happens when standards bodies have perverse incentives to confuse. See the SD Association for even worse branding.


This isn't entirely true. The point of the manufacturers together agreeing on a standard is to remove some of these perverse incentives; here's how it works in theory:

Without the standards group:

If manufacturer A is confusing, then they can sell more than manufacturers B and C. So in turn manufacturers B and C will be confusing as well in order to keep up. This is bad for consumers and generates ill will against USB in general (globally bad for A B and C)

With the standards group:

Manufacturers A B & C all agree on something non-confusing, and in order to use the USB trademark, they can't deviate from what they agreed upon. USB is now sunshine and rainbows so USB customers are happy in general (globally good for A B & C).

Something has broken down in the system if they make names that are confusing since USB being confusing is bad for everybody, and none of the manufacturers get a leg up on the others if they all use the same confusing language.


I really loved USB-C then they screwed it up so badly. Can we have 1 cable that is required to have every exact feature available, no exceptions? Works on any device, supports everything from storage to video transfer (Monitors) and whatever else Apple would add to it.


Apple sells one:

https://www.apple.com/shop/product/MQ4H2AM/A/thunderbolt-3-u...

It does Thunderbolt 3, USB-C 3.1 Gen 2, and up to 100W of power delivery. It's a shame it's only available in one kinda-short length, but I think that's a technical limitation (any longer would require an active cable for Thunderbolt).


Curious how compatible it is with other non-Apple devices? The worst thing I've heard of is phones getting burnt out by overpowered USB-C chargers which I think is crazy and why the spec should enforce one and only one type of cable. I've been buying USB-C cables from Anker and from Belkyn that I've seen at Walmart cause they're brands I trust not to screw things up in a ridiculous way, thankfully my phones have survived both so far. I rather pay a little more if it means my phone wont get damaged.


Look at Thunderbolt, where a cheap-ass cable still costs $20, for an idea of how that would turn out.


And USB-C 10 Gbps cables aren't much cheaper, especially if they do 100W power delivery. Both probably could be cheaper if there were more demand for them, but siren song of cheap slow cables seems to have made it impossible to scale production up.


Well, realistically, you don't need 100 watt power delivery for a keyboard, so you might want a less robust cable.

The big mistake IMO was a lack of obvious keying. Very little stops me from using that 2w-capacity keyboard cable to try to charge my laptop. If we're lucky, there's signaling at the device to tell me what's wrong, but if they had made a different plug on the 100W cable, it would eliminate a lot of cockpit errors.

Having one cable for everything is a huge red herring. If you need a keyboard and a power brick and an external hard drive, you need three cables. It doesn't matter if they're different, as long as the OEM isn't pulling some Apple-style "one single port is enough" BS.


So you'd want a low speed connector, a high speed connector, a high wattage connector and also a high speed + high wattage connector? So 4 plugs, or more if you want to add even faster speeds in the future. And then you need cables which have one of each of those connectors on each end, so we're talking about dozens of variants.

No thank you. The drama around USB-C is insanely overblown; the vast majority of use cases are fine with the 60 watt + 480 mbps required minimum configuration for USB-C cables. In the few rare exceptions where you need >60 watts or higher speed, you're generally going to have purpose-built hardware anyway, so just devoting a special cable to that isn't such a big deal.


I think the drama is mostly about the fact that there is no way to tell what a cable (or a port) "supports" by looking at it.


> Well, realistically, you don't need 100 watt power delivery for a keyboard, so you might want a less robust cable.

But on the other hand doubling the power wire thickness and leaving the signals wires untouched shouldn't add more than 10-20% to the price of the cable.

> The big mistake IMO was a lack of obvious keying. Very little stops me from using that 2w-capacity keyboard cable to try to charge my laptop. If we're lucky, there's signaling at the device to tell me what's wrong, but if they had made a different plug on the 100W cable, it would eliminate a lot of cockpit errors.

Every cable supports at least 60 watts. It's not a big deal if you use the wrong one to charge with.


That's a licensing issue/cost.


Is it? They’re complicated active cables.


true. and, with a licensing fee ;)


Regular USB A/B had the same problem. I have been frustrated many times plugging in a micro B cable only to find that it carried power but not data


I don't think there was ever a USB spec that allowed you to skip the data lines. That is purely a chinese cost optimization. Meanwhile, the confusion with USB 3 (C?) is designed in.


You say cost optimization, I say safety feature.


That "safety feature" prevents the client device from negotiating a higher current limit than 100mA. The host controller may generously provide more, but there's no guarantee of this without the data lines.


The host is required to not be damaged by a short circuit on the power lines.

That means a device can ignore the signalling entirely and just keep taking more power till the supply turns off, and then scale back 10 percent.

More and more Android phones do that, and it works very well.

Only disadvantage is if you plug it into a hub, it can cause the whole hub to go out for a bit. Devices solve that by not using the above algorithm when they detect any active signalling.


I remember the brain bliss of seeing the first usb-c article. Funny how the world goes


> For reasons that remain hard to understand, the decision was made to retroactively rebrand USB 3.0: 5Gb/s 3.0 connections became "USB 3.1 Gen 1," ...

> USB 3.2 doubles down on this confusion. 5Gb/s devices are now "USB 3.2 Gen 1."

So when they eventually release USB 3.3, my rusty old memory stick will be auto-upgraded to USB 3.3 Gen 1? I like it!

(/s)


Honestly, the USB-IF guidance is pretty sane here. The "USB 3.0" or "USB 3.1 gen 1" or whatever simply should not appear anywhere in the consumer space. Everything should be labeled and using "SuperSpeed USB X Gbps".

I'd go as far to say that articles like this exactly help keep the confusion up because the way they keep bringing up the term "USB 3.2" etc as something that consumers should care about, when it is something that should be avoided.

> What this branding meant is that many manufacturers say that a device supports "USB 3.1" even if it's only a "USB 3.1 Gen 1" device running at 5Gb/s. Meanwhile, other manufacturers do the sensible thing: they use "USB 3.0" to denote 5Gb/s devices and reserve "USB 3.1" for 10Gb/s parts.

Both manufacturers are doing the wrong thing: referring to the USB standard version at all. Devices should not say that they support "USB 3.0" or "USB 3.1 gen 1". They should say they support "Superspeed USB" or "Superspeed USB 10 Gbps".

Superspeed might be a silly name, but the branding there is almost perfectly unambiguous, and nicely explicit about the data rate in a way that "USB 3.2 gen 2x2" is not.

What is the most surprising is that USB-IF isn't able to keep the vendors in line despite being industry consortium and owning all the relevant trademarks and other IP. They really should start enforcing their policies more aggressively.


That's really no excuse for USB-IF's terrible versioning scheme.

"SuperSpeed" is useless in terms of branding because even technical users, much less your average consumer, have no idea what that means. Is "SuperSpeed" faster than "High Speed" or "Full Speed"? (Answer: "Full Speed" < "High Speed" < "SuperSpeed".)

"20 Gbps USB" is better than "USB 3.2 Gen 2x2", sure, but that's not exactly a high bar. It just means USB has failed so badly with version numbers that people would rather just quote a spec figure.

Wishing for everyone to switch over to "SuperSpeed" and "SuperSpeed+" will not make it so. Fact is, both users and manufacturers are accustomed to "USB 3.0", "USB 3.1", etc. And it's not hard to see why people prefer simple version numbers as opposed to meaningless labels.

USB-IF should have simply released USB 3.1 and USB 3.2 without this ridiculous business of retroactively renaming older versions.


The USB versioning makes perfect sense when you think it as the version of the standard, literally the document itself, which encompasses all the modes. And it does make sense to have single standard instead of having the information scattered across dozen historical standards that have been partially superceded.

I would point out that this is not a new thing. The situation was exactly the same with USB 2.0, technically it was perfectly correct to call compliant full speed device "USB 2.0", because 2.0 did indeed include full speed mode (as does presumably all the USB 3.x standards)


I just want to add that this is something anyone can help with to make the situation clearer. Just drop "USB 3" in all it's different forms from your verbiage, and start using "SuperSpeed USB" instead. I bet if everyone on HN would switch then that would already make a difference in the tech community and eventually trickle to the general populace.


Is it USB 2 or USB 3? Is it 3.0, 3.1, or 3.2? Gen 1, Gen 2, or Gen 2x2? Do I need USB-A or Micro USB or USB-C? Do I need USB-A on one end of this cable and USB-C on the other, or some other combination? Does it support Quick Charge or Power Delivery? If it's Power Delivery, how many Watts?

Those are just some of the questions non-technical consumers are expected to know the answers to when they are buying USB products.

This is absurd.


Lightning, although proprietary, is an amazing port compared to the mess that is USB. I am going to miss it once Apple phases it out.


It's easy to make a universal port when your universe is a small handful of your own devices with no numbers to live up to.


The "small handful" adds up to about a billion devices, but it's true that they were made by a single manufacturer.


A billion devices, but just 24 different ones, over 3 product lines, plus a couple of trivial, slow ones like keyboards.


Something has to be done to stop these people.

What, I don't know. But their Joker-like approach to USB standards and practices is costing us all dearly.


I just had a thought USB is now more of a clusterf^H^H^H^H^H^H^H complicated than the mess of cables interfaces it was meant to replace back in the mid 90's.


It's amazing how much that mess of cables either wasn't solved or wasn't solved by USB. My power, network, sound and display cables weren't replaced by USB, my printer was replaced by a phone/tablet. It replaced the PS2 ports for keyboard and mouse, saving about 2 seconds of setup time at the cost of more latency.

It his (or has been) handy for connecting gadgets and drives over the years, but maybe we should go back to PS2 and just use it USB for "front of the machine" ports.


Counterpoint: A mix of USB-C, Bluetooth and Wifi have replaced my power, network, sound, microphone, printer, keyboard, mouse, phone charging, and display cables.


What has to be done is consumers need to refuse to buy components with non-standard USBs.


How will that help? The incompetence of the standards body is the problem at hand.


Shameless plug: if you want to become less confused about USB-C my primer is at https://superuser.com/a/1200112/41259 and also https://www.reddit.com/r/thinkpad/wiki/newdocks contains some helpful intro.


So they decided to put some maths behind the new Spec to simplifies the branding.

USB 3.2 Gen 1 = (3 + 2) x Gen 1 = 5Gbps

USB 3.2 Gen 2 = (3 + 2) x Gen 2 = 10Gbps

USB 3.2 Gen 2x2 = (3 + 2) x Gen 2 x 2 = 20Gbps

Pretty Clever ( Or Dumb ).

To make matter worst they still have not make USB-C mandatory for USB 3.2. So you could have a USB 3.2 Gen 2 Host with a USB-A Cable that didn't not support the previous USB 3.0 SuperSpeed ( 5Gbps ) so it will fall back to USB 2.0 Speed ( 480Mbps ) , Or a USB-A Cable that did not support the USB 3.1 SuperSeed+ ( 10Gbps ) which is also called USB 3.1 Gen 2 and is now called USB 3.2 Gen 2 so it will fall back to USB 3.0 SuperSpeed ( 5Gbps ) and was called USB 3.1 Gen 1 and now called USB 3.2 Gen 1.

You also have crappy USB-C cables that does not confirm to the USB-C spec. And the consequences depends on the quality of the cable. [1] And in some extreme cases, you could have a USB-C cable that do not support USB 3.2 Gen 2 x 2, because some how only pins on one side are working as intended.

Although things has gotten better since his post and Google has been putting some pressure into fixing it. USB-C is still a bloody bag of hurt. So in case anyone is still rooting for iPhone using USB-C without MFi. I seriously hope people will reconsider. Using MFi will defeat the purpose of USB-C since you will need a MFi Cable, which is one way to say current lightning to USB-C cable spec and cable are fine. And in case you are wondering about iPad Pro, you should take a look at the controller size, it is roughly 5 times bigger.

Or we make USB 4.0 that standardise on USB-C and force a much more stringent specification on Power Delivery, Speed and Quality. But as we all should be able to tell this isn't a technical problem, more like political, design by committees, marketing problem.

[1] https://bensonapprovedcom.wordpress.com


At least WiFi figured out how to improve.

Would love to see "USB 4" .

No ".0" afterwards, just the simpler numeral.

And if you need a bigger connector, then buy the "USB 4 XL connector" cable.


"Would love to see "USB 4" ."

I'd rather go for a (possibly faster) Ethernet with a redesigned smaller connector supporting POE. Smaller devices such as mice and keyboards could talk at PHY level while smarter ones with beefier microcontrollers could employ more and more layers of the network stack so that routing/tunneling data over the network when needed would become trivial. Security wise, there would be no physical difference between an external local disk and a NAS, but software could be used to filter devices from the network. The upside is that there would be a single stack for everything: from storage to security cameras, audio systems, data acquisition gear, sensors and actuators etc. All ports would be electrically insulated, the stack is already open and wonderfully documented, the hardware is already near realtime and there are no royalties to be paid to use that.


That was one of the original goals of Zeroconf. In 2002, Stuart Cheshire said [1]:

> My hope is that in the future — distant future perhaps — your computer will only need one wired communication technology. It will provide power on the connector like USB and FireWire, so it can power small peripheral devices. It will use IP packets like Ethernet, so it provides your wide-area communications for things like email and Web browsing, but it will also use Zeroconf IP so that connecting local devices is as easy as USB or FireWire is today. People ask me if I'm seriously suggesting that your keyboard and mouse should use the same connector as your Internet connection, and I am. There's no fundamental reason why a 10Mb/s Ethernet chip costs more than a USB chip. The problem is not cost, it is lack of power on the Ethernet connector, and (until now) lack of autoconfiguration to make it work. I would much rather have a computer with a row of identical universal IP communications ports, where I can connect anything I want to any port, instead of today's situation where the computer has a row of different sockets, each dedicated to its own specialized function.

[1]: http://www.stuartcheshire.org/TheIdeaBasket.html

Sadly, instead of using existing successful networking systems and automating the configuration, we've created a new union-of-all-possible-protocols with a new connector (or 5) and made understanding the compatibility matrix a nightmare.


I didn't know about that interview, thanks for sharing. It seems so natural to me to go for the one technology that fits all needs (Ethernet+POE and possibly a smaller connector). All it would need is a big player to implement that, then release the changes for free with no strings attached, but probably most parties involved can't squeeze the same profits out of open technologies.

Here's an idea to push for this technology. I have no way to develop this, neither the technical knowledge nor the money required, but comments welcome anyway. Essentially I would:

  1- design a smaller connector that could host all Ethernet pairs plus power supply,
say a smaller RJ45 with contacts on both sides, no need to reinvent the wheel here.

  2- for testing purposes, build a small bridge board (a matchbox sized black box) between the socket and the
 original Ethernet plug and a nearby USB port for power alone. Ethernet pairs would pass untouched (or possibly
 replicated through a switch chipset+magnetics) while the bridge board would contain any necessary circuit to power
any load connected to the updated Ethernet port(s). Remember we're aiming at ditching USB for that purpose, that's why I wouldn't map Ethernet on an USB-Eth chip even for testing; and all power would be taken by the device supply asap.

  3- Now we have some cheap hardware which is easy to replicate to use on every platform, including SBCs where the 
USB connector could be spared because we take all power from the board. Let's move to software.

Some proof of concept would be needed to demonstrate that, so here's my question: how hard would be to satisfy point [1] and build some hardware (mice, keyboards, audio etc) that can use the new network standard to show its benefits to the industry?


Aah, that's a beautiful dream of a unified, wired communication technology standard.

That's what we need more of, a coherent vision to guide design and development. Early Apple had that, it feels like, then lost the coherence in the 2010's. The phrase "design by committee" is often used to characterize what not to do, so it seems design by a small focused group or individual tends to go in the right direction. But there must be real-world examples of successful "design committees" or "standards bodies", who avoided the gravity/entropy towards producing monsters of complexity and compromise..


I was speaking branding, not tech stack. Sorry if that caused any confusion.


I had the pleasure of working with a girl who referred to 'ethernet cables' as 'server cables' as that was the functionality provided by ethernet, the wifi on her Macbook meant the internet was always on but lacked 'server access', hence the name of the 'ethernet cable' as a 'server cable'.

In this spirit I think that 'USB 4' really needs to be 'USB 4G'. Then you can have phones and the connectors use the same number. My former workmate would understand it. A '4G' cable for a '4G' mobile phone signal, that would be simple and the least techie people could understand it whilst people on HN faceplant in perpetuity.

USB '5G' would be along with '5G' phones. Everyone could be happy and cables could be fairly random, just try a selection until you get one that works, the clue being that cheaper ones for power adaptors don't let you connect your '5G port' on your computer to your screen or move data to your phone, much like mini USB now. Who reads labels anyway?


What about devices that don't have interet access ?

This seems like a very bad idea, the two technologies don't move at the same speed and are totally unrelated. I think it would be even more confusing for consumers.


I think you missed the sarcasm in the post you're responding to.


"SuperSpeed USB" is your "USB 4".


But what do they name it when they upgrade Superspeed?

Ethernet at least had 10Mb, 100Mb 1000Mb/Gig, etc.


They already did upgrade SuperSpeed. Twice.

> USB-IF’s recommended nomenclature for consumers is “SuperSpeed USB” for 5Gbps products, “SuperSpeed USB 10Gbps” for 10Gbps products and “SuperSpeed USB 20Gbps” for 20Gbps products


Don't know if it's my dyslexic brain, but did anyone else have a hard time reading that article due to trying to keep sense of the different standards names?


It's not just you — it's virtually everyone who will be purchasing a "USB 3.2" cable in the future!


Experienced EE here.

Yes.


I feel USB workgroup is having its Winamp moment. Skip USB4 and bring us version 5: as lean as the v2 but with the good bits of v3.


Wouldn't make sense to just use the speed as version ? i.e USB 5GB, USB 10GB-C, USB 20GB etc


They do that, in the "consumer branding". This is an odd case where the marketing is better-thought-out than the spec.

You have to live with calling it a "SuperSpeed" cable instead of a USB cable, though. :p


Where the USB-IF messed up - they should have used their trademark to standardize these cables.

They should have visibly differentiated charging and data cables, and required all cables with USB connectors at both ends to be rated 100W, and this new 20 gbps (2x2) speed if also supporting data.

Cables with USB-C and another USB port at the end should be required to meet both the power and data requirements of that connector. (so no charging compatibility connectors). USB-C to USB-A should have been restricted to 3.0 only.

Alt modes can use a plain USB-C cable (like Thunderbolt 3 does), but cannot place additional restrictions on the cable if they want USB connectors on both ends.

If the cable has any other connector on the end (such as DisplayPort), it would be up to the creator of the DisplayPort alt mode to indicate if there are any differing requirements.

Likewise if the cable is integrated into a device, it can meet its needs - be charging only, only use USB 2,etc.

This imho would solve nearly all the problems - the problem is that I can pick up a USB-C to USB-C cable and it could literally be one of (at least) six things internally - with no markings to differentiate.


BTW, here is the actual guidance/press release from USB-IF

https://www.usb.org/sites/default/files/usb_3_2_language_pro...

To me it seems the "press" is distorting this press release.

Excerpt :

USB 3.2 Gen 1

o Product capability: product signals at 5Gbps

o Marketing name: SuperSpeed USB

USB 3.2 Gen 2

o Product capability: product signals at 10Gbps

o Marketing name: SuperSpeed USB 10Gbps

USB 3.2 Gen 2x2 o Product capability: product signals at 20Gbps

o Marketing name: SuperSpeed USB 20Gbps

USB 3.2 Key Messages

• Defines multi-lane operation for new USB 3.2 hosts and devices, allowing for up to two lanes of 10Gbps operation to realize a 20Gbps data transfer rate, without sacrificing cable length.

• Delivers compelling performance boosts to meet requirements for demanding USB storage, display, and docking applications.

• Enables end-users to move content across devices quickly, conveniently and without worrying about compatibility.

• Backwards compatible with all existing USB products; will operate at lowest common speed capability.


Why don't they simply change usb-c cable identification into a couple numbers, a bit like trousers?

One for power, one for data. It could be any scale, as long as products like monitors would need data > x, phones will charge faster up to power y, laptops need power z, etc. Two numbers are easy enough to emboss, engrave, print into a connector, cable or device...


> For reasons that remain hard to understand, the decision was made to retroactively rebrand USB 3.0: 5Gb/s 3.0 connections became "USB 3.1 Gen 1," with the 10Gb/s connections being "USB 3.1 Gen 2." The consumer branding is "SuperSpeed USB 10Gbps."

Does anyone care about Gbps for USB? I've never hit the bandwith limitation on USB and see USB type-C as just an annoyance making me buy new cables.

> And 20Gb/s devices will be... "USB 3.2 Gen 2×2.

Pretty straight forward: USB 3.x supports 5 * 2^x Gbps. What we really need is MORE retrobranding!

- USB 2.0 becomes USB 3-3.4

- USB 1.0 full speed becomes USB 3-8.7

- USB low speed becomes USB 3-11.7


> Does anyone care about Gbps for USB? I've never hit the bandwith limitation on USB

You have if you've tried to used multiple external monitors over USB (a la macbooks).


HDMI over USB sound like hardware abuse to me. All systems I've only ever owned had write_mostly interfaces for video display.

Really, USB speed limits are mostly bound by flash memory limits so manufacturers struggle to justify a replacement. I guess faster charging is nice?


> Does anyone care about Gbps for USB? I've never hit the bandwith limitation on USB and see USB type-C as just an annoyance making me buy new cables.

You haven't met new Windows (Dell, etc.) laptops yet? If you want multiple monitors with a single connector (there's often just one), it makes all the difference whether you can connect two 4k monitors @60 Hz or not.


But that's a capability on top of USB that isn't even implied by the Gbps branding.


Ah right, it's Thunderbolt. Dell is doing some weird things with it, like apparently transporting way more power than the standard allows. Shrug. Dell TB16 (thunderbolt) dock is not the nicest experience.


> Does anyone care about Gbps for USB? I've never hit the bandwith limitation on USB and see USB type-C as just an annoyance making me buy new cables.

External SSDs are probably the only time it matters, or even multiple drives connected through a USB hub.


You don't need an SSD to get benefit from the transfer speeds of USB3. Decent hard drives support linear read speeds of ~160MB/sec. USB2 will throttle that to 40MB/sec after protocol overheads.

I've also run into this using external ethernet adapters, which I still want when I'm copying large files or streaming video via airplay or steam. My last 3 laptops were missing inbuilt ethernet ports. USB 2, while much faster than the wifi speeds available at the time, only supports half the bandwidth of gigabit ethernet.


Do you need to go from 10 Gbps USB3 up to 20 for a hard disk?


So the very same device can be accurately labelled "USB 3.0", "USB 3.1 Gen 1" and "USB 3.2 Gen 1". Those are all actually the same thing? Do I have this right?


No, it shouldn't be labeled any of those. Instead it should be labeled "SuperSpeed USB" and nothing else. And that hasn't changed over the different USB revisions.


Except the list of speeds is ridiculously arcane. People can easily figure out that 3.1 beats 3 beats 2. Good luck getting them to guess whether "full speed" beats "high speed", or why there are three different things all called "Superspeed+"


There is not supposed to be any superspeed+ branding. From USB-IF guidelines:

> NOTE: SuperSpeed Plus, Enhanced SuperSpeed and SuperSpeed+ are defined in the USB specifications however these terms are not intended to be used in product names, messaging, packaging or any other consumer-facing content.

> USB-IF’s recommended nomenclature for consumers is “SuperSpeed USB” for 5Gbps products, “SuperSpeed USB 10Gbps” for 10Gbps products and “SuperSpeed USB 20Gbps” for 20Gbps products

Those are hardly "ridiculously arcane". Full speed is not really relevant, so only thing consumers need to know is that superspeed > high speed, which I don't think as very high bar.


Did they delete the 1x2 mode? If not, the answer to "Does this cable support SuperSpeed USB 10Gbps?" becomes significantly more complicated than it should be.

But fine, if we actually see everything branded as "SuperSpeed USB XXGbps" then it's overly long but it's understandable.

I am not convinced we won't see a hideous hodgepodge of different branding guides, though. Some saying "SuperSpeed USB __Gbps", some saying "USB 3.2", some saying "gen __", some saying "SuperSpeed+"


Oh, I see. I guess that's... clearer.


USB with and without thunderbolt was already confusing enough for your normal joe. Now you have 3 version of USB 3.2 with require you to do a detail specsheet lookup before plugging in?


They should just rebrand it as SB and be done with it, there's nothing universal about it anymore.


Probably the same person who made decisions for Java branding back in Sun Micro days.


One of the really bad branding issues / borderline scams I have dealt with is devices saying USB 3 compatible when what they mean is usb 2 but it happens to work with usb 3 ports using usb 2.


* USB 3.2 Gen 2 10 Gbps * USB 3.2 Gen 2x2 20 Gbps

Therefore, If a motherboard has two USB 3.2 Gen 2 10 Gbps port, can it be labeled as USB 3.2 Gen 2x2?


It's like they're marketing pain relief, all new formula, super rapid fast action, targets pain directly....


Ok, guess I'll go back to RS-485


And then there's USB-C, which is an attempt to cram all the ports on the back of a desktop into one tiny connector. It has USB! It has HDMI! It sometimes has DisplayPort! Maybe audio, too! It has power! In both directions, even! Sometimes it even has JTAG, but we don't talk about that.

How does all this interact with UCB 3.2?


Good luck finding a cable that supports what you want. It might only be a matter of time before we have to have chips in the cable to indicate what it can and cannot do!


You thought USB-C was going to solve all your problems?

But they need a new reason to sell you more cables, dongles, and adapters.

Maybe they can call the next one USB 5G! That not only helps confuse the branding even more, but . . . It worked for AT&T.




Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: