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.
Apple Thunderbolt Display Teardown Reveals Plethora of Hardware Inside
- Thread starter MacRumors
- Start date
- Sort by reaction score
I'm wondering the same thing. *looks at his 27" 2011 iMac*
Same.
http://www.chargedpc.com/news/2/New-2011-iMac-Teardown-Part-1.html
http://www.chargedpc.com/product_images/uploaded_images/dsc-0668.jpg
I'd rather pay another $100 and get the Dell UltraSharp U2711. Its literally the same panel but without the stupid Apple Mirror Coated Screen
It is a matte screen... with a piece of glass in front of it. Just do as iFixit and Anand did an pull the glass off with suction cups. It'll look less attractive, but hey, that's why Apple put the glass in front of in the first place, for aesthetics.
Someone should make bezels for these things out of some type of ferrous metal that will just stick to the magnets that are already there for holding the glass on. Cheap as hell 2 minute conversion to matte display to shut all the whiners up.
You can petition Apple all you want, and I'm sure they'll give your concerns due consideration. Unfortunately, sales of Apple devices with glass clad displays are up several million units over their matte predecessors, and those that petitioned with their dollars probably have the stronger lobby.
Or buy suction cups for $5 and take the glass off yourself
Or apple could just sell a proper display with matte finish instead of hacking away at a poorly executed display?
----------
You got any links or images as to how the monitor looks without the glass covering?
Im just curious. But as someone else mentioned, these glossy screens have no place being in front of any panel, especially IPS and others.
Apple made a display that they hope will sell well in the market—like most for-profit companies that produce goods. Their whole current design language revolves around aluminum and glass, which is being received quite well compared to all the plastic out there. There is no good way to make a piece of glass matte. You're welcome to apply any of the antiglare films which are readily available, but there are good reasons why Apple doesn't do this. There's also plenty of competition in the display arena, so if you think this display is generally poorly executed, look elsewhere. If you love this display, but really can't live with the glass, then just remove it.
You could try the iFixit post that the original article links to for one, or the Anandtech teardown that was also referenced earlier: http://www.anandtech.com/show/4832/the-apple-thunderbolt-display-review/11
U2711: 6000 SEK (7800 Dell Price)
Apple Cinema Display: 8000 SEK (ex. 9500 Apple Store)
Apple Thunderbolt Display: 8900 SEK (9500 Apple Store)
U3011: 10000 SEK
I think I'll continue.
Last edited:
Apple Thunderbolt Display Teardown Reveals Plethora of Hardware Inside
What, exactly, did you expect to find? This has got to be one of the dumbest post titles evah!
What, exactly, did you expect to find? This has got to be one of the dumbest post titles evah!
Actually, it's a glossy screen with a piece of glass to make it even more glossy.
Dells are just matte. With multiple switchable inputs. So, yeah. More features, less money. But I don't think Apples displays are remotely directed at a rational buyer.
Unfortunetely, you're using the line speeds in your calculations, and not including the overhead.
There's a reason that USBIF's specs say that USB2.0 will never reach any more than 40MB/sec (320Mb/sec) in actual data transferred. Likewise, they expect USB 3.0 to get around 3.2 Gb/sec.
Given that PCIe 2.0 uses the 8b/10b encoding, your overhead is at least 20%, so a 5 Gb/sec lane will never get more than 4 Gb/sec.
Now, you need to encapsulate XHCI packets for the USB controller and run them over PCIe. Which incurs more overhead, of which I'm not sure how much there is, but there's definitely latency involved. Using a guess of 10% overhead for this encapsulation results in 3.6Gb/sec available from PCIe 2.0 1x to transport the expected 3.2Gb/sec of a single port of USB3.
But wait, there's more. Since I don't know the encoding for Thunderbolt, there's another layer of overhead. We certainly have the bandwidth, but the latency is going to be pretty bad, which slows things down. (think about why 2G cell service sucks so bad for web browsing and games, even though it's got more bandwidth than a 56k modem. latency matters)
At this point in time you'll have several possible problems to look at:
1) We might have hit a bandwidth restriction that makes it not possible to get USB 3.0 at full speeds once passed through Thunderbolt.
2) We might have hit latency problems which make the USB port unusable for a certain class of devices (greetz to all you audio and video engineers).
3) We're also assuming that the USB 3.0 chips are flawless. Fat chance
4) All the calculations also assume we've only got 1 USB 3.0 port. And given what the display has, that's the most we could possibly get. Another back-of-the-envelope calculation: Given that Thunderbolt's line speed is 20Gb/sec, you'll need to subtract out the Displayport signal (8-10Gb/sec). Leaving you with about 10Gb/sec before overhead. If you're trying to keep all the stuff in that the display has, you don't have enough for more than 1 USB 3.0 port.
I'm definitely aware that I don't have measured numbers for every stage of getting a USB 3.0 port on to a product like this. But the quick and dirty calculations say that it's either a really tight fit or compromised.... assuming perfectly implemented 3rd party controllers (yeah right).
If you want USB 3.0 over thunderbolt, might want to start without the external monitor or wait until thunderbolt channels get faster.
Sure, if they put it on there compromised, people will be happy that "it's got USB3"... until they benchmark it against a future desktop and ask why it's slower. We all understand how Apple releases products, sometimes they'd rather just skip out on the feature if it isn't perfect. And some other times, there's MobileMe....
This would make replacing a damaged monitor cable MUCH easier for a repair tech. No soldering or undoing tiny ribbon-connectors involved. Might be something done to make sure a poor solder job or other labor-related task doesn't effect TB performance. Bad cable comes off, new one snaps in, cover replaced. Good-to-go.
Edit: Of course, I also note that Apple could have just put the TB port on the back of the display and supplied the cable as a separate in-box item. But then if you damaged your cable you might be able to (GASP!) replace it yourself instead of having to pay someone if you're squeamish about cracking open your display.
The awesome thing about being able to charge your portable macs off the screen is that they need charging. You can set it down in front of the display, hook it up, and you're good. If you needed to charge, you're charging, if you didn't, you're not going to drain your battery. You don't need to hunt for your charger, or look for a free outlet, or climb behind your desk to plug in your charger, only to do the reverse to unplug it when you leave.
I'd say for at least 95% of people, a mini is not going to be used as a portable computer, so you don't have to worry about frequently messing around with the power cord. A mini doesn't have a battery to fall back on, either, so when the magsafe is tugged and detaches (as it's designed to do) your mini would fail in the middle of whatever you were doing. That isn't good.
No, it looks like a glossy screen. Compare the MacBook Air next to a glassy MacBook Pro. That's what a Mac looks like without the glass. A significant reduction in glare.
And I say you're wrong. I like my glossy screens. All of them. The one on my MBA, the one on my iMac, and the two on my 23" Dell Ultras on my desk at work. Guess what. Obviously a lot of other people do, also. A lot more people are buying them and using them (without issue) than whining. Go buy and ugly matte screen in another brand. But do us all a favor and give it a rest.
It's really quite simple - instead of all the whinging about the glossy screen on the Thunderbolt display, if you don't like it then don't buy it, after all you are under no obligation whatsoever to do so.
I bought one on Monday and I'm delighted with it.
I bought one on Monday and I'm delighted with it.
Very informative as always. I always look forward to iFixit teardowns following new hardware releases
Well, here in the US, which is where the commentor is from I think:
Dell U3011 - $1,499 (currently disounted to $1,299)
Dell U2711 - $1,099
Apple Thunderbolt Display - $999
So for us, the Apple unit is cheaper.
In the iFixit post they use the word "plethora" and then draw attention to the fact that they have done so. The title was merely riffing off of that.
Actually, it's a glossy screen with a piece of glass to make it even more glossy.
Dells are just matte. With multiple switchable inputs. So, yeah. More features, less money. But I don't think Apples displays are remotely directed at a rational buyer.
After looking at those pictures again, it think it may be somewhere in between. Almost all panels ship with a protective film of some sort applied to them, which traditionally was always matte or antiglare. Then the "glossy" fad came about and some OEMs started using glossy films instead, arguing that at higher pixel densities the matte films blurred small text (reflections and fingerprints be damned). Thus the whole glossy vs. matte war was launched. I'm guessing the panel in the ATD has no protective film on it, because it doesn't need protection when it's trapped under glass. The Dell's definitely have a matte film applied, which some owners have disliked enough to remove. The MBA has a glossy film, which I assume is just to make the cheaper, non glass-covered TN panel look consistent with the rest of the product line.
As for the Dells having more features, that all depends on what you really consider a feature. Dual Mode DisplayPort devices allow OEM's to add HDMI, DVI and DisplayPort connectivity easily and with little additional cost. If you consider having all three of these input ports, plus an additional DVI-D input, plus VGA, component, and composite analog inputs all on your monitor a feature, even though only one of them can actually drive the display at a time, then yeah. I guess you won't need to use an adapter very often, but it doesn't add any additional functionality for the modest amount it adds to the cost of the display. Height adjust and swivelyep, I'll agree that those are features. A built in 4 port USB hub and memory card readerO.K., I guess. But when you plug a MacBook Air into a Dell monitor, it sure as hell doesn't give you Gigabit Ethernet, FireWire 800, a 720p video camera, microphone, 49W 2.1 sound system, a second USB 2.0 host controller, and a second power adapter.
And as for Dell's being cheaper, the U2711 has been out for some time now and the current list price is $1099, or $100 more than the ATD. The cheapest etailer pricing I see is $840 for the U2711 vs. $935 for the ATD, so only $95 (10%) less. If you're the owner of a 2011 MacBook, which would you go for?
Thunderbolt cables are active and essentially house the PHY in the cable connector, thus you really need them to be terminated at both ends for them to work. I totally agree about making the cable detachable. I hate carrying displays with damn cables dangling off of them. I even asked an employee of my local Apple Store if both cables were detachable and he assured me that they were...
Is it just in my backwoods area, or are all Apple Store employees universally devoid of any product knowledge? I was buying a toslink cable there the other day, one that comes with the 3.5mm mini optical adapters so you can use it with your Mac, and when I asked an employee if they had any other lengths available he replied, "No, we just carry that one as a matter of convenience to our customers, since there aren't really any Apple products that use that type of connection." Umm... yeah, buddy. I also love how 80% of the questions I ask there are met with "Let's Google it and find out!" Which inevitably leads to them showing me a MacRumors thread with 350 posts that I've already read and saying, "See, the answer is right here!"
Unfortunetely, you're using the line speeds in your calculations, and not including the overhead.
There's a reason that USBIF's specs say that USB2.0 will never reach any more than 40MB/sec (320Mb/sec) in actual data transferred. Likewise, they expect USB 3.0 to get around 3.2 Gb/sec.
Given that PCIe 2.0 uses the 8b/10b encoding, your overhead is at least 20%, so a 5 Gb/sec lane will never get more than 4 Gb/sec.
Now, you need to encapsulate XHCI packets for the USB controller and run them over PCIe. Which incurs more overhead, of which I'm not sure how much there is, but there's definitely latency involved. Using a guess of 10% overhead for this encapsulation results in 3.6Gb/sec available from PCIe 2.0 1x to transport the expected 3.2Gb/sec of a single port of USB3.
But wait, there's more. Since I don't know the encoding for Thunderbolt, there's another layer of overhead. We certainly have the bandwidth, but the latency is going to be pretty bad, which slows things down. (think about why 2G cell service sucks so bad for web browsing and games, even though it's got more bandwidth than a 56k modem. latency matters)
At this point in time you'll have several possible problems to look at:
1) We might have hit a bandwidth restriction that makes it not possible to get USB 3.0 at full speeds once passed through Thunderbolt.
2) We might have hit latency problems which make the USB port unusable for a certain class of devices (greetz to all you audio and video engineers).
3) We're also assuming that the USB 3.0 chips are flawless. Fat chance
4) All the calculations also assume we've only got 1 USB 3.0 port. And given what the display has, that's the most we could possibly get. Another back-of-the-envelope calculation: Given that Thunderbolt's line speed is 20Gb/sec, you'll need to subtract out the Displayport signal (8-10Gb/sec). Leaving you with about 10Gb/sec before overhead. If you're trying to keep all the stuff in that the display has, you don't have enough for more than 1 USB 3.0 port.
I'm definitely aware that I don't have measured numbers for every stage of getting a USB 3.0 port on to a product like this. But the quick and dirty calculations say that it's either a really tight fit or compromised.... assuming perfectly implemented 3rd party controllers (yeah right).
If you want USB 3.0 over thunderbolt, might want to start without the external monitor or wait until thunderbolt channels get faster.
Sure, if they put it on there compromised, people will be happy that "it's got USB3"... until they benchmark it against a future desktop and ask why it's slower. We all understand how Apple releases products, sometimes they'd rather just skip out on the feature if it isn't perfect. And some other times, there's MobileMe....
Both USB 3.0 and PCIe 2.0 use a nominal data rate of 5 Gbps and 8b/10b encoding, so they both offer 500 MB/s to the lower layers. Most, if not all, of the currently shipping standalone USB 3.0 host controllers have 2 or 4 downstream ports and a PCIe 2.0 x1 back end. The overhead and latency associated with the USB SuperSpeed protocol is generally considerably higher than that of PCIe, so there shouldn't be any issue there. The only time PCIe becomes a problem for USB 3.0 is if you're using a PCIe 1.0 x1 connection, or a PCIe 2.0 x1 lane that is coming off of an overtaxed PCH that is bottlenecked by a 20 Gbps DMI. Thunderbolt provides 10 Gbps per channel to the underlying protocol layers (sans 8b/10b) and has ridiculously low latency, so it shouldn't be an issue either.
I was previously convinced that each Thunderbolt device could only make use of a single channel, and that the DisplayPort and PCIe packets from the ATD were confined to one 10 Gbps pipe, thus making USB 3.0 a non-starter. I now believe this to be incorrect. I think that each of the protocol adapters within the Thunderbolt host controller is connected to the built in Thunderbolt switch via a single channel. This would mean that display and data are being switched separately, and USB 3.0 would have the necessary headroom. However, despite the fact that all of the Thunderbolt controllers we have seen thus far have a PCIe 2.0 x4 back end, I'm pretty sure they only have a single PCIe protocol adapter, and thus all attached devices are limited to a combined maximum throughput of 1250 MB/s. While still plenty for USB 3.0, it's less than I had hoped. With the bare minimum amount of PCIe overhead on current Sandy Bridge systems that works out to 1028 MB/s.
There are still plenty of good reasons to wait for Ivy Bridge before jumping on the USB 3.0 bandwagon though. The available controllers are steadily dropping in size, power usage and cost, but the drivers are still super spotty. I'm pretty sure Apple wants to make sure it's USB 3.0 drivers are flawless when Ivy Bridge hits in a few months, rather than pour resources into developing and releasing drivers for early, third-party host controller silicon.
Remove Glass ≠ Matte screen
Having removed the glass from hundreds of iMacs for service. I can guarantee that removing the glass does not reduce the reflectivity. The face of the LCD inside is also glass (it also scratches VERY easily). Matte LCDs have a plastic layer or coating over this glass to provide the matte finish and scratch resistance. The glass layer on the LCD is usually about the thickness of 3-5 sheets of paper, leaving it also prone to cracking under any stress. While I agree that all of this glass bouncing light around in front of our LCDs is hideous, it does provide some protection from breaking the actual LCD (I replace the glass on MacBook (Pro) and iMac MUCH more often than replacing the LCD, which is usually a multi-hundred dollar part. New glass is cheaper to replace if/when damage occurs.
No, the only way to get a real matte finish iMac would be to replace the internal LCD with one manufactured as matte and then leave off the glass. IMO, however, you're setting your self up for breaking your new LCD whenever you remove the glass. When I worked for Apple and the new aluminum iMacs were released, the Geniuses were scratching and cracking the LCDs on nearly every one until they had a few under their belt. Even then, only a few of us would fix them because of the potential for damage during repair.
I would not recommend that anyone remove the glass from the ATD, ALCD, or iMac, unless you enjoy replacing $999 LCDs.
Having removed the glass from hundreds of iMacs for service. I can guarantee that removing the glass does not reduce the reflectivity. The face of the LCD inside is also glass (it also scratches VERY easily). Matte LCDs have a plastic layer or coating over this glass to provide the matte finish and scratch resistance. The glass layer on the LCD is usually about the thickness of 3-5 sheets of paper, leaving it also prone to cracking under any stress. While I agree that all of this glass bouncing light around in front of our LCDs is hideous, it does provide some protection from breaking the actual LCD (I replace the glass on MacBook (Pro) and iMac MUCH more often than replacing the LCD, which is usually a multi-hundred dollar part. New glass is cheaper to replace if/when damage occurs.
No, the only way to get a real matte finish iMac would be to replace the internal LCD with one manufactured as matte and then leave off the glass. IMO, however, you're setting your self up for breaking your new LCD whenever you remove the glass. When I worked for Apple and the new aluminum iMacs were released, the Geniuses were scratching and cracking the LCDs on nearly every one until they had a few under their belt. Even then, only a few of us would fix them because of the potential for damage during repair.
I would not recommend that anyone remove the glass from the ATD, ALCD, or iMac, unless you enjoy replacing $999 LCDs.