iPhone share of U.S. traffic hits 69%

I don't think CDMA phones actually transmit anything until they've heard a tower they can talk to as well. This is at least as important for CDMA as for GSM since CDMA operates in bands shared by other modes, and might inadvertently step all over someone else's licensed band if it transmitted before hearing the thing it expected to receive the signal (I think CDMA phones may also need to time-synchronize with the tower before transmitting something to it, which it can't do if it can't hear the tower). I think the battery draw is instead caused by the computing the phone needs to do while scanning for a signal to synchronize to.

I think scanning for a GSM signal might be cheaper than that, though I suspect scanning for a WCDMA signal might be equally difficult. As for the difference in practice, I've noticed that my GSM/3G phones tend to give up scanning and just stay idle after a period with no signal (until I push a button or something to prompt it to try again) while the CDMA phones I've had seem to just keep searching until the battery is gone. I don't know why that is.

Dennis Ferguson

On Wed, 01 Jul 2009 19:28:42 GMT, Dennis Ferguson wrote in :

That's not it works, and both GSM and CDMA mobiles check in (register) periodically when on, whether they hear traffic or not, so the network knows where they are. Otherwise the network would have to "ring" all cells for an incoming mobile call, which would be wildly inefficient and impractical.

Why should that be? There's a huge (okay, I'm dating myself) chuck of NAND FLASH in every iPhone. What settings do you expect to find cleared?

Steve

The clock, although most phones get this from the network.

Oddly the iPhone doesn't seem to, if you put the iPhone on a blocked data plan without wifi it doesn't seem to update the time, so I'd hazard a guess that it's using NTP from Apple's servers, although I've never bothered to packet sniff to find out.

There are apps that attempt to interface with various content sources despite the lack of any official standard way to do this.

They tend to a) need special-case code for each site they interface with, and b) break frequently, sometimes because content distributors are deliberately trying to break them.

Whether or not the battery (or any other part) is easily user-replaceable typically has no relationship to whether it's covered by the manufacturer's warranty.

You're creating an arbitrary and ad-hoc standard specifically to attack Apple for not living up to it, one of the more common tricks we see in CSMA.

I'm sure I know you could manage to interpret things that way. Such an interpretation doesn't seem to have much relationship to reality. Indeed, for the last few years, Apple has given battery life estimates for devices substantially more inline with real-world results than the estimates many other vendors provide.

That doesn't really make any sense. Apple provides a battery replacement service, and it's reasonably priced as compared with what Apple has historically charged for replacement batteries.

Putting a five year warranty (or whatever) on the internal battery wouldn't make anything less "difficult" for people, it would merely give some people a free battery, that under either the current arrangement or in an alternative world where Apple didn't have built-in batteries, they'd have had to pay for. I see absolutely no reason why Apple should be obligated to buy people free replacement batteries merely because replacing one requires same-day service at an Apple store (where you very well might have gone to buy the battery anyway) rather than being something you can do at home.

This makes sense, actually. Most GSM networks worldwide don't broadcast the time since GSM base stations (unlike CDMA) aren't required to know what time it is. It is only in North America, where the GSM networks compete with time-capable CDMA and where GPS receivers are often deployed with GSM base stations to help with E911 positioning, that the facility is ubiquitous. The mechanism to optionally broadcast the time from GSM base stations was only added to the GSM standard around the time that US networks began to deploy GSM.

NTP hence seems like the lowest common denominator. It works everywhere as long as the phone has an Internet connection, which the iPhone assumes will be all the time.

Dennis Ferguson

You are changing the topic. The phone registers with a tower when it can hear it, and once attached to the network exchanges messages periodically. When the phone is not registered with the network it transmits nothing until it hears a tower it might want to talk to.

Dennis Ferguson

Methinks not. In the US, licensing and allocation of cellular bandwidth is by major cellular provider (AT&T, Verizon, T-mobile, Sprint, etc) and therefore by modulation method. You will never hear GSM and CDMA signal on the same frequencies. You will hear them on the same band (i.e. 850MHz, 1900MHz in the US), but not on overlapping frequencies. Besides, CDMA, being spread spectrum, has no way to detect a GSM TDMA carrier. The spread spectrum deconvolution process turns a non-spread spectrum carrier, such as with GSM, into wide band and low level noise. That's how spread spectrum gets its fabulous interference rejection. Similarly, GSM treats CDMA as just noise.

FCC Cellular Band Plan for 850Mhz:

FCC Cellular Band Plan for 1900mhz:

A Block US map:

B Block US map:

This is why you won't find more than two primary providers using

850MHz in one geographic area. With only two sub-bands available, the 850MHz band can only handle two providers, especially if they use different modulation methods. 1900MHz can handle 3 major providers and about 3-5 minor providers. Each can have their own favorite modulation method.

I don't really know how GSM or CDMA handles the initial scanning for base stations. I tried to Google for this info and failed. However, I did find this page on hacking GSM:

It will take a while to digest.

System time broadcast is ubiquitous in Africa, the Middle East and portions of the Caribbean.

Cell ID is also broadcast in many places, including Africa.

Yes, I probably overstated that (though I've been to South Africa and the carrier I used there didn't seem to do it). It seems to be unusual where I've been in Europe and Asia, though in the UK I've seen a different mechanism in use (on Orange, I think) which only works when you use GPRS. Telstra in Australia does it, though I'm pretty sure they didn't when I first went there.

I've only ever seen Cell ID in use in China, on China Unicom, but that network didn't broadcast time even though the facility shares the same GSM message.

Dennis Ferguson

On Thu, 02 Jul 2009 16:13:39 GMT, Dennis Ferguson wrote in :

So you really think cell phones don't run down their batteries much faster when on standby in a weak-to-no signal area as compared to a strong signal area? ;)

It's a bit of a pain when you're traveling though, if you rely on the no-data-while-roaming switch then you don't get accurate time on your device even though time is available from the network.

OTOH, given how badly some networks (I'm looking at you, Rogers) bungle their time settings, maybe it's not all bad.

My observation on longish runs of no signal (chunnel England-France

35mins or longer on a bad day) is that several models of phones do keep scanning but not in a battery-draining way. As soon as the train comes up for air the phone wakes up and its GSM buzzes can be heard again on audio gear (especially if roaming - which net do I use). If you try a manual selection in dead air space, it immediately replies with No Signal, no delay, no "maybe I ought to check first".

They do gobble battery resources on weak links as expected since the tx level is not backed off, and those breakthrough buzzes shriek higher.

the iphone definitely gets the time off the phone network. i disabled data entirely leaving only voice, and when i got off a plane in a different time zone, the phone obtained the correct time.

On Fri, 03 Jul 2009 13:58:39 -0500, nospam wrote in :

Sorry, but this is Usenet, reserved for rumors, speculation, and flaming

-- actual facts are off-topic. ;)

that would explain your posts.

Ahh good point -- It pulls the time *zone* from the network, but the time from data, or did in 2.2.1, anyway.