Slightly Smaller Apple TV 3,2 (A1469) Contains A5X SoC, BCM4334 Combo
by Brian Klug on January 29, 2013 6:33 PM ESTYesterday along with iOS 6.1 being pushed to iPhones, iPads, and iPod Touches, Apple released an update to the Apple TV 2,1 and 3,1, and curiously enough released another image for an unannounced Apple TV 3,2 product with model A1469. Originally it was assumed this was an international release with minimal changes, however it turns out there's much more inside, including an A5X SoC and BCM4334 (up from BCM4330) WLAN and BT.
Matthew Panzarino from The Next Web pinged me about the FCC documents posting and whether there were any changes inside, which piqued my interested and made me dig a little deeper. First, the Apple TV 3,2 model is slightly smaller, as noted by Engadget. Inside however from what we can glean from the FCC documents there's no longer a two antenna solution for the BCM4330 WiFi+BT combo, but rather a single antenna solution with different gain and utilizing a BCM4334 combo instead. This is the same WLAN+BT combo as used in the iPhone 5 and basically all of the newly refreshed iDevices, and is built on a 40nm RF CMOS process. Since the previous antennas were a part of the Apple TV PCB, and the FCC disclosure notes that the new Apple TV uses a single PIFA (Planar Inverted F Antenna), it's highly likely we're looking at a completely different PCB.
That fact intrigued me and made me take a look inside the Apple TV 3,2 IPSW, which contained exactly what I was looking for. Instead of the A5R2 SoC (S5L8942) inside the Apple TV 3,1, this new device contains an A5XR2 SoC (S5L8947) as shown in the screenshots I've taken of the Restore.plist file, though there are numerous others. It's entirely possible that Apple is again using different bins of the A5X, it's not possible to tell whether CPU or GPU cores are fused off at this point from my digging through the IPSW. Update: I should note that S5L8945 is the original 45nm Samsung produced A5X, S5L8947 is the A5XR2 that is in this unannounced Apple TV. Note that previously the A5 underwent a similar plus 2 offset, from S5L8940 (A5) to S5L8942 (A5R2) which changed process geometry from 45nm to 32nm HK-MG at Samsung. It's possible we're also looking at a process node change with the A5X to A5XR2 here.
This is an interesting silent update and development for the Apple TV. Given the IPSW release onto the web before an actual hardware announcement, I originally suspected this was a silent upgrade of the Apple TV platform to newer hardware, however the presence of an A5X SoC inside makes things a bit more curious. Hopefully we will see an official announcement in the next few days.
Facebook
Twitter
RSS
16 Comments
View All Comments
boostern - Tuesday, January 29, 2013 - link
Maybe Apple is going to teasform the AppleTV into a gaming console?Bit then the question is...why not A6X? I think you are correct when you say recycling...
name99 - Tuesday, January 29, 2013 - link
ONE model of Apple getting into gaming would be that AppleTV is the console, and you play games with dumb controllers --- whether D-Pad, or Wii, or even Kinect.But that's an old model, and why reinvent the wheel?
Another model is that the AppleTV is essentially a co-ordinator/server/display controller, but the bulk of the computation is happening on each user's controller (which is of course an iPad or iPhone). In such a model, there is less need for ultimate CPU/GPU performance in the Apple TV box because most of the work is done in the iPad/iPhone.
Obviously this second model requires all users to own an iPad/iPhone/iPod Touch (which makes Apple happy), but it also allows for much more total compute power to be thrown at the problem. On the other hand, it's harder to program for (though this type of programming has become pretty mainstream in gaming, I'd imagine) and, bigger problem perhaps, it does's immediately lend itself to unusual game-specific peripherals (rifles, steering wheels, rock band instruments, etc). However it would not be hard to have those peripherals connect to the "owning" iOS device by BlueTooth.
Point is, to the extent that Apple wants to go down this path (of being a living room game hub), I see they will do it by requiring each user to own their own iOS device, and so they don't need as much computational capability in the Apple TV as a traditional console.
inighthawki - Tuesday, January 29, 2013 - link
Such a programming model is not mainstream, and would be very difficult. The ability for each controller to be the powerhouse would effectively require the main console to be a server and have all of the players as network clients. Hardly an easy feat for indie developers. Sure your big name game companies won't have much of a problem making their engine work, but the hassle of ensuring a solid network layer and bug-free gameplay is far from ideal.smalM - Tuesday, January 29, 2013 - link
As S5L8945 is the original 45nm A5X, S5L8947 is most likely the die shrunk 32nm "A5XR2".ltcommanderdata - Tuesday, January 29, 2013 - link
I wonder if this is Apple's 28nm HPM TSMC test chip? Assuming the iPhone 5S will have a June WWDC and will be using a 28nm HPM TSMC A7 SoC, the A6/A6X design might have been finished too late to be used as a test vehicle and have the results fed back into A7 development. The A5 wouldn't be a very good test since it'd be a very small chip at 28nm. The A5X is a good choice since it's a known design while being a big die chip for a harder test.On a side note, I think the realignment of the iPhone and iPad launches are largely driven by SoC development. Unlike the iPad 2/iPhone 4 A5 SoC days, the resolution difference between the devices means they can no longer share a chip. It makes sense that the iPad uses the iPhone's SoC with an enlarged SoC. That necessitates the iPhone launching first each calender year and the iPad following. So I think the iPhone launching mid-year with an A_ SoC followed by the iPad launching September/October with a derivative A_X SoC will be the pattern going forward.
Brian Klug - Tuesday, January 29, 2013 - link
You are correct, I forgot to note the +2 change and the corresponding possibility of a die shrink/process node change. I've noted it above. :)-Brian
AdamChew - Wednesday, January 30, 2013 - link
Yes apple have stopped innovating and now only doing recycling whereas others are innovating leaps and bounds.Sad isn't it 80k people at Cupertino don't know what to do and all the pundits are innovating like mad.
rangerdavid - Wednesday, January 30, 2013 - link
An acorn hits you on the head, and the sky is falling? I've seen other statements like this, and they puzzle me. Let me see if I understand this correctly: Apple will never release a new, innovative, or competitive product again.Great. Can I buy your AAPL stock now?
But seriously, Apple will remain competitive for years and years. And you should thank them for initiating this wonderful technology horse race, Apple vs. Google vs ?Microsoft? (marginally). Apple is doing just fine.
I'm waiting for my wearable Siri^2-enabled Oakley-cobranded augmented-reality iSunglasses.
darwinosx - Wednesday, January 30, 2013 - link
He is joking, making fun of the pundits...lmcd - Friday, February 1, 2013 - link
Probably excess stock.