You can implement improving solar technology on anything, even on a toaster if you'd like.
That doesn't mean it's useful.
Your Casio watch has a handful of transistors on an ultra low power, ultra slow speed process. I'd be pleasantly surprised if its clock speed is part of the MHz club.
If Apple is going to create a watch, you'd at least expect it to be in constant contact with your phone over Bluetooth. That means a pretty beefy (by Casio standards) processor. And a fancy screen etc.
For the small size that it will have, I just question how much it charge it could really add.
My guess is: not much. I'd love to be proven wrong.
Your guess is absolutely correct. Given the amount of incident solar radiation and the conversion efficiency of commercially available solar panels, the amount of power generated is really low.
Lets take a panel to be used in a phone for example, say a 10cmX5cm panel (best case IMO..in reality it'll probably be smaller). The amount of incident solar radiation at any point on the earth's surface, at sea level, on a clear day and assuming the rays are completely perpendicular is about 1000 W/m2. So our 10cmX5cm panel can theoretically generate 5W of power. In real world conditions, considering location, clouds, shading, time of day affecting the angle and intensity of the rays, etc, the average figure is around 250 W/m2. So our panel is now down to 1.25 W theoretical. Conversion efficiency is in the high teens at best, so even if we take 20%, the panel can generate all of 0.25W. Hoping all my math is correct, that's barely anything. Add to this the fact that a phone is likely to be in a pocket or purse most of the time, solar panels on a phone are pretty much pointless.
It is not a watch but a mobile phone, this is in the first place and because you didn't understand, a speculation for solar powered iPhone based on existing technology and potential to be implemented by Apple only if they wish.
See above.
Direct charging of the phone isn't possible with todays smartphone power draws and usage patterns (ie, typically stuffed into pockets or purses as opposed to lying around in sunshine).
For those who desire solar charging, a number of panel-to-battery solutions exist, and then the user can fast-charge battery to battery, where "fast" is in contrast to what it would take if the phone had to trickle charge from the solar panel directly.
The concept generally doesn't work for phones. Too much has to change in order for it to be functional. Even for watches the arrangement is a bit iffy, with mechanical movement to capacitor solutions arguably working better in practise.
Competely agree with all your points. And heck..charging a power bank at home and carrying it around with you is much more practical than a panel to battery solution anyway.
Back on topic. Today's announcement by Qualcomm that its flagship SoC up to 1H 2015 will be 20nm, not 16nm FinFET and its Snapdragon 410 (ARM A53) due by Q4 2014 will be a 28 not 20nm part, does this make anyone believe that the iPhone 6 will be fabbed on a 28nm process, as the IP6 will require huge wafer volumes, and at the very least TSMC isn't ahead of schedule.
For that timeframe, Qualcomm had no choice but to go for 20nm (I also believe the decision to go with A57 was not their original plan and the part is later than it should have otherwise been). 16FinFET will not even be entering mass production until Q4'14 so I wouldn't expect any products to be out until H2'15. Snapdragon 410 is a budget part and apart from the timing and availability issues with the 20nm process, the per transistor cost was likely too high anyway (Note that even on 28nm, it is on 28LP and not on 28HPM). I wouldn't be surprised if Apple stayed on 28nm but it wouldn't be outlandish to think that they could go to 20nm by Q3'14 either.
TSMC is predicting that 20nm will ramp faster this year than 28nm did in 2012. I believe they are predicting something like 20% of their revenue from 20nm during the 4th quarter.
Absent a 20nm SoC from Qualcomm, that kind of revenue almost has to be coming from Apple.
That would be quite a ramp. For reference, TSMC's share of 28nm revenue for the last quarter of 2013 was 34%, and this is more 2 years into production. In Q3'12, which is about a year after they announced mass production, it was 13%. Besides, for a iphone to ship in Q3, the silicon will have to enter production in at least Q2 right? I suppose that its possible that Apple may be going to 20nm for the next iphone..but like I said..I wouldn't be surprised in the least if they stayed on 28nm.
From the 16th of January, 2014:
At the time, and up until Qualcomm announcement now, no one had announced any SoCs on 20nm.
So either
a, TSMC have been producing a SoC for someone who does not pre announce their products, or
b, Wei doesn't know what he is talking about.
I'll go with option a.
Couple of points:-
1. We do not know how many wafers TSMC is producing
2. Apple is not the only company which does not pre-announce their products. TSMC was expecting something around 10 tapeout's on 20SoC by H2'13 so there could be a lot of other suspects.
But I'm not completely discounting Apple..I'm just saying its not a sureshot that it has to be Apple.
Apple hires two ex-Broadcom modem employees and suddenly they'll have no need for any wireless chip/module made by a third party? Sounds like jumping to conclusions. Maybe if they bought all of Broadcom this would have some weight to it.
My main question is, even if they do try to develop their own baseband, how would they get around the patents? (Gotta love the irony for Apple here
)
