Slashdot Mirror
Intel Breaks Qualcomm's Hold On Apple's Baseband Chips (wsj.com)
Long-time Slashdot reader randomErr writes: In a big blow to Qualcomm, Apple plans to incorporate Intel baseband chips into at least some models of the new iPhone 7. The selection of Intel chip means that in newer iPhones Apple will no longer support CDMA technology popularized by Qualcomm. The Wall Street Journal states that many industry analysts believe Intel could be supplying as many as half of of baseband chips for Apple's handsets.
This was the last key iPhone component that didn't have two sources, and the Journal estimates that Intel's revenues could now increase by up to $700 million before the end of 2016.
This was the last key iPhone component that didn't have two sources, and the Journal estimates that Intel's revenues could now increase by up to $700 million before the end of 2016.
UMTS (the 3G GSM standard) uses a CDMA-based modulation scheme, yes, but that is totally unrelated to what is discussed in this article. What the new Intel chips do not support is the IS-95 standard and derivatives, which also uses a CDMA-based scheme, but is otherwise unrelated, and which is misleadingly referred to just as "CDMA". UMTS is supported just fine by either modem.
Verizon, Sprint, etc. use IS-95 and successors (CDMA2000, EV-DO, etc.) for 2G and 3G service. The GP is correct that Verizon and Sprint would need a complete LTE rollout to turn that off and use only the standards that form part of the GSM family tree (GSM, UMTS, LTE).
GSM was based on TDMA - everyone gets an equal timeslice of the bandwidth, even if they don't actually use it. In CDMA, everyone gets an orthogonal code and broadcasts whenever they want to. Broadcasts by other phones raise the noise floor for your phone. SNR then scales depending on how many people are transmitting at any given time, and all the bandwidth gets distributed automatically and equally between only those transmitting at that time.
TDMA was fine for voice. But when it came to high-speed data, GMS simply couldn't compete with CDMA's superior bandwidth allocation. They threw in the towel after a year - most implementations of 3G on GSM used wideband CDMA. They just named it UMTS, HSPA+, etc. because of sour grapes. This is why you could talk and use data at the same time on GSM phones - they had a TDMA radio for voice (still do), and a CDMA radio for data. CDMA phones used the same radio for voice and data (which were built on different protocols since voice was about a decade older) so couldn't do both simultaneously.
Most LTE implementations are OFDMA - does the same thing as CDMA, except using orthogonal frequencies instead of orthogonal codes. OFDMA requires more processing power to separate out the individual broadcasts, which is why it came after CDMA. Early OFDMA implementations like WiMax sucked up too much power with processors of the time, and would drain a cell phone battery in about 2-3 hours. It wasn't until a few years ago that low-power processors allowed us to implement OFDMA while not requiring a recharge halfway through the day. But CDMA was pretty much the proof of concept needed to make OFDMA a reality. Before CDMA, nobody knew if a real-life cellular network with hundreds of devices broadcasting simultaneously using orthogonal signaling would actually work or scale like theory said it would.
If the people saying the U.S. should've adoopted GSM had gotten their wish, our cellular data speeds today would probably be down below 1 Mbps. When a competitor introduces a far-superior product, it forces the other players in the market to improve, instead of sitting on their asses not improving things because people are paying them anyway. Now that LTE is becoming ubiquitous, loss of CDMA would be less of an issue. But any phone built without CDMA will not be able to fall back to 3G data in most areas of the world.