Sign in with
Sign up | Sign in
Your question

Q9450 High Temps

Tags:
Last response: in CPUs
Share
June 25, 2008 3:15:08 AM

I built my first system approx. 1 month ago, and after reading countless forums, I have noticed that my temps are higher than most, especially core #3. I have called Intel's tech support, and they said that they do not recognize ANY hardware monitoring programs EXCEPT what your bios reads. Which is only the cpu temp. They also said that I need not worry about core temps, only what the cpu temp is as a whole. My temps are as follows at idle.

IDLE:
http://i287.photobucket.com/albums/ll129/bryan240g/Base...(CPUID Hardward Monitor)
http://i287.photobucket.com/albums/ll129/bryan240g/Temp... (REALTEMP 2.6)

100% LOAD:( prime 95, small fft)
http://i287.photobucket.com/albums/ll129/bryan240g/OC32...

My room is normally at a temperature of around mid 70's right now. Hotter during the day. Here is a link to pic of my case and the arrows illustrate the airflow.
http://i287.photobucket.com/albums/ll129/bryan240g/Case...

My heatsink is the Xigmatek HDT-S1283 , with a 120mm Scthye fan on it.

I ran prime95 for 16 hours straight and didn't have 1 failure. I would like to OC higher but at 3.6 I failed 2 tests. But then again it's the first time I have ever done anything like this. I have heard that quads sometimes have internal sensors that get suck, which if that's true how come of the four, I still see #3(the highest one) moving and not stuck on any 1 temp.? I have reseated my heasink 3 times, and cleaned it thoroughly each time. I am using the bolt on kit from Xigmatek, and I have it tightened down until it stops. I am just not sure what to do from here. Not worry about it? Everybody who see's these temps says JESUS that is High Temps. So it makes me worry especially being a noob at this. So I am calling all hands on this, that are willing to help me that is. I need some opinions. I would appreciate any help I can get. Thanks.

More about : q9450 high temps

June 25, 2008 3:21:33 AM

heat sink is installed wrong or the compound is wrong or the readings are wrong

at 1.2v at max at 3.2 you should see 55c max

this is normal 3.6ghz - it says lower its just the only screen shot i have handy
http://s63.photobucket.com/albums/h138/4rothrocks/?acti...

this is an actual messed up system where the voltage is supper high due to a number of issues - but the temps still are in the 60c at full load and even thought this only ran 1 min they stayed under 70c

June 25, 2008 3:23:02 AM

bryan240g said:
I built my first system approx. 1 month ago, and after reading countless forums, I have noticed that my temps are higher than most, especially core #3. I have called Intel's tech support, and they said that they do not recognize ANY hardware monitoring programs EXCEPT what your bios reads. Which is only the cpu temp. They also said that I need not worry about core temps, only what the cpu temp is as a whole. My temps are as follows at idle.

IDLE:
http://i287.photobucket.com/albums/ll129/bryan240g/Base...(CPUID Hardward Monitor)
http://i287.photobucket.com/albums/ll129/bryan240g/Temp... (REALTEMP 2.6)

100% LOAD:( prime 95, small fft)
http://i287.photobucket.com/albums/ll129/bryan240g/OC32...

My room is normally at a temperature of around mid 70's right now. Hotter during the day. Here is a link to pic of my case and the arrows illustrate the airflow.
http://i287.photobucket.com/albums/ll129/bryan240g/Case...

My heatsink is the Xigmatek HDT-S1283 , with a 120mm Scthye fan on it.

I ran prime95 for 16 hours straight and didn't have 1 failure. I would like to OC higher but at 3.6 I failed 2 tests. But then again it's the first time I have ever done anything like this. I have heard that quads sometimes have internal sensors that get suck, which if that's true how come of the four, I still see #3(the highest one) moving and not stuck on any 1 temp.? I have reseated my heasink 3 times, and cleaned it thoroughly each time. I am using the bolt on kit from Xigmatek, and I have it tightened down until it stops. I am just not sure what to do from here. Not worry about it? Everybody who see's these temps says JESUS that is High Temps. So it makes me worry especially being a noob at this. So I am calling all hands on this, that are willing to help me that is. I need some opinions. I would appreciate any help I can get. Thanks.


dude - i just posted to 5 people who really ask really dumb questions with out doing the minimal amout of research - AT LEAST THIS IS AN INTELLIGENT THREAD!

do this post, your system specs (mobo etc) and post settings. your temps are high and i do not use that program - it could be wrong. use core temp and repost.


the 1283 rocks read my post (again non stop shelf promotion by of myself, i apologies in advance!):

http://www.tomshardware.com/forum/250728-28-xigmateck-t...
Related resources
Can't find your answer ? Ask !
June 25, 2008 3:28:02 AM

Will do, I will take pics of my bios and repost in a few minutes. Thanks for the reply.
June 25, 2008 3:42:33 AM

Hey how do I insert pics into my post without, having to provide a link?
June 25, 2008 3:47:28 AM

  1. [img]http://link here[/img]


So like from your links above:



Heh... I need to go to werk... bbl
June 25, 2008 3:52:49 AM

System Specs:
Silverstone TJ-09 Case(All Aluminum)
EVGA 780i Mobo
(4gb) Mushkin pc-8500 ram
Intel Q9450 @ 3.2
Xigmatek HDT-S1283 Cooler(sychte 120mm fan attached)
(4) Scythe case fans 3 exhaust 1 intake, all 120mm
Using Artic Silver 5

Here are pics of my bios, if they tell you anything else.





" alt="" class="imgLz frmImg " />
June 25, 2008 3:53:35 AM

Thanks I got it now. Hope these pics help make sense of my issue, I am really sweatin it.
June 25, 2008 4:02:50 AM

At 3.2GHz on my Q9450, I have never seen my temp over 15C above ambient (I've never even seen it break 40C with fans on low). I have a similar air-flow in my case, too.

The temps in bios are quite high, for those voltages/clocks, I think. Try setting it back to stock clocks/volts and try the same tests. See if the temps drop down to something reasonable.
June 25, 2008 4:05:41 AM

Well I forgot to mention that earlier, you see at sock setting, with my stock cooler on I had the EXACT same temps before I added the Xigmatek cooler, right then i knew something was wrong. Like I also said I have reseated it 3 times, cleaning it in between. I swear that the exhaust air coming from my case, is almost cool. Not in the least bit even warm, feels like a nice cool breeze. And it's 70+ degrees in my room right now.
June 25, 2008 4:12:41 AM

Core Temp does not play nice with the Q9450, Real Temp does.

That said, RMA right now, one of your cores is off. The Intel sheet specifies core diffs, I recall a 5c diff ok, anymore, and time to send back.

Ht and PCIe need to be disabled as well...............
June 25, 2008 4:15:51 AM

Ok, I will take that into consideration, I just need to hear that from a few others and that is exactly what I will do. Anybody know how long it takes to RMA a cpu from Intel?
June 25, 2008 4:25:33 AM

I would rma to the company you got it from.
June 25, 2008 4:32:30 AM

I bought it at Fry's and unfortunately, they have a 15 day return policy on Cpu's and I am over 30 days. I think Intel is my only shot. But to be honest before I go that route. I do not disagree with Granite3, but I would just like to be reassured by others that my cpu is f**ked., before i send it off and be without for god knows how many weeks.
June 25, 2008 5:00:23 AM

Fry's FTL. Bump?!?! LOL. this isn't the runescape forums buddy.
June 25, 2008 5:01:51 AM

I know sorry, but I am getting angry and jst want people to see my thread. Fry's FTL????
June 25, 2008 5:20:18 AM

for the lose = ftl
for the win = ftw
June 25, 2008 5:21:20 AM

Nothings wrong with your CPU, just the thermal sensors, you can try RMA if you OC. If you don't, then don't worry at all.
June 25, 2008 5:43:16 AM

I would try intel, but just don't mention you are using the xigmetek heatsink and if they ask if you overclock say no. Apparently intel doesn't rma some stuff if it is not using their stock heatsink.
June 25, 2008 6:31:44 AM

evga mobo is hard to tune since its all auto compared to asus

your nb can go as high as 1.6v as long as you cool it - since you have the stock fan that hard. is it on?
with asus, i add a 40mm fan to nb heatsink/pipe contraption


cpu is 51c in bios at 3.2ghz with a quad core thats normal


i suspect you have a bad case and bad air flow in the case?
June 25, 2008 6:37:08 AM

Yeah, you wouldn't want to say you overclock. This is a situation where lying is good.
June 25, 2008 6:45:47 AM




dude! 1500 rpm fan speed! what is that crapola?????

get a 2500 rpm fan and throttle fan to 1900-2000, nvidia has poor fan controls on the mobo so add a zalmam controller - problem solved!

if not get a 1900 rpm fan those are the perfect 120mm fans they are quiet.

1500 sucks azzz its too low!




first always make one change at a time and recheck. make notesl if the system boots faster or hangs then you made a bad change - note it!

lock all your voltages to default settings, you see the nb is 1.5v so do not go lower and bring it up to 1.6v and get that baby up to 3.55ghz. you want to got just 900/1800fsb. 895 and 1790 fsb gets you 3.55ghz and its the sweet spot for the combo with stiker ii, at least!

when you look in bios if your 1.2ht is 1.2 then set it too 1.2

enable speed step and disable the othe crap c1e - i run speed step since it drops the multiplier and keeps your cpu cool under low load - that way you crank up the voltage and get years and years out of it!

bring your Vcore to 1.4-1.46 in bios, if you have good cooling keep it under 55c in bios

the GTV are for cpu right? then set them to .3v and the one that is .65v manually set it for that

turn off all spread spectrum


note: the last 780i q9450 i shipped was today! the system ran at 3.55ghz and 65c max - unfortunately it was a striker ii

with the stiker you need to lower the ldc multipler to 4 and the raise the GTL to .3mv

but there is only one GTL for all 4 cores

your high GTL tells you that core 1 is all pissy and need more voltage - remember these settings are to oc the system with stablity not get the temps down
June 25, 2008 7:01:20 AM

you definitely don't want Vcore to be at 1.4v, as it will cause your chip to slowly die.
June 25, 2008 1:33:58 PM

Just got back from werk... now to mention my thoughts:



I use the Cool Master 590 case, which has 2 fan vents for the top. They are designed for exhaust, but your HS is designed for an airflow to go into the fan, out the rear. So the problem I see, is airflow disruption from the top fan. Now I have only one fan (didn't need both to be occupied) on the top to force air down into the airflow of the HSF, which is a Tuniq 120 for my E4400. The only problem with that, is dust. :lol: . o O (but it should help decrease load temps.)

So having allot of fans can make things worse. You would need to try different things with the fans to see what works best.

Edit: I noticed the VGA cards, kinda wonder how much heat comes from that... do you have a side fan? If so, maybe you can force air out from the side to draw more air from the front, making more neg pressure to force more in from the front.



So this would be your load temps. First off, your using 2 different apps, RealTemp, and Hardware Monitor. Those 2 apps depend on the correct Tjuction max temp for the cores in order to determine its approx temp. RealTemp is going to be set at 95C, while Hardware Monitor will be set at 100C. So your going to get different temps there.

Since Intel will not publish the correct Tjmax, the software makers will make an assumption to what it maybe. That is why SpeedFan and CoreTemp will give higher/lower readings depending on its preset Tjmax value. All the programs execept Hardware Monitor will allow you to offset or change the Tjmax, so you can have a more correct reading, when you try to calibrate it.



Using RealTemp, test sensors, should indicate if any of the sensors are sticking. And it looks like Core1 is sticking (Core0/Core1/Core2/Core3 is its label). There isn't much you can do about that, but it should still give correct temp values when it is under load, and above 61C, if I understand correctly, but in this case, even the max temps are weird. The only real fix you have is to RMA it like the others said.
June 25, 2008 9:39:54 PM

Well to reply a few posts back my case is not and does not have bad airflow. Infact it is a very highly rated one at that. I do have a zalman fan controller, and the fans are at wide open throttle all the time. I ordered a Rexus NMB-MAT (Panaflo) 120mm Case Fan. We'll see if that paste and fan makes any difference. I don't think it will, considering the stock cooler also had practically the same temp issue. And maybe if adjust the fan directions, it may help out. Ok so I set my core voltage, and fsb voltage to 1.1250.

Core Temp shows me in the low to mid 40's with 1 care still showing 70ish(it fluctuates)
Real temp shows me in the high 30's, with the 1 core in the low 60's

I ordered some artic cooling mx-2, and found a good tested way to apply thermal paste to a direct touch heatpipe see this link.
http://benchmarkreviews.com/index.php?option=com_conten...

And once again I am going to reseat the cooler, adding the 2500rpm, 120mm case fan I bought from the egg, and see what happens. In my bios system monitor, at stock settings, with lower frs voltage and lower bpu core voltage, I am hovering around 39 degrees. So that's where I am now.
June 26, 2008 3:04:22 AM

Anybody here think that lightly sanding down the heatsink base with 1500-2000 grit sand paper, is a bad idea? And then applying my ac mx-2 paste is going to make a difference? MY heat sink is the Xigmatek hdt-s1283.
June 27, 2008 3:24:21 AM

UPDATE on my high temp issue. So today I lightly lapped my xigmatek hdt-s1283 cooler. I made it pretty shiney, but not a mirror finish. I also applied artic cooling mx-2, and followed the directions for applying it, from benchmark reviews, which was recommended to me by Daniel who is a Xigmatek support tech. It only dropped my temp 1 degree. They are all still in the mid to high 30's, and the core with the high temp is around 60, at idle according to RealTemp 2.6. At this point I am tired of reseating that cooler, and I am just going to OC a little until I I hopefully get to 3.4-3.6, and let it ride. I am finacially stable, and if I fry it then i will just have to get another one. Thanks for everybody's help and recommendations on here. If anyone here finds out any more info on the Quads high temp issue, post it. I geuss there are alot of people on many different forums having the exact same issue. Thanks again.
June 27, 2008 10:52:24 PM

shadowthor said:
you definitely don't want Vcore to be at 1.4v, as it will cause your chip to slowly die.



that is correct, just with each breath you too are slowly dieing! In fact turning on any computer will slowly kill it!

if you sleep 10-12 hours a day vs 8 you might add a few days to your life, what have you gained and what have you lost?
Same with running your cpu at 1.4-1.46v - yes you will kill it a little faster, but for 99% of the users the cpu will become insignificant ( are using that P3 still?, are you using a 286 or 386?) before the cpu dies.

what kills it is heat and voltage - moderate voltage with good cooling may take 10% or even 30% of a 7-10 year life span at 24/7 server output which is 20-30 years of usage for normal people!

bottom line is you gain alot with overclocking and you loose little just like if you sleep the proper 7.5-8.5 hours per day with a mini nap!

lets say i am wrong and cpu life is 50% shorter - how much will that cpu cost in 5 years? yep about 10 bucks on ebay!

so the worst case is your out $10 years from now but your system is far faster and more responsive!

Warpedsystems 6 years of only overclocked systems and still "zero failures"!
June 27, 2008 10:57:19 PM

bryan240g said:
UPDATE on my high temp issue. So today I lightly lapped my xigmatek hdt-s1283 cooler. I made it pretty shiney, but not a mirror finish. I also applied artic cooling mx-2, and followed the directions for applying it, from benchmark reviews, which was recommended to me by Daniel who is a Xigmatek support tech. It only dropped my temp 1 degree. They are all still in the mid to high 30's, and the core with the high temp is around 60, at idle according to RealTemp 2.6. At this point I am tired of reseating that cooler, and I am just going to OC a little until I I hopefully get to 3.4-3.6, and let it ride. I am finacially stable, and if I fry it then i will just have to get another one. Thanks for everybody's help and recommendations on here. If anyone here finds out any more info on the Quads high temp issue, post it. I geuss there are alot of people on many different forums having the exact same issue. Thanks again.



thats is an excellent project - you should take notes on temps and test it in an house with AC so the room temps is the same.

a good lap will give 3 c reduction tyipcal - if the hs was bad more. some heat sinks are disked on purpose - thermalright.

thermalright need lots of pressure and extra compound but they work best when seated correctly --- sorry i am rambling here.....

3.6ghz is cake if you tune the system right remember the nb is key - cool it! you want 60-62c max and if you can get 1.6v give or take .05v at 50-55c you get your 3.6ghz and long life!







---------------------------
thermal compound

first rule: more is good too much more is bad! compound will flow and will find the right thickness - too much is bad

ever use alot of compound and pull off the hsg what do you see? compound all over the edges rolls of it! why is that? it flowed out!

second rule: how to apply to all hsf: your put on the cpu and trowl it even smooth from center to edge this mounds in the middle

third rule - more is good, add a second amout to the center if your using thermalright or direct touch and spread it out in circle making a shape like q quarter - why? that compound will flow and fill the gaps

fianlly, turn the cpu hsf as you tighen it or after its tighen this drives out air and starts the flow


why not apply it too the hs as shown in your link above? air gaps thats why! when you put the hsf on the cpu you might get air trapped.

heat and pressure cause the compund to flow - heat comes from the cpu and thermal dynamics drives the flow out
a b à CPUs
June 27, 2008 11:09:36 PM

dragonsprayer said:
that is correct, just with each breath you too are slowly dieing! In fact turning on any computer will slowly kill it!

if you sleep 10-12 hours a day vs 8 you might add a few days to your life, what have you gained and what have you lost?
Same with running your cpu at 1.4-1.46v - yes you will kill it a little faster, but for 99% of the users the cpu will become insignificant ( are using that P3 still?, are you using a 286 or 386?) before the cpu dies.

what kills it is heat and voltage - moderate voltage with good cooling may take 10% or even 30% of a 7-10 year life span at 24/7 server output which is 20-30 years of usage for normal people!

bottom line is you gain alot with overclocking and you loose little just like if you sleep the proper 7.5-8.5 hours per day with a mini nap!

lets say i am wrong and cpu life is 50% shorter - how much will that cpu cost in 5 years? yep about 10 bucks on ebay!

so the worst case is your out $10 years from now but your system is far faster and more responsive!

Warpedsystems 6 years of only overclocked systems and still "zero failures"!



Keep in mind that a 45nm CPU (like the 9450) will have more problems with a short life at high voltage than 65nm, and should not be pushed as high. I would probably not go over about 1.375-1.4 volts on a 45nm for normal usage. (the Q9450 is actually specified at 0.85-1.3625V, compared the the Q6600 at 0.85-1.5)
June 27, 2008 11:19:29 PM

the new 45nm due do hafnium enriched gates are subject to faster degradation - this is where the 50% factor came from.

while a pre-45nm cpu i.e. 65nm q6600 will get 10-20% loss of life from oc under 1.5v - remember good cooling reduces the process too.

the 45nm will get 30-50% .........but

there thermal controls to help that, and bigger Vdroops are build into the 45nm bios.

at 1.5v with 65nm your get a droop of .5v with intel or .8 with nvidia but those upto 1.2 with nvidia

so 65nm at 1.48v running 1.44v in orthos may run 1.4v in orthos at 1.48v with the newer bios

this explantion is short and may not be exact but its close enough


bottom line is with most systems running 1.4-1.45 normal usage your system will run for years and years at 24/7 ouput - most of which is sleep or idle






---------------------------
intel is very conservative and they warranty the product for 3 years for 24/7 usage with an expect life span of 7-10 years

under normal usage, computers will last a long time.

email me in few years i just shipped a few at 1.54v with q9450, i warranty the cpu 3 years, if i have to replace them i buy some on ebay for $25!
June 28, 2008 2:08:48 AM

Thanks for chimming in and helping me out on this guys. Being a noob half of what you guys said went way over my head, but I am learning stuff daily, and trying to apply it. Thanks again I will update some more this weekend.
June 28, 2008 4:17:31 AM

Hi Guys

I just wanted to ask a couple of quick questions about my Q9450 also and temps. Below is a screenshot of the temps I get when using 'Real Temp' and after running prime95 for 5 minutes at full load (i know it's not very long).

Anyway, should I be treating these temps as pretty accurate?

My Setup is a Q9450 on a GA-X48-DS4 Mobo
Corsair TX750 PSU
Corsair TWIN2X4096-6400C4DHX 4GB (2x XMS2 2GB) PC-6400 (800MHz) 4-4-4-12
Leadtek 9800GX2 Vid Card
Antec 900 Case
Arctic Freezer Pro 7 Cooler with the stock thermal paste it comes with. (I live in AUS and getting hold of the Xigmatek is no easy feat)



I've noticed that Core 0 sticks at 34c unless I put a load on it. This is ok though?

One other question in regards to overclocking. I overclocked it by leaving it at 1.25v which is the stock and just changed the FSB to 400. This brought it up to 3.2GHz. I then noticed that my RAM Frequency changed from 400MHz to something like 560MHz so I just manually changed it back to 400MHz. Am I right in saying that it the CPU and RAM is 1:1 as you can see in my screenshot it is currently running 5:6. Is this bad If I wanted to leave it all stock? As far as the RAM goes, when I first powered up the PC after building it, then RAM timings were 5-5-5-18 however I manually changed this to 4-4-4-12 as this is what the RAM is apparently capable of. Was this the right thing to do. It runs like a dream. I just don't want to stuff anything up and there is a lot of this technical stuff I just can't get my head around no matter how many times I read it over.

Anyway, after I changed those settings in the Bios, I also change the PCI setting to 100MHz rather than auto. I then started up my PC and all worked great. Games were great. Stress testing was great. and the Temps were only a little higher than what they are now. Hardly any difference.

Now the problem is I don't know much about overclocking and wasn't sure if I had change the settings correctly or forgot to change something. Is that all I pretty much needed to do? Does it sound right?

Anyway, Thanks for reading this and helping out.

a b à CPUs
June 28, 2008 4:38:52 AM

That sounds right to me.
June 28, 2008 7:32:59 AM

You have a sticky sensor on one of the cores, you can check by pressing test sensors in realtemp.
June 28, 2008 9:08:56 AM

thatguy2001 said:
Hi Guys

I just wanted to ask a couple of quick questions about my Q9450 also and temps. Below is a screenshot of the temps I get when using 'Real Temp' and after running prime95 for 5 minutes at full load (i know it's not very long).

Anyway, should I be treating these temps as pretty accurate?

My Setup is a Q9450 on a GA-X48-DS4 Mobo
Corsair TX750 PSU
Corsair TWIN2X4096-6400C4DHX 4GB (2x XMS2 2GB) PC-6400 (800MHz) 4-4-4-12
Leadtek 9800GX2 Vid Card
Antec 900 Case
Arctic Freezer Pro 7 Cooler with the stock thermal paste it comes with. (I live in AUS and getting hold of the Xigmatek is no easy feat)

]http://img230.imagevenue.com/loc57/th_25425_mycpu_122_57lo.jpg

I've noticed that Core 0 sticks at 34c unless I put a load on it. This is ok though?

One other question in regards to overclocking. I overclocked it by leaving it at 1.25v which is the stock and just changed the FSB to 400. This brought it up to 3.2GHz. I then noticed that my RAM Frequency changed from 400MHz to something like 560MHz so I just manually changed it back to 400MHz. Am I right in saying that it the CPU and RAM is 1:1 as you can see in my screenshot it is currently running 5:6. Is this bad If I wanted to leave it all stock? As far as the RAM goes, when I first powered up the PC after building it, then RAM timings were 5-5-5-18 however I manually changed this to 4-4-4-12 as this is what the RAM is apparently capable of. Was this the right thing to do. It runs like a dream. I just don't want to stuff anything up and there is a lot of this technical stuff I just can't get my head around no matter how many times I read it over.

Anyway, after I changed those settings in the Bios, I also change the PCI setting to 100MHz rather than auto. I then started up my PC and all worked great. Games were great. Stress testing was great. and the Temps were only a little higher than what they are now. Hardly any difference.

Now the problem is I don't know much about overclocking and wasn't sure if I had change the settings correctly or forgot to change something. Is that all I pretty much needed to do? Does it sound right?

Anyway, Thanks for reading this and helping out.



yes that is all right - lock as many things as you can - make sure to lock the pci-e too

yes running ram at 4-4-4 at ddr800 is faster then 5-5-5 at 1066 in most cases

3.2ghz at 1.2v is nice - you cpu will last until after armageddon and you can use it too research planting crops, making stills and passive soar heating!

seriously that is correct and nice spot 3.2ghz
June 28, 2008 11:03:12 AM

Lol Thanks. As far as the testing sensors in real temp goes, there was no movement for core 0.

As far as overclocking goes, I guess I will put it back to 3.2GHz in that case if there is no real issue with it.

Thanks heaps guys.
June 28, 2008 4:11:40 PM

thatguy2001, while I can appreciate that you have a similar issue. I do not appreciate you hijacking MY thread. If you wanna post a thread then do so, but it aggervates thread starters when someone comes in and takes over there thread. Now I have to sift thru all your threads to get back to mine. Please have some respect. Thank you.
June 29, 2008 3:54:04 PM

No problemo, just remember for the future.
!