Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Thunderbolt 3 disguised as USB-C

ifraaank said:
Thoughts?
Click to expand...

Well several people (including myself) speculated on the possibility of that some time ago.
(https://forums.macrumors.com/thread...s-that-leave-nmp.1854817/page-2#post-20828674 though that's not a very good link).

Basically, yes, its a very sensible decision for both USB-C and Thunderbolt - Intel were planning to use different connectors for Thunderbolt 3 anyway, but connector confusion could have been the last straw for Thunderbolt when its facing serious competition from USB-C. ISTR that Intel (and Sony, at one stage) were originally going to use a USB-A compatible connector with extra pins for Thunderbolt, but the USB consortium wouldn't let them.

The only downside I see is that Thunderbolt could have done with some sort of cable-locking mechanism so that the cable to your Petabyte RAID array.

So... that seems to set the stage for the Skylake Macbook Pros (whenever they arrive) to switch to an all-USB-C design. Hope the market for USB-C plugs and adapters has picked up by then... it should do.
 
blacka4 said:
wonder if you will be able to add a USB-C card with thunderbolt to a older pro.
Click to expand...

Do you mean an "classic" pro or a pre-TB3 "new" pro?

AFAIK nobody has produced a Thunderbolt 1 or 2 card for a classic Pro yet - and a USB-C card would need the same functionality. Thing is, Thunderbolt really isn't a "must have" on a tower system with space for half a dozen hard drives and a bunch of PCIe slots.

As for the "new pro" - A TB2-to-TB3/USB-C adapter might be possible, but it would be bottlenecked to TB2 speeds, no DisplayPort 3, no charging (not important on a Pro, but might be a consideration on MacBooks). It would have to be a full-blown TB peripheral with a USB controller, TB daisychain and DisplayPort circuitry so again, the question is, where's the demand when you've got plenty of TB and USB3 ports?

If TB3 peripherals have 'daisy chain' ports then, presumably, they'd need to include a USB3 host controller in order to offer full-blown TB3/USB-C on the 'output' port.


T'hain Esh Kelch said:
When plugging USB3 stuff into a Thunderbolt3 port, it'll work fine. The other way, not at all?
Click to expand...

Yes, USB3 or USB-C devices should work on a TB3 port - the port will detect that whatever is plugged in isn't a TB cable and switch back to being a USB port.

Thunderbolt 1 or 2 devices won't work at all on a regular USB-C port (and I doubt an adapter is possible, either).

However, when Thunderbolt 3-enabled drives, displays etc. start appearing, it seems feasible (barring some non-obvious technical obstacle) that they could also have dual-purpose "input" ports so they'd work with either a Thunderbolt or a USB-3 "output". There are already drives etc. with both Thunderbolt and USB-3 connections - perhaps with TB3 they'd share a connector.
 
I mean classic tower pro. I would love to have the ability to daisy chain HHD enclosures like the LaCie 8big thunderbolt rack mount hardware using my classic pro.
 
Well that solves Apple's problem with the Thunderbolt connector being too big.

While I see where it's confusing to have similar ports that might be incompatible, if USB works in TB and not the other way around that's still dramatically better than the current landscape of ports—essentially you'd get 1 or 2 or n extra USB ports even if you don't use TB.

Also, while MVC can probably eat his crow about Apple ditching TB, he can enjoy the suggestion that eGPU support might actually be coming...
 
Its best thing that could happent to both standards.

And worst thing that could happen to People owning first gen Mac Pro.
 
koyoot said:
And worst thing that could happen to People owning first gen Mac Pro.
Click to expand...

Nah …
I mean we've all known for over a year that TB3 would get new ports. But assuming that TB3 is backward compatible in the first place, there's no reason to assume that the different port is a problem. They made mixed cables for FW400 and FW800 to solve the same problem.

But it's a good indication why the Mac Pro hasn't been updated. They're clearly waiting for this to ship.
 
  • Like
Reactions: PhiLLoW
foobarbaz said:
But it's a good indication why the Mac Pro hasn't been updated. They're clearly waiting for this to ship.
Click to expand...
Completely Agreed. Tommorow AMD is supposed toshow new cards, so lets hope all of this will be included in new Mac Pro presentation at WWDC.
 
blacka4 said:
wonder if you will be able to add a USB-C card with thunderbolt to a older pro.
Click to expand...

Essentially, No. The thunderbolt controller needs three inputs types. PCIe , DisplayPort , and GPIO. The older machines don't have the GPIO infrastructure to support it. Same with older mainstream Windows PC motherboards. Either lacking GPIO or the specifics aspects of low level firmware/GPIO that Thunderbolt requires.

There is a certification path now for workstation TB cards now. It hasn't represented the explosion in TB deployments that all the major proponents of that claimed it would represent.
 
T'hain Esh Kelch said:
Please explain this to me.

When plugging USB3 stuff into a Thunderbolt3 port, it'll work fine. The other way, not at all?
Click to expand...

The Type C USB connector ("USB C") is a connector not necessarily a protocol. All guaranteed to get when connect to a Type C connector is USB 2.0. That is it. Normally will get USB 3.1 ( gen 1 ,really USB 3.0, or gen 2 , the newer faster) when connect a USB 3.1 device.

The alternate modes are all purely optional. DisplayPort , Thunderbolt , and anything else that gets a "alternate mode" designation. Plugging into a Type C connector may or may not work for the other types. It depends upon what non USB infrastructure is put inside the system. There is a "mode" switch inside. It is kind of like a KVM switch so that one keyboard , video , mouse can be used with multiple computers. Only it is a single computer and multiple peripheral connectors protocols. Multiple protocols don't run over the same wires (except with Thunderbolt).

Type C with the "alternate modes" enable requires some sort of switch sitting behind the connector. One of the core basic functions of the TB controller is to be a switch. It is a relatively natural fit. That is on the computer system side.

For TB peripherals they expect there to be a TB controller in the computer system. Just like a DisplayPort (DP) only monitor is going to expect to be talking to a DP source inside the computer system. They don't "speak" USB 3.0 at all. There may be some corner cases with some single drive enclosures with using Type C in pure USB 3.1 gen 2 pass thru mode eventually, but generally TB externals need something besides USB.
 
  • Like
Reactions: theSeb
koyoot said:
..., so lets hope all of this will be included in new Mac Pro presentation at WWDC.
Click to expand...

Did you even read the article???

" ... Intel expects the first Thunderbolt 3 products to start shipping by the end of the year and to hit their stride in 2016. ..."

There is nothing to ship in two weeks as far as TB v3 is concerned. It is technology primarily targeted at 2016. Apple will probably hoard the initial production so decent chance will see it in the late Fall. Obviously, the only one port MacBook that is missing TB will probably be high on the priority list to get it.

By late Fall ( Nov-Dec ) might as well wait for Xeon E5 v4 since it will arrive in Jan-Feb 2016.


The Mac Pro 2013 was announced ~6 months more so as the "retirement" of the old form factor and that the Mac Pro wasn't dead overall ( as the Mac Pro was at that point banned in the EU because such an old, out of compliance design ) more so than a normal Mac Product announcement. There was nothing to buy. There is zero rational reason to invoke the Osborne Effect on the newer model.

Tommorow AMD is supposed toshow new cards,
Click to expand...

The higher end AMD HBM cards aren't likely coming to the Mac any time soon. AMD is probably going to have enough problems fulfilling mainstream gamer orders let alone anything Apple is going to pile on top. There is no reason to give Apple a major price break on these GPUs in the first 5-6 months of availability. Also doubtful the drivers would be ready before OS X 10.11 either ( but that isn't as remote a possibility).
 
Last edited:
  • Like
Reactions: edanuff
deconstruct60 said:
The higher end AMD HBM cards aren't likely coming to the Mac any time soon. AMD is probably going to have enough problems fulfilling mainstream gamer orders let alone anything Apple is going to pile on top. There is no reason to give Apple a major price break on these GPUs in the first 5-6 months of availability. Also doubtful the drivers would be ready before OS X 10.11 either ( but that isn't as remote a possibility).
Click to expand...
And why do you thought that I was talking about Fiji and HBM chips? ;)

Ive said in other thread before, that what I expect for Apple in Mac Pro is the new GPUs like Tonga, and Hawaii, or respins of them on new process, and updated with new Features, and technology.
 
By the time TB3 is ready for the Mac Pro the current rev would be best case scenario three years old. I feel like it makes sense to update it in the meantime, even if it's not some super leap.
 
koyoot said:
And why do you thought that I was talking about Fiji and HBM chips? ;)
Click to expand...

Because the AMD sales pitch machine had run out teasers about HBM in all the tech spec porn press around two weeks ago.

http://anandtech.com/show/9275/amd-teases-upcoming-video-card

Probably to blunt the impact that Nvidia's 980Ti would have around now. Their counter "strike" is probably coming in a couple of days with more info.

Ive said in other thread before, that what I expect for Apple in Mac Pro is the new GPUs like Tonga, and Hawaii, or respins of them on new process, and updated with new Features, and technology.
Click to expand...

Tonga as a speed bump to the current Mac Pro design? Possible. It would help cut down on some of the moaning and groaning if Apple offered an GPU upgrade option for the current systems alongside some minor speed bumps. ( that wouldn't be WWDC stage show time, but a press release ). If Apple is waiting until 2016 they could use something to serve as a stopgap. (similar to the stopgap update to the MBP 15" that is probably killing time for something later).

But the notion that coupled to TB v3 because it is a major move and the Mac Pro can only make major jumps in tech then HBM would be something to extensively delay for. The rebadges ..... there is no good reason at all to hold up the Mac Pro release cycle on that.

I think you are overselling the differences in the Global Foundries tech. That stuff is likely just e mid market filler for the upcoming AMD roll out.
 
fuchsdh said:
By the time TB3 is ready for the Mac Pro the current rev would be best case scenario three years old. .
Click to expand...

The vast majority of Mac Pro 2013 systems didn't ship until 2014 (there was a multiple month backlog on orders). It was really a 2014 system. Intel mainly targeted 2014 for TB v2 just like they are mainly targeting 2016 for TB v3.

If the Mac Pro is going to be tightly coupled to major TB updates they are on a 2 year cycle.
 
koyoot said:
And worst thing that could happen to People owning first gen Mac Pro.
Click to expand...

The worst thing that could happen to first gen "new" Mac Pro owners is that it became the last gen Mac Pro.
That would be worse.

A $25-50 dongle to go from miniDisplayPort based connector to Type C based connectors isn't the end of the world. Expensive but not the end of the world. TB v2 only bandwidth cables should be substantively cheaper than the TB v3 ones.


You are also presuming that there is a rapid and massive shift to the new Type C based connector for TB devices. I wouldn't bet on that. Vast majority of Type C connector devices coming to the market are going to be USB only with no support for the alternate modes. It is cheaper so the "race to the bottom" USB market is going to largely follow that path.

The TB v2 design knowledge is more mature and more broadly spread out. If Intel sells the TB v2 controllers at a deeper discount I suspect there will be more than few peripheral vendors that follow TB v2 for another couple of years.

TB v3 is gong to have an even longer and deeper Intel certification/validation testing sequence to go through. The port is going to have to support TB v1, 2, and 3 along with "legacy" (from TB perspective) DisplayPort and USB (for a subset of implementations) pass thru modes. The Type C connector was designed with 20Gb/s in mind so the stretch to 40 Gb/s is going require attention to detail ( not the cheapest chop-shop components you can find ). Before had PCIe , DisplayPort, and GPIO inputs. Now probably have a fourth of USB 3.1.

( the https://thunderbolttechnology.net/blog/thunderbolt-3-usb-c-does-it-all mentions there is a 10GbE direct point-to-point mode now. For systems that could be another layer in the cert process. Wonder how much hardware assist is in there. )

Using Type C doesn't mean TB devices will become "just as cheap as" USB devices. Not going to happen.
Very narrow dongles like FW , Ethernet , don't buy much more with a TBv3 controller as TB v2 was overkill for speed already.

The higher end of the TB market will certainly move up quickly, but the parts of the market that were trying to work closer to the affordable range probably are not. Just like there are way more USB 2.0 devices than USB 3.0 ones... same thing will be relatively true for Thunderbolt for at least a couple of years.
 
Last edited:
theluggage said:
...
So... that seems to set the stage for the Skylake Macbook Pros (whenever they arrive) to switch to an all-USB-C design. Hope the market for USB-C plugs and adapters has picked up by then... it should do.
Click to expand...

I would be more 'scared' of them dropping MagSafe from the MBP than USB. That would lock down one of the TB ports in many situations for very little reasonable gain.

If Apple resists the ridiculous march to maximum thinness maybe not (they have other max thin models... the MBP doesn't have to be entirely redundant) . The current MBP has USB sockets on either side. The TB v3 controller is still maxes out at two ports. To replace the two USB Type A sockets with "Thunderbolt" enabled Type C sockets would require two TB controllers per MBP. That is a bit of a stretch at x4 PCIe v3 each. Plus the board space required (just plain USB 3.0 is likely smaller consumer).

It appears that you get a "free' USB 3.1 controller with the new TB controller.

"...
Alpine Ridge also integrates its own USB 3.1 (Superspeed+) host controller, which in turn serves dual purposes. ..."
http://anandtech.com/show/9331/intel-announces-thunderbolt-3

Type A and Type C sockets could be used to easily distinguish between the 3.1 gen 2 provisioned sockets (Type C with Superspeed+) and the 3.1 gen 1 ( the USB controller on the internal PCH chipset. ). Kind of like when had a FW400 and FW800 port on MBPs for a while. You get a faster USB socket for 'free' by adding TB. They can leave the far more physically compatible physical ones.

But yeah some super duper minimalist design would put two TB Type C ports on either side and just drop everything else ( HDMI , USB Type A , MagSafe , regular SDXC ). Current MBP design doesn't require a HDMI adapter or a SDXC adapter.
 
  • Like
Reactions: Xteec
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back