Canadian Mac Forums at ehMac banner

A G5 iMac Repair Story

12K views 8 replies 6 participants last post by  PCBONEZ 
#1 · (Edited)
iMac G5: Up In Smoke

by Matt Neuburg <matt@tidbits.com>
http://www.tidbits.com/tb-issues/TidBITS-777.html


This is the story of how my iMac G5 joined the legion of machines that recently have spontaneously failed, and how the problem was resolved.

I purchased my 20-inch iMac G5 at the end of November, and was deliriously happy with it from the start. It crunches numbers in sprightly fashion, runs GarageBand without hiccupping, and even scrolls Microsoft Word documents fairly quickly. It writes DVDs. It has a huge hard disk. The screen displays two full pages of text and is drop-dead gorgeous; it feels worth the price of the entire computer. And then of course there's the astounding form factor: in essence, the computer consists solely of a two-inch-thick monitor, with all the works inside it.

<http://www.apple.com/imac/>

Even before my purchase, I had been hanging out on Apple's discussions boards, where I proceeded to acquire quite an education. For example, I learned that the iMac G5's bus throughput is faster if RAM is installed in matched pairs; therefore I replaced the stock 256 MB of RAM with two 512 MB sticks early in the game. Removing the iMac G5's back cover and installing the RAM was astonishingly easy. Indeed, one of this model's most remarkable features is that it is highly user-repairable. Four internal LEDs assist in diagnosis, and the parts are so ingeniously arranged and connected that, if need be, the user can easily replace the hard drive, the optical drive, the power supply, the inverter, the display, and even the midplane (essentially a sheet of metal to which everything else is attached - replacing the midplane replaces the logic board and fans, and involves removing all the other user-serviceable parts).

<http://discussions.info.apple.com/imacg5/>
<http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G5/iMacG5/04_Expansion/chapter_5_section_1.html>
<http://docs.info.apple.com/article.html?artnum=86812>
<http://docs.info.apple.com/article.html?artnum=300205>
<http://www.info.apple.com/usen/cip/pdf/imacg5/033-2494.pdf>

Intimations of Mortality -- About a month ago I started seeing anecdotal reports, on Apple's discussion boards and elsewhere, that repeated in essence an identical story: iMac G5 users would notice that the machine was giving off an acrid smell, like plastic melting or tires burning, and shortly thereafter the machine would fail. I made a mental note, but I also thought, "This can never happen to me."

<http://discussions.info.apple.com/webx?128@@.68ab823e>
<http://discussions.info.apple.com/webx?128@@.68a4aa56>
<http://www.macintouch.com/imacg5part05.html>
<http://www.macintouch.com/imacg5part06.html>
<http://www.macintouch.com/imacg5part07.html>

But of course it did. On Tuesday, 29-Mar-05, the iMac gave off an ominous smell. It was a foul, slightly nauseating smell, rather like burnt tires; I had to open all the windows just to remain in the room. The fact that I was prepared by the similar reports from other users was suddenly useful. I expected the computer to fail soon, but at that moment it was still running, so I immediately backed it up, twice - once to four DVDs, and again by synchronizing it with my trusty iBook G3/600 that sits in the living room, hooked up to the stereo system. I also ran the Apple Hardware Test (by starting up from a special CD that comes with the computer), and the computer passed all the tests. The next day, Wednesday, the iMac was still running in the morning. I went out to lunch with some friends, and when I returned in the afternoon, the iMac was in a deep sleep from which I could not wake it. I shut it down and couldn't start it up again.

A Doctor in the House -- My first response was (using the iBook) to go onto Apple's iMac G5 support page, where a link leads to a sequence of pages that act as a diagnostic assistant. These pages guided me through an analysis of the situation. At each step, you're presented with a set of choices or questions or instructions, and so you proceed, page by page, to a solution.

<http://www.apple.com/support/imac/>

The assistant elicited from me that the computer was not powering up, that there was no chime, and that the power outlet at the wall was working. It told me to remove the iMac's back cover, plug in the power cord, and examine the four internal LEDs. If the first LED had been off, this would mean I needed a new power supply. But the first LED was on, so the assistant told me to press the internal power button, and then the System Management Unit reset button, to see if the computer would power on. It didn't. The assistant gave its final diagnosis: the mid-plane needed replacing.

<http://www.apple.com/support/imac/assistant/nopower/>

(Incidentally, this series of diagnostic Web pages is both instructional and entertaining. You can learn a lot about your computer just by pursuing an imaginary scenario. For instance, if my computer had started up when I pressed the internal power button, but would not start up when the back cover was replaced, the diagnosis would have been that I needed a new back cover.)

On the last page of the diagnosis, I encountered a pleasant surprise - a link leading seamlessly into the ordering of a new midplane. It looked as though I would solve this entire problem without ever directly contacting a human being. I clicked the link, provided the computer's serial number, and ordered the midplane. You have to supply a credit card number to be charged in case you fail to return the original midplane, but if all goes according to plan, since the computer is under warranty, the entire operation is free. Apple pays for the midplane itself, for shipping the new midplane to you, and for shipping the old midplane back in the same carton (by means of a second label, self-addressed and pre-paid, underneath the carton's label addressed to you).

<http://www.apple.com/support/imac/assistant/nopower/52.html>

Thus, although somewhat disturbed that my computer had failed, I went to bed feeling that Apple's system for handling the situation was commendably efficient.

Second Thoughts -- The next day I woke up and started to worry. I was remembering some more of what I'd read in the user reports on the Internet about various spontaneous failures. Some users had described swollen capacitors, and there was some expert explanation on MacInTouch and elsewhere of how a batch of capacitors with a bad electrolytic formula had been manufactured through industrial espionage, and why this might cause them to swell. But I had seen nothing wrong with any capacitors. Furthermore, those who reported the capacitor problem were also generally reporting flickering displays, along with failure of the Apple Hardware Test, which my computer had passed.

<http://discussions.info.apple.com/webx?128@@.68a31fdd>
<http://www.pbase.com/johncoggi/image/40667990>
<http://g5support.com/group/viewtopic.php?t=4124>
<http://www.oliver-kreuzenbeck.de/iMac_problems/iMac.html>
<http://www.badcaps.com/causes/>
<http://www.spectrum.ieee.org/WEBONLY/resource/feb03/ncap.html>

(continued...)
 
See less See more
#2 ·
In addition, those who talked about the smell nearly always mentioned the power supply. Some seemed to be saying that replacement of the midplane alone had not helped. Others, in fact, seemed to be saying that replacement of the power supply had been sufficient. My hunch was that there might be two different problems, one involving capacitors, another involving the power supply; I might, I feared, have been misled by my memory of the capacitor stories into accepting the diagnosis that the midplane was at fault, whereas the real problem might be the power supply. As a further complication, some users seemed to be reporting that the failure of the power supply might take down the midplane as well, perhaps simply because soot from the burning power supply is blown into the midplane. In any case, despite the online diagnosis with the internal LEDs, I was no longer confident that replacing the midplane would solve the problem, and I decided to contact Apple directly.

Phone Tag -- Getting through to Apple by phone turned out to be no easy task. Whenever I tried, I got a busy signal. I thought perhaps I was calling out of hours, but no hours are posted on Apple's Web site (as far as I could find), so I couldn't be sure. Thus it was Friday before I finally got through to a human being at Apple. I started by routing myself through the voicemail system to a customer service person who might be able to tell me the status of the mid-plane order, which I had not been able to learn from Apple's support Web pages. She was very nice, but she couldn't give me any information, which I found odd. She then transferred me, not without some difficulty, to a technical support person.

At this point the story turned positive again. I had described the situation a little to the customer service person, and it seems she had passed this information on to the technical support person before he came on the line. Thus when he picked up the phone he was completely ready to deal with the problem. He seemed to know all about these iMac G5 failures, and I had the sense that he really didn't need to listen to my story at all, but was just waiting for me to stop talking so that he could tell me the answer he'd been prepared with all along. My hunch was absolutely right: the power supply was probably working well enough to light the internal LED but not well enough to power up the computer, and he'd have a new power supply sent right out. He took my credit card information again; this was all he needed, since he already had my computer's serial number and, through it, my address.

Now began a waiting game. During the weekend nothing happened. On Monday, much to my surprise, the midplane arrived; it was lucky I was home, as a signature was required. But I didn't open it, because it was sealed with a piece of tape that read, "Don't break seal unless using parts." Well, I wouldn't know whether I'd be replacing the midplane until I saw whether replacing the power supply fixed the problem; and I didn't want to give the impression that I'd used the replacement midplane if I hadn't.

On Tuesday and Wednesday the power supply didn't arrive, and I was getting nervous. I was particularly distressed by the fact that although I could see on Apple's support Web pages that the order for the power supply had been entered, those pages provided no further information. You'd expect that, like Amazon and similar operations, when the part shipped, that fact would be noted, perhaps along with a tracking number so you could estimate the time of arrival; but no such information was forthcoming. I regard this as a flaw in Apple's fulfillment system.

So early Wednesday afternoon I tried again to phone Apple. Getting through was nearly impossible. I was on hold in the voicemail system for 45 minutes and then, just as it seemed I was about to speak to a real human being, I was disconnected. I phoned back and waited another 45 minutes. But at last I did speak to someone, and after begging him not to disconnect me, I found that he was able to tell me exactly what I wanted to know: the power supply had not shipped, it would ship that same day, and he had a tracking number for me.

Anticlimax and Afterthoughts -- There was a simple, quick, and happy ending. The power supply arrived the next day (Thursday). It took about five minutes to replace it; a practiced hand could have done it in two, as the attachment and cabling of the power supply on the midplane are ingeniously designed to make it easy. (Apple supplied a printed version of the replacement instructions, and had even sent along a Phillips-head screwdriver.) I put the back cover onto the computer, plugged in the power cord, pressed the button, and after a heart-stopping pause it started right up. I ran DiskWarrior, just in case the spontaneous shutdown had caused any damage to the hard disk's directory structure (there didn't seem to be any), and synchronized once again with the iBook G3, which had been valiantly serving as my primary computer for a week. Everything was now as it had been before.

<http://www.info.apple.com/usen/cip/pdf/imacg5/033-2497.pdf>
<http://www.alsoft.com/DiskWarrior/>

I handed the two boxes - one containing the used power supply, the other containing the unused midplane - back into the DHL system for pre-paid return to Apple. The question of that midplane caused me some misgivings; perhaps, I thought, as long as I've got it I should install it. But I didn't; my old midplane seems to be working fine, and if it, too, is going to fail, I'll just have to wait for that to happen and deal with the problem then. Logical considerations must prevail over emotion; being without the computer for a week and a half had been wrenching and I wanted to avoid having this happen again, but when all's said and done I had no evidence that it would happen again if I didn't replace the midplane - or, for that matter, that it wouldn't if I did.

Indeed, the tentacles of irrational emotion remain insinuated into my thought processes: it's hard, in the aftermath, to separate fact from fantasy. I regard the iMac G5 with a certain mistrust. I back it up daily instead of weekly. I sniff the air for traces of that ominous smell. I tend to shut down the computer when I leave the house, instead of putting it to sleep (because it was asleep when the smell started, and because if the power isn't on, the power supply can't burn out). But of course none of that makes rational sense: there isn't the slightest evidence that the computer isn't good as new, and the mistrust can be expected to fade away over time as the iMac continues to function normally.

Response and Responsibility -- What are the implications of this little adventure for Apple Computer, Inc.? From my personal perspective, it was Apple's own discussion groups that apprised me in advance of the possible impending danger; that's good. On the other hand, the online diagnostic tool, though comforting, gave the wrong answer; at the height of the crisis, Apple was vexingly difficult to reach by phone; parts did not ship very promptly, and shipping dates and tracking numbers were not provided on the Web.

More broadly, how widespread are these failures, and what are they costing Apple? In its recent financial results conference call, Apple made no mention of such incidents, and gave no hint that its bottom line was being adversely affected. This might be disingenuously creative bookkeeping, but perhaps the number of failed iMac G5s is really not that large, or perhaps, even if it is, the costs of replacement and dealing with the public is insignificant in comparison with Apple's overall profits. Since all reports are anecdotal, and since failures are vastly more likely than non-failures to be reported on the discussion groups, it is impossible even to guess what percentage of iMac G5s are spontaneously expiring. Still, one user described an 18 percent failure rate in his shop; another said that the Genius Bar minder at the local Apple store spoke of five similar failures having been brought in that week; a CompUSA employee has seen 20-odd failed machines in the last couple of months. MacInTouch continues to reflect a steady stream of failures. Apple must know far more about the problem than it is telling, and one wishes that they would just come clean and reveal it: "Here's what went wrong, here's what we are doing about it, here's what we have learned, here are the chances are of your computer failing, here's what will happen if it does." That, however, will almost certainly not happen.

<http://www.apple.com/quicktime/qtv/earningsq205/>
<http://www.macintouch.com/imacg5part08.html>
 
#6 ·
iMac G5 power supplies and bad capacitors

The iMac G5 power supplies that fail, usually have electrolytic capacitors that fail in the same way as the ones on the logic board--their tops expand and sometimes leak, often (not always) killing the capacitor. They were made in China, using a formula stolen from a Japanese capacitor maker, but minus a critical stabilizing chemical. But the ones in the power supply were made by a different company (Ltec) than the ones on the logic board (Nichicon), unless Ltec is a subsidiary of Nichicon (I Googled for web pages that contained both names, but didn't find anything that would indicate this). I'm sure there are other parts in these power supplies that sometimes fail, or parts that fail as a result of the bad capacitors, but I don't know which parts those might be, but it would be worth replacing bulged capacitors and see if that helps.
 
#8 ·
No thread too old



Yes, I know that, Dad. Your point? People use search engines to look for answers to things they're having problems with, as I did to find this thread, and often find old threads referring to what they're looking for, so if it helps someone who finds an old thread, for me to add something useful to it, so be it. It's my choice whether to start a new thread, and not as necessary or "proper protocol" as you may think, for someone to create a new thread instead of adding something useful to an old one. Most people may read just new threads, and they're welcome to do that.
 
#9 ·
Well, if this is to be left here 'to help people' then we need to correct all the misinformation John Sawyer left here.
--

Matter 1.
A scientist working for Rubycon defected to LTEC.
He was recreating a formula Rubycon was using in some of their caps.
That scientist's staff at LTEC stole an incomplete version of the formula and ran off and started manufacturing and selling it.
Caps manufactured by companies that bought this incomplete electrolyte resulted in capacitors in which the electrolyte broke down rapidly, formed gasses inside the can, and then they bloat and pop the can open.
-
Rubycon, LTEC, and Nichicon did not buy that electrolyte.
BECAUSE THEY MAKE THEIR OWN ELECTROLYTES. ~ DUH!
--

Matter 2.
Nichicon HM and HN series ONLY, for part of 2002 thru 2004 ONLY, and only those produced on one shift in one factory,,, were defective due to a manufacturing error.
This is well know and was admitted by Nichicon through several press releases.
--

Matter 1 and Matter 2 - ARE NOT RELATED other than they have to do with bad capacitors.

--
Then there is Matter 3 which is STILL going on and not directly related to either of the above:
-
A study done by the University of Maryland and published WAY BACK in 2001 or 2002 showed that capacitors manufactured with Aluminum produced in Taiwan (or China) are prone to a failure similar to the failure caused by the problem in Matter 1, but for a different reason. That aluminum has excessive levels of certain impurities that 'leach' into the electrolyte causing it to break down, and form gasses, and so on just as Matter 1,,, only it takes a bit longer because the impurities have to leach into the electrolyte first.
-
Roughly 30% of all the worlds electrolytic capacitors are produced in Tiawan so this is a BIG DEAL.
-
This is also EXACTLY why Japanese capacitors are preferred over Taiwanese or Chinese capacitors. However, in all fairness, not all manufacturers in Taiwan and China use locally manufactured aluminum. Those that go out of their way to use good aluminum (they import it) get a bum rap simply because of where they are located and what their 'neighbors' are doing.
--

The worst of the worst caps were those that suffered from both the Matter 1 electrolyte and the Matter 3 aluminum but those haven't been manufactured since (probably) sometime in 2002/2003 when it became clear and widely known in the industry what the problem with Matter 1 electrolyte was. [Mind you, they have shown up from time to time since then remarked as legitimate good brands of caps. (aka counterfeits) ]
.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top