CPU high temp alarms


Recommended Posts

Just replaced my Intel Xeon E3-1230 v3 with a E3-1241. Now I'm getting CPU hi temp alarms. I'm not really experienced on the hardware and its possible I did something wrong. The CPU fan is running and appears to be installed correctly but I'm not really sure how to tell. The CPU is used and could be defective I guess. Would welcome any suggestions.

Link to comment

Did you clean the heatsink and CPU mating surfaces thoroughly so none of the old hs compound was left? Did you apply the correct amount of compound before you assembled it? Does the heatsink appear evenly and tightly seated against the CPU? What type of attachment hardware is it? Perhaps take a close up picture of a 3/4 view of the heatsink on the board and post it here. Depending on the type of attachment hardware, there are a couple rookie mistakes that could result in a less than perfect clamp down, where only one edge of the hs is actually touching the CPU.

Link to comment

The new CPU didn't include a fan so I'm using the fan that came with my 1230. 

 

I cleaned the the old compound off the base of the fan using the cleaning compounds from the arctic silver kit and applied what looked like an equivalent amount of arctic silver. Also cleaned the top of the 1241 although that may not have been necessary.

 

The mounting posts for the cooling fan appear to be completely inserted. Found a YouTube video at Intel tech support and used that for a guide. I removed the fan after the first incident and I could see that the heat transfer paste had been flattened between the two surfaces, so I think there isn't anything wrong with the fan install. I'll try and post some pics later today.

 

The CPU temp stays fairly low at first when I boot (about 52C), but as soon as I put a load on it the temp shoots up to about 100C and the alarms start. CPU activity is also very high. 90% plus on all cores using a single stream on Plex to test. I would expect the load to be less than 25% for this with only one stream.

Link to comment

Those hold downs are problematic on reuse sometimes. Are you sure that the clearish plastic part that pokes through the motherboard holes are fully seated on all 4 posts? The way they grip means that pushing them back through a second time sometimes catches one of the "petals" on the top side of the board instead of going through the hole. The only way to know for sure whether that is the issue without taking it back off is to look at the back side of the motherboard, if you can see it.

Link to comment

I'm fairly sure the posts are fully inserted. If I get a chance I may pull the motherboard but honestly I'm not sure the cpu is good. Can cooling problems effect cpu load? The fact that the load on it is so much higher than it is with the old 1230 that it replaced makes me doubt that I'd ever trust it now. I may just go ahead and return it for a refund and shell out the money for a new 1241.

Link to comment

Any Intel retail cooler will cool a 1241 no problems.  I'm of the mind that it's not seated correctly too.  

 

On the last picture it looks like the right-hand little black catch hasn't caught on the ratchet thing on the clear part of the clip.  You need to push those black clips down HARD until you hear the sharp click.  They should be approx 2mm further down.  

 

When you twisted the black bit to release the clips, did you twist the clip back?  It looks like you haven't on one of those pictures.

Edited by HellDiverUK
Link to comment

I pulled the motherboard last night and checked the reverse (bottom) side. The clips are fully inserted and mounting legs are fully compressed. The eBay seller states that a stock intel fan isn't good enough and in fact a water cooler fan is required for a 1241. Since the seller is willing to accept a return I've decided this is really the best option. The fact that CPU load reaches nearly 100% when running one stream on plex was also concerning. Went ahead and ordered a new 1241 from newegg. Ordered a noctua fan as well since it appears intel no longer includes the stock fans with their cpu's (according to newegg customer service).

 

Thanks to everyone for the support. It's possible I only needed a new fan, but I doubt I would have ever trusted the CPU, and every time something went wrong I would regret purchasing a used one.

Link to comment

I wouldn't believe the seller. They appear to have the same TDP (80w), so logically, the old fan should have been fine...

 

Another question: if upgrading your cpu, why not go with something that would be more of an upgrade? You appear to be gaining less around 700 points on cpu mark, and no extra cores etc. Seems like a big expense for little return...

 

[Edit] Just looked at your options for that socket, and there doesn't seem to be that much more out there... Looks like you'd have to go with new mobo and/or ram as well to see a significant performance jump... That makes it more difficult (and pricey) for sure...

Edited by DoeBoye
Link to comment

The replacement isn't really intended as an upgrade. I'm really doing this to solve an incompatibility between the 1230 and unRAID 6.2+. You can see the details in this thread if you're interested. Since everyone reporting this problem was using a 1230/1231 I figured replacing the CPU would have a good shot at fixing the issue. It appears I was right about that. Booted 10 or 12 times without the issue showing up.

Link to comment

 

1 hour ago, wgstarks said:

The replacement isn't really intended as an upgrade. I'm really doing this to solve an incompatibility between the 1230 and unRAID 6.2+. You can see the details in this thread if you're interested. Since everyone reporting this problem was using a 1230/1231 I figured replacing the CPU would have a good shot at fixing the issue. It appears I was right about that. Booted 10 or 12 times without the issue showing up.

Ahhhhh! That makes sense (Though why it would affect the 1230 and not the 1241 is beyond me. They appear to be almost identical. Only difference I could see is the 1230 supports DMI and the 1241 DMI2). For those that, like me, who had no idea what DMI was: :)

Quote

DMI: direct media interface (DMI), which is a point-to-point interconnection between an Intel integrated memory controller and an Intel I/O controller hub on the computer’s motherboard

 

When does the new CPU arrive?

Edited by DoeBoye
Link to comment
15 minutes ago, DoeBoye said:

Only difference I could see is the 1230 supports DMI and the 1241 DMI2)

 

DMI is the same for both, it's just referenced different in the ARK site, DMI=2.5GT/s, DMI2=5GT/s, in fact it's the same for all CPUs that use the same socket.

 

The xxx1-v3 models are Haswell refresh, usually 100MHz faster than the equivalent xxx0-v3 model.

Link to comment
1 hour ago, DoeBoye said:

When does the new CPU arrive?

Wednesday

 

1 hour ago, DoeBoye said:

Ahhhhh! That makes sense (Though why it would affect the 1230 and not the 1241 is beyond me. They appear to be almost identical.

You got me. I really don't know much about the under-the-hood stuff. Whatever the problem is it must be fairly rare. I only found 4 reports besides mine that I could really identify as being the same issue. Three of them had 1230's and one had a 1231. Getting the 1241 was really a shot in the dark. That's the only reason I went with a used CPU. Wanted to not spend too much without knowing for sure if it would help.

Link to comment

In case it's helpful, the correct orientation of the latch for the Intel cooler is rotated fully clockwise before locking.  Looking at the photo, I have to agree with others' comments - they appear unlatched.  Even if they are going through the PCB, they will not be pressing the heatsink down onto the CPU with enough pressure to ensure good thermal contact.  The slot in the retainer should point to the center of the fan when in the lock position. 

 

Details here from Intel...  http://www.intel.co.uk/content/www/uk/en/support/processors/000005852.html 

Link to comment
7 hours ago, S80_UK said:

In case it's helpful, the correct orientation of the latch for the Intel cooler is rotated fully clockwise before locking.  Looking at the photo, I have to agree with others' comments - they appear unlatched.  Even if they are going through the PCB, they will not be pressing the heatsink down onto the CPU with enough pressure to ensure good thermal contact.  The slot in the retainer should point to the center of the fan when in the lock position. 

 

Details here from Intel...  http://www.intel.co.uk/content/www/uk/en/support/processors/000005852.html 

 

Absolutely correct, didn't notice that, and at least in the second photo it's clearly visible it's in the unlocked position.

Link to comment

Just so you know:

- There is no compatibility issue between unRAID and the 1230

- There is no need at all for water cooling on a 1241.

 

You have either:

- Not fitted the heat sink correctly

- Your motherboard is f**ked (either hardware issue, BIOS setting, or temp reporting).

 

I have a E3-1275v3 running just fine on a random aluminium Intel heat sink, even running 100% CPU load it doesn't go much over 70C.  

 

At work I run an old E3-1240(v1) on an Asus server board, again with a stock heat sink and have had no issues, and it's running a heap of VMs.

 

 

Link to comment
7 hours ago, HellDiverUK said:

Just so you know:

- There is no compatibility issue between unRAID and the 1230

Curious how you have determined this? The 1230 ran 6.1.9 without any issues but would fail to boot 6.2+ in several different servers (multiple reports of this happening). Installing the 1240 fixed this issue.

 

I can't say if there's a compatibility issue or not. If I was the only person having this problem I'd think it was user error, but I'm not. I have to suspect that there might be an issue with the CPUs.

Link to comment
19 hours ago, johnnie.black said:

Like I said it's not a typo, it's just referenced differently, 5GT/s DMI= DMI2

I'm going to have to respectfully disagree. It's a typo. Not that it really matters, but for the record, every other chip of that generation states "5GT/s DMI2", while the 1230 states "5 GT/s DMI". Please refer to the screenshots below :).

 

 

1241v3.PNG

1230v3.PNG

Link to comment
4 hours ago, wgstarks said:

Curious how you have determined this? The 1230 ran 6.1.9 without any issues but would fail to boot 6.2+ in several different servers (multiple reports of this happening). Installing the 1240 fixed this issue.

 

I have a 1230v3 running 6.3.2 just fine.  And a 1230Lv3.  Both machines at work which run solidly.

Link to comment
On 4/11/2017 at 4:11 PM, HellDiverUK said:

I have a 1230v3 running 6.3.2 just fine.  And a 1230Lv3.  Both machines at work which run solidly.

@HellDiverUK

You're absolutely correct. Wasn't the cpu causing the boot problem. Was doing a lot of testing with the server un-racked and running it via IPMI and worked great. Re-racked the server and reconnected the old APC console that I normally use and the boot problem is back, so it appears that having the console connected causes the boot process to freeze in 6.2+. No idea why.

 

Probably won't tell the wife about this though.:D

  • Upvote 1
Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.