Inside the iPhone 3G dropped call complaints
Experts and analysts of all stripes are trying to explain what's wrong with the iPhone 3G, but their answers are frequently supported by bad science, outlandish claims, and pure speculation. Here's what's wrong in the reports, and why a simple firmware update is likely to solve the current issues.
While nobody has formally studied the problem, lots of iPhone 3G users are complaining that they can't find 3G service, can't maintain 3G service in areas where other 3G phones can, witness wildly fluctuating signal strength bars on the phone, or conversely can't use 3G because it consumes battery life too rapidly.
Many articles on the subject are referencing Apple's support forums, where some discussions have gotten so long that forum moderators have had to lock the original thread and create a new overflow discussion.
Clearly, there are real problems. How widespread and common those problems are is more difficult to pinpoint. Apple said it sold a million iPhone 3G units on its opening weekend, and Piper Jaffray analyst Gene Munster reported that each of the company's US retail stores are now selling an average of 95 iPhone 3G sales per day. He expects the company to sell 4.47 million this quarter. That indicates that well over two million iPhone 3G units have already been sold to users.
Even if the 3G issues were only affecting one percent of the phones sold, that would leave twenty thousand users with problems. If only a tenth of those users posted comments online, that would easily account for the two thousand messages on Apple's discussion boards.
Blame the provider?
In the US, AT&T has been fingered in the iPhone 3G's reception problems due to the telco's relatively new and limited service coverage of its 3G network. Even in urban areas where AT&T's service maps indicate there should be 3G service, the iPhone 3G frequently fails to find it or maintain a strong enough signal to complete a call.
Compared to Sprint and Verizon Wireless, which both have wider 3G service coverage in their more mature 3G EVDO networks, AT&T is building out its 3G network using UMTS, a worldwide standard. AT&T is also forced to use different radio frequencies than other UMTS providers, which results in less technical maturity for AT&T's 3G network than those overseas.
AT&T primarily uses the 1900MHz band in the US, but is working to expand its use of its 850MHz band, a lower frequency that allows radio signals to spread farther and penetrate walls easier. Europe uses the even higher 2100MHz band for 3G, but there is also more dense network coverage there.
While AT&T's network is still experiencing some growing pains, the iPhone 3G's reception issues are also being reported in other countries too, even in Europe where 3G UMTS networks have been built out for some time. In those locations, the iPhone's dual band 3G radio uses the standard UMTS frequencies, making it hard to blame AT&T for more than just its limited coverage.
Dropped calls by provider
An article on the iPhone 3G by BusinessWeek cited unnamed sources to report, "the problem is affecting 2% to 3% of iPhone traffic, the people say. That compares with a dropped-call rate of around 1% for all traffic for AT&T." A source for the dropped call rate at AT&T wasn't given.
Studies on dropped calls are difficult because users don't report their dropped calls, and providers would be challenged to know whether phones on their networks ended a call on purpose or not. Further, calls may be dropped for a number of reasons, from poor service coverage or intermittent signal interference to phone set problems to users walking into a elevator or bank vault.
A study on dropped calls published by mindWireless in February 2007 ranked US providers on dropped calls by analyzing 80 million calls on 130,000 wireless accounts over a the first six months of 2006. It defined a dropped call as any two calls placed to the same destination within two minutes, without a call in the middle. This would not identify dropped calls where the user did not call their party back immediately, or where they were called back by the dropped party. It also excluded voicemail calls.
The company reported that Sprint had a dropped call rate of 5.4%, AT&T Wireless 5.7%, Verizon 8.0%, Cingular 11.3%, T-Mobile as 13.8, and Nextel at 14.6% (not including push to talk calls). AT&T Wireless was bought by Cingular in 2004, but the company was still in the process of merging its networks when the study was underway; that merger combined the GSM towers operated by both, strengthening Cingular's signal. Over the next year, Cingular subsequently rebranded itself as AT&T. Sprint has also since merged with Nextel, although those two companies operated incompatible networks (CDMA and iDEN) that couldn't help each other in terms of signal.
Those numbers indicate that the reported "2 to 3%" dropped call rate on the iPhone 3G, as well as the 1% drop rate for "all traffic on AT&T" are not likely to be anywhere close to reality. They are also not the product of any scientific study, since the iPhone 3G as only been out for a month and during that time the firmware has been updated.
Incidentally, Sprint and AT&T began fighting over the ad line "fewest dropped calls" last year, and AT&T was separately sued by subscribers over its claim as false advertising. AT&T no longer makes that claim, but now advertises "more bars in more places." That promise hasn't solved iPhone 3G reception issues however.
Blame the components?
Nomura analyst Richard Windsor kicked off the iPhone 3G panic when he published a research note suggesting that the iPhone's problems were due to a faulty industrial design using Infineon chips, and suggested that Apple might have to recall the faulty units.
The problem is that Windsor isn't a technical expert; he's a financial analyst. More problematically, this isn't the first time he's described a speculative hardware problem and sounded a false alarm for a possible recall based upon erroneous guesswork. Last year, he claimed that the original iPhone was plagued a faulty design for a film on its screen that used "a chemical deposition to provide touch sensitivity based on heat."
Windsor wrote that the design had failed in earlier attempts to make it work after just a few months, and suggested Apple might have to accommodate a massive recall after iPhones suddenly stopped working in the first three to six months. That never happened, but more importantly the iPhone also never used a heat sensitive film. It has always used an entirely different multitouch technology based on sensing capacitance that lays under the iPhone's glass screen, not on top of it.
The chips used in the iPhone 3G are similarly not unique nor the likely subject of a massive recall. Guenther Gaugler of Infineon told BusinessWeek, "Our 3G chips are, for example, used in Samsung handsets and we are not aware of such problems there."
Blame the production?
Some have blamed Apple's phone manufacturing instead. NyTeknik ("New Technology"), a Swedish publication, said that problems associated with the iPhone 3G may be due to problems in high production manufacturing, and notes that similar "normal childhood illnesses" have affected phones from Samsung, LG, Sony Ericsson and Nokia.
Testing each iPhone during manufacturing would cost more than its actual components cost, according to Claes Beckman, a professor of microwave technology at the University of GÃ¤vle. The site performed its own testing on an iPhone 3G and found results for nominal sensitivity of 3G radio signals that were below the minimums set by the ETSI standards body. However, it also noted the iPhone 3G design has passed the CE mark, which means that it originally met the ETSI standards in testing. This led the group to believe that the problems cropped up in manufacturing after production accelerated.
Users reporting problems with their iPhones have been asked by Apple to provide their "build week," represented in the fourth and fifth digits of the unit's serial number. This could mean that Apple is tracking problems with phones manufactured between specific dates; the company has been swapping out phones for users with complaints.
The number of faulty devices may fluctuate slightly during manufacturing, but there is yet no clear suggestion that problems have accelerated with new production. Some users report having exchanged out several new iPhones without seeing any difference in exchanged models between different build weeks.
Blame the firmware?
The two sources cited by BusinessWeek indicated confidence that Apple would be able to address reception issues in the upcoming iPhone 2.1 software update, expected next month in September. Earlier in the month, Apple released 2.0.1, a bug fix that also included updates to the iPhone 3G's baseband firmware. That update had some impact on the signal strength display that users were seeing, but no details were provided on what the release actually fixed.
Earlier this year, Apple released a 1.1.4 update which also addressed a problem with dropped calls that some users were experiencing at the time on the original GSM iPhone. It too was only described as being a bug fix without offering any specifics. The iPhone 3G's UMTS technology is more computationally complex than the original iPhone's GSM radio. While its chipsets are also used in phones by other makers, the firmware Apple is using to drive the its hardware is unique and has plenty of room for maturity and optimization.
The good news is that Apple is selling millions of iPhone 3G units all of the same design; other manufacturers, such as Motorola, Samsung, HTC, and others not only sell fewer smartphones than Apple but also offer a range of different models, ensuring that each model gets less focus. All of Apple's attention is going into optimizing the iPhone 3G. RIM, which sold twice as many smartphones as Apple last fall, and Nokia, which sold just over 8 times as many, similarly split their development resources across a wide number of different models.
Credit iPhone 2.1?
When it arrives, the iPhone 2.1 software is expected to combine firmware optimizations with higher level software updates, including tools to enable developers to work with more accurate GPS data for turn-by-turn directions, as well as the notifications system for third party apps that Apple described at WWDC. The notification feature was reported missing from the fourth iPhone 2.1 beta released to certain iPhone developers just days ago.
Features are frequently added or removed during beta build testing, but the removal of the notifications system from the 2.1 build may relate to an effort to deliver its anticipated low level firmware updates as soon as possible and perhaps sooner than planned, leaving the notification service to be distributed as part of a separate release.
Apple has not publicly connected the notification system with the iPhone 2.1 release; it delivered the details of both under a non disclosure agreement intended to prevent speculation and panic as changes occur in its deployment schedule.