i5-3570K High Temps With 240mm Radiator

[D3L3T3D]

Hi guys,

 

Context

CPU : i5-3570K

Thermal paste : Noctua NT-H1

Motherboard : ASUS P8Z77-V LK

Cooler : Corsair H100i

Pump speed : 2215 rpm

Cooler fans : 2x Noctua NF-F12 running at full RPM under load (around 1400 rpm)

Vcore : 1.200 V (very constant, Ultra High LLC)

Frequency : 4.0 GHz

Load temps : Average 82 C, max 86 C when running Prime95's Small FFTs torture test.

Idle temps : 45-50 C

 

As you can see, the OC is very conversative considering my setup. I definitely lost the silicon lottery as my CPU is unstable at 4.2 GHz with 1.25 V, but it should still be running at decent temps on 1.2 V.  What could cause such high temps?

 

I tried reseating the cooler and reapplying the TIM a couple of times to no avail.

 

Thanks a lot!

[Imakuni]

1 minute ago, Reich said:

Hi guys,

 

Context

CPU : i5-3570K

Thermal paste : Noctua NT-H1

Motherboard : ASUS P8Z77-V LK

Cooler : Corsair H100i

Pump speed : 2215 rpm

Cooler fans : 2x Noctua NF-F12 running at full RPM under load (around 1400 rpm)

Vcore : 1.200 V (very constant, Ultra High LLC)

Frequency : 4.0 GHz

Load temps : Average 82 C, max 86 C when running Prime95's Small FFTs torture test.

Idle temps : 45-50 C

 

As you can see, the OC is very conversative considering my setup. I definitely lost the silicon lottery as my CPU is unstable at 4.2 GHz with 1.25 V, but it should still be running at decent temps on 1.2 V.  What could cause such high temps?

 

I tried reseating the cooler and reapplying the TIM a couple of times to no avail.

 

Thanks a lot!

Uh... what's your ambient temps? It could help justify the high temperatures.

[Hecknologist]

The problem with the 3570k is the Thermal Interface Material under the heatspreader is not great, and has degraded over time. I have the same CPU and had the same issue.

The only way to fix the temps is to replace that TIM, which means delidding your CPU.

There are some great guides online, just google it, but be warned: it is somewhat risky if you go with the razor method and you are voiding your warranty.

I used the vice method and would highly recommend that option.

[D3L3T3D]

18 minutes ago, Imakuni said:

Uh... what's your ambient temps? It could help justify the high temperatures.

Good point, should have mentionned it.

 

Room temperature : 24 C

Mobo temperatures :  around 30-40 C

 

And even more interesting : liquid temperature is at 31 C. Huge difference between liquid and CPU temps.

 

15 minutes ago, xVIP said:

The problem with the 3570k is the Thermal Interface Material under the heatspreader is not great, and has degraded over time. I have the same CPU and had the same issue.

The only way to fix the temps is to replace that TIM, which means delidding your CPU.

There are some great guides online, just google it, but be warned: it is somewhat risky if you go with the razor method and you are voiding your warranty.

I used the vice method and would highly recommend that option.

Yeah, that can definitely be the reason. Not sure I would want to delid my CPU, sounds like risky business... What difference did you notice after changing the TIM?

[Tyrosen]

15 minutes ago, xVIP said:

The problem with the 3570k is the Thermal Interface Material under the heatspreader is not great, and has degraded over time. I have the same CPU and had the same issue.

The only way to fix the temps is to replace that TIM, which means delidding your CPU.

There are some great guides online, just google it, but be warned: it is somewhat risky if you go with the razor method and you are voiding your warranty.

I used the vice method and would highly recommend that option.

Great guides? There is non greater than THE delidding guide:

I am not to be liable for direct, indirect or consequential damages or for any loss of your processor arising in connection with using this video, including, but not limited to, medical, and physical. The information contained within this Linus Video is provided 'as-is', without warranties as to its accuracy whether expressed or implied.

 

Good Luck!  

[PRTI]

7 minutes ago, Tyrosen said:

I am not to be liable for direct, indirect or consequential damages or for any loss of your processor arising in connection with using this video, including, but not limited to, medical, and physical. The information contained within this Linus Video is provided 'as-is', without warranties as to its accuracy whether expressed or implied.

 

You really love to be sue free ?

[Tyrosen]

2 minutes ago, keNNySOC said:

 

You really love to be sue free ?

 

[Hecknologist]

11 hours ago, Reich said:

What difference did you notice after changing the TIM?

Here is a link to the results I got from it:

Yes, it can sound quite scary and I would definitely recommend doing your research. If you do go ahead with it you just have to take your time and not do something stupid, like use a chisel. If you don't feel confident with tools like a bench vice maybe get some help or look at alternatives.

 

11 hours ago, Tyrosen said:

Great guides? There is non greater than THE delidding guide:

I would ABSOLUTELY recommend against the razor method, as one little scratch can/will kill your CPU. The vice method has no sharp objects involved and requires just a little pressure on the sides and the heat spreader will pop right off. It also means you can use masking tape to give some padding while you hold it.

[Tyrosen]

2 minutes ago, xVIP said:

I would ABSOLUTELY recommend against the razor method, as one little scratch can/will kill your CPU. The vice method has no sharp objects involved and requires just a little pressure on the sides and the heat spreader will pop right off. It also means you can use masking tape to give some padding while you hold it.

All jokes aside, he's right, IF you want to delid, use the vice method. Blunt objects, more surface area, less error

[Hecknologist]

I just re-read through my old thread about how I did it. I used an old celeron off ebay as practice so I knew what I was doing before the real deal. I'd recommend doing that first.

More info about the details of it can be found here:

http://www.overclock.net/t/1313179/official-delidded-club-guide

Or just look at this list:

Here is a list of CPU's that should be able to be delidded and some are very cheap so great for saving your chips.

IHS Not Soldered To Die

Hyperthreading/Single Cores

-(S-478) Pentium 4 HT (Northwood "A" and "B" Core)*
-(S-478) Celeron
-(S-775) Celeron
-Celeron 420
-Celeron 430
-Celeron 440
-AMD Athlon 64 3200+*
-AMD Athlon 64 3700+*
-AMD Athlon 64 3800+ (Venice core)

Dual Cores

-AMD X2 5000+ BE (Brisbane core)
-Celeron Dual Core E1200
-Celeron Dual Core E1400
-Pentium Dual Core E2140
-Pentium Dual Core E2160
-Pentium Dual Core E2180
-Pentium Dual Core E2200
-Pentium Dual Core E2210
-Pentium Dual Core E2220
-Pentium Dual Core E6300
-Core 2 Duo E4300
-Core 2 Duo E4400
-Core 2 Duo E4500
-Core 2 Duo E4600
-Core 2 Duo E6300 (L2 stepping)*
-Core 2 Duo E6400 (L2 stepping)*
-Core 2 Duo E7200
-Core 2 Duo E7300
-Core 2 Duo E7400
-Core 2 Duo E7500
-Core 2 Duo E7600

 

Hope this helps.

[Wiflare]

I've been thinking of delidding my 3570k for a while. the temps are a bit high for a kraken x61 (peaking at 55, idle in 35s range), it's not great but nothing to potentially destroy my system over.

Lol, maybe once I get my CNC mill, I'll make my own heatspreader assembly; with extra fins.

[D3L3T3D]

Thank you guys for all your great tips. I appreciate it very much!

 

I re-seated my H100i and so far the results are great. Like I said in my main post, this isn't the first time I tried re-seating it, but this time I put more TIM (and spread it more instead of just putting it all in the middle) and made sure that the waterblock was tighter on the CPU. My TIM (Noctua NT-H1) does not spread easily and that might definitely have caused a poor contact between the waterblock and the CPU lid.

 

I also decided to keep the OC at a conservative 4 GHz but managed to lower the vcore at a stable 1.184 V, which definitely helps with keeping the CPU cooler, although that did not make much of a difference prior to re-seating my H100i.

 

Now, at 100% load using prime95's Small FFTs torture test, my temps average around 65 C, with a maximum of 69 C. This is by no means great, but I can tolerate it if my CPU can stay around 60 C while gaming. I will gladly avoid delidding my CPU if it can stay at safe temps.

 

Furthermore, the temperature delta between the CPU package and the H100i's liquid is now around 30 C, instead of the previous 63 C. This is a major difference and the previous discrepancy was likely an indicator of poor waterblock seating.

 

As a side note, though, I'm thinking that the NT-H1 might be degrading very quickly when under heat. It is definitely not the first time that I try fixing my temps and notice an improvement after re-seating my watercooler, only the see the temps increase again over time.

 

I will try to remember to update this thread if that happens so as to help any one finding themselves in a similar situation.