Thunderbolt 3 embraces USB Type-C connector, doubles bandwidth to 40Gbps | Hacker News Reader

Thunderbolt 3 embraces USB Type-C connector, doubles bandwidth to 40Gbps

arstechnica.co.uk

214 pointsrbii191 comments

I'm kinda disapointed about the whole Type-C scenario.

Now when you've got a Type-C Port, you just don't know what it can do. It could be anything from a USB 2 Port, up to a USB 3.1 + 20A Power + Thunderbolt Port.

You can't see this from looking at it, and Spec Sheets today have a problem of listing the version numbers of HDMI and DisplayPort Ports, I don't see them suddenly becoming very concise about this.

The promise of USB has always been one of standardization. If I knew my mothers notebook is from around somewhere after 2008, if it has a USB Port, it will be USB 2 compatible.

In 2020 i won't be able to know if the monitor i recommend my mother will be able to connect to her laptop over the USB Type-C Port or not, or if it will run at full resolution or not, if it will be able to charge the notebook over it, or not... etc.

Its just a mess. If you make a standard, please make it a standard, not a pick and choose affair where the result ist just confusion.

I disagree, provided that the standard provides for some degree of backwards compatibility.

I think it's great how a USB 3.0 drive will still work on a USB 1.0 computer from 1999. Sure, it'll run slower, but it'll work, and even without an adapter! With Apple's previous various high-speed ports (Firewire 400, Firewire 800, Thunderbolt), they have no backwards compatibility -- from FW 400 -> 800 you could use an adaptor, but for Thunderbolt you're out of luck completely.

The point is, there will be Plug-Compatible Devices that are not Backwards compatible, as evidenced by the Thunderbolt Port switching to it.

Of course other Devices will follow that come without external power supplies and other features that don't allow for graceful degradation. Thats my point.

If it all were backwards compatible as it has been until now, things would be great, but that door has been shut now.

Ah, I see your point. At the same time, though, there do exist some USB 3.0 peripherals that require of a lot of bandwidth (like docking stations), that, while theoretically backwards compatible, wouldn't actually work with a USB 1.0 port in practice due to bandwidth constraints.

> Of course other Devices will follow that come without external power supplies and other features that don't allow for graceful degradation. Thats my point.

I don't get your point.

The Type C connector will support USB 3.1, which AFAIK is backwards compatible with older versions of USB.

The Type C connector will also support Thunderbolt 3 (and the datatypes that Thunderbolt supports, aside from perhaps DisplayPort 1.3).

Support for the protocols is done at the chipset level. All Type C connectors in Intel computers will support everything that the Type C currently supports, as the support is built in to the Alpine Ridge chipset. Other device makers will support all the USB 3.1 stuff at a minimum out of the gate, and potentially support Thunderbolt down the road also.

Where exactly is the backwards compatibility issue? Do you mean forward compatibility?

Just so we're clear, you'd rather have to carry around 10 different dongles just so you can know, visually, what a port is capable of?

I'm sorry - when I buy the laptop, I look at what the port can do, and I remember it. If I'm in a meeting and I have to use someone else's, I ask if they know. If they don't, I try it, and if it doesn't work, we move on.

I can't come up with any scenario in which this is a bad thing, outside of sheer laziness or poor planning.

But he's OK that you can't tell USB v1 from USB v2 visually. Somehow he's able to discern or recall the date of manufacture but won't be able to remember that it's a Thunderbolt port.

They probably should come up with some way to simply mark things though, so you can easily tell by looking at what you're plugging in that everything is good.

Many USB 3.0 ports have blue inner connectors to indicate that they're 3.0 ports, and some USB ports will have yellow connectors instead to indicate a variety of things (from better charging rate to supplying power even when the host device is powered off). I think the logical extension of this would be resistor-style color code bands to indicate different sets of functionality (yellow for Thunderbolt, red for charging, etc.).

    from FW 400 -> 800 you could use an adaptor, 
    but for Thunderbolt you're out of luck completely

No you aren't:

http://www.amazon.com/Apple-Thunderbolt-to-FireWire-Adapter/...

That only works in one direction. You can use it to connect FireWire devices to a computer with Thunderbolt, but you can't use it to connect Thunderbolt devices to a computer with FireWire.

Contrast with USB. I can go out and buy a USB storage device brand new off the shelf in 2015, plug it into an iMac from 1998, and it will work. It will be slow but it will work out of the box.

I don't have a thunderbolt device yet, but I wouldn't expect the behavior you're talking about. Being able to connect my old firewire devices might be nice, but I wouldn't expect to be able to use new devices on an old port.

There's no reason to expect it. But USB does it anyway, and it's really nice to know that I can buy USB 3 devices and use them with my USB 2 computers as long as I don't need top performance.

He said he was "out of luck completely". He's not.

That comment is entirely about using newer peripherals with older computers. He is.

Even today - a USB 2 port might not have enough juice to power your hard drive, so we're already in a state of "non standardization."

I look very, very forward to the day when a USB C port means that I have a very strong chance of getting 15 watts for my tablet/phone from any vendor - and the sooner apple adopts USB C for their iphones, the better.

It even happens at the cable level. I've had a couple of devices (Galaxy S3, Mophie) that required me to the use the proper higher-quality cable, despite being standard micro USB.

If i recall correctly, that's because the basic signaling for USB charging is resistance based. Too high a resistance on the wire, and it may heat up (even melt the plastic wrapping) if they try to draw more than the default 500mA USB amount.

I believe that higher currents are negotiated with digital signalling. Does that allow for the two ends talk to each other to determine the performance of a cable between them? The Apple lightning cables themselves include a chip.

Anyway, an interesting thing about USB C (and maybe 3.0, too) is that it can also negotiate 12 or 20 volts and even up to 5 amps. For the same wattage, a higher voltage will heat the cable less.

Its a bit of both.

If the data pins are shorted facing the device (so that the device detects its own outgoing signal on the incoming pin) it will consider the port a dumb charger port and attempt to charge from it based on the cable quality.

If there is a proper port response, and the port claim to be a charging port, the device can ask for more than 500mA in increments of 100mA. There will be limitations depending on there being ongoing data traffic or not.

USB 3 cables identify their current capacity through resistive values on the ID pin. Voltage negotiation is also permitted for USB 2 although I doubt it's used much outside of niche products designed to work together.

The logic is actually simple:

If you're trying to plug in a USB device, just plug it in. It'll work, regardless of whether it's TB 3 or just plain USB 3. (update: this is probably for 99% of cases since USB is much more common than TB)

If you're trying to plug in a thunderbolt 3 device, it's very likely you know if you have thunderbolt.

Will I? I've seen software developers (tech people!) return their MacBooks to Apple store because they wouldn't drive a Thunderbolt display (with the issue of course being that older MBs had only DisplayPort output, which is physically identical).

This will cause confusion (especially with added confusion around DisplayPort standard support over the USB connector) and it's unfortunate that a new connectivity standard is a regression.

Yea this is a valid point, but since TB 1 and 2 carried DisplayPort signals, there's no technical reason why Apple had to limit TB displays to just TB ports (that I'm aware of... happy to be corrected). They should've just worked.

But these ports won't have that problem. Although to be honest, the new Macbook has a pure USB3 port and not TB3 :) So Macbook owners might still have that issue. Dammit Apple!

I'd imagine that Apple has learned from history that its proprietary & non-widespread (niche) connectors have not seen the type of industry-wide adoption that they would have liked to have seen (a la ADC, firewire, iPhone 1st gen, etc) regardless of technical merit. They have diverted from industry standards when the current off-the-shelf components did not necessarily meet their design desires/criteria.

I'm guessing that they found a pretty decent strategy when thunderbolt took on the mini-dp port and added thunderbolt to the mini-dp connector. Sure there's some confusion in plugging thunderbolt into a mini-dp port, but as Apple was concerned, they were not taking away any functionality, but merely just adding the thunderbolt feature to an existing port.

Apple's potential adoption of Thunderbolt using the Type-C connector would be a similar strategy to the mini-dp/thunderbolt strategy. This finally allows them to unify ports on their whole line up macbook/pro/iphone/ipad/atv/accessories/power/displays, which with their scale would probably translate to a pretty massive reduction in cost for maintaining inventory of the many assortments of ports and manufacturing.

Based on the direction they've taken with the new Macbook and this announcement, It's pretty obvious that this is the direction they would like to go with and lays out the foundation for the expansion of the retina iMac into a retina thunderbolt display.

I'd say it's definitely more of a feature than it is a regression. The world of consumers has shown that it's been able to cope with USB 1 -> 1.1 -> 2 -> 3, with very little pain. For most people it will be the Type-C connector running at speeds varying from 12Mbps, 480Mbps, 10Gbps, 40Gbps... If you have the higher speed, you can supper everything below it (kind of like 10G/40G ethernet). Mac's will probably be sold with 40Gb Type-C ports, iPhones with 480Mb Type-C ports, Retina displays with 40Gb Type-C ports, all with the consumer being none-the-wiser as to whether they are running in USB mode or Thunderbolt mode.

>iPhone 1st gen

Actually iPhone 1st gen got HUGE industry adoption, if we're talking about third party peripherals etc.

If you mean other companies using it in their phones, computers etc, Apple never allowed that in the first place.

What's the case for Thunderbolt not just being USB 3.2? Now that it's using the exact same connector, it really seems like a solution in search of an actual use-case. I thought the external GPU thing would be amazing, except that totally failed to materialize, and I've even heard that it's being actively blocked.

Patents and trademarks?

I fail to see how this is any different then the way that thunderbolt piggybacked on the display port interface. Yeah it's less then ideal but it's not a deal breaker.

Good thing computer prices are so ruthlessly deflationary!

It just doesn't really make sense to worry about this kind of stuff from a practical point of view. Sure, maybe it'll be a pain for historians and conservationists. But for everyone else, we'll just be constantly buying new computers anyways, right?

that's very similar the home computer boom. if you look at the offerings at the time, everyone was running in a different direction. so many wonders. and in the end, cheap and available won.

you mean, just like a dvi port?

>In 2020 i won't be able to know if the monitor i recommend my mother will be able to connect to her laptop over the USB Type-C Port or not, or if it will run at full resolution or not, if it will be able to charge the notebook over it, or not... etc.

2020? Even in 2015 you can just look up the laptop model on Google and get results about whether the monitor will work with it or not, the resolution etc.

Apple, and by extension most other OEMs, will likely label the ports in the same way Apple already differentiates Thunderbolt ports from the physically indistinguishable Mini DisplayPort. Most likely scenario:

- Type C with Thunderbolt gets the Thunderbolt zig-zag logo

- Type C with DisplayPort alternate mode gets the "screen" rectangle

- Type C with USB only is unmarked

or perhaps the USB logo for USB only.

The USB type-c cables are also just a mess. They can support 5 different USB PD profiles and can support either just USB 2.0 or USB 3.1. This means 10 different cables. The cable included when buying the new Macbook for instance only supports USB 2.0.

People will soon wonder why their device is charging so slowly/having such low data transfer rates.

Now there are 2 additional thunderbolt cables (powered and unpowered) with type c on both ends. People will think it's an USB cable and wonder why their connection isn't working at all, or try to replace it with an USB cable

> an USB

how do you pronounce usb?

Usually similar to ooh-ess-bay because I'm not a native english speaker, but here it was just an error.

Hmm, I don't understand the description fully - does that mean that it can happen that I can get a device with USB-C plug, plug it into computer and it won't work because the device uses Thunderbolt 3 and the computer isn't an Apple?

Or will all Thunderbolt 3 devices be able to scale back and communicate with devices over USB if the controller is not available? How will that work with displays over USB-C connectors? Is there a possibility that now instead of just checking for a USB port, we'll have to read the list of controllers/protocols available in devices connecting over USB-C ports?

USB Type C provides a standardized protocol for protocol negotiation. I assume that Thunderbolt uses that.

It's already used in Type C to enable plugging in USB, chargers (PD spec), HDMI or DP adapters (https://www.chromium.org/chromium-os/dingdong) and I think even to route PCIe over Type C. Adding Thunderbolt to the mix shouldn't be too hard.

I think the issue is, that if you have a thunderbolt over USB-C display, and plug it into a normal USB-C port on your laptop (say the latest macbook), it won't work, since the laptop doesn't support the thunderbolt signaling. So it's a new point of confusion - USB devices which are physically compatible but not electrically.

It's something that could possibly be solved through the markings on the devices etc and I still think doing thunderbolt over the USB-C port is great. Perhaps the display example could fall back to supporting a display-over-usb standard, it won't work as well but could help.

I'd hope they gave up on their own DP signalling and just use the type C one.

I don't think we have enough detail about Thunderbolt 3 yet to know whether devices will be able to detect that they are connected to a USB port and not a Thunderbolt port. If they can detect that, they may well be able to choose to degrade.

It seems like an ideal way for Apple to be able to ship a display which can be connected to both thunderbolt and USB-C hosts, without the user having to care.

Obviously some Thunderbolt devices (e.g. PCI-E chassis) wouldn't reasonably be able to degrade to USB, but a good proportion of them, would.

I very much hope this is what is going to happen!

The Type C spec also mentions a PCIe alternate mode. The main complication is to support both wrapper protocols given that thunderbolt pushes twice the bits per second over the same wire (if using active cabling)

Wrapper protocol?

The alt-mode stuff is a physical signal multiplexing, not a protocol encapsulation.

I don't think you're allowed to sell a device with a USB jack that won't work with USB, because the consortium has a trademark on it. Or perhaps you just can't advertise it as USB?

Thunderbolt isn't an Apple interface, it is an Intel interface. They were the first to use it, because they picked the best and fastest interface to combine video with network, USB, firewire, etc into one cable. However, several other brands have caught on, including the Dell m3800 I bought recently.

I didn't mean to imply it's an Apple interface - but I haven't seen TB 2 in the wild anywhere else then MacBooks/iMacs and most Thunderbolt devices refused to work with Windows (at least as I tested on my rMBP) due to lack of drivers. I did not mean to cause confusion, I apologise.

I have been seeing Thunderbolt as an option on motherboards when I do builds for at least the last three years. But I am generally looking at motherboards in the $200 price range, as the lower end ones are usually lacking connectors and features in ways that are much more important to me than Thunderbolt.

Manufacturers I believe don't usually include it because most PC buyers would not use that feature as a purchasing criteria.

For the record,

1) Apple was the first to use USB even though it wasn't an Apple tech

2) Apple was the first to use Thunderbolt even though it wasn't an Apple tech

3) Apple is now the first to use USB 3 Type C (aka "USB-C") even though it wasn't an Apple tech (although rumor has it that Apple engineers had a lot of input on it)

The technology you're probably confusing it with is FireWire, which WAS Apple tech.

Hmm, I think you misunderstood my post - I did not mean to imply Thunderbolt is Apple tech, neither did I do so for USB-C. I'm not sure how FireWire is relevant here.

> 3) Apple is now the first to use USB 3 Type C

Didn't the Chromebook Pixel 2 come out slightly before the MacBook?

Yes, the MacBook was released April 10, but the Chromebook Pixel was already shipping USB type-C in March. Someone in the other Thunderbolt type-C thread also mistakenly claimed that Apple was the first company to use type-C.

I believe the MacBook was technically the first USB-C device announced (and rumors about it go all the way back to last year), but the Pixel shipped earlier.

Actually, the Nokia N1 tablet was announced late 2014, shipped in China earlier this year, and featured a Type C port.

Color me corrected.

It's not a big deal. No doubt both products were being developed long before they were announced, so Apple and Google were both ready to be the early adopter.

Now I'm just waiting for something to connect to the Pixel via a Type C <-> Type C connector, other than another Pixel ;)

How about a 2TB SSD with a performance rating of 850 MBps?

http://betanews.com/2015/06/01/sandisk-announces-2tb-ssd-in-...

This reminds me that Apple didn't work on it alone. I have a friend whose dad was an engineer at IBM. Back in high school (or maybe early college—late '90s at any rate), I was at his house one time and he unrolled a huge schematic… it was some kind of controller chip for Firewire. He's actually on a bunch of the patents for it.

Thunderbolt uses the Mini DisplayPort connector which was and is Apple tech, although DisplayPort itself is not.

Not sure why, but calling the mDP connector 'technology' vs. 'engineering' seems wrong to me. They are electrically identical as far as I can tell. The only innovation was in scaling the connector size down.

* 1) Apple was the first to use USB even though it wasn't an Apple tech

Not actually true. USB was shipping on PC motherboards for about a year before Apple went whole hog on USB. At the time, however, almost no one actually used USB (I think the first primary use of USB was keyboards and mice, and the first generation of USB keyboards and mice were notably inferior to their PS2 equivalents). Once Apple backed USB, other companies started to do the same.

More information here: http://skeptics.stackexchange.com/questions/2785/did-apple-j...

The first USB-C device was the Nokia N1 released only in China and unveiled on 18 November 2014

When Thunderbolt was first released it caused a lot of confusion in my office for people with DisplayPort who couldn't hook up to Thunderbolt monitors, bought for them by our IT group that really didn't understand the difference. I see this problem coming on strong AGAIN with this. This port is just USB-C while this other port is USB-C and Thunderbolt... That sounds like a load of confusion waiting to happen.

I think the Apple TB Display is comparable here.

If you have a rare TB device, you're probably a professional, and yes, you'll need to ensure the TB logo is on the port. If it's an extremely common USB device, it'll work on any port, without exception.

The Apple Thunderbolt monitors are Apple's fault. There was no reason they couldn't have accepted a raw DisplayPort signal, but they still didn't.

Confused. I have an Apple TB display and plug it into a ThinkPad mini DisplayPort port. Works fine.

It really doesn't work - I really really wish it did.

Your ThinkPad has a Thunderbolt port (which doubles as a DisplayPort).

Either your ThinkPad has Thunderbolt as was suggested, or you have a DisplayPort Cinema Display.

FWIW, one reason would be that they'd needed to have somehow worked a USB connection into the mix, for pure DisplayPort ports, otherwise you have no way to control the brightness, use the camera, or the audio output.

It's a shame they didn't allow DP users to connect a USB cable to the Thunderbolt Display and get access to all the other devices. Not aesthetically wonderful, but would've made some old customers happy.

To be fair, Thunderbolt is backwards-compatible with DisplayPort. You can use a DisplayPort display with a Thunderbolt port, but the reverse isn't true.

Ugh Intel. We just fixed the USB problem of "I can't figure out which way to plug this thing in", why do you have to go and add the complication of "Wait, is this a Thunderbolt or USB C device/port?"

This is why we can't have nice things.

Not really. This is a smart move. Why can't the hardware can't support both? If it did, you wouldn't need to know which protocol was in use.

I think that's the point. Only one port, multiple protocols.

Well, almost all devices will probably just use USB 3.1 anyway. The few exceptions are probably for professionals and they simply have to be plugged into a port with the thunderbolt symbol using cables that have a thunderbolt symbol, too.

I think you've got it backwards - if I'm reading this correctly, and you own a (for the sake of argument), 2016 MacBook Pro - your Thunderbolt 3 ports will support everything. USB/Thunderbolt are now using the same physical interface (no need for dongles), and Thunderbolt 3 will support USB 3.1. In the incredibly rare (will likely never be seen by an average consumer) scenario that you have some high-end equipment that actually requires Thunderbolt 3, and not USB 3.1 - you'll definitely know whether your device supports it, and on which ports.

Put in regular terms - you can now use your USB 3.1 device in any port that you can plug into.

But, ostensibly, there's going to be thunderbolt devices soon which, while having a USB Type-C plug, won't work in a regular USB port despite the Type-C plug.

Given that USB 3.1 + Type C ports support USB PD (100 Watts), and 10 Gigabit signaling rate (7.1 Gigabits demonstrated transmit) - the need for thunderbolt devices will be so excessively rare for the average consumer as to be a non-issue.

For the few Pro Users who Specifically paid a small fortune for their thunderbolt device - they will certainly know which port to plug it into, there will be zero confusion.

What's nice (awesome) about this, though, is that when Apple/Microsoft/Sony/Asus/whoever decides to differentiate their high-end products by offering Thunderbolt 3 ports that people will rarely have a use for, they can actually make use of them immediately as USB 3.1 ports. Right now I have a MacBook Air w/2 USB + 1 Thunderbolt Port. I have too carry Ethernet Dongles for both Thunderbolt and USB to cover the situation in which I run out of ports and need ethernet. And I frequently run into a situation in which I've used up my two USB ports, and my third adapter isn't thunderbolt (I know, I should carry a Thunderbolt -> USB adapter, but god, I carry so many cables already).

In this new world - I would have three, fully functional USB ports, with one of them also capable of functioning as a thunderbolt port should I have a Thunderbolt peripheral.

Awesomeness.

Source? Sure, it's theoretically possible that a TB3 5K@60Hz monitor won't work when plugged into a USB Type C port. But there would be no good reason for that, it should downgrade smoothly to either 5K@30Hz or 4K@60Hz.

I'm running off of the assumption that Thunderbolt continues to use a different physical layer (which it pretty much has to to go faster than USB 3.1), and is piggy backing on USB Type-C's "Alternate Mode" which lets you negotiate different physical layers that you have a transceiver for.

EDIT: Looks like my assumption is correct:

http://en.wikipedia.org/wiki/USB_Type-C

Sure, but why wouldn't the TB3 transceivers be designed to support USB 3.1 as well? Those transceivers have a wide range of dynamic adaptation to support high speed transmission, so it doesn't seem like much of a stretch to adapt to USB 3.1 signalling levels & protocol.

Because even above the physical layer, the two protocols are very, very different to the point that you're nearly doubling the amount of work that you have to do. USB at it's core is a host polled interface that looks like a network, and Thunderbolt is at it's core a multimaster RDMA interface. You'd basically be designing two different devices.

Sure, but USB is a very cheap protocol client-side. I fail to believe that requiring it in all TB3 clients would be prohibitive.

I admit i'm not super clued-up on this, but i just looked up the new MacBook, it turns out that the sum total of its ports is one (1) USB-C and one (1) 3.5mm headphone jack. Does that mean i have to choose between charging, connecting a video projector, or using some USB hub i'll have to lug around to be able to read someone's presentation off a USB key while simultaneously connected to a projector? I wouldn't want to be in a position where i'm forced to run on the battery in certain circumstances, either (i'm a bit paranoid about the battery running out -- this is not impossible if one is giving a presentation after having spent a few hours on a train [in a backwards country without power outlets, of course] working hard)...

Assuming you dislike carrying splitter dongles, it sounds like the new MacBook isn't for you.

(It's not for me, either.)

The new macbook is for people who are 100% focussed on mobility, and nothing else. That is a small niche of people. Think of it like an non-touch iPad with a keyboard, and you kind of have a sense of its purpose. The vision for the Macbook is "Wireless".

Just looking at your use cases though:

o Charging - the idea is that you have battery life in the 8+ hour range, so that when you are connected to a projector, you don't need to be charging.

o Reading a presentation off a USB key while connected to a projector - I've been doing that exact thing for the last 3 years, on an almost weekly basis - and 100% of the time I borrow the USB key for a second, and then hand it back to the owner (or put it in my USB wallet). The important thing there is sufficient local storage.

But yes - The MacBook is probably not for you, and it's definitely not for me.

I'm just psyched that my next MacBook Air (or whatever they call it - maybe I'll have to move upstream to the Macbook Pro if they dump the "Air" line) will have three USB capable ports instead of the current 2 + 1 Thunderbolt.

Your USB device is guaranteed to work. If you have a TB device, then yes, you'll need to actually look at the port to see if the TB logo is near it.

Since TB devices are so uncommon, I don't see a problem.

Sadly, the blocker for external graphics isn't the power draw, it's that Intel steadfastly refuses to license Thunderbolt for external GPU enclosures.

That's why you can't buy a macbook air and plug in a little $300 box with a GeForce 970 and play high-end games on it right now. There are no technical reasons why this won't work-- in fact, people have hacked together solutions that work great.

Intel doesn't want to let you do it.

Slide #15 in the presentation actually has "External graphics" as a bullet point, so maybe that's about to change?

So it does! I hope you're right, that would be amazing!

I suspect this is because they think thunderbolt 1 and perhaps even 2 do not have the bandwidth to give satisfying graphics performance. This might change with thunderbolt 3.

TB1 and TB2 might not have had enough bandwidth for high-performance stuff but it did have more than enough bandwidth to blow IGPs out of the water: DYI external TB2 systems had 80~90% of "internal" desktop performances on synthetic benchmarks[0], even with a TB — EC chain you could reach very, very respectable performances[1]

[0] http://www.anandtech.com/show/7987/running-an-nvidia-gtx-780...

[1] http://forum.techinferno.com/implementation-guides-apple/427...

Not anymore with TB3:

From: http://www.anandtech.com/show/9331/intel-announces-thunderbo...

>>> Meanwhile gamers will be happy to hear that Intel is finally moving forward on external graphics via Thunderbolt, and after more than a few false starts, external GPUs now have the company’s blessing and support. While Thunderbolt has in theory always been able of supporting external graphics (it’s just a PCIe bus), the biggest hold-up has always been handling what to do about GPU hot-plugging and the so-called “surprise removal” scenario. Intel tells us that they have since solved that problem, and are now able to move forward with external graphics. The company is initially partnering with AMD on this endeavor – though nothing excludes NVIDIA in the long-run – with concepts being floated for both a full power external Thunderbolt card chassis, and a smaller “graphics dock” which contains a smaller, cooler (but still more powerful than an iGPU) mobile discrete GPU.

Seems to be trading all our different ports with the problem of having all levels of different cables?

This port is going to be very expensive for the manufactures. If it does everything I'm going to need a bunch. Does anything stop OEMs making a row of identical ports that only 1 charges my laptop, only a couple take can use the fastest cable and I'm sure all sorts of potential shenanigans,

Is there any word on whether the cables will be different? I suspect the cables will match, but the protocols the port support will be determined by the controller chipset.

I can't see how the cables will look much different between each other but even so they will all fit the same connector. There's already planned the 20gbps copper, 40gbps active copper and optical.

What mechanism can stop the £0.99 charge only cables people expect to plug in their portable hard drives through.

> What mechanism can stop the £0.99 charge only cables people expect to plug in their portable hard drives though.

This has nothing to do with this Thunderbolt announcement, the situation would be the same even if only USB 3.1 was supported. However, the solutions are the same as ever:

1. Don't use charge only cables, buy cables that support charging and data. They're normally affordable.

2. Keep the charge only cable plugged in to your power source.

3. Mark the charge only cables with a sticker.

As for the Thunderbolt 3 announcement, it looks like there are three proposed cables, and it looks very likely that the passive and active versions will support both USB 3.1 and Thunderbolt 3. Perhaps the light based ones won't, but they'll be niche at first, unless Thunderbolt 3 takes off rapidly.

USB-PD can check the cable to see what it's capabilities are so it won't push too much current over the wrong type of cable. I imagine any scheme to use TB3 over a Type-C connector will be able of checking the cable as well.

Thunderbolt 1&2 connectors have active electronics in the connectors, and that seems likely to stay in 3, so yes, the cables will be different.

They specifically call out in the release that 20Gb will use passive cables and 40Gb will require active cables.

I think we will see symbols next to the port like the thunderbolt symbol and a charging symbol. And I guess most manufacturers will enable all of their ports at least for charging to keep things simple (and customers happy).

The high current required for charging makes that one of the more difficult things to spread around to all the ports, I'd imagine.

After the MacBook came out, but before the new Chromebook Pixel, I was telling people that Apple only put a single type-C port on their MacBook for this reason. I said "When PC's come out with type-C, they'll have multiple ports, but you'll only be able to use one of them as the power port. Apple would rather have a single port than have such a confusing setup."

Of course, then the Chromebook Pixel came out and proved me wrong. I still expect some PC makers to try and shave costs by only supporting power input on one of the type-C connectors, but the Chromebook Pixel example will ensure that such an arrangement will get slammed by reviewers.

Is that definitely the case? It's still not unknown for laptops to have exactly one charging downstream USB port, unlabelled, so that some devices will charge from one port and not another. I wouldn't be at all surprised if the same is happening with USB-C.

Yea, this is a problem that could be solved by color coded stickers.

Why do you need a bunch? Most will be able to get by with a single port, to one-two monitors with a couple USB-A ports (and ethernet) on the monitors... or via a Thundebolt/USB-C dock.

To run multi-monitor you'd need the monitor to embed a DP-MST hub or embed a USB 3 video device if you only used a single port.

The alt-mode capability cannot traverse a hub.

More likely you'd have dual cables running to dual monitors. Each monitor could provide internal USB 3.1 hubs and associated devices plus feed 100W back up the line if needed.

I would think that the first monitor would have a thunderbolt passthrough for a second monitor connection... IIRC, the apple thunderbolt displays have things like ethernet, etc on the monitor itself, though I don't think gen-1 supported multiple displays on a single connection, it wouldn't surprise me if this one changed said passthrough.

This already happens. My laptop has three USB ports. One with just the USB symbol, one with the superspeed USB symbol ("SS") and one with the superspeed symbol and a lightening bolt. I'm not sure exactly what that lightning bolt... Who knows what that means - maybe support for the Battery Charging Specification (5V, 1.5A)?

We'll just have to get used to ports having a load of obscure symbols next to them.

The port with the lightning bolt might charge devices even when the laptop is turned off.

It's a sad kind of funny we'll finally get a port I can plug anything in either way up but now have to look even harder at the icons to make sure it's the correct identical port. Can't even feel for the USB embossed logo side up for that one.

And yes the lightning bolt will charge things without powering your laptop on. Very handy.

It's a good point. With my current USB devices, I know that the short cable will charge my phone properly and the long one won't. I know that I've seen a 'charge only' cable. I'm not sure this will be fixed with the newer USB standards.

In my experience, many micro-USB cables that won't charge simply have worn plugs. That is, if you wiggle it just right, it will charge, but when you leave it lying on the desk it won't. However a new cable of the exact same SKU will fit so tight you can hang the phone from a wall socket with it. I think this is good engineering, because something will wear out, and it's much less bother to replace a $2 cable than it is to fix a worn-out USB jack in your phone.

And finally, the future has arrived. If the iPhone 7 (doubt the 6S) sports USB-C as well, there will be a truly universal connector. Imagine screens, laptops, TVs, phones, mp3 players, docks, hard drives and toasters all using the same plug (well, maybe not the last one). There will be a painful time of transition until we are there, but hopefully it will be the last one.

(Conveniently this also saves Apple on the new MPBs, they can now have both USB A and USB C ports without it being weird)

They will find a way to ruin it. Don't worry

Apple using a standard connector, dream on.

Apple led the way with the original USB in the late 90's. PC companies and users dragged their feet. "We don't like change." It took almost 5 years to gain traction.

A couple months ago on HN I asked about PC's getting the new USB C connector and no one seemed in a hurry. It won't be a standard until PC's ship with at least one port.

> It won't be a standard until PC's ship with at least one port.

Phones (and tablets) getting a Type-C port may very well coerce the PC market into faster adoption.

Apple wasn't even involved in the creation of the USB spec. Every major PC manufacturer was, and had committed to the spec already. Apple did "lead the way' by shipping a computer with no legacy ports, but the PC industry hardly "dragged their feet". It takes years to churn so many PCs with (legacy ports) which dominated the market.

> Apple using a standard connector, dream on.

Apple is clearly pushing heavily behind the scenes in both the USB-IF and with Intel/Thunderbolt to drive USB Type C ports. Reportedly their engineers did much of the work designing the port. Looking at their one-port Macbook, they’re obviously heavily invested in USB Type C’s success. I assume every future Mac is going to be mostly USB Type C ports. Can you clarify what you mean with your comment?

I’m guessing iPhones will stick with Lightning on one end for at least the near future though.

I just want to thank you for qualifying your description of Apple's contribution to USB-C in a correct manner, as opposed to perpetuating the myth that Apple invented USB-C. It's initial conception having come out of the USB 3.1 development by AMD, HP, Intel, and Microsoft. It is refreshing to see someone not spreading the revisionist history that has been popular on this subject.

My external non-expert impression (which could be totally off-base; I don’t have inside sources) is that folks making devices like phones and tablets were jealous of the reversibility, faster battery charging, and size advantage Apple could obtain with Lightning jacks, compared to the ridiculous extended USB 3.1 Micro-B jack, and started clamoring for a better connector sometime in ~2012. Once Apple realized such a new connector would be coming with or without them, they decided to throw their weight behind it and make sure that the connector met all of their own needs.

Apple certainly isn’t the only party (or even the main party) responsible for the new connector, but from what I’ve read they do deserve some applause for putting their resources in and making sure the new connector is the best it can be.

That all agrees with how I understand the situation as a fellow external non-expert. I just got really frustrated a while back by all the 'Apple invented USB-C' and even the 'Apple basically invented USB-C' headlines. So wanted to applaud someone for choosing the phrasing 'much of the work' which I believe to be a correct statement, and if incorrect only errors slightly on the side of too much credit.

I have literally never seen anybody claim that Apple invented USB-C, and I read a lot of stupidity. Was this really common?

I saw an awful lot of of articles similar to these search results a while back, and found it a tad annoying. Then again I guess I must of fallen in with a bad crowd if this is the type of stuff I was exposed to. https://www.google.com/webhp?sourceid=chrome-instant&ion=1&e...

> Can you clarify what you mean with your comment?

For several years Apple has been the one phone company using a non-standard connector on their phones. I mean, if they switch to USB-C then great - but it's worth saying that Apple is already the one problem company in this area. Everyone else has switched to standard USB and got on with it.

I admit i'm not super clued-up on this, but i just looked up the new MacBook. Does that mean i have to choose between charging, connecting a video projector, and using some USB hub i'll have to lug around to be able to read someone's presentation of a USB key? I wouldn't want to be in a position where i'm forced to run on the battery in certain circumstances (i'm a bit paranoid about the battery running out -- this is not impossible if one is giving a presentation after having spent a few hours on a train [in a backwards country without power outlets, of course] working hard)...

In the short term, yes, except adapters with multiple ports like apple's USB-c To HDMI adaptor that has both an HDMI port and another USB port to plug in power. Long term.. I would expect projectors/displays to supply power to their connected laptop.

Apple replaced its proprietary MagSafe with USB-C for the latest MacBook, so there's precedent.

(Aside: but I wonder if magnetic connectors are even appropriate for a cable that can transmit data as well as power. With MagSafe, if the cable was disconnected, the laptop battery would take over and you'd be fine, but with thew new MacBook you'd potentially have data loss if you had a hard drive daisy-chained in. So perhaps that's a factor in why they got rid of magnetic connectors?)

> they can now have both USB A and USB C ports without it being weird

I’m expecting them to kill off USB type A ports entirely on future laptops. We’ll see what happens to Magsafe, HDMI, and SD card slots. I wouldn’t be too surprised to see a Macbook Pro with 4–6 USB type C (Thunderbolt 3) ports, a headphone jack, and nothing else.

> Imagine screens, laptops, TVs, phones, mp3 players, docks, hard drives and toasters all using the same plug (well, maybe not the last one)

I don’t think toasters are a good fit, but USB Type C with its 100W DC could be great for powering other small appliances (LED desk lamps, small fans, printers, scanners, video cameras, small TVs, routers, modems, electric toothbrushes, ...), if USB Type C starts showing up in outlets in homes/cars/airplanes/airports/classrooms/...

I could even see them ditching the head phone jack eventually. I can't imagine it will be long before someone starts making USB C headphones.

I hope they don't do that, I don't want to go around with a DAC hanging off my headphones. And making them more expensive.

It wouldn't need a DAC, just a special plug. Type-C connectors support an analog audio mode.

Nice, I wasn't aware Type-C could route anything analog besides power. I assumed it just piped digital serial protocols.

Well, considering that USB A headsets are already pretty common...

USB Type C actually includes a special mode allowing analog audio output specifically to allow "USB C" headphones (and adaptors).[0] Whether anyone will actually implement it is another question.

[0] See pages 24-26, 29, 34 of this PDF: https://intel.activeevents.com/sf14/connect/fileDownload/ses...

You can do DP to HDMI, DVI, or VGA. So a USB-C DP plug supersedes all other display ports.

I agree on SD cards though, just because most SD Card slots can impress the card into the frame of the machine for portability.

On the balance, having one connector and hand-waving everything else as "it's all software" is probably better than the bewildering combination of ports and protocols we have now. But it will come with its own new and interesting issues as well.

For example, when can/can't you plug your USB-C webcam into your USB-C monitor and expect your computer to detect it? If I plug my USB-C toaster into a USB-C computer, will it power up? Does a USB-C hub (like the ones we use with USB-A when you don't have enough ports) necessarily support every possible protocol, or can you only plug a subset of computer peripherals into it? Will all USB-C cables have the amperage ratings to safely power my laptop from the wall, or will there have to be different cable types for different uses?

That's already the case with USB ( and to some extend Bluetooth and NFC )

What do you get when you plug you USB phone in your USB computer ? What about the tablet ? Why can I read files from my phone via USB but my phone can't read file from my USB HDD ?

Why is my Bluetooth headphone mono with the computer, but stereo with the phone. Why can's my NFC yubikey work with my NFC phone.

Right now, people mostly ignore it because there is generally broad support in devices provided by manageable number of generic protocols. So you mostly only have to care about the few cases when stuff don't work rather than thinking what works.

There is no reason to be overly pessimistic about USB-C, it is an incremental improvement over technologies already suffering from the same issue.

I was hoping the USB-C would actually fix this issue - after all, USB was mostly universally compatible for vast majority of devices and I think people started to expect that if you can plug it in, it works.

> Imagine screens, laptops, TVs

I have no doubts they will continue to have VGA ports until 2025...

Intel's upcoming Skylake architecture is rumored to not support VGA. That might speed along its demise somewhat.

I actually use the VGA port on my ThinkPad somewhat regularly—it's still the most common display connector in many offices. However, I do carry dongles for HDMI and DVI (adapting from the built-in DisplayPort) just in case. I'm looking forward to having USB-C replace all of them.

Any theories on when TVs will use USB-C over HDMI?

> there will be a truly universal connector

Well, the connector will have a universal shape. But the cables won't work universally (you need special ones for high-speed Thunderbolt), and two things sharing a connector won't mean they'll work together due to different protocols (USB 3, DisplayPort, Thunderbolt, others).

The linked site seems to be completely toast, but here are some secondary sources which I’d guess have approximately the same content: http://www.cnet.com/news/thunderbolt-3-and-usb-type-c-join-f... http://www.pcworld.com/article/2929798/thunderbolt-3-to-work... http://www.theverge.com/2015/6/2/8704067/thunderbolt-3-usb-c...

Can someone explain whether I should care that they are still using DP 1.2 rather than 1.3? I had thought 1.3 was needed for non stitched 4k 60hz, but this should have plenty of bandwidth, so I'm now not sure what the issue with 1.2 vs 1.3 is.

I was disappointed in the lack of 1.2a, having access to adaptive sync would be a nice feature for gaming.

Displayport daisy-chaining for screens above 1980x1080p requires 1.3. (Using two daisy-chained dell screen's on my 2011 macbook air does not work because it requires 1.3 and mba only supports 1.2.

The article claims dual 4k @ 60Hz over 1 cable; so that seems to rule out stitching? That, or it's an awful lot of DP1.2 channels + stitching, which I didn't know worked with multiple physical heads.

It appears that this cable has 8 lanes of DP 1.2 connectivity. 4 DP 1.2 lanes is enough for a non-tiled 4K 60hz display. So TB3 can supply enough bandwidth for two of them. A 5K 60Hz display requires either DP1.3's higher bandwidth lanes or more than 4 DP1.2 lanes teamed in a tiled arrangement. TB3 with DP1.2 will be able to handle single tile 4K displays or double tiled 5K. It will not be able to handle single tile 5K or the other features of DP1.3 (high refresh rate/8K/alternate bit depths)

So now you can't even put glue in the thunderbolt port anymore to prevent DMA attacks?

I don't think DMA attacks have been fully solved yet via software, or am I not up to date here? I guess you could blacklist the driver.

Just glue over your USB-C ports with tinfoil.

You do realize that this is a very easy to pull off attack? http://www.breaknenter.org/projects/inception/

Do you also think people who use screenlocks are tinfoil hatters?

It's more that your computer should be considered compromised regardless if you allow someone to gain physical access to it. You should know where your device is at all times.

I'm not willing to throw away my device's resale value to block a physical compromise attack.

It's about noise. The physical lock on your door doesn't prevent anyone from breaking in. But you can clearly tell someone broke in and it creates noise.

It's not always practical to take your notebook to the toilet with you, so you want to make it as hard as possible for someone to compromise it in a limited time frame. Security is never absolute, it just about buying time to notice an attack. This applies to the physical lock on your door just as much as it does on your notebook

If the thunderbolt part (more accurately, the pci-express lanes) could be disabled without losing the USB 3.1 and display port capabilities the port still might be "good enough" for most purposes.

And yeah, we really need device firewalls that isolate everything via IOMMU and don't allow drivers to do any memory mappings until the user confirmed the device.

The USB C connector is awesome and this is great news.

I still wish the industry would have standardized on the 2.5mm jack. It can be plugged in any direction. You could pull a cord out of a mess of cables and it would not snag as there is nothing to snag on.

Not very suitable for high-speed impedence-matched communication, and not ideal for power delivery: the 'ring' prong in the socket touches the tip on insertion.

And for a 12-wire USB Type-C equivalent you'd need a TRRRRRRRRRRS connector.

Hence why we prefer Speakon connectors instead of those lame jack plugs they insist on putting on all guitar amps (and a great way to blow up a Marshall amp - just half pull that speaker output cable out and watch the valves go a funny colour eventually)

And they leave out the one thing that would be interesting, a wiring diagram.

the C port is interesting in that it provides both USB pins and a set of pins that can be used for a number of functions (Displayport being its out of the gate use).

If this leave the USB pins alone, and make use of the supplemental pins for Thunderbolt, there will be no port confusion. Especially as Tunderbolt is already set up to carry Displayport data anyways.

Then it just comes down to the chipset to negotiate the right protocols.

That's basically how it works. There's a pair of wires that run USB 2.0. Those are then used to negotiate what the 8 high speed wires and 2 sideband wires are used for.

If you read the spec sheet for USB 3.1 and the Type-C connector you'll see they also talk about PCIe in alt-mode as well. DP in alt-mode is the one getting all the press though.

Wiring diagram and lots more details here http://www.anandtech.com/show/9331/intel-announces-thunderbo...

So, will Apple finally be able to refresh the Thunderbolt Display?

I got an old one I want to sell before it gets obsolete. I guess the time is now.

Edit: "Thunderbolt 3 integrates USB 3.1, optional 100W power delivery, 5K @ 60Hz display."

Optional power delivery? Optional nowadays just mean "will be removed before even reaching production"

Whenever this makes it's way to the Macbook Pro (probably next year?), this will definitely enable them to make the upgrade. Right now, they don't have any TB3-capable computers, so no one would want to buy that monitor.

And I guess the sticky part will be all those 12" Macbook's with Type C connectors that wouldn't work with such a display...

Those 12" Macbook's aren't capable of driving a 5K display at 60Hz, period, even if they had a TB3 port. However, a TB3 display plugged into a standard type-C port would negotiate a DP1.2 connection so it could run 5K@30 or 4K@60.

I'm not saying they should. I'm just saying there will be confusion about there being a Type C connector on the monitor and the laptop, and that they don't work together.

But I'm saying that they probably will work together. Downgraded, sure, just like when you plug a USB3 hard drive into a USB2 port. It's not fast, but you can get your files off.

Universal Serial Bus. With optional universalness.

I think Roritharr is right, this is the beginning of the end for the great era of standardization that USB brought. If there was a port and a cable, you pretty much knew it would work for everything the port should do.

DisplayPort and HDMI port/cable versioning have been a colossal pain for me. GPU spec lists on shopping sites don't even list what version they support, so you have to constantly cross reference things to manufacturer datasheets.

http://i.imgur.com/ffMR5gj.png

What's sad is that we had the same sorts of differentiations in the past, like with DVI-I (digital and analog) and DVI-D (digital only) and Dual-link DVI (for big screens). So what did we do? We called them different things and then actually advertised what it was capable of.

But apparently saying "DisplayPort 1.2" is too complicated now or something? Maybe there are too many different optional features that may or may not be supported by a GPU or the cable or the display, so they just assume none of them will work? I don't even know.

That screenshot isn't an old graphics card either, it's a GTX 970.

Not looking forward to USB turning into the same mess.

EDIT: This isn't all new, I recall seeing some laptops with strange things like yellow USB ports that were still powered for charging when the computer is sleeping. But that's a minor feature that most people wouldn't even notice. Major capabilities missing from some ports is different, and can get more confusing than "the blue ports are faster but all of them will work".

That is completely inaccurate. USB-C is backwards compatible with older versions of USB. And just like older versions of USB, the version directly determined capabilities. They did make an effort with USB3 to make the ports blue, but even that isn't universally true.

USB is already "that mess" depending on what it is you need the port to do. Let me know how that USB-powered device that needs 90W does when you plug it into a port that's only 1.0 capable.

Optional power delivery? Optional nowadays just mean "will be removed before even reaching production"

Your point is taken, but we also have to consider if that would even be possible. 100W is not a power level that a laptop device is going to be able to output, so it clearly can't be standard. It's good to have a standardised option, however.

Given that it supports dual 4K displays at 60Hz with one cable, or a single 5K display at 60Hz with one cable, I think it's safe to say Apple are waiting for this to announce their new cinema display.

Anyone think they'll pre-announce it at WWDC (like they did the Mac Pro)?

With any luck that will mean the return of a ~24" 4k screen along side a new ~27" 5k

I'm interested in how backwards compatibility works - will old Thunderbolt devices work with just an adaptor cable? As someone who does a lot of professional video/audio work, I have a bunch of Thunderbolt devices so I hope that's the case.

This article ( http://www.theverge.com/2015/6/2/8704067/thunderbolt-3-usb-c... ) explicitly mentions an adapter for the older Thunderbolts. I assume the adapter would be passive.

I have a Thunderbolt Display I have been meaning to sell. I think a.s.a.p is probably the time to do it, since I expect a release of a USB-C display at WWDC.

Is this finally one port to rule them all? Please let it be so so that we can keep things the same for more than 12 months at a time.

Intel just cant let go the idea of collecting IP royalties from every single device you plug into your computer :/

Yeah, they should just stop developing USB technology for other people to use.

One port to rule em all

One port to find them

One port to bring them all and in the darkness bind them.

Considering the DMA vulnerabilities of thunderbolt, that's not too far off.

At least until we have some sort of IOMMU-based hotplug device firewalls in our operating systems.

USB 3.0 has DMA too.

But is that remote DMA? Controller/driver-managed DMA transfers in the fashion of "shovel the next X incoming kilobytes to this memory range" is not the same as arbitrary writes to host memory initiated by a device.

Ah, maybe it isn't, fair point.

On USB only Host is allowed to be a bus master. Thunderbold does PCI/PCIe and by extension allows any plugged gizmo to take over whole computer.

Isn't that already the case? Current-gen Intel and AMD platforms have IOMMUs; Linux supports them. Linux and Windows support disabling DMA without disabling the rest of e.g. Firewire.

That's for a specific firewire protocol.

Thunderbolt exposes PCIe-lanes, so you can pretty much attach any pcie-device that has drivers. If there are any drivers that do not use the IOMMU properly (just pass through everything?) then we're back to square one. And I don't think all drivers are iommu-aware.

The system should only hotplug devices/drivers with user approval.

Reply on news.ycombinator.com