Intel Q9550 c1+Xigmatek 1283+P182 = High Temperature! HELP!

Post by **NeilBlanchard** » Mon Feb 23, 2009 3:49 pm

Hi,

Did you try taking out the front dust filter?

Kate · Post by **Kate** » Mon Feb 23, 2009 6:48 pm

Hi Neil,

Ya Im pretty sure that the fans are facing the correct position, noctua fans do not have the arrows pointing the direction but I used an old thermaltake fan to figure out the position.

The CPU fan is blowing air through the heatsink, as per xigmatek instruction sheet... But I changed the fan from the stock xig to a scythe kama... And used the same orientation...

I am not home at the moment, but as soon as i get there, I will try covering the top fan and testing the temps... Ill post the results here...

Thanks!

Acook,

I was seriously thinking of asking for my dads help to actually add another 120mm fan to the top... but I am really thinking about getting into watercooling just for the SB/NB and processor..

Im thinking of buying swiftech mcp350 pump, the micro res rev2 and the water blocks by bitspower, inlcusing the NB/SB solution for the 790i ultra and a good cpu water-block, I emailed swiftech if it would be enough to cool the components with this setup but they still havenÂ´t answer.

Thank you!

ACook · Post by **ACook** » Tue Feb 24, 2009 6:10 am

bearing in mind I have no experience with watercooling...

have you ran any of your intended applications for a sustained period yet, not any of the heating up benchmarks?
I'm just wondering what your real max temps are, over a sustained period, not just a short spike, which the mb can handle no trouble.

besides, if the mb isn't unstable after 24h of various benchmarks, I really thing you're just hypochondrying...

before you go any further with expensive things, I'd really try to troubleshoot the diff in case open/case closed, by using some high cfm loud fans, just to see if they make a difference, putting them in various positions in and out blowing.

Eric · Post by **Eric** » Thu Feb 26, 2009 9:46 am

I have similar 'problems' with a Q8200 cpu.

First I downclocked the cpu to 2000MHz (333x6) with 1,5000 VID (1,088v stressed)
After 30 minutes Prime 95, the core temperatures are 66-61-64-61c and Tcase is 38c. When I stop Prime95, 15 minutes later the temperatures are down to 56-51-53-51c and Tcase is 24c. The total system power is 98 Watt.

After two hours cooling down (pc off), I ran the cpu at default speed (2333MHz, 1,2375 VID (1,168v stressed)).
After 30 minutes Prime 95, the core temperatures are 70-65-67-65c and Tcase is 43c. When I stop Prime95, 15 minutes later the temperatures are down to 57-51-53-51c and Tcase is 24c. The total system power is 112 Watt.

When I do some math, the powerdifference is 112w-98w = 14w. If I assume a PSU efficiency of 84%, the real difference is 11,8w. With an average Tcore raise of 3,8c, this makes an C/W of 0,32.

The airtemperature right in front of the cpu cooler fan is about 27c. So the cores are almost 40c warmer than ambient. If all measurements are correct and so are the thermal diodes, than this implies a cpu power draw of 40/0,32 = 125W, which cannot be posible because the total system power is only 112w (including PSU inefficiency).

There were no additional casefans installed. The cpu cooler fan is blowing the air through the back of the case.

I've reseated the cooler twice. The first time I made a layer of thermal paste on the cpu. The second time I removed some of the thermal paste, with no effect. The third time I cleaned everything and applied a line of thermal paste. The core temperatures lowered by 1-2c.

When I touch the heatsink during load, the heatsink feels not very warm. When I put a thermometer on the alu base, the temperature reaches 36c during load (about 7c lower than Tcase) and 28-30c during idle (4-6c higher than Tcase). When I stop the fan and stress the cpu for a couple of minutes, the heatsink really warms up (42,5c measured, probably higher), so I think the heatsink does make good contact.

My setup:
Q8200 cpu with Xigmatek HDT-S1283, default fan (@ 820rpm), Xigmatek crossbow and AS3
Gigabyte EP45-UD3L motherbord
Antec Solo case with Enermax Pro82+ 385W PSU
Sapphire HD4670 with Accelero S2

When I search the internet ("Q8200 temps"), I find more similar problems.

So the bottomline is: can I count on the thermal diodes or is the specified TJmax not correct?

Kate · Post by **Kate** » Thu Feb 26, 2009 3:21 pm

ACook,

If you dont, neither do I!

I already ordered some parts for watercooling, meanwhile I am conducting some experiments using various fans to see what will happen to temperature if I "add" more air inside the case...

Somehow the P182 is very intake deficient, I found that adding a side fan (made a hole in a cartboard and added the fan), with the fan blowing air IN, the temperatures dropped amazingly well...

With the fan blowing air out, well.. I didnt get the same result.

I believe that adding a Scythe Kama Bay to the P182 would help immensily, but the more fans we have we bigger the noise, thats why I am seriously thinking of trying watercooling!

Nothing is unstable at the moment, but I am afraid that when I add all the cards I want and have the 8GB total of ram, the MB might overheat!

Hello Eric,

Welcome to the forum!

I am starting to believe that the HDT-S1283 needs a lot of circulating air to cool, I got a table fan (Yes I know it is ridiculous), blowing air IN the case and the temperatures dropped ridiculously.

Then I tried to play with the apartment AC a little bit, put it all way down to 15C, the room temperature dropped from 25 to 19-20C that alone made the temperature of my Q9550 drop a lot, even with the case closed and door closed.

5 degrees less in the ambient temperature made the CPU get to high 30s idle... Amazing!

So I decided to try watercooling, maybe I can get lower temps, at lower noise...

-----

I started a thread to discuss my watercooling setup... meanwhile the air cooling tests will continue.

Check here:
viewtopic.php?t=52612

Wibla · Post by **Wibla** » Thu Feb 26, 2009 5:11 pm

I added a Kama Bay and saw some improvement. Also removing the filter doors helps alot.

Kate · Post by **Kate** » Thu Feb 26, 2009 9:16 pm

Wibla wrote:I added a Kama Bay and saw some improvement. Also removing the filter doors helps alot.

Nice tip! Thats what I heard that adding Kama bay will help, and I belive so... The P182 needs better airflow!

I honestly do not want to remove the filters as they are there for a reason!

Also, removing it would not increase the airflow significantly, maybe it adds 5% to 10% more air? I dont really think it makes a lot of difference.

I have tried checking temperatures when I removed the air filters, and I didnt see any improvement!

Thank you!

Post by **NeilBlanchard** » Fri Feb 27, 2009 3:54 am

Hi,

I think that some of you are tilting at windmills? What are the the temperatures that you are worried about? Can you confirm them with physical evidence?

Several posts have spoken about erroneous thermal sensors in some CPU's.

By all means, do things to improve the noise levels in you systems, but do this based on the sound you hear -- but I highly doubt that in a (properly built) system, that you are going to have temperatures that are too high. What do you think all the other folks around here have done?

Low temps are not the goal, and they don't help anything. I have not read anything that indicates that you need to worry about the temps. If the system is too noisy, then work on that.

ACook · Post by **ACook** » Fri Feb 27, 2009 6:15 am

I think by now she's satisfied the cpu temps are within spec, but the mb/nb temps worry her.

I know nforce4 chips ran hot and were designed to run hot, and have always had the impression nvidia chips are designed to run hot anyway, so I don't really think there's a problem. But as long as there's no clear data on the safe temps of these chips, and given the amount of heat generating hardware kate wants to include, can't blame her for being cautious.

btw, 25C room temp, is that normal for you, or would dropping it increase your bill proportionally and that's not worth it?

(I think 21C is getting too warm for myself, but I don't have AC to lower anything during the summer. luckily .nl isn't too hot, although lately.....)

Nick Geraedts · Post by **Nick Geraedts** » Fri Feb 27, 2009 9:01 am

I'm surprised your temperatures are that high as well. I've got my Q9550 overclocked to 3.4GHz @ 1.25V (BIOS) and my load temperatures are 62-57-57-57 while running BOINC projects and playing games (there's a ATI 4870 and 4850 in the case as well). Mind you, I've got a TRUE cooler in my system and I'm using Arctic Cooling MX-2 thermal paste.

I'm using 800RPM slipstreams in the front, rear, and top vents, as well as one on the CPU heatsink itself. I'm planning on taking some tin snips and a dremel to my case this weekend though. Hopefully I'll remember to snap pictures along the way.

Kate · Post by **Kate** » Fri Feb 27, 2009 9:43 am

NeilBlanchard,

I got the CPU undercontrol, its a bit too high for my liking but it is perfectly safe, my worries at the moment is the MCP & SPP temperatures, at evga forums they say that the MCP should not exceed 60C! and i am very very close to that my current MCP temp is 58-59.

My worry is that when I add a bigger GPU and or sli/tri-sli cards the air inside the case will become even warmer, and it will cause the MCP go to above 60C, that is what worries me the most.

The physical evidence is pretty clear, when I leave the case fully open, the motherboard temperature drop around 5-6 degrees, thats why I am investigating possible ways to add a side fan or increasing air circulation inside the case.

The case is perfect in all senses, and I really like it, the problem lies in my MB and CPU, especially the MB which is designed for gamers who actually choose a more "open" case to increase air flow!

As I said previously the idea is to have a silent computer, but 1st I need to find ways to keep the components cool enough so I can have a fail safe computer! ^_^

ACook,

yes you are right the CPU temp is pretty "ok" even tho it is not in the level I would like, but its ok! The MCP & SPP temps worries me!

I like the 24-25C room temp, its not warm and its not cold, you see I live in a tropical country, temps outside is usually 30-35C (during summer).

Nick Geraedts,

I am not really sure if my temp sensors are working properly, so I really cannot make any comments, but yes I have high temperatures, things gotten a lot better latelly, when I added the itnake fans, and used AS5.... But it is not "the" way I really liked. I was expecting high 30s (idle) for q q9550 undervolted, non overclocked with HDT-S1283.

But hey, if I am within safety then I am happy

niels007 · Post by **niels007** » Sat Feb 28, 2009 3:34 am

Never ever ever ever trust what random people say on forums.. There are only very few exceptions to that rule, some of which you can find here. You're trying to cure non existent problems.

On the ''net'' you'll find people who say 40c is hot, some say 60 is hot, some say 1500 is hot, though that is probably in a topic discussing the melting of steel.

This forum is often a sane place; we tend to say temps are irrelevant provided the system runs stable, for example running Prime95 for 24 hours and loops 3dmark 2006 for 24 hours and perhaps thirdly running 3dMarkt while ALSO running Prime95.. Graphics cards can get toasty, cpu's can get toasty, mobo chips can get toasty.. Thats part of what they are designed to take for many years. Only in really extravagant situations (poor cooler, poor mounting, no airflow at all) might temperatures become so high that the system becomes unstable.

The amount of time and money you're spending worrying and buying watercooling kits is, sadly, very likely to be objectively not justifyable!

I'd say forget the temperatures, install a good cpu cooler, some low RPM fans, soft mount the harddisk and enjoy near silent stable computing!

Kate · Post by **Kate** » Tue Mar 03, 2009 6:12 am

HI Niels007,

Indeed that what I should do! I usually do not trust people randonly, the problem is some of the comments in the forum is correct, if you take a closer look in the picture below... you will notice that..

1- the CPU is undervolted # 1.032v;
2- Real Temp shows the CPU temp as 52-39-46-45;
3- Everest shows CPU temp as 39C and the core temps as 51-38-46-35;
4- Everest also shows that SPP & MCP temp are 53 & 61;

The comments I got from the people in the forums:

Your CPU has stuck sensors, the correct temp is shown as the core 2 temp....
This makes sense, because the MB temperature is 35C and using my thermometer the INSIDE CASE temp is 31C since its a high end cooler the CPU should be around 7-8C above room temp...

You should RMA your processor, it is broken!
I contacted Intel, they said the reported temp are not real temperatures just a way for the processor to know when to activate the thermal protection, and it should NOT be used to measure CPU temp.

So indeed, I rarely believe people in some strange forums...

But, the SPP & MCP temps are really HIGH... I contacted EVGA, and asked they said 70C for MCP is acceptable, however it would be best to have the temp UNDER 60C.

Since the P182, is not designed for airflow, the MB suffers... However I love the Antec P182, for me its the PERFECT case, but the only way I found to cool the MCP&SPP is using watercooler, and keeping the case.

Now... My POV: "Antec P182 is an amazing case, but it not really designed for quadcore CPU + tri-sli + hot nvidia chipsets, so another cooling solution has to be used."

Overall I am happy, things are stable and working VERY well, hopefully in a week or two I will have my watercooling system helping the MCP&SPP temps!

Thank you!

Eric · Post by **Eric** » Wed Mar 04, 2009 9:01 am

Last saturday I reseated my cooler again. Now I applied the thermal paste as descriped at benchmarkreviews (link at first page). The temperature dropped a little bit (2 C).

I did also a test with a big (standing) fan blowing air directly into the case at maximum speed. The room temperature was 20,5 C. The cpu temperature (Tcase) dropped to 15-16 (yeah right) and the core temperatures reported by Realtemp 3.00 were at minimum 48-44-44-42 (based on a TJmax of 100).