360 Falcon: 50% cheaper to manufacture?
- Thread starter 22psi
- Start date
Love & Sacrifice
Newcomer
More than likely the final phase of the falcon augmentation will signify some form of newness to the consumer. Whether it be a new color. New Design. Or a special sticker on the box.
A smaller form factor dosn't necessarily do away with the original HDD housing and the way the faceplate fits on the system.
I can't see them going with a smaller form factor with this 65nm drop. At the moment the current case can barely handle 90nm chips, I think 65nm should fare better but not to the point where MS will want to start closing in the walls, so to speak.Perhaps at 45nm we might see a 360 slim, though as a previous poster pointed out it would be interesting to see how they would fit the hard drive into any smaller system. It'll be interesting to see if MS implement a new cooling system on the new units, or indeed motherboard layout.
Personally, I'd be very happy if MS managed to source some quieter DVD drives for their 65nm units, one that do not spit disks back onto the disc tray in the event of a crash. Several of my games have light surface scratching from this now. Though I haven't got any fubar'd games from serious scratching.
Personally, I'd be very happy if MS managed to source some quieter DVD drives for their 65nm units, one that do not spit disks back onto the disc tray in the event of a crash. Several of my games have light surface scratching from this now. Though I haven't got any fubar'd games from serious scratching.
Goto commented on that a while ago (though pretty much else of that article turned out to be different). He suspected that integration would not happen until 45nm. So chances for that are pretty slim, I guess.
I kind of agree, I get the feeling that the current form of the 360 is more designed to accomodate the 65nm chips and just barely the 90nm ones...
It still wouldn't be 50% though, because as I said in my first post, we'd need to know the actual die sizes and where they are on the process maturity curve, because it normally takes some time to get the yields up towards their theoretical maximums, which is what would have to happen to yield the arbitrary 50% on those parts... it takes time. Now, if the ancillary components have themselves been reduced to under half their original cost, then ok - but I doubt that's happened. Reduction yes, but to what extent we won't know until we open one of these things up (and that includes the power supply).
For some reason a press base that doesn't regularly cover (or fully understand) process shrinks has latched onto the Falcon and the idea of 50%; yet the Falcon is no different than the multiple shrinks which have taken place with most every major console throughout its life. What captures the imagination here is the codename associated with it, and the fact that it might promise a fundamentally improved user experience.
Yeah, I think they want to get the power brick in the box first as thats probably a very relevant cost savings. Even if the smaller chips in the box would allow a smaller form factor, that space is probably taken up by putting the PS inside.
I'd imagine that the yeilds will be very good right off base. Both CPU and GPU are clocked to sensible yeilds for 90nm processes so they are likely to fall easily withing 65nm's clock ranges, even at lower voltages. The main variance would be defect densities, but overall its not inconcievable that yeilds could actually go up from where they are on 90nm fairly quickly.
Well die defects was definitely what I was referring to, and I'm with you in expecting yields to be higher right off the bat leaving 90nm, just not sure about *twice* as high at this early stage. That depends largely in part on how TSMC's yields on complex ICs is doing at 65nm, and the yield learning on these two products in particular. Though I imagine they're very good overall, I would be surprised if they were pushing the outer bounds this early into its life.
(I'm under the assumption that TSMC has taken over from IBM/Chartered at 65nm for the XeCPU.... or is it still Chartered?)
50% is the straight ratio of gross die per wafer increase assuming a linear shrink - it won't be that good, but it'll be around that range. And when I say yeild increasing I mean in terms of useable die vs gross die per wafer.
WRT to the graphics part, TSMC's 65nm process isn't that new. We're alreeady pushing out more complex chips than Xenos on 65nm and Xenos is just a shrink from an existing architecture.
that should come as absolutely no surprise. I'd imagine that they will dual source between 90nm parts and 65nm parts just in case there is any issues; only when they are sure of the 65nm supply will they cut off wafer orders for 90nm, deplete 90nm inventories and fully transition to 65nm. Its also likely that GPU and CPU are on slightly different tracks so some units may have one part at 65nm and another at 90nm.Apparently, the migration to Falcon is a gradual process that culminates in a smaller form factor.
Love & Sacrifice
Newcomer
You are correct---The GPU 65nm chip process is a little ahead of the CPU 65nm tracks by a couple of months. With the CPU tracks only now started ramping up.
RancidLunchmeat
Veteran
...interesting. I do a check of inventory at my local retailers every time I hit a Target/Walmart/Best Buy/etc.. just to see what's what. I have yet to catch a Wii in stock, and haven't seen a single 360 in stock since sometime back in June. (PS3s are in great abundance.)
I found it interesting that the 360s were all gone with no price cut considering I hadn't read a story about them clearing the channel.
A smaller form factor dosn't necessarily do away with the original HDD housing and the way the faceplate fits on the system.
Fair enough. I wouldn't call it 'small' then though.
Fair enough. I wouldn't call it 'small' then though.
It would certainly make for an interesting redesign considering that the HDD and the faceplate are perpendicular on the box.