Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Hi - I have a similar problem so am researching the solution. I'm pretty sure its a temperature/overheating issue.

Has anyone performed an Apple Hardware/Diagnostics Test to see if the computer can tell you exactly whats wrong with it. ( I'm actually having trouble running the test which is a question for another forum!!)

I've added a pic with my temperatures, are these normal or abnormal!?

Thanks
 

Attachments

  • Screen Shot 2020-03-16 at 12.47.33.jpg
    Screen Shot 2020-03-16 at 12.47.33.jpg
    57.4 KB · Views: 276
I searched this but found different people say different things. I recently had my logic board replaced by a mac technician and now a red LED light is always on next to the processor. Even when the Mac is turned off.
View attachment 465728
The memory seems to be alright showing all 24 GB in About This Mac and the mac runs properly it's just this red light wasn't on before the logic board was replaced.
Hi

Did you wind up resolving this? I have almost the same problem except, my red light is off when the Mac is turned off. Mine powers up but doesn't boot at all with no signal to a monitor.
 
This unfortunately did not fix the issue. Any other suggestions? Has anybody finally diagnosed and found a solution to this issue???
 
Last edited:
Hey all! Pretty new here and did my best to exhaust all the forums but I can't seem to get to the bottom of it.

My issue is very similar to you guys -

Running macOS Mojave 10.14.6 on a 4,1 flashed to 5,1 cMP
2 x 3.46ghz 6-core Intel Xeon
32GB 1333 MHz DDR3
RX580 GPU


So my issue is, I was originally running 10.14.4 via DosDude's hacked method - running a slow 5870 powering 2 monitors and 2 x 2.26ghz.

I went ahead and followed @tsialex walkthrough for installing mojave via mac updates, doing a clean install of High-Sierra to clear the hacked install of Mojave. All seemed well, GPU and New CPU's running well but then ran into a red light on the CPU board and only showing 28GB instead of 32GB. Can't really place the cause since it's my first time dealing with it.

I reseated all the RAM, reseated the CPU's and the light went away, showed all 32GB of RAM - but then, about 10minutes in, the computer froze and restarted, and red light is back on along with only showing 28GB of RAM.

Issues that accompanied it is the CPU load, for something as simple as exporting 20 images to desktop from Lightroom, is 98%! That's new. RAM pressure is sky high as well - did a PRAM reset and it worked out, so don't have any screenshots of that - but hopefully someone can provide a bit of insight.

Provided a screenshot of where the light is showing up. It looks to line up with dsd706 but no idea that is either, so I could be wrong on that correlation.

Thanks in Advance!
I'm using my cMP 4,1>5,1 as an unRAID server and I encountered exactly the same issue. Everything was running fine until that light suddenly showed up.

I rebooted the box and got this...
IMG_2024.jpg


I removed the RAM in Slots 1+5 and the light went away but the error still appears upon restart.

It looks as thought all the processor cores/threads are working fine however.

I haven't removed/cleaned the cpus and heatsinks yet.
 
I had the same red LED lighting up. Turned out be be a defect RAM stick in slot #1, swapped it and now it's all okay.
 
I had the CPU #2 light up after a processor replacement. This is after a long story:
  • I decided to upgrade from 3.0GHz Dual Xeons to 3.46Ghz. -> Don't - it's really not worth it. But for 60 bucks, I could not resist.
  • I have a brace on my thumb for an injury and became clumsy. I dropped the chip while trying to insert CPU #1. Two socket pins bent, CPU tray GONE. $125 later, I am at try number two.
  • This time, I am super careful, but I after partial assembly, I try to test the setup without paste or tightening down the heat sinks. I was not sure whether those cheap CPUs were ok. Those Xeons cannot get that hot in 20 seconds, right?
  • Get the dreaded red failure LED.
  • Try the known working chips. Same thing!
  • Now I am sweating!!!
  • I decided to do one last test and go all the way. I applied my tube of Kryonaut paste with the 3.46Ghz chips, tightened down the heat sinks, and put in all six RAM sticks.
  • And it ... just worked!
The lesson here is that the thermal system will cause failure very quickly, if not set up right or not working properly.
 
I only test CPUs without heat sink until the machine bongs (and so tell post ok) and do suddenly kill power. Without drives of course.
 
Last edited:
don't post with one faulty CPU, but you can left the 2nd cpu empty.

Don't go further than post without a heat sink. I'd also do not recommend doing this. I have a half defective cpu board what I use for this tests.

The voltage regulators are without cooling, also. It's a very rude method. I just do that before delidding.
 
Check your Northbridge Spring Retainer Rivets on the processor tray .. and yes. . de-dust the heatsinks.
Sorry to necro an old thread, but I was dusting my daughterboard (2009 Gainestown) a minute ago and a spring fell into my lap. A brief search shows me that it’s apparently this exact spring, but only the spring fell out, not a screw. Where does it go? I’m not turning this thing on again until I can get it back in place.

EDIT: Oh god, I can’t do that repair. That’s it, then. I can’t use my computer anymore. For ****’s sake.
 
Last edited:
Sorry to necro an old thread, but I was dusting my daughterboard (2009 Gainestown) a minute ago and a spring fell into my lap. A brief search shows me that it’s apparently this exact spring, but only the spring fell out, not a screw. Where does it go? I’m not turning this thing on again until I can get it back in place.

EDIT: Oh god, I can’t do that repair. That’s it, then. I can’t use my computer anymore. For ****’s sake.
The repair is doable. If its a single CPU its easy, if its a dual upgraded CPU it is advanvced.

If you can not do it, maybe others can.

Dont use the machine without the spring, the Northbridge heatsink cant make full contact to the chip, toasting it.
 
The repair is doable. If its a single CPU its easy, if its a dual upgraded CPU it is advanvced.

If you can not do it, maybe others can.

Dont use the machine without the spring, the Northbridge heatsink cant make full contact to the chip, toasting it.
I have PTSD regarding applying thermal paste. Everything else seems easy enough, but since I‘m too afraid to reapply the heat sinks…

I’ve recently been experiencing total system lockdowns that freeze what’s on the display. I thought it was my GPU at first, so I did software tweaking (nothing). Then I figured it was the GPU being dusty, so I popped it out and air’d it (it was perfectly clean). While I was in there, I sprayed down my daughterboard and out came the spring, and then I found put about this plastic retaining screw fault. So I’m guessing an overheating northbridge is what has been causing my system freezes, right?
 
I have PTSD regarding applying thermal paste. Everything else seems easy enough, but since I‘m too afraid to reapply the heat sinks…

I’ve recently been experiencing total system lockdowns that freeze what’s on the display. I thought it was my GPU at first, so I did software tweaking (nothing). Then I figured it was the GPU being dusty, so I popped it out and air’d it (it was perfectly clean). While I was in there, I sprayed down my daughterboard and out came the spring, and then I found put about this plastic retaining screw fault. So I’m guessing an overheating northbridge is what has been causing my system freezes, right?
could be, if the pastic pins break, the Northbridge heatsink could get loose and can easily overheat.

However, even if that was not the culprit, you cant run the system without Northbridge cooling. As you say, heatsinks, I guess thats a dual one.

If it was upgraded and properly done (delidded processors, polished surface, plastik spacer in place, etc... there is no big thing, taking the right heatsinks off and redo the Northbridge.

But some play russian roulette with the CPUs on dual 2009s, lidded, half torqued screws, gambles with the voltage stabilisator's cooling...
 
If it was upgraded and properly done (delidded processors, polished surface, plastik spacer in place, etc... there is no big thing, taking the right heatsinks off and redo the Northbridge.

But some play russian roulette with the CPUs on dual 2009s, lidded, half torqued screws, gambles with the voltage stabilisator's cooling...
Does this tutorial accurately describe a long-term fix, in your mind? Up to and including the type of screws, washers, and nuts to be used as replacement? Thanks for the replies, by the way!
 
Best post about northbridge heatsink repair:

 
Does this tutorial accurately describe a long-term fix, in your mind? Up to and including the type of screws, washers, and nuts to be used as replacement? Thanks for the replies, by the way!
no, drilling a hole in the backplate, just to mount easier is mad :)

Also we dont know, in your case, if its a MP4,1 single or dual CPU machine. MP4,1 Dual CPU machines need special care for the processors.
 
no, drilling a hole in the backplate, just to mount easier is mad :)
I was going to say. I ordered 3x14mm screws rather than 16 so I wouldn’t have to do that.

Also we dont know, in your case, if its a MP4,1 single or dual CPU machine. MP4,1 Dual CPU machines need special care for the processors.

It’s a 2.93 Gainestown, so dual. What’s the skinny on them?
 
It’s a 2.93 Gainestown, so dual. What’s the skinny on them?
It isn't that hard if still with the original delidded CPU.

But you still need to be careful when remove the heatsink, because the CPU most likely will stick to the heatsink (by the thermal paste). If the CPU "fall" onto the socket and damage the pins, then the tray will be rendered useless.
 
Last edited:
It’s a 2.93 Gainestown, so dual. What’s the skinny on them?

So that are the original CPUs. IF people provide details, that would ease life in forums much more...

The most upgrade the processors.
Dual 4,1 Gainestown use lidless processors.

The processors to upgrade have heat spreaders on, they have to go, to make them fit in the Dual 4,1 CPU boards.

Some brave ones just let the lid on, and fiddle with the sensor/fan plugs and gamble with the heat pads of the voltage regulators (or simply let them in the air, prevent cooling).

Others just rip the lid off and let all the residues on the CPUs. After that, they are too thick and do not make proper contact to the heat sinks. Sometimes the pressure, caused by added thickness, is too high and they damage the sockets.

Thats why it is so important to mimic the original delidded Gainestown processors physically.

Luckily I never had an (original) procesor fell off the heat sinks, when removed. You need rather force, to separate them.

But of course, it can happen. Especially, after someone worked on it.
 
Last edited:
But you still need to be careful when remove the heatsink, because the CPU most likely will stick to the heatsink (by the thermal paste). If the CPU "fail" onto the socket and damage the pins, then the tray will be rendered useless.
Luckily I never had an (original) procesor fell off the heat sinks, when removed. You need rather force, to separate them.

But of course, it can happen. Especially, after someone worked on it.
Yeesh, thanks for telling me that can happen. Now I know to be paranoid about it. I have very little experience with building (full) computers otherwise, but it seems dangerous to me to not have that locking bar mechanism you see on every other motherboard. Is that just something all the Xeons forego, or was it an Apple decision?

Also, is thermal paste brand just a matter of opinion, or is there a standout I should be using to repaste the CPUs and northbridge once I get it rebuilt?
 
Yeesh, thanks for telling me that can happen. Now I know to be paranoid about it. I have very little experience with building (full) computers otherwise, but it seems dangerous to me to not have that locking bar mechanism you see on every other motherboard. Is that just something all the Xeons forego, or was it an Apple decision?

Also, is thermal paste brand just a matter of opinion, or is there a standout I should be using to repaste the CPUs and northbridge once I get it rebuilt?
For the cMP, thermal paste branding / model is not critical.

The cMP's cooling system was over engineered. Even you use some very low end thermal paste, and apply it poorly (but still consider correctly), then worst case should be just the CPU will run a few degree warmer (e.g. within 10°C), or the fan noise increase a bit (less than 1000 RPM difference).

However, by considering you are lack of experience, when you choose thermal paste, please make sure the paste is non conductive (most of the silicon base paste non conductive anyway, you just need to double check that).

IMO, Artic Silver 5 is one of the easiest paste to be used on the cMP. The current model is non conductive, and very easy to apply. Just draw a large "stop sign" onto the CPU, then you are good to go.
Screenshot 2024-12-04 at 13.09.57.png


If you really worry about the CPU fall into the socket, you may pull out the heatsink horizontally, or even turn the CPU tray upside down before you pull out the heatsink. Of course, in this case, better to have another person to help you to hold the tray / heatsink.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.