Late 2013 imac 27 no power

[jsg68]

Hoping somebody here can help with ideas on what might be wrong....

One line summary: imac spontaneously shut off while in use, won't power back up at all, doesn't seem to be the power supply, so what is it likely to be?

More details:

I have a Late 2013 iMac 27, which has been trouble-free since I bought it at the very end of 2014, until two weeks ago, when it spontaneously shut off while I was using it, and then could not be restarted. No noise or anything when it shut down, just the main display, and an external display connected to it, suddenly went black. For a moment I wondered if I could have moved the cursor into a hot corner that just shut down the display. (But I hadn't, of course.)

Pressing the power button on the back, either momentarily produced nothing -- no chime, no display, no fan -- no indication of anything at all, except a USB-connected superdrive would make a clicking noise after the power button was held for 10 seconds or so. So some indication of 5V USB power. (I subsequently discovered that the same click occurs if the drive is connected, and the iMac is first plugged into a power outlet, so apparently the drive is reacting to power that is present on USB whenever trickle power is present, even with the system off, and the power button depressed for 10 seconds must result in a momentary interruption of that power.)

Googled and tried an SMC reset, no help. Since it was trivially easy to test, also removed RAM (had added 16 GB or RAM to bring the machine up to 32 GB about a year ago), but unsurprisingly, that made no difference, either.

Haven't worked on one of these before, and being a critical machine, preferred not to start experimenting. No appts available for a week at the local Apple Store, so I took it to Microcenter, listed as an Apple-authorized repair center. Long story short, after being strung along and lied to for 8 days, I finally gave up and went back to pick the machine up. (They hadn't even cracked the case.)

Looked online, found info on the four LEDs indicating power supply and motherboard functionality. Ordered a kit and opened it myself, found that just LED 1 was lit, indicating trickle power in an off state. Pressing the power button, long or short, did not produce any change. In particular, LED 2, indicating proper operating power to the motherboard never lit, flashed or anything. Figured that confirmed the power supply was the problem. (This testing done with the display removed and disconnected.)

Ordered a replacement power supply from iFixit (used -- only thing I could order and have shipped on an expedited basis). Got it today. Looks fine except that the plastic is broken/missing on one side of the bracket that receives the data ribbon cable. Don't think that's likely to be a problem, the cable still seems to seat ok. Anyway, I swapped the power supply, then before reinstalling the left speaker, reconnected the power button, plugged the unit in, and tried booting it. No change. Exactly the same symptom I saw previously: LED 1 indicating trickle power lights shortly after plugging the unit in. Pressing the power button, short or long, produces nothing. Plugging in the superdrive I get the same clicking behavior. (This is how I discovered the click occurs just from plugging the iMac into 120V power.)

I have not attempted to test for a bad power button, but the symptoms here suggest against that being a problem. The unit spontaneously lost power while in use--don't see how that could be a power button issue. There's something still happening when the button is depressed for ~10 seconds. And if I unplug the system from 120V, LED 1 stays lit until I press the power button, which indicates it's triggering discharge of a capacitor (if I understand correctly, that's expected behavior). So seems like the power button isn't the issue.

Seems unlikely that the power supply I received is bad and produces exactly the same failure as the original power supply.

So is there something beside the power supply that would seem likely to explain this failure, whereby the unit doesn't even get to the point of trying to boot?

Very much appreciate any insights anyone here might have.

Jonathan

 

[Nguyen Duc Hieu]

The Power Supply Unit maybe the problem, but after replacing it, the problem persists, then the power receiving unit (logicboard) is left with the problem, isn't it?
We all try to avoid thinking about the worst-case scenario, but it does happen sometimes.

 

[jsg68]

Sure. But it would be nice to understand what the failure model actually is, if that's the case. Does anybody understand how power up actually works? You press the power button...but apparently that doesn't directly switch the power supply operational power on? Rather the signal passes through to the motherboard,a nd it has to signal back to actually trigger operational power out?

Seems like it would have to be something like that for a motherboard failure to drive this symptom.

Does anybody definitively know how this works? And if so, know what component on the motherboard actual controls such?

 

[DeltaMac]

Power Supply (assuming good) provides power to the logic board. With power cord plugged in (and power button not pressed yet), you get "trickle", or "standby" voltage from the power supply. That is indicated by the #1 LED. Press and release the power button. That causes an "power-on" signal from the logic board to the power supply to turn on the power supply. And, #2 LED comes on, indicating power in the system. And, you don't get that #2 LED. You have already tried replacing the power supply, and it sounds like you have also eliminated the power button. Next is the main Logic Board -- and I would now decide if it is worth the money to repair.

 

[Fishrrman]

Agree with DeltaMac above.
Think LONG and HARD about putting any more money into a 7-year-old iMac.

If you can't fix this with what you already have "on hand", I'd say it's time to be looking for a replacement. If not new, perhaps Apple-refurbished.

Or... consider a 2018/20 Mini (new or Apple refurbished). Mini's have proven to be just about the most robust and reliable of all Macs, over the long-term.

 

[jsg68]

Assuming I can confirm that the replacement power supply is, indeed, operating correctly, I will indeed need to decide if it is worth investing in a replacement logic board, but generally, I don't think I agree with your logic.

While this unit is a 7 year old design, the actual machine is closer to 5.5years old. The processor (3.5GHz quad core i7) is perfectly fine for my needs, and ought to be for at least several years to come---I'm almost never CPU bound. Memory had become an issue at 16GB, but after bumping up to 32GB, it's been fine, and that too ought to be adequate for some time still.

It looks like a replacement logic board can be had for as low as $200 (if I remove and re-use my existing CPU and heat sink). That would be dropping down from my current 780M/4GB graphics to 775M/2GB, but that's probably not a big deal since, other than very occasional video editing, I don't do much that's graphics heavy.

On a machine this age, I have to figure on replacing/upgrading the SSD and HDD (3 TB fusion drive) now, or in the near future, but that too is a pretty manageable expense in the few hundred dollar range. There's not a lot else that you would *expect* to be likely to die. (Frankly, in 40 years of having personal and business computers, this is the first one that has died on me, vs. I ended up retiring because the hardware became too obsolete. And that includes machines I kept running for >8 years.)

So is it worth it to invest $500 in this machine for a logic board and drives? Even if I could only squeeze another year out of it, probably yes (but I expect I should be able to get several years, at least). With tax and applecare, it originally cost $3400. That's north of $600/year at this point.

A new Mac Mini probably is the most likely alternative if I'm going to invest in a new machine, but if I do that I'm looking at $1500 just for the Mini *and* having to immediately to breach the warranty by opening it up to install a few hundred $ worth of RAM. (No way am I paying Apple $600/$1000 to upgrade to 32/64 GB.) Add a monitor, external disk, and tax, and I'm probably looking at 5x the cost.

Yeah, I could spend less and go buy a used/refurb unit, but it seems questionable that I could do so and get more bang for the buck than repairing my existing machine, and would I have any real reason to expect such to be more durable than this one, repaired?

 

[Nguyen Duc Hieu]

Spending 200$, you save more than just a beautiful 27" 4K display.
I'm with you,jsg68
Remember to buy a set of display adhesive + proper tools as well.

 

[DeltaMac]

Be cautious about replacement logic boards. "good price" on 6 or 7 year old Macs often lead you to parts that are not currently available (the dreaded "out of stock" can be frustratingly common at iFixit, for example) - so I wish you luck on that.

 

[jsg68]

Ah, yes, it is indeed out of stock at iFixit presently. Which looks like it would bump the cost up to ~$400 (for logic board with processor/heat sink included). Still likely a reasonable choice vs. the alternatives.

At the moment, I'm still waiting to get a response from iFixit, so I can feel confident the power supply they sent is go, such that it must, indeed, be the logic board.... :/

Anybody have documentation on what the pinout voltages are supposed for the power supply? Do these power supplies actually output only 12v? (Voltage dividers are done on the motherboard for all other required voltages?)

 

[Nguyen Duc Hieu]

Ah, yes, it is indeed out of stock at iFixit presently. Which looks like it would bump the cost up to ~$400 (for logic board with processor/heat sink included). Still likely a reasonable choice vs. the alternatives.

At the moment, I'm still waiting to get a response from iFixit, so I can feel confident the power supply they sent is go, such that it must, indeed, be the logic board.... :/

Anybody have documentation on what the pinout voltages are supposed for the power supply? Do these power supplies actually output only 12v? (Voltage dividers are done on the motherboard for all other required voltages?)

iMac Intel 27" EMC 2429 Power Supply voltage values? - iMac Intel 27" EMC 2429

Where can I find the pin diagram with voltage values so that I may test the power supply? thank you Is PIN 1 the bottom left? - iMac Intel 27" EMC 2429

www.ifixit.com

 

[jsg68]

I had come across that article, but it's a different class of power supply, right?

2429 dates to 2011 and A1312 looks to be 2009-2011. This is a late 2013 2639/A1419, which I understand to use a power supply common to 2012-2019.

The power supply in that article has 14 pins, whereas mine has 12.

 

[jsg68]

Just to close this out, another power supply swap with no change made it seem very unlikely to be the power supply. I ended up buying a parts imac off ebay (working, but with a smashed screen -- matched my top of the line CPU/GPU, but was actually cheaper than buying the logic board alone, and I figured I could recoup some of the cost by selling remaining parts).

And I can now confirm it was indeed the logic board.

DriveDX says both my internal drive (3TB fusion) are in good shape, so I guess I won't proactively replace them. (Though the parts imac turned out to have a 1TB SSD -- had been listed as an HDD -- so I'm _slightly_ tempted to switch to a non-Fusion config between that and the 3TB HDD. But most likely that will wait for some time in the future....)

 

[fhopper]

Sorry for your ordeal. I am just now cleaning my stuff off my 2013 27" maxed out BTO because my 2020 is here. The idea that Big Sur will not be supported made me nuts so I got a new machine even though as you said these 2013 machines are still good powerful Macs. (I split the 1TB Fusion drive several years ago so there is a 1TB spinner and a 120GB motherboard SSD blade. Boots very fast on the internal SSD.)

917
Single-Core Score
3314
Multi-Core Score for the 2013 and:
1220
Single-Core Score
8808
Multi-Core Score for the new i7 is a lot of difference.

If you wanted to buy my 2013 machine we could discuss it.

 

[jsg68]

Lol, I'm kind of the opposite. I actually didn't upgrade to Mojave until a few months back, and I was upgrading from Sierra! To my view, Apple making it difficult or impossible (especially frequent on iOS) to downgrade counsels patience before upgrading, and I usually only do so if there's some specific functionality I feel is sufficiently compelling to warrant such.

In this case, I think it was that cross-platform app functionality had gotten sufficiently out of sync with my iPhone that I decided to go ahead and upgrade. However, I still have fully-licensed business versions of Mac Office 2011 which are 32-bit only, so didn't want to go all the way to Catalina, and accelerate having to pay anew for Office. And accordingly, at this point I'm not at all concerned about not being able to upgrade to Big Sur.

 

[Nguyen Duc Hieu]

Just to close this out, another power supply swap with no change made it seem very unlikely to be the power supply. I ended up buying a parts imac off ebay (working, but with a smashed screen -- matched my top of the line CPU/GPU, but was actually cheaper than buying the logic board alone, and I figured I could recoup some of the cost by selling remaining parts).

And I can now confirm it was indeed the logic board.

DriveDX says both my internal drive (3TB fusion) are in good shape, so I guess I won't proactively replace them. (Though the parts imac turned out to have a 1TB SSD -- had been listed as an HDD -- so I'm _slightly_ tempted to switch to a non-Fusion config between that and the 3TB HDD. But most likely that will wait for some time in the future....)

Isn't faster to just move your good LCD panel to the iMac you just bought, then clone the OS+ personal data?

 

[jsg68]

Isn't faster to just move your good LCD panel to the iMac you just bought, then clone the OS+ personal data?

Turned out that wasn't an option. The screen is actually crushed in one corner, and when it arrived I realized that the rear housing was deformed along with it. I don't know what happened to this machine, but the effect is as if the upper right corner was placed into a vice, and then a lot of pressure was applied to the other end to actually bend the aluminum (and shatter the glass). So the rear housing is unusable. You couldn't mount a good display to it.

Even if that were not the case, the time/effort is probably closer to 6 of one, half dozen the other. The replacement unit had a 1TB SSD only, vs. my 3TB Fusion drive. So I'd have had to fully remove the logic board anyway, in order to swap out the SSD for mine, and install the SATA and power cables for the HDD. Then also remove the HDD bracket from mine and install it in the new unit.

(I don't think it would have been easy for me to convert then and there to the 1TB SSD + a non-fused 3TB HDD, which is what I would do if I convert, because my Time Machine backup is for about 2.4 GB of data on the 3TB fusion drive, and I don't know if I could restore a boot disk as only a partial restoration. I think I probably need to restructure my data first to get a <1TB boot drive image that I can restore from.)

On the other hand, it would have saved me the trouble of swapping the stands (which I did do, so that the serial number printed on the bottom of the stand would match the serial number of the swapped logic board).

Like I said, six of one half dozen the other, though on balance, I prefer sticking with as much of the machine I have, given I know how I've treated this machine. No idea of whatever the prior owner may have done with that one.

I also wanted to swap out as little as was practical, just to definitively confirm the source of the problem. I even swapped the power supply data/signal cable, just to confirm that wasn't the actual problem. (I had already tested all seven wires for continuity, so it would have been surprising if that were it.)

(I didn't bother to swap the wireless card, and I didn't want to start disassembling the heat sink and CPU, but it also didn't seem like a problem with either of those would be likely to cause a complete failure to even attempt to power up -- i.e., to get to the point of LED 2 lit.)