Roughly Drafted is claiming a source close to AT&T has spilled the beans on what’s really going on with the iPhone and its 3G connection problems, and what 2.0.2 did to fix it.

In a nutshell? An iPhone 3G running 2.0 or 2.0.1 tries to pull too much power from the network, so when multiple iPhones connect, a tower can actually run out of juice and start dropping calls and losing data.

Why hasn’t upgrading to 2.0.2 already fixed the problem? Simple: some people haven’t yet upgraded, so their 2.0 and 2.0.1 iPhones keep pulling too much power, causing the same problem even for people who have upgraded but are stuck on the same tower (or same high-density city like San Fran or NY). Only when most users have patched to 2.0.2 will people stuck on high-demand towers see improvements.

Earlier reports and theories have lain blame on everything from the 3G radio and antenna, to the Infineon chipset and Apple firmware, to the carriers themselves. We here at TiPb have long been saying the problems were likely a combination of factors, and firmware that pulls too hard on networks that aren’t that hardy seems a far better explanation than any one previously offered. It also goes a long way to explaining why Bluetest didn’t find any hardware issues, and why both Apple (via their website) and AT&T (via SMS) have really stepped up the push for this update.

So, do we finally have our answer, or is this just the next “shot in the dark”? Are you still having 3G problems? Is your neighbor still on 2.0 or 2.0.1? Tell them to upgrade now and then let us know if it helps!

Cool. Who else loves this cat and mouse game that Apple and the iPhone dev team have with one another? First one goes and the other quickly follows, you gotta admit, at the very least it makes for good theater. At best? Well, a jailbroken iPhone of course!

This time around 2.0.2 has been pwned and is available to us via Pwnage Tool 2.0.3. It’s Mac-Only so Windows users will have to use QuickPwn. PwnageTool 2.0.3 includes the newest build of Installer.app. Heads up to those waiting for QuickPwn to come to the Mac (I am), they state that it’s about a week away.

But bad news for those still looking for a baseband unlock of the iPhone 3G since there are no significant updates on that front. Well, I guess we can’t have everything right?

Read

Apple seems to have taken the unprecedented step of posting a special page for iPhone software updates, highlighting the very recent 2.0.2 firmware:

The iPhone 2.0.2 Software Update is the latest version of iPhone 2.0. It’s available now and it’s easy to install. Just follow the step-by-step guide below.

Said guide features large but tasteful graphics, featuring “Steve’s iPhone” no less. Taken together with the MobileMe News blog, could this finally be a sign that Apple is taking customer communication at least a tiny bit more seriously?

(Via iLounge)

Gotta love Steve Jobs and his blunt-force emails. This time, it’s a lucky AppleInsider reader who sent it on a complaint about crash-prone applications, a problem which has plagued the iPhone 2.0 pretty much since launch. And what did the drive-by-Steve’ing say?

This is a known iPhone bug that is being fixed in the next software update in September

Will that be the still-in-beta 2.1? Another hotfix like 2.0.2? Jobs, of course, didn’t elaborate. Smart money, however, would be on a 2.0.3 rev. so that Apple doesn’t have rush 2.1 out prematurely, with just more of the same issues.

I know my Apps experience the dreaded Home Screening of Death (HSOD) semi-regularly, especially the new/updated ones. What about you? Smooth sailing or lots of crashes? If the latter, how far away must September seem? (And is it just us, or is the unresolved bug list for 2.x pretty dang scary compared to the relatively stable 1.x?)

Ed Baig over at USA Today (via Daring Fireball) is reporting that:

Apple (AAPL) acknowledged Tuesday that a software update for the iPhone partly fixes the connection snags that have caused a global firestorm for the new iPhone 3G. Though mum on details, Apple spokeswoman Jennifer Bowcock said on Tuesday, “The software update improves communication with 3G networks.”

Really?

As I mentioned previously, yesterday in Downtown Montreal, it looked like I was unable to connect to the 3G network at all. Last night in the suburbs of Montreal, however, I was able to connect (though it took a while). Today, downtown again… nadda. Or so I thought. I switched back to WiFi and still couldn’t connect… to Gmail.

Yup. While I’d tested Gmail, MobileMe, ActiveSync, and MobileSafari yesterday, sometime since then I’d made the mistake of just hitting Gmail in MobileMail.app to see if a connection would pop up. Turns out that was really shoddy testing on my part. See, Gmail on iPhone says I haven’t had any messages since 6pm last night. Gmail on the desktop however, while continuously giving me “Server error: too many simultaneous connections (Failure)”, shows 50+ more, right up to this very minute. Now, I’ve Twittered nearly constantly about problems with Gmail IMAP lately, from invalid certificate errors, to server connection problems, to the mail outage they had a week or so back (not coincidentally the same time MobileMe was out… again).

So what’s going on? Are their network connection problems or is Gmail IMAP that really buggy (according to Twitter again, it’s buggy enough to make some iPhone developers abandon it entirely)? And has this been adding to, or merely confusing my 3G network connection problems?

My guess is the former. Intermittent 3G network connection errors, and Gmail IMAP still really isn’t ready for prime time. (And why that doesn’t get the blog-focus MobileMe gets, aside from the admittedly free nature of the beast, is a bit perplexing).

I plan to run more (and better) tests today, and hopefully get something of a less obscure picture.

So, this weekend I had a lot of problems connecting to the 3G network. Bars showed full. 3G icon was lit up. But email and web browsing — any type of network activity really — either took forever to resolve or timed out completely. Today was even worse. Couldn’t get on for most of the day. Zip. Zero. Zilch. And this was AFTER installing yesterday’s hot new 2.0.2 firmware (once I got it to download...). So what’s going on?

Are there carrier issues resulting from less mature 3G networks? Is there an Infineon 3G chipset hardware problem? Is Infineon dragging their heels about writing better drivers? Is something in Apple’s iPhone 3G software stack that’s just not connecting well, or timing out too quickly? Or is it a horrible confluence of all of the above, making it an especially tough — and frustrating — bug to squash?

Given the lack of any apparent, or at least successful, fix in 2.0.2, Engadget says Apple is “shooting in the dark” trying to resolve the 3G issues. I don’t think so. I think, as one of our commenters mentioned, 2.0.2 was scheduled to add support for the addition 20+ countries and carriers coming on line this week, and crammed in whatever minor improvements Apple had ready. Rewriting the 3G drivers, especially if Infineon isn’t moving at Apple-required speed, isn’t likely to happen before the rumored September 2.1 release (which, as mentioned in the post on turn-by-turn GPS, has already jettisoned Push Notification Server support, hopefully because Apple is laser-focused on delivering an actual, gosh-darn real stable release in 2.1).

I don’t know about you, but at this point, that’s the priority I want them to have moving forward. Do one thing at a time, do it very, very well, and then move on…

UPDATE: Fourth time was the charm! Download finished quickly and update took under 10 minutes. What’s in it so far? Commenter Sarah Healey, below, says support for the additional carriers coming on line this week, which we should have expected (thanks!), Craig Hockenberry on the Twitter says an iPod Touch/Xcode glitch was corrected for devs, and iLounge says “update all” is gone from the App Store. Oh, and, some 3G bugs might have been squashed…

No sooner had Dieter broken the story about the immanent iPhone 2.0.2 firmware drop, then I tried to download the sucker and update. And got nothing because Dieter had said 5pm and it was only 4:45. D’oh. Tried again at 5pm. Still nothing. 5:09, however, and BOOM! Update engines fired on full. 248.7MB of “Bug Fix” goodness began to flow down my slow, expensive, likely throttled Canadian cable pipe.

Or so I hoped. Seems the Lords of Kobol hate me. (Or at least the Lords of Cupertino as transfered by the intertubes and, ultimately, Videotron). My download generated an unknown error (-3256).

Tried again, same result.

Anyone else experiencing this?

UPDATE: Former Apple employee Chuq Von Rospach has just blogged about a meeting with an unnamed current Apple insider who let slip that:

90% of the disconnects are initiated inside the phone, which would exonerate AT&T. Most of the disconnects are being generated by crashes in the driver code for the 3G chip, which comes from the chip vendor, not something Apple written and outside of Apple’s direct control. Complicating this — even though Apple is handing over “here is the bug, here is the fix, update the driver”, the turnaround from the vendor on driver updates is on the order of 2-3 months. Said, um, lack of urgency not exactly making people inside the projects happy.

Understated much? If he’s not, as he says, being lied to, Chuq thinks this lack of responsiveness may be why Apple went ahead and bought PA Semi a few months back: to bring the chipset in house and more fully under their own control.

Remember that 3G network connection glitch we mentioned a few days back? The one that might be a hardware problem with a software fix? (Apple Insider weighs in today that this could, in fact, be likely). MacRumors is reporting that Steve Jobs, as he or someone acting on his behalf is sometimes want to do, has responded to an email inquiry about it, and provided the following:

We are working on some bugs which affect around 2% of the iPhones shipped, and hope to have a software update soon.

If this is authentic, and 2% is a solid number, given that the iPhone 3G sold 1 million units its first weekend, and may be over 3 million units now, that’s a staggering 60,000+ users potentially affected, — never mind 20 more countries set to launch later this week.

Unless Apple pushes out a 2.0.2 hotfix, firmware 2.1 has been in beta for a while already, and in general rumored to be heading towards a September-ish release (probably to coincide or follow up Apple’s tradition fall iPod and Mac product Special Event, which last year introduced both the iPod Touch and the WiFi Music Store).

Casey recently posted that the current 2.1 beta removed support for the Push Notification services (Apple’s conceptual replacement for multitasking functionality) to allow for more internal development, but perhaps also to fast-track the 3G fix?

That could leave the 2% in the slow lane for a while still…