Speedfan System Temp very high - advice?

[Heartborne]

Hello all!

I have a pc I built myself a couple of years ago. I haven't really paid attention to the operating temps because I have never had any issues. However, one of my case fans has died and another was connected by a swing-away bar which was left elsewhere.

The system is thus:
Raidmax Smilodon Gaming Mid-tower
PC Power and Cooling 610 watt PSU
EVGA 780i mobo w/nforce 4 chipset
4 GB corsair dominator 1066
Intel core 2 quad q6600 kentsfield
PNY XLR8 Nvidia Geforce GTX 260 (new)
WD 300 GB HDD
64-bit Windows Vista

This machine has always run a little on the warm side, mainly because of cables interrupting the airflow and the case being a little smaller than what it probably should be. However, the temps have never been extreme until now.

I have noticed in particular that under load the system temp shoots up to 104 degrees celsius. At idle it is sometimes over 80 degrees celsius. I'm not entirely sure what the system temp refers to, but this seems to be really high. Does anyone have any advice on what this temp is and how I can reduce it?

Thanks!

[123456]

104C? O_O That's dangerously high! I'd suggest going into your BIOS after doing something intensive like video encoding or gaming and then read your PC temps under "System Health" or similar. Those may give us a better indication of your temperatures.

[jlreich]

I agree with 123456. Check temps in the BIOS. That temp is way too high and should have burned up the north bridge before getting to that temp. Sounds like Speedfan is giving a false reading.

If there is a fan on the NB make sure it is functioning.

[Heartborne]

Thanks for the advice, guys. I took a look at my "system monitor" in the BIOS and it gave me the following:

CPU 62 degrees c
Board: 47 degrees c
MCP: 63 degrees c

I think I may have figured out what is causing the very high reading. Some time ago, I removed the stock cpu fan and replaced it with an aftermarket one with a nice massive heatsink... but I had a problem! It was too large to install properly in the case. It was summer, I had no A/C and I had to make a call... risk burning out my cpu (which was idling around 80) or install the heatsink with the fan pointed down. I installed it with the fan pointed down and monitored the temps for a while, found that the cpu was running much cooler and that the other temps were not affected. The fan points down at the northbridge, which has its own fan and also has a case fan blowing on it.

Recently an 80mm fan that blows directly onto the cpu heatsink stopped working. I imagine this helped to cool down the heatsink so that the air blowing out of the fan was cooler.

My theory is that there is a little pocket of hot air sitting right between the cpu and the graphics card, over the northbridge. The heat sensor is picking up this hot spot and that's why I'm getting such a high reading.

Of course, that doesn't explain why the reading for board in the bios is so much lower. I will replace the 80mm fan and see if I notice any immediate improvement.

All of my other temps are a little high but ok. Most of them are in around 60 at idle. GPU goes up to around 80 under load, but cpu stays under 70. A little higher than I would like, but I've been running it 3 years at these temps and it seems to be just fine. Ideally, I should have built this machine in a larger case. Next time I build I'll go for one of those big expensive cases that have more room inside to set things properly as well as allow better air flow.

[jlreich]

but I've been running it 3 years at these temps and it seems to be just fine.
As I recall you have had performance issues with this system since you built it. Everything from low benchmark scores to crashes. But if you are happy with the performance and don't want or can't do anything about it, it is up to you.

[Heartborne]

JLReich, I'm afraid you misunderstand the history I've had with this build. At the VERY beginning when I first built the machine I had two major issues. The 8800 GTS that I used in the original build did not have the kind of power I thought it did. I fixed that by installing the 9800 GTX.

The other major issue was a crash issue caused by my misunderstanding of how the RAM could be overclocked. My motherboard BIOS has a feature that allows the RAM to be overclocked "safely." What I didn't know is that it only supports up to 2 modules. I had 4. When I turned the feature off the crashes stopped.

So, once I fixed those two issues I have had no problems with the system up until last week, when the 9800 GTX died on me. I replaced it with the GTX 260 and now it works fine. I've run it with no issues in the meanwhile. So yes, I had 2 issues at the very beginning of the build and I took the necessary steps to resolve them both. Benchmark and gaming performance are both fantastic and I have had no crashes since then.

The issue I'm looking at now is just that "system" temp reading so high in speedfan. I think (and hope!) it's probably a false reading, but there are some veterans on this forum who are much better able to understand this issue than I am.

Does speedfan have a different way of reading the sensors than the motherboard system monitor does? Is the high temp due to the absence of a sensor? I've read that can be an issue. I imagine if the northbridge were really that hot my motherboard would have died by now. Am I mistaken? Do modern systems have a failsafe that causes them to shut down if they start running too hot? Also, wouldn't one extreme temperature cause a lot of heat inside the case and raise all of the temps?

[jlreich]

I'm afraid you misunderstand the history I've had with this build
That's what I get for trying to remember. :p

Does speedfan have a different way of reading the sensors than the motherboard system monitor does? Is the high temp due to the absence of a sensor? I've read that can be an issue.
It's a matter of the program having the right drivers, and well written drivers, for your particular sensor chhips. Also some monitoring programs read them more correctly than others. And that is not to say one is necessarily better than the other, but that one may be better at reading your particular sensor chips, while the other may be better at another set. Another thing is there usually seems to be at least some discrepancy between readings in the BIOS and software readings. It's rare that they are actually dead accurate. If there is ever any doubt always trust the temps in BIOS over any software readings. ;)

Try HWMonitor (http://www.cpuid.com/softwares/hwmonitor.html)from CPUID to see if the readings are the same or different.

I imagine if the northbridge were really that hot my motherboard would have died by now.
If the NB was really over 100C it would likely be transistor goo by now.

Do modern systems have a failsafe that causes them to shut down if they start running too hot?
Yes, but either it is disabled in the BIOS, or they simply don't work as they are supposed to. I have seen both cases.

Also, wouldn't one extreme temperature cause a lot of heat inside the case and raise all of the temps?
Yes, especially if you don't have good airflow in your case.

Another thing to consider is do you know for certain the 100+ reading is your NB? It is not unusual for a PSU to be around 100C. After all even an 85% efficient PSU loses quite a bit of power in the form of heat.

Also, the sensors can sometimes get stuck. It happens. I have a sensor for core #3 on my CPU that always reads 38C. It's been that way since I got it. No big deal, I know the other cores run in the high 20's to low 30's, so it's all good.

If the NB was really over 100C you would not be able to touch the heatsink without burning yourself. So ground yourself and if you can hold your finger on it for five seconds or more the temp should be fine. If it is too hot to hold your finger on it even for a couple of seconds then it is too hot and there is a problem.

[Heartborne]

JL, thanks for all your advice. I wonder if it is the psu that is getting that high reading? That would make sense! I thought "system" was a pretty vague description for it.

Here is the data I got from Hardware Monitor. This is after about 2 hours of playing WoW, so the system was under load for a while. these temps are, obviously, before the system began to cool down just after I closed the game.

http://img802.imageshack.us/img802/751/hardmon.jpg

[123456]

Those temps are fine, a little high for CPU temp, but not too concerning, IMO. There must be a faulty sensor/reading for that 126C reading.

[mjc]

If you played a game and those were your temps just after you quit...something is lying to you. A 'system' temp hot enough to fry an egg should SMELL hot...

[jlreich]

It is odd that it would get labeled as system, but it does look like that 126 temp is the PSU.

but not too concerning, IMO.
If those were the temps in my system I would be freaking out! :eek: The CPU as high as 73 and the cores reaching 80. :eek: Heck, I wouldn't even turn my system on until I came up with a solution. ;)

But if there has been no performance issues, no shutdowns, then these have to be false readings. Either the sensors chips are bad or you got a bunch of bad thermal diodes. Something is wrong. :confused:

[Heartborne]

Thanks for the input everyone. I think I can rest easy about that "system" temp, as JL said it's probably the PSU.

As far as the other readings go, either they are false readings or things are running just a little higher than they should be. I've always had readings like this and they've never caused me any significant problems. I have looked up the "safe" temperature range for all of these components and these readings do fall within them, but many users say they have much lower temperatures. I wonder if ambient heat in the room could have something to do with it?

In any case, since I don't have any performance issues right now I think I can let it rest. thanks for taking a look everyone. I think I will house my next build in a full tower that has better air flow and maybe I'll work on better cable management, as I'm sure having a mess of cables hanging in the case is not helping the air flow.

Thanks again!

[mjc]

I wonder if ambient heat in the room could have something to do with it?

Yes...

Let's say one system is in a room kept at 68 F...and an identical system is kept at 75 F....the one in the 75 F room is always going to run hotter than the on at 68.

[jlreich]

You can't cool something down very well with hot air.