USB-IF is made up of manufacturers. The primary goal is lower BOM. The confusion is the point. Making everything optional is the point.
USB-IF is not going to adopt a standard that requires every manufacturer to retool or redesign their Widget in order to slap a USB 4 label on it. Why would they vote for that? The whole 2x2/Gen2/etc thing was so some of the members could do some minor revisions then keep shipping their existing chips and products without them being considered "old", with a bit of "design by committee" mixed in.
USB-IF doesn't give two shits what you or I or anyone else thinks about it, nor what our actual experiences with it are. They care about how much it costs. And by cost I mean an $0.01 difference per unit is considered a significant savings.
Yep. If you start looking at it from the point of view of who benefits and who makes the standards, it makes a lot more sense.
The physical connector is terrible and ends up being the first thing to break on a lot of electronics. The labelling is confusing and no one knows what to buy, so you always end up buying more than you need. Motherboard manufacturers were happy to re-label their marketing materials on old boards to say USB 3.1 so they could sell them to unsuspecting consumers that were tricked into thinking 3.1 > 3.0. It goes on and on.
Look at the big picture and it starts to feel like the goal is to create a bad design because it means the manufacturers can sell us more junk.
The USB consortium already botched that badly with mini-A, so learned from that in the design of micro-A and then C. C was also influenced heavily by Apple's experience with lightning (Apple is a core member of the USB Consortium).
The physical part of type C is quite good: the nice grounding shell (preferable IMHO to lightning) high frequency design, and most importantly (per your comment) failure is mostly directed to the cable, not the jack (one of mini-A's problems) since a bad cable can be replaced while a damaged jack these days means replacing the device.
AFAIK the Essential phone was brutal for jack failures. My friend (repair shop owner) gave me an LG V60 this year and basically said it’ll be great until the USB-C port wears out.
My newest iPad with USB-C seems like the jack is already flaky while old iPads with lightning ports are all good even with the abuse they have to endure from kids they were handed down to.
I have so little faith in USB-C ports that I’ve started trying to buy magnetic adapters for all my stuff.
People break the stupid wafer trying to clean them.
They suck IMO.
Sadly it's patented so no one else could implement in their laptop.
Then, we guarantee their safe release given one condition: Anyone whose grandma mails us a fully compliant USB cable gets to go free.
I guarantee the standards body will come to their senses shortly.
But since this is HN:
I'd die in captivity if this was me. Others would get out right away.
From the "Falsehoods Programmers Believe About Grandparents"[1]:
- People have grandmothers
[...]
- Grandmothers aren't technically competent
[1]: I made this up here and now.
Wireless charging or dedicated cables are just less painful than power over USB. The idea of running video signal over USB is just cringe inducing.
Besides, wireless charging is much slower and video over wireless introduces lag.
As to recharge speeds, it’s easily comparable with USB. Sometimes USB charging is noticeably faster, but use the wrong charger and USB will be far slower.
https://debugger.medium.com/wireless-charging-is-a-disaster-...
Wireless charging can do 15+W depending on the phone.
As to the scare quote about 3.5 Billion phones, the charger they mentioned consumes 0.25W when the phone isn’t attached or 6Wh per day that’s the same energy it wastes charging a battery 12Wh per day. You can find many wired wall chargers that waste more than 0.5W when the phone isn’t charging so using one would actually take more energy than using that wireless charger for a single phone.
Wireless is close enough to the fastest possible USB speeds not to worry about it and much faster than the slowest USB speeds.
I travel a lot for business and I will soon be doing the digital nomad thing flying across the United States. I use a portable monitor that is both powered via USB C and does video over USB C.
https://www.amazon.com/dp/B095GG31KX
I also have a few of these truly “universal” USB C cables.
As to wireless, an iPhone 13 pro max can do 15W wireless vs up to 27W with the right USB charger or 2W USB if you happen to plug into the wrong charger without thinking about it. In practice I only charge overnight or while driving so it’s hard to see a practical difference.
Although the cable can passively fall back all the way to USB 2.0, apart from that the only thing that these cables have in common with USB is the name. It's a very high speed serial bus with insane tolerances and complex protocols (even the cables have CPUs in them). So no need to cringe, except at the absurd naming conventions of the USB consortium.
Of course this also means there aren't many chipsets, unfortunately.
That's an unfortunate truth and it happens to other standards as well, e.g. HDMI:
> "Products can no longer get certified for 2.0 only for 2.1, and also 2.1 features are optional to implement, so popular features like 4k120, ALLM, VRR are not required," Brad Bramy, VP of marketing and operations for the HDMI LA, confirmed to Ars. "Manufacturers could only implement eARC, for example, and claim to be a 2.1-enabled device."
https://arstechnica.com/gadgets/2021/12/the-hdmi-forum-follo...
Could someone ELI5 this line of thinking? Exactly how is a $0.01 saving so consequential? Is a better-performing $9.99 cable too expensive for some people but a $9.98 cable within their budget somehow? Is each manufacturer manufacturing 100 million cables and saving $1M here? What is the logic?
Quick math:
Let's say 1 million cables at $0.50. That costs you $500k. If you get the cost down to $0.49, you just enabled another 20k cables to be manufactured, and another $200k in profit.
1 million for 490k and sold for 10 million is a 9.5 million profit.
10k is a rounding error.
For a typical cable, it's probably around $0.30 to $0.50 to make, and they tend to sell between $5 and $15, depending on the spec.
For simplicity's sake, we'll assume $0.50 cables that sell for $10. Now, if you save $0.01 in manufacturing, so the cost is now $0.49, and if you sell say 100k cables (feels like a believable amount) then you've saved a cool $1000 in making them. So your profits are $1000 larger now, and you can pat yourself on the back. Good job!
The kicker then is how those profits can be used to make MORE cables. And $1000 of extra profits means $1000 of extra cables. Just over 2000 more, in fact, which you can also sell for $10 each.
Those extra 2000 cables have now earned you a cool $20k of profit ALONE, from only $1000 saved dollars, which itself was from only a single saved cent! Amazing job!
If you keep repeating this, you can probably start to see how this begins to run away:
Even starting with only $1 to make $0.49 cables, you end up earning over $1M on your 19th batch, even if you were spending only 10% of your sales to make new batches of cables and pocketing the other 90%! You actually pocket $1.5M on that batch, and then only spend $167k on the next batch, which promises to earn you even more!
Meanwhile, your competitor making $0.50 cables, starting with a similar $1 and 10% to new batches scheme, would only get to pocket $1.1M on their 19th batch! That $400k is a pretty tidy bonus you've earned.
On your 20th batch, you're selling 80k cables, but your competitor can only afford to sell $60k. We've not even reached the 100k a batch mark, let alone a million cables, and you're already way ahead.
And that's compound interest, folks! From $1, saving only a single cent, we can earn $400k more than we would have otherwise. Outstanding.
EDIT: I took so long to write this that I see andrewstuart2 has summed up the first half of the story far more concisely than I. Nicely done! I feel that compound interest is important for seeing the bigger picture of profit over time, so I'll keep this comment as is.
So I use brand (Anker, amazon basics, Apple, etc) and store (amazon.com, best buy, target, etc) to narrow down the countless cable options.
On the device side, the ability to get those few percent extra profitability is the difference between a failed product and a successful one to higher ups. If you give a TPM the option to choose between $0.05 off the BOM and theoretical improvement to reliability, they'll choose the former almost every time.
I remember one product launch that turned into a disaster because it had been cost optimized so hard that the circuit would go slightly out of spec under physical strain and it turned into a yelling match over the few cents of additional plastic in the housing to mitigate the issue.
A penny here, a penny there, and soon you're talking something like $0.15/cable.
If you sell a million cables, that's $150k.
But there's also the fact that people usually buy stuff based on price. If you go on Amazon, as a consumer, you're likely to buy the $3.99 cable over the $4.99 cable.
And with the profit margins necessary that $0.15 cost savings can easily translate to a $1-$1.50 price difference for the end consumer.
Seems like there is a market for cables that are easy to identify.
Bother devices should then have the colors required IN the socket/connector.
This problem can never be solved. Where there are improvements over the same connector, you need some way to identify. Either with a simple solution such as colors or markings, or with on-screen / audio feedback.
For the record, USB PD 3.1 has 28V, 36V, and 48V fixed voltages enabling up to 140W, 180W and 240W:
It's totally possible to deliver 3kw in the form factor of USBC safely. Obviously there needs to be extra circuitry in both ends of the connection, but that extra circuitry should still be cheaper than today's thick dumb AC cables.
Wouldnt that take like 400V? Not sure how safe that would be.
With high voltages, it's all down to how many joules you allow to be dissipated in a human. Ie. Current*voltage*time. Specifically, anything over about 0.01 Joules is dangerous. To use high voltages in a consumer product you need to limit either the current or the time, and you need to do it in a way that is very robust to being chewed on by a baby in a bath. And it needs to be robust to that even when built by a corner-cutting factory.
One approach to that is to have either end of the link 'test' the other end - ie. Simulate a human touching the wire. If power isn't turned off fast enough to be in spec, then the remote end or cable isn't working correctly and shouldnt be used.
To supply 3000W with a 110V supply would require around 28A, which will trip any 10A or 15A breaker commonly found on most house lines going to electrical outlets, but it's well within the realm of something that could be supplied by a circuit equipped with a 30A or higher breaker and qualified wiring.
Of course at 30A or higher, we're talking about things that connect via those bulky connectors you would find on dryers and other high load electronics. Bulky for your safety, mind you.
The utter, utter majority of the world runs on something around ~230V [1], except America and Japan.
[1] https://en.wikipedia.org/wiki/Mains_electricity_by_country#/...
* https://www.youtube.com/watch?v=jMmUoZh3Hq4
My stove and clothes dryer are certainly not 120V.
Washing machines too I've heard.
That's the crux. Your first two points already exist, but no one is required to use them, so no one does.
The only way to mandate this is making USB a standard not open as-is (ideally patent-protected). But that is the reason USB succeeded and changing this isn't a good idea.
* They do specify power tiers, 15W is the minimum for chargers, 30W I believe is the minimum for power delivery cables, with cables being additionally certified (and marked) for 60W, 120w and 240W.
* A certified cable with USB-C ports on both ends should be a USB cable, full-stop. Negotiation of alt modes is up to the devices on both ends.
The problem is that there are a lot of vendors that use garbage terms like USB 3.2 2x2, which were never meant for the market at all. I suspect this is usually a consequence of those products not actually being certified, and thus avoiding using the official marks.
IMHO, that there was anything publicly talking about "USB 3.2 2x2" is a side effect of the USB-IF not having the teeth to enforce use of the "USB" mark. Vendors would start using the right terms and running basic conformance if, say, they found they were no longer able to get their products imported. But the Implementers Forum is made up of the same vendors who are shipping broken devices, cheap charging cables, and motherboards with spec-invalid ports.
When many people talk about USB 3.2, people assume there is some sort of semantic versioning of the port or cables itself. Thats not how specs work - everything which was USB 3.0 compliant is USB 3.2 compliant. If you want to mandate new things or break old ones, you create a new version like USB4.
Many people have the same problem with HDMI 2.1a - there is no device certified with HDMI 2.0 that is incompatible with HDMI 2.1a. Thus, advertising 2.1a support means _nothing_ compared to advertising HDMI 2 support in general. It was a marketing failure that the HDMI licensing organization allowed vendors to use version numbers to imply _anything_, vs advertising that the devices you are connecting support features like 8K/60 or VRR.
In the current generation I've seen devices with "DP" used to mean "Power Delivery" instead of "DisplayPort"
Any mandate that is possible to ignore or misinterpret will just add noise.
* Somehow get all manufacturers to only create quality cables complying with the standard whose capabilities match what they say they do.
That won’t happen. Some manufacturer will start making a cable that’s longer than what the spec says is possible, or will try to get away with a bit less shielding or a bit less conductor in the cables than what the spec requires, or will decrease quality control, so that their cables are fine, on average, but 1:1000 buyers will get a melon that’s sub-spec.
I can't believe this is a modern standard. It looks like marketing deception, not a standard.
If you buy a certified USB4 cable, it has the markings and should work as you expect it to work.
If you buy a cheap Chinese cable, it won't have any markings or it has wrong markings. It won't be possible to mandate that every cheap Chinese cable has correct markings. It would require Amazon or Walmart to do some kind of quality checking for all their sellers, which won't happen. They allow SSD scams, so they also allow USB scams. https://news.ycombinator.com/item?id=32628381
That is, until the Chinese get on this fancy new internet thing.
The logos for power ratings are visually indistinguishable from the logos for data rate.
So you tell your family member that they need a 40-gigabit cable, they get a 60-Watt cable, and call you back to say it's still not working.
At this point, the USB-IF is beyond incompetence and well over into malice. They hate you all.
Is there anything the requires them to not all support the max feature set, apart from cost - some use case that would break, or be dangerous, or be mutually incompatible with another feature?
It isn't just a question of cost -- if you want a long cable to charge a phone, you'd probably prefer it to be thinner and more flexible, and you don't need the extra power.
TB4 cables over 1.5ft require inline active signal boosting/filtering chips.
lol there are still HDMI cables without an HDMI logo on them. There are still plain old full speed USB cables without the USB logo on them. China alone will destroy this single requirement and laugh as they sell hundreds of millions of cables per year.
I really do wish it were possible to mandate this. Things would be easier.
* USB3.1Gen1? USB3 (because it literally is).
* USB3.1Gen2? USB4.
* USB3.2Gen1? USB3 (because it literally is).
* USB3.2Gen2? USB4.
* USB3.2Gen1x2? USB5.
* USB3.2Gen2x2? USB6.
* USB4Gen2x1? USB7.
* USB4Gen2x2? USB8.
* USB4Gen3x1? USB9.
* USB4Gen3x2? USB10.
* USB4v2? USB11.
Power Delivery? Likewise one whole version number (eg: PD100W) based off watts carried, and append that version number to the USB version number.
So for example: USB9 PD100W, USB7 PD30W, etc.
Dead. Fucking. Simple.
The U in USB stands for Universal. The first step to being universal is to be so dead fucking simple even your technologically illiterate grandma can tell things apart.
I hate the Chrome Versioning System, but it is by far dead fucking simple, and USB could use some dead fucking simpleness with its version numbers.
I choose to believe that there is still time to fix all this.
Here is what they need to do:
* Specify a set of power tiers: 15W, 30W, 100W.
* Specify a set of bandwidth tiers: 10Gbps, 20Gbps, 40Gbps, 80Gbps
* Specify a set of icons representing protocol capabilities (e.g. “DP” for DisplayPort)
* Mandate that every port and cable is labelled with those two numbers and any associated icons.
* Call it USB 5.
* Drop all the “2x2” and “version 2” and “Gen2” and “SuperSpeed+” terminology.
Apologies if I forgot some dimension of USB functionality.