Apple's new Thunderbolt cable sports internal firmware, chips

Posted:
in General Discussion edited January 2014
A teardown of Apple's just-released $49 Thunderbolt cable has revealed an "active cable" with transceiver chips on each end.



iFixit took apart the new cable on Wednesday and discovered two Gennum GN2033 Thunderbolt Cable Transceiver chips, other much smaller chips and "tons of little resistors" tucked into the metal connector.



The teardown experts were prompted by a tip from ArsTechnica, who had been told by a support technician for storage maker Promise that Apple's Thunderbolt cable is a "smart cable" with internal firmware.



Apple released the Thunderbolt cable on Tuesday, alongside the first compatible peripherals--external RAID systems from Promise, which range in price from $999 to $1,999. Early tests of the Promise systems have revealed blazing fast write speeds of up to 700MB/s, as much as 21 times faster than FireWire and USB 2.0.



Gennum's website describes its transceiver chips as a requirement for the cables due to "the unprecedented speed of the new Thunderbolt technology places unique demands on the physical transmission media. The GN2033 provides the sophisticated signal boosting and detection functions required to transfer high-speed data without errors across inexpensive Thunderbolt copper cables."



Gennum's GN2033 Thunderbolt Cable Transceiver | Source:iFixit



Sources within the telecom industry told ArsTechnica that active cables are usually used at data rates in excess of 5Gbps. Chips at either end are calibrated to the attenuation and dispersion properties of the wire in order to "greatly [improve] the signal-to-noise ratio."



Intel has also reportedly chosen to use active cabling for "future optical-based iterations of Thunderbolt," the report's source added. Though passive optical cabling is more common, active optical cables could allow fiber optics to be coupled with electrical cabling for power transmission. Additionally, "current electrical ports can be forward compatible with future optical cables" if active cabling is used, said the source.



Teardown of Apple's $49 Thunderbolt Cable | Source:iFixit



On the more consumer end, LaCie has promised a Thunderbolt solid-state drive is coming this summer. A recent demo of the drive reached read speeds of 827.2MB/s.



Intel and Apple worked together to develop the specification, with Intel providing its "Light Peak" technology and Apple offering its Mini DisplayPort standard. Thunderbolt drives two separate 10Gbps links, one for displays and one for PCI-Express devices, and could reach speeds of up to 100Gbps when the cables transition from copper to optical.







Similar to Apple's experience with its in-house developed FireWire standard, the company has a fine line to walk in making Thunderbolt a unique value-add for Macs while still driving widespread adoption of the standard to ensure a large enough market for a range of third-party peripherals. Currently, the Mac maker is the only supplier of Thunderbolt cables.



Earlier this week, Sony announced a new VAIO Z laptop that implements a proprietary version of Thunderbolt. The electronics giant pulled a similar maneuver with its custom version of FireWire, called i.Link.
«13

Comments

  • Reply 1 of 42
    felmanfelman Posts: 21member
    Nice! Imagine: 2012 "Please plug in your Thunderbolt cable for a firmware update." ... "Thanks! Your cable is now up to spec with a new max speed of 20GBps on each channel! Enjoy!"
  • Reply 2 of 42
    hmurchisonhmurchison Posts: 12,419member
    Wow ..now I actually feel like $50 is a steal.
  • Reply 3 of 42
    I hope I am wrong, but I fear a repeat of SCSI and FireWire, where cost held back adoption, which in turn kept the cost high. For many people, USB3 is going to be fast enough, without costing $1000 for peripherals and $50 for a cable.

    I was excited when TB was released, and pulled the trigger on a new MBP. That was months ago, and we are only now seeing anything of TB, and too much for my wallet. I find myself wishing my MBP came with eSATA and USB3...
  • Reply 4 of 42
    tonkintonkin Posts: 42member
    Had the same thoughts regarding USB3 & storage.
  • Reply 5 of 42
    tallest skiltallest skil Posts: 43,388member
    Quote:
    Originally Posted by Applecation View Post


    For many people, USB3 is going to be fast enough, without costing $1000 for peripherals and $50 for a cable.



    You realize this is the first cable ever made, right? USB cables were expensive in 1997, too.
  • Reply 6 of 42
    8corewhore8corewhore Posts: 833member
    Monoprice will charge less than a third of what Apple is charging.



    BTW, I recall buying an HP that had USB on it when most people hadn't heard of USB

    and there were no peripherals for it. It will come and it will be affordable.



    That HP motherboard failed - warranty center wanted to replace it with a MB that did not have USB - said nobody needed it and it would be years before anyone did. I insisted on the USB which they then had to special order.



    It sure was nice not having to connect my game controllers to sound cards anymore and messing with device manager conflicts.
  • Reply 7 of 42
    hmurchisonhmurchison Posts: 12,419member
    Quote:
    Originally Posted by Applecation View Post


    I hope I am wrong, but I fear a repeat of SCSI and FireWire, where cost held back adoption, which in turn kept the cost high. For many people, USB3 is going to be fast enough, without costing $1000 for peripherals and $50 for a cable.

    I was excited when TB was released, and pulled the trigger on a new MBP. That was months ago, and we are only now seeing anything of TB, and too much for my wallet. I find myself wishing my MBP came with eSATA and USB3...



    The important part to realize here is the "difference" between SCSI and Firewire as opposed to eSATA, USB and now Thunderbolt.



    With SCSI you had a technology created by a consortium that that expensive because there was significant intelligence in the controller. Ditto for Firewire.



    The thing that kept SCSI and Firewire expensive where they that were never integrated into mainstream boards.



    Now compare this to



    eSATA - External SATA is just taking a standard SATA bus and hardening it enough to survive outside the case. SATA was already an integrated items on millions of mainboards.



    USB - same thing...integrated into mainstream mainboards and the costs go down.



    So what makes Thunderbolt different than SCSI or Firewire.



    A. It's an Intel technology which means it's going to hit mainstream mainboards

    B. It leverages existing PCI Express technology that is already proven (a la SATA tech)



    So while Thunderbolt "feels" like older technologies that didn't catch on a deeper look into who and what are behind the format yields significant differences. Differences that will make Thunderbolt more of a mainstream product than SCSI or Firewire could dream of.
  • Reply 8 of 42
    Quote:
    Originally Posted by hmurchison View Post


    The important part to realize here is the "difference" between

    ..... Differences that will make Thunderbolt more of a mainstream product than SCSI or Firewire could dream of.



    True. Once Intel integrated them into their next motherboard revision, it will get mainstream. Also significant, so far only devices are external devices for enterprises hence RAID solutions. That's why they are not cheap. I know tweeter gives you instant reply etc. but just bide on this one. Soon every Macs will have these which means it won't feel and cost like early adopters penalty.
  • Reply 9 of 42
    jeffdmjeffdm Posts: 12,951member
    Quote:
    Originally Posted by hmurchison View Post


    So while Thunderbolt "feels" like older technologies that didn't catch on a deeper look into who and what are behind the format yields significant differences. Differences that will make Thunderbolt more of a mainstream product than SCSI or Firewire could dream of.



    Still, the integration into main boards doesn't help that there's a lot more technology required in the cable and on the peripheral end. Also, USB was built into main boards long before it actually got any traction, the low price of the host connector wasn't enough, it required a considerable kick start elsewhere.



    Another issue is this is the first try at a consumer connection that requires electronic circuitry in the cable. Being a complicated and high spec cable and protocol, it seems like it's going to take considerable time to establish itself.



    I'm hopeful, but it pays to limit one's enthusiasm on a new technology, it will take time to get sorted out.
  • Reply 10 of 42
    addaboxaddabox Posts: 12,665member
    Quote:
    Originally Posted by Applecation View Post


    I hope I am wrong, but I fear a repeat of SCSI and FireWire, where cost held back adoption, which in turn kept the cost high. For many people, USB3 is going to be fast enough, without costing $1000 for peripherals and $50 for a cable.

    I was excited when TB was released, and pulled the trigger on a new MBP. That was months ago, and we are only now seeing anything of TB, and too much for my wallet. I find myself wishing my MBP came with eSATA and USB3...



    Quote:
    Originally Posted by Tallest Skil View Post


    You realize this is the first cable ever made, right? USB cables were expensive in 1997, too.



    Hell, a 6' USB 2 cable from BestBuy costs $25 right now.



    Sure you can get them for far less online, but $50 for the first cable from Apple, including the extra silicon, doesn't sound bad at all, considering USB 3 cables are probably going to be going for near that at places like BestBuy, at least at first.
  • Reply 11 of 42
    jb510jb510 Posts: 129member
    Why is this surprising people? I thought Apple said that these copper cables would eventually be replaced by fiber cables using the same connector because the transceivers would be in the cable...



    So presumably could by a 30m 'optical' thunderbolt cable someday and plug it into the same connector as used by the 2m copper cable.
  • Reply 12 of 42
    hmurchisonhmurchison Posts: 12,419member
    Quote:
    Originally Posted by JeffDM View Post


    Still, the integration into main boards doesn't help that there's a lot more technology required in the cable and on the peripheral end. Also, USB was built into main boards long before it actually got any traction, the low price of the host connector wasn't enough, it required a considerable kick start elsewhere.



    Another issue is this is the first try at a consumer connection that requires electronic circuitry in the cable. Being a complicated and high spec cable and protocol, it seems like it's going to take considerable time to establish itself.



    I'm hopeful, but it pays to limit one's enthusiasm on a new technology, it will take time to get sorted out.



    It's not going to be cheap overnight but it'll get there. It's too flexible being based on PCI Express there's not a whole of of software expertise that's going to be required but it "is" an intelligent connection so it's never going to be USB cheap (which is good because USB just uses your CPU for most read/write tasks)



    Intel has confirmed that Ivy Bridge in 2012 won't have integrated Thunderbolt . 2013 is a good bet (I added that last part)
  • Reply 13 of 42
    dunksdunks Posts: 1,254member
    Quote:
    Originally Posted by jb510 View Post


    Why is this surprising people? I thought Apple said that these copper cables would eventually be replaced by fiber cables using the same connector because the transceivers would be in the cable...



    So presumably could by a 30m 'optical' thunderbolt cable someday and plug it into the same connector as used by the 2m copper cable.



    Smart move on Apple's part. I initially thought that going with copper now would cause all kinds of headaches in the future. But if the current ports will truly be forward compatible this will let them get the technology out now at price that can compete with USB3 and allow them to defer the optical technology until a time dictated by the market.
  • Reply 14 of 42
    sflocalsflocal Posts: 6,092member
    Quote:
    Originally Posted by jb510 View Post


    Why is this surprising people? I thought Apple said that these copper cables would eventually be replaced by fiber cables using the same connector because the transceivers would be in the cable...



    So presumably could by a 30m 'optical' thunderbolt cable someday and plug it into the same connector as used by the 2m copper cable.



    It only makes sense. If the circuitry was on the motherboard, the optical interface would most certainly fail over time due to constant plugging, unplugging and being susceptible to dirt. By placing the transceivers in the cable and eventually integrating the optical components internally, and sealed, it really creates a durable environment. Me like a lot. Makes perfect sense.
  • Reply 15 of 42
    jeffdmjeffdm Posts: 12,951member
    Quote:
    Originally Posted by hmurchison View Post


    (which is good because USB just uses your CPU for most read/write tasks)



    I've seen this repeated a lot, but let's be honest here, is that actually a noticeable problem? The CPU usage has always been negligible at best whenever I check.
  • Reply 16 of 42
    jeffdmjeffdm Posts: 12,951member
    Quote:
    Originally Posted by jb510 View Post


    Why is this surprising people? I thought Apple said that these copper cables would eventually be replaced by fiber cables using the same connector because the transceivers would be in the cable...



    Maybe because your assumption doesn't necessarily follow in the way you think it does. Also, there was a paltry amount of useful information available. I've never seen such a comment made by Apple reps or engineers.



    A transceiver in the optical cable would have been required to make it work without an optical connector, optical connectors probably would have been prohibitively expensive and unreliable for consumer devices that get connected and disconnected a lot. Until yesterday, I've never seen anyone mention anything about the electrical cable requiring a transceiver built into it. It is usually the case that electrical transceivers were in the device, not the cable.



    I am somewhat frustrated that there's very little in between the sparse press releases and blog posts and the dense data sheets and developers notes. I always feel like I fall in between them, I don't really care to know enough to know how to build them, but I want to know more than what's readily available too.
  • Reply 17 of 42
    hmurchisonhmurchison Posts: 12,419member
    Quote:
    Originally Posted by JeffDM View Post


    I've seen this repeated a lot, but let's be honest here, is that actually a noticeable problem? The CPU usage has always been negligible at best whenever I check.



    All things being considered i'd rather my CPU handle applications and UI and lead the read/write stuff to a smarter controller.



    Where I've noticed it is in connecting a broadband device to my computer via USB vs Ethernet. The USB always had a bit of a lag. I don't know if it was poor drivers or what but the difference was there.
  • Reply 18 of 42
    jeffdmjeffdm Posts: 12,951member
    Quote:
    Originally Posted by hmurchison View Post


    All things being considered i'd rather my CPU handle applications and UI and lead the read/write stuff to a smarter controller.



    Where I've noticed it is in connecting a broadband device to my computer via USB vs Ethernet. The USB always had a bit of a lag. I don't know if it was poor drivers or what but the difference was there.



    I would say there are too many variables at play in that example to blame the connector protocol. It may be a factor, but nailing it down looks pretty hard.



    When the CPU usage doesn't even make a blip at full speed and doesn't perceptually alter the rest of the computer's performance, it's really hard to justify using CPU usage as knock against the USB protocol.
  • Reply 19 of 42
    guinnessguinness Posts: 473member
    I think TB will have a harder time gaining traction, because you'll need a whole new motherboard to support it.



    USB or FW support could be added with a PCI card, Expresscard...this is useless because I don't plan on upgrading my computer (PC) for another 3-4 years, and I'm dubious with my Mac (3.5 year old Mini), because I'm not sure where Apple is going with the Mini, or that I even want a new one.



    As for firmware in cables, it reminds me of something out of the pages of Bluray ("you must update your cable's firmware before a connection can be established...").



    Eventually, it could catch on, when everyone buys a new computer, but ports come and go, see FW (for most consumers), and it took ages for USB to get established.
  • Reply 20 of 42
    One aspect of this new Thunderbolt [TB] cable/connector, at least from the images posted, is that the plug end itself is rather long. Of course, because it has circuit board and chips it must be larger than a typical plug.



    But my concern is that the plug will now exert a non-trivial amount of torque or leverage or up/down stressing onto the TB socket motherboard connector.

    Not just when plugged in or removed, but from general cable movement where the weight of the cable or shifting of devices may start to cause either loose signal connection and/or wear and tear on the TB socket's attachment to the motherboard.



    That's been a concern of mine with my Mini's video adapter to my DVI monitor cable, so I've been glad to have the adapter which uses a short cable length between plugs. I have worked on a number of computers (mostly notebook/laptop) that have had connector ports fail because of physical wear and tear on the socket.

    (and I remember the 'old days' where most every connector --vga,serial,parallel,etc-- had two little screws to hold them onto the computer socket. For both PCs and Macs.)



    I am hoping that the Apple mobo designers are making the socket connection especially resilient and strong to take care of the issues related to the inevitable extra stress.
Sign In or Register to comment.