Intel launches Thunderbolt 3, will use USB Type-C connector

Scorpus

Posts: 2,162   +239
Staff member

intel thunderbolt usb type-c usb connector type-c thunderbolt 3

At Computex 2015, Intel has announced Thunderbolt 3, a new iteration of the company's high bandwidth hardware interface. And for the first time, Intel has ditched Thunderbolt's proprietary connector, instead opting to funnel everything through the soon-to-be-ubiquitous USB Type-C connector.

Thunderbolt 3 doubles the bandwidth of Thunderbolt 2, providing 40 Gbps of throughput through the small, reversible Type-C connector. Due to the increase in bandwidth, Thunderbolt now supports a whole range of new protocols and features, essentially making it a better version of USB.

For starters, Thunderbolt 3 will fully support USB 3.1, which only requires a measly 10 Gbps of bandwidth for full functionality. It will also support power delivery based on the USB standard, allowing a Thunderbolt 3 device to receive 100W of power through the port, or to send out 15W of power to connected devices.

Thunderbolt 3 also supports PCI Express 3.0 x4, meaning it can act as a connector to external graphics boxes for notebooks in a similar way to Alienware's proprietary solution. It also supports eight lanes of DisplayPort 1.2, allowing two 4K 60 Hz displays to be connected to the one Thunderbolt 3 port and used simultaneously.

Networking is supported through Thunderbolt 3, providing a 10 Gb Ethernet connection between computers. And as always, daisy chaining allows six devices to be chained together through a series of Thunderbolt 3 cables.

Intel states that Thunderbolt 3 products will begin to hit the market before the end of the year, with wider availability slated for 2016. Now that it supports USB 3.1, PCIe 3.0 x4, and DisplayPort through a USB Type-C connector, it could just become the hardware interface of choice.

Permalink to story.

 
That is one great news, I would do a macarena, if I weren't fat and lazy.

I hope the markings are going to be very clear on TB3 connectors, not to confuse the cable with USB 3.1 Type C.

This is especially great news for Apple users, as this one will soon appear in the updated products. I expect some of them to be announced in June (already this month). Expecting external 5K monitors on a single TB3 cable - that'll be awesome!

As for the PC, I'm not so sure. It took way too long to see TB2 solutions integrated. And I don't expect any video card to feature such port over DisplayPort 1.3 or 1.4a.

My question is, TB3 throughput is sufficient to allow DisplayPort 1.3, so why no word on that? We are talking 40Gbs vs 32.4Gbs.

And what about 4K@120Hz? This one is supposed to be a given.
 
Last edited:
So can we expect a single port to run both TB3.0 and USB3.1 cables/devices? In other words... can a TB3.0 port run a USB3.1 device with a USB-C cable? or vice versa? I guess probably not. Maybe if there is some kind of a switch implementation to verify with the device/cable first, before deciding how to complete the connection.
 
Now if they would only come out with Thunderbolt adapters for non-compliant hardware...
 
That's an impressive cable. Instead of having multiple different ports of different form factors on a laptop, you could effectively replace the power port, the USB ports, video output port(s), and network port (or add it back), with a bank of Thunderbolt 3 enabled USB-C ports. In addition, if the motherboard supports you'd be able to add one of those external GPUs with no addition hardware.

You would not have to use each port for a specifc purpose either, where any of the ports could act as AC in, or video out. Of course, the implementation I created in my head is probably expensive and you'd still have to deal with cabling, though I guess you could get away with a half dozen USB-C cables with adapters at the end of them...
 
So can we expect a single port to run both TB3.0 and USB3.1 cables/devices? In other words... can a TB3.0 port run a USB3.1 device with a USB-C cable? or vice versa? I guess probably not. Maybe if there is some kind of a switch implementation to verify with the device/cable first, before deciding how to complete the connection.

If they're doing this is because the port is capable of that bandwidth but it's limited by the losses in the USB cables. The Thunderbolt cables have integrated electronics in them and thus can program the USB port, just as any other device, to support its functions. If you have electronics embedded in the cable you may make it optic and just power the transceiver at both ends of the cable.

If they're using what is available, I highly doubt they'll be changing the spec of USB 3.1. Engineering is not limited by something trivial as differentiating a Thunderbolt cable from a USB one connected to a USB 3.1 port, nor it will strive for the hardest solution.
 
"I hope the markings are going to be very clear on TB3 connectors, not to confuse the cable with USB 3.1 Type C."

It is completely compatible with USB 3.1 so it doesn't matter. You can plug in your USB devices and they'll work fine.
 
"I hope the markings are going to be very clear on TB3 connectors, not to confuse the cable with USB 3.1 Type C."

It is completely compatible with USB 3.1 so it doesn't matter. You can plug in your USB devices and they'll work fine.
Yep. But not the other way around. You aren't going to have TB3 devices work (or work at full speed) with a USB 3.1 cable.
 
Looking at those figures makes me wonder. Why we need PCIe(NVMe using SATA ports) for replacing SATA 3.0?

Thunderbolt 3.0 could replace internal power and data cables for SSD's, while allowing NVMe on PCIe compatibility. Or be used as USB connectivity for front panel connections. Why have a difference in internal and external device connections?
 
Back