Complications after iMac Upgrade

[JHart]

Hello all! I purchased my first Mac back in August 2011. It is a refurbished 2011 27" model that came with an i5-2400 processor, the standard 4Gb of RAM, 1Tb platter drive, and Radeon HD 6970M graphics. The day I received it, I upgraded the RAM to 16Gb, and added a 120Gb Vertex 3 SATA3 SSD with two 60Gb partitions for OSX and Bootcamp.

Just this weekend I bit the bullet and ordered an i7-2600K processor to get a little more performance out of it. I ordered some Arctic Silver 5 for the new processor, but it still hasn't been delivered, so I ended up picking up some cheap silver thermal paste from a local PC repair shop. The tear down went smoothly, but when I uncapped the thermal paste and tried to squeeze it out on the clean heatsink, I got a watery mess where the paste liquid had separated from the rest of the compound (kind of like the liquid that collects over ketchup, squeeze jelly, or peanut butter). I wiped everything off and started again with just clean paste. I got the new processor in and went about reassembling the computer. Caught up in the excitement of the new addition, I wasn't as careful as I should have been, and ended up snapping a capacitor off while re-installing the main board.

Judging by its location, I believe it had something to do with SMC functions on the board. Here's a look:

Since the connection is pretty darn small, I figured the board would likely just have to be replaced if it no longer worked. I proceeded to patch things up and tried firing her up. Everything booted normally with just a slight delay that I attributed to the hardware change.

The first thing I noticed was the CPU fan gradually raise, then slowly settle back down. I immediately opened iStat to check out the temps. After 10 minutes or so, everything appeared stable, with the CPU hovering between 27 and 30 C at idle. The CPU fan continued to surge very slowly, going between 900 and 1800rpm every 30 seconds or so. Nothing too concerning, almost like gentle ocean waves, haha. After running some benchmarks and Prime95, the temps averaged highs around 65 C, with a peak of 74 C while Prime was running 8 threads.

When I booted into Win 7, the temps seemed to be a bit warmer than OSX. The machine idles around 40 C, and stays fairly steady around 74 C when running BF3 at max settings. My concern arises when I run Prime95. As soon as the test starts, temps start to climb, and continue all the way up to 90+ if I don't stop the test.

All things considered, I'd say I came out pretty lucky, but I would still like to pursue fixing these issues. I'll likely try re-installing the heat sink with the Arctic Silver when it arrives and see if my temps drop down a bit. The temps still bother me in Windows, though I haven't seen them get that high under anything but synthetic benchmarks. The fan does pick up and stay steady around 2000rpm when the temps get above ~60 C, so I believe they are still working somewhat correctly. Does anyone have an educated guess what the purpose is of that capacitor or how much a new MB costs? Sorry to be so long winded, I just didn't want to leave anything to speculation and invite impractical responses. Thanks in advance.

-J

 

[talmy]

Virtually all capacitors in digital circuits are "bypass" capacitors that filter the high frequency noise on the power busses. There are always more needed than required for operation. Knocking off the capacitor almost certainly is not your problem. It sounds thermal to me but I don't know if its bad thermal paste or greater power consumption of the new processor (very bad if it is -- the computer is designed to dissipate only the power consumed by the OEM processor.) New logic boards are outrageously expensive, IMHO replacement usually means it's best to just buy a new computer.

Hopefully somebody else can chime in with more detail. I'm an electrical engineer and even designed computers in the past, but I've never opened the back of my iMac.

 

[JHart]

Thank you for the input! After everything booted up, I figured the capacitor (if that's in fact what it is) couldn't have been vital to the system's operation. Thankfully that seems to be the case!

I had hoped that since the i7 I upgraded to is an upgrade option directly from Apple, that power consumption shouldn't be an issue. They're both 95w TDP class processors after all. I'm sure others have accomplished this exact upgrade since there are many write-ups across the net from reputable sites.

At this point, I don't feel it's necessary to replace the board while I still have other options to explore. Thanks again for your time.

-J

EDIT: You don't suppose there's any harm in my running the 2600K (unlocked) processor versus the standard 2600 processor, do you? I haven't yet seen a utility to monitor voltage for fluctuations, but nothing I have currently shows any sort of variance in processor clock. It just sounds like I have small oscillating fan sitting on my desk.

 

[iMikeT]

EDIT: You don't suppose there's any harm in my running the 2600K (unlocked) processor versus the standard 2600 processor, do you? I haven't yet seen a utility to monitor voltage for fluctuations, but nothing I have currently shows any sort of variance in processor clock. It just sounds like I have small oscillating fan sitting on my desk.

After reading your initial post, this is exactly what I was thinking at first. That is you installing the 2600K as opposed to the 2600 that Apple puts in. But your guess is as good as mine.

Also, I remember reading somewhere on the web about swapping out the processor in an iMac and needing to install some custom EFI afterwards. This may be a factor of the issue, then again it may not. Like I said, your guess is as good as mine.

What ever the case may be, good luck with your iMac. You're braver than I am for attempting this. Keep us posted on how things are going.

 

[wonderspark]

Well, if it still works without it, I guess you're ok... but am I crazy to think you couldn't just solder that sucker back on there? If you had two solder guns and a friend heating one side, you heating the other, you could place it back on in one shot.

Edit: maybe even silver epoxy would be easier, like this guy's repair.

 

[JHart]

After reading your initial post, this is exactly what I was thinking at first. That is you installing the 2600K as opposed to the 2600 that Apple puts in. But your guess is as good as mine.

I still haven't ruled this out, but of course there is very little info out there on this. I researched the difference between 2600 and 2600K thoroughly before buying it. I opted for the 2600K because it was only $5 more, and the 2600 was out of stock. From what I've read, the 2600K needs a P67 chipset in order to access the clock features. Since my Mac has a Z68, I assume it would run at default settings.

Also, I remember reading somewhere on the web about swapping out the processor in an iMac and needing to install some custom EFI afterwards. This may be a factor of the issue, then again it may not. Like I said, your guess is as good as mine.

I haven't seen anything about this. If you happen to find the article/info again, I would love to read it.

Well, if it still works without it, I guess you're ok... but am I crazy to think you couldn't just solder that sucker back on there? If you had two solder guns and a friend heating one side, you heating the other, you could place it back on in one shot.

The + terminal would be manageable, but the - terminal is TINY! If you look in the photo I posted, the contact is barely visible on the edge of the little circle. This photo is about 4x life size, depending on your resolution.

I haven't yet had the time to re-seat the heat sink with the Arctic Silver, but will try this weekend. I'll update you guys with the results. Thanks again for all the helpful input!

 

[JHart]

Some Updates!

Last night I removed the 2600K processor, and installed an i7-2600 with some new Arctic Silver 5 thermal paste. Right from boot-up I could tell there was a difference. The 2600K ran just fine, but always seemed to take a considerable amount of time to boot (around 1 minute), and 5-10 minutes when 'probing hardware' before the Apple Hardware Test. The 2600(non-K) boots to the OS X desktop in around 19 seconds, and only takes one or two minutes to perform the probing in the AHT. Something was definitely going on there.

The new processor/paste idles between 26* and 30* C. Running 1080p conversion in Handbrake under OS X maxes out around 66* C on the CPU diode. On the other hand, the fan is still iding up and down slowly, and during video conversion or stress tests in Win 7, temps creep up to 92* before the fan steadies around 1100rpm, bringing the CPU down to 85-86* and holding it there.

I went ahead and ordered the MG silver epoxy that wonderspark suggested trying, and it came in the mail today. I'd like to give it a shot this weekend, but I'm still concerned about how small the connection is. I might try using a scalpel to lightly expose more of the terminal, but dang that thing is tiny! I'll post updates once that's done.

 

[JHart]

Finally got around to tearing the iMac down again to attempt resoldering the capacitor.

I attempted to file and scrape away the paper that had stuck to one of the terminals on the capacitor, and in doing so, snapped the terminal off the cap. So after some careful evaluation, I decided to scrape away a tiny bit of the board to expose the terminal and placed a minuscule drop of the silver epoxy on there. On top of that, I layed the solder joint that had broken off the capacitor, then put a small dab of the epoxy on the end of that to contact the capacitor's terminal. The other terminal of the capacitor was relatively easy. The only concern I had was getting some of the epoxy on the capacitor's metal shell. You can see in the photos what I'm talking about.

I setup my borrowed hair dryer next to the new connection and placed a cheap thermometer in the air's path. Things got up to 50* C before the thermometer when black. I went ahead and let it cure for roughly 35 minutes just to be certain it was secure.

Set in place:

Secured:

Heating up:

Finished Product:

I carefully reassembled everything and fired her up. Nothing. Wait... a slight fan spin, then silence. Tried again, same thing. As I sat there at the table at 2 in the morning looking around in dismay, I noticed four sticks of RAM on my bench where I kept the parts as I pulled them. :Facepalm:

Popped two of the ram sticks in and gave it another shot. Sucess!! Waited a few seconds for OS X to boot up and noticed there was no hint of any fans revving up. I've run stress tests in both Windows and OS X now and confirmed that the fans now idle up according to the thermal readings of the components.

I'm very thankful for all the ideas and insight, and especially thankful to wonderspark for providing the link with MG Silver Epoxy. This stuff really is awesome despite the high temps required for quick curing! Thanks again everyone!