Why couldn't apple just put those chips in the computer themselves.
My gut feeling was so they have an excuse to charge an arm and a leg for the cables.
My gut feeling was so they have an excuse to charge an arm and a leg for the cables.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Thunderbolt Cable Teardown Reveals Electronics and Firmware
- Thread starter MacRumors
- Start date
- Sort by reaction score
Do you know if it is possible to go Macbook Pro -> iMac -> iMac, in serial, where each iMac is in target display mode?
I suspect that 10Gbps won't drive dual 2560 at 60Hz, but technically is it possible? Perhaps with a reduced refresh?
Sorta. There's nothing optical about your MBP's ThunderBolt port. What they're referencing is a comment from someone at Intel indicating that when they do start to use optical cables for ThunderBolt, the cables can or will use normal, non-optical connectors on the ends, and the components needed to translate to and from an optical signal will be on the cable itself -- not on the devices.
No. The signalling on copper cables is not the only limiting factor: the whole ThunderBolt system has a limited amount of bandwidth assigned to it (a certain number of PCI-Express lanes) by the system's chipset, and these potential optical cables would still be connected via an electrical mini-DisplayPort style connector on either end, and the deployed circuitry for those might also have some limitations.
The only difference such an optical cable is likely to make for current ThunderBolt-enabled machines is that optical cables can be much longer. It might also help push them closer to the current limits of the spec, but that's all.
The other point is that there might come a day when devices are using a newer, faster version of ThunderBolt, that requires the cables use optical signaling to achieve higher data rates. If Intel sticks to its plan, what this means is that you'd be able to connect a newer optical-required-for-full-speed device to a device from today, and they could still communicate, but you'd be limited to today's speeds.
So, basically the same as every other backwards-compatible bus connector that's ever been made.
Neither ThunderBolt nor fibre are magic. They won't be able to make your computer suddenly faster several years from now when next-gen ThunderBolt devices start to appear.
Because the chips help with handling attenuation on the cable, and it's part of the Thunderbolt spec. Without the chips, reaching 20Gbps over copper is incredibly hard.
PS: Read http://news.ycombinator.com/item?id=2711918 for a better understanding.
The chips are tuned for the length and electrical characteristics of the cable. Longer cables need different firmware. They could have put the firmware in the cable and left the chips in the computer and peripherals, but then they would have problems when they shift to optical cables.
wow, I really wasn't expecting that one. At 10Gbps those are not particularly low power. Or cheap.
This reinforces my theory that Intel developed Light Peak with some amount of involvement with Apple. SJ sees this product and thinks "this is amazing - if I'm going to have cables than just this *one*" and forces Intel to bring out Thunderbolt, a copper implementation of Light Peak. Intel didn't back USB3 initially, because they figured they could get Light Peak to work (optically), but Intel couldn't find a cheap/reliable optical way of doing it, so they design a copper based solution, with the assumption that in the future someone will figure out how to do optical interconnects cheaply, thereby making it necessary to put a O/E converter in the cables.
But this doesn't explain why they didn't integrate the data recovery/equaliser into their Thunderbolt chip... The current solution seems like an expensive (and power hungry) kludge.
For those that are interested, National Semiconductor has been doing what I guess are similar chips for years:
http://www.national.com/mpf/DS/DS50EV401.html#Overview
(the Gennum datasheet is under NDA)
This reinforces my theory that Intel developed Light Peak with some amount of involvement with Apple. SJ sees this product and thinks "this is amazing - if I'm going to have cables than just this *one*" and forces Intel to bring out Thunderbolt, a copper implementation of Light Peak. Intel didn't back USB3 initially, because they figured they could get Light Peak to work (optically), but Intel couldn't find a cheap/reliable optical way of doing it, so they design a copper based solution, with the assumption that in the future someone will figure out how to do optical interconnects cheaply, thereby making it necessary to put a O/E converter in the cables.
But this doesn't explain why they didn't integrate the data recovery/equaliser into their Thunderbolt chip... The current solution seems like an expensive (and power hungry) kludge.
For those that are interested, National Semiconductor has been doing what I guess are similar chips for years:
http://www.national.com/mpf/DS/DS50EV401.html#Overview
(the Gennum datasheet is under NDA)
Last edited:
Ok. It works.
Now give me a freaking multi-port TB adapter so I get my firewire back. And keep it under $1000 Apple.
Now give me a freaking multi-port TB adapter so I get my firewire back. And keep it under $1000 Apple.
Still a not $50!
I'm still not convince that this is worth $50! Look at the iPod Shuffle it has more chips and has a memory inside. Those chips are probably added to make us believe that Apple should sell this for $50. That Intel Light Peak demo showed on IDF back in 2009 is just a regular cable as far as I could remember.
I'm still not convince that this is worth $50! Look at the iPod Shuffle it has more chips and has a memory inside. Those chips are probably added to make us believe that Apple should sell this for $50. That Intel Light Peak demo showed on IDF back in 2009 is just a regular cable as far as I could remember.
I think we'll see USB 3.0 show up in Macs when Intel adds native support in their chipsets. It's about practicality, not animosity. It's just another standard.Intel 2012 chip will have USB 3.0 support and Apple will be forced to adopt, once this happens TB will live the FW800 status. The cost alone will be the deciding factor along with availability of product on the market.
If it was a good deal yesterday and was something you needed, then it's still a good deal today. FW800 isn't going away anytime soon.
That said, I'm using complete 500GB FW800 drives for about the same money as your enclosure. YMMV.
Because they're fine-tuned to match the exact physical properties of the wire they're directly attached to. That's the whole point: not all wires will be the same.
Also, this sort of signal-correction has to be done at either end of the wires themselves (i.e., inside the cable) -- you can't correct that from the other side of a connector port. It would be like trying to clean the outside surface of a window by washing it from the inside.
Only old DisplayPort monitors have to go at the end of the chain. TB compatible monitors can be anywhere in the chain.
Not really any different than an HDMI cable selling for $79. I'm sure they put the chips in there just to boost the price of the cables.
Wow, you could tell what was inside of a cable by watching a demo? Impressive.
Nobody who knows what they're talking about ever said something like that.
There's a difference between "you can do anything using this connector" and "you will do everything using this". The sales pitch was the former, but you may have misheard it as the latter.
ThunderBolt is never going to replace USB for the things USB was originally designed for -- like connecting a mouse and keyboard -- because there's no way it will ever be remotely close to cheap enough, and connecting things like that over ThunderBolt offers no advantages (and several disadvantages!) over USB.
Think of ThunderBolt as a replacement for FireWire and ExpressCard, not low-end USB.
Gutsy Call
Pretty gutsy call by Apple and Intel! Will this port have a 3 year life? Or will it have the 10+ year life of USB. Will it survive the shift to fibre cables, with the same end ports? Or will it get replaced by a dumb fibre cable with smarter ports for the end devices, the computers and drives?
I guess I will get a TB port for free when I buy my next Mac, so I shouldn't care, but I do have quite a collection of cables that are useless now, and that problem will likely grow forever.
Pretty gutsy call by Apple and Intel! Will this port have a 3 year life? Or will it have the 10+ year life of USB. Will it survive the shift to fibre cables, with the same end ports? Or will it get replaced by a dumb fibre cable with smarter ports for the end devices, the computers and drives?
I guess I will get a TB port for free when I buy my next Mac, so I shouldn't care, but I do have quite a collection of cables that are useless now, and that problem will likely grow forever.
+1
I love OWC's enclosures because they consistently use Oxford chipsets. In this case it's Oxford Semiconductor 944SE
Did they tear apart the cable? A 2009 demo would have likely used an optical cable and transceivers, but I did not look that close..
Fundamentally, it's probably a question of whether or not a current iMac in target display mode implements that by acting as a DisplayPort-only device, or as a ThunderBolt-aware device. And, if it's the latter, does it support using the second TB port to daisy-chain? I have no idea.
But that still wouldn't actually work for this scenario, because according to their specs, current MBPs are capable of supporting a maximum of just 1 external display.
It might be possible that Apple's just saying this to keep things simple, or that a future firmware update could change it. The AMD GPUs in the mid and higher end MBPs are capable of outputting to something like 6 or 7 displays at once thanks to DisplayPort, so I'm not sure where the limitation is.
How many times have we seen this kind of stuff from apple in the past? Not again?
Optical Revolution
There was a company in late 2009 who made a general purpose optical processor, they said scaling it for mass production wasn't difficult. It was running sort of 17,000Ghz but doing simple things - with that speed does it matter if it is a 6502 clone?!?)
Anyway, gallium arsenide CPUs still use silicon and optical will have I/O restrictions* however if there is going to be this sudden lift above the 3.4Ghz barrier then I/O is key and I am glad Intel are thinking ahead.
*My business partner used to program Cray super computers, he said they used a Vax cluster between the Cray and PC network simply as a data buffer! Another friend at London's Imperial College said the Cray would be given 100Mb of data and would spit out a 9Gb text file which they'd spend the next month analysing.
So a bit off-topic but you see where this is going!
There was a company in late 2009 who made a general purpose optical processor, they said scaling it for mass production wasn't difficult. It was running sort of 17,000Ghz but doing simple things - with that speed does it matter if it is a 6502 clone?!?
)Anyway, gallium arsenide CPUs still use silicon and optical will have I/O restrictions* however if there is going to be this sudden lift above the 3.4Ghz barrier then I/O is key and I am glad Intel are thinking ahead.
*My business partner used to program Cray super computers, he said they used a Vax cluster between the Cray and PC network simply as a data buffer! Another friend at London's Imperial College said the Cray would be given 100Mb of data and would spit out a 9Gb text file which they'd spend the next month analysing.
So a bit off-topic but you see where this is going!
If find it funny that people are bitching about price, when I can't see that many people using this technology outside those who really have to move large data around.
Look at how excited people are about iCloud and wifi, so considering this is just another device with a cable I would expect it be used more by Pro shops and not the rest of the population of mac users. I know I have no need for this, its not like having something transfer that fast will make different for the kind of work I do.
Sound like the fangirls are excited about spec but most will never really get much out of this. For people who needed speed for backup it would have been easy to use Ethernet based backups, but I figure most where just using USB2 with a smaller subset using firewire. But its always nice to read how much fantasy and how then how it just to expensive for something most will never get any use from.
Look at how excited people are about iCloud and wifi, so considering this is just another device with a cable I would expect it be used more by Pro shops and not the rest of the population of mac users. I know I have no need for this, its not like having something transfer that fast will make different for the kind of work I do.
Sound like the fangirls are excited about spec but most will never really get much out of this. For people who needed speed for backup it would have been easy to use Ethernet based backups, but I figure most where just using USB2 with a smaller subset using firewire. But its always nice to read how much fantasy and how then how it just to expensive for something most will never get any use from.