The technology inside Apple's $50 Thunderbolt cable | Hacker News Reader

The technology inside Apple's $50 Thunderbolt cable

arstechnica.com

51 pointsmtviewdave23 comments

Apple should just sell the cables at cost (or near-cost) so that they maximize adoption of Thunderbolt. They can make money elsewhere and therefore afford not to profit, but they would gain better adoption in the long-run.

That only looks like a cable. If you think of it as a pair of 10gps NICs and a wire between them it seems like a bargain.

The chips are doing some kind of signal conditioning; they're hardly NICs.

If I had to make an analog to any kind of networking device, it'd probably be a MAU, or possibly an AUI+MAU. https://secure.wikimedia.org/wikipedia/en/wiki/Medium_Attach...

I wonder if it's possible to increase the speeds, without having to revise the chip-on-the-motherboard. Then users could upgrade their speeds by buying a new "faster" cable.

It looks like that's what they're doing (unless I misinterpreted this quote):

> Our telecom source noted that Intel made an unusual choice in also using active cabling for future optical-based iterations of Thunderbolt. Passive cabling is more common, but active cabling could offer some advantages. For one, active cables could combine fiber optics with electrical cabling for power transmission. Another good reason to use active optical cables, according to our source, "is that your current electrical ports can be forward compatible with future optical cables."

Those future optical cables won't be faster. Some people are optimistically misinterpreting Intel and thinking that they'll get 100 Gbps out of existing Thunderbolt ports, which is definitely not true — those ports are already going as fast as they can. Optical only gives you longer distance.

Previous discussion: http://news.ycombinator.com/item?id=2711918

I'm going to go out on a limb and say that Thunderbolt will be more successful than Firewire. It won't trickle down to the cheapo stuff, but it's not designed for that. What it really does is solve the external storage problem for low-profile devices (Air, and eventually maybe even iPad) allowing professionals to finally leave their Mac Pros behind and move to an Air + Thunderbolt Display "docking station" setup that is an order of magnitude increase in freedom and convenience.

Other vendor adoption is not as much of a problem as it was for Firewire (or SCSI for that matter) because Apple's position is so much stronger today. Thunderbolt can go far on Apple's ticket alone. I know I'll be willing to pay a premium for active cabling if it means the data rates hold up to their advertised potential.

I agree that's going out on a limb, a bold move. It seems to me the adoption curve and history of Apple dropping fringe technologies they grow bored with after significant investment by a few firms will keep adoption far below FW rates. USB 3 has supplanted USB 2 for high speed disk access already. That battle is over and done with before Apple has even entered the race. Very high speed external disk access uses eSATA.

For video cameras, it will be USB 3 for everything other than Sony which has its own special precious non-compatible variation Thunderbolt that only works on VAIOs since Sony loves to play the special precious protocol game with their little princess friend Apple.

In 5 years no one will support Thunderbolt and Apple will be on to some new thing. Between now and then there will be at most a dozen devices supported, and all massively overpriced. The failure to adopt USB 3 and BluRay, in addition to their locking down their OS to be iOS with Big Screen will lead to Apple losing significant market share. After the big plunge in a few years, people will say that things all went wrong when Tim "Cost Cutting at any Cost" Cook was foolishly put at the helm rather than someone, anyone, with the slightest design sense, taste, or concern for the customer's experience.

"For video cameras, it will be USB 3 for everything other than Sony which has its own special precious non-compatible variation Thunderbolt that only works on VAIOs since Sony loves to play the special precious protocol game with their little princess friend Apple."

I love this so much.

Your post is very entertaining, but let's bookmark and come back in 5 years.

The reason I don't buy your argument is because I think in practice Thunderbolt is going to be well more than twice as fast as either USB 3.0 or eSATA, and the smart cabling is a future-proofing engineering move that will make sense as drives and busses become faster. If that pans out, then Apple will be in good shape with this technology to not have to fuck with it for the next 5-10 years, and the pros will happily pay the premium.

As for the Lion hate, well, my biggest fear is that people eat it up and they are emboldened to move quickly to an App-Store-only model. Until that day though you still have a respectable Unix + The Best Commercial Apps OS that led so many geeks to Apple over the last decade.

I fully expected the article to show a monstercables-like revelation of nothing but plain copper wire inside. Happily disappointed now.

Any pointers to how these active cables actually work and why it is better to do signal processing in the plug instead of doing it after the plug in the devices?

You can use thinner copper and make longer extensions, because active cabling help compensate for attenuation and crosstalk. Really you can put any kind of DSP you want in there: skew, equalization, etc. DisplayPort is only spec'd for 3 meters over a passive cable, but goes up to 33 meters over an active cable. Also, although I can't remember a source, the Thunderbolt over copper uses about 25% of the power compared to the original optical version.

Also, the chip inside a cable and/or its firmware can be tuned to that cable, and the chip could offer some overcurrent/spike/whatever protection.

Blowing a $50 cable is no fun, but blowing a chip in your laptop is even less fun.

[I do not know whether see cable do either of those]

Basically, the cable is tuned to its specific length. It also reduces cost for people who never use Thunderbolt.

Having active cables also leaves the door open to optical Thunderbolt cabling in the future, without the need for new hardware.

Someone mentioned that you can tune the transceivers to the individual cable, and that's a big part of it. Another thing is that discontinuities in the physical layer are a big deal at these speeds (think connector). It's easier to deal with that 'inside' the plug just a short distance from the discontinuity and then have a much more reliable and predictable link between the two transceivers in the cable. Especially if the transceivers are bumping up the power significantly (I don't know if these ones are), because the discontinuities cause reflections and transmitters will be damaged if the signal reflected back at their output is too high powered. The reflections can now happen at the plug/jack interface, where it doesn't need to be so high powered because the signal will not need to travel as far.

Always get a kick out of the markup on cables. USB cables in bulk are less than $0.50, HDMI cables aren't much more. Profit margins on them are ridiculous.

Did you read the article?

Place some patented IP in the cable to shut out third parties. That's classic Apple innovation.

I thought lightpeak was from intel?

"Our telecom source noted that Intel made an unusual choice in also using active cabling for future optical-based iterations of Thunderbolt ... Intel would not say when official specs would be released to other manufacturers" -- from the article

Reply on news.ycombinator.com