Hello,

A few weeks ago I bought a new PC with a dual core processor (3800+). In the last few days I checked the temperatures with a few programs like everest, Sisandra, SpeedFan etc but each say soemthing else. what I mean by that, is that they all detect the same C numbers but associate it with a different part. for example, Everest reports the MB is 32C and the CPU is 70C while PC Wizard 2006 reports CPU as 25C. SpeedFan I dont even know which part is which since it only names them as Temp1, Temp2, Temp3 etc.

Problem is, that Everest, SensorsView and Sisandra (partly and I'll explain why) think its the CPU at 70C. when I say Sisandra is partly saying its the CPU, its because under "Mainboard information" it says CPU 1 temp 34C but CPU 2 temp 70C while on the "CPU & BIOS" section it says each core is 32-34C. and the tip it gives me is that the mainboard is too high temperature. so as you can see they're conflicting their own readings.

When I go to the BIOS it always says the CPU temp is around 35C and I even set it to alert me at 60C.
Last night, while being on windows I started to hear short one second beeps coming from the inside of the PC and I had to shut the PC down to make it stop. when I turn it on a minute later it was fine as usual. this morning, I guess after PC was turn on for max an hour, it started with the beeps again. and again now as I type this message and I was clicking on "Mainboard information" in Sisandra it started again so I restarted the PC.

I have an Award BIOS and looking here it says "If an Award BIOS detects that the CPU is overheating it may play a high pitched repeating beep while the computer is running." - by running I assume they mean while being on windows because I dont get any beeps at startup (except the one long beep that tells me all is OK). anyway, according to that page it sounds like a CPU problem. the CPU fan is spinning fast as usual and the other day when I touched the side of the heatsink, it wasn't very hot. so I dont know what to believe, some temp programs say its the CPU overheating at 70C while the BIOS say 34-36C. I came across a forum where someone thought maybe its a problem with dual cores doubling the temp. for example if the temp is 35C then Core 1 35C + Core 2 35C = 70C. it does make sense in a way but still I cant count on that, esp. with the beeps I started to get last night.

I was wondering, is there a way that one core can overheat more than the other? I thought they are close to one another but Sisandra reports the second core as 70C while the first 34C. its a huge difference. I dont overclock at all. I have a ThermalTake Soprano Black case with one 120 fan at the back, 120 fan at the front and 80 fan on the side panel. the Videocard is passively cooled (Gigabyte 6600GT - NX66T128VP model) as well as the MB (Gigabyte K8N-SLI). here are some screenshots of the temps in different programs, any help would be appreciated!

Everest:


Easy Tune and SpeedFan:


Sisandra


Sisandra Warning


SensorsView:

Have you checked the cd that came with the mobo for any monitoring software? If there is any on it then it may be more accurate!!!

Thanks for replying,

Yes the software that came with my gigabyte motherboard is Easy Tune 5 which you can see in the second pic I posted. it says its 33C but the problem is knowing if it counts the second core as well or if its defective. I find it weird that in Sisandra for example it does show the same Celsius as Easy Tune but only for the FIRST core. the second is 70C which is too strange. also, Easy Tune only shows VCoreA voltage - why not listing the second core voltage too? the beeps started last night, I think it is a temperature warning. I wish the BIOS also listed the motherboard/Chipset temperature but I only see the CPU. for all I know it could be the motherboard overheating but I have no way of knowing that.

I just ran Burn-In Computer in Sisoftware Sandra and it still shows CPU2 at 70C and 2.56V unlike the first core at 36C and 1.34V. it doesnt look normal to me

Burn-In-Temps


Burn-In-Voltage

Go into BIOS (CLICK DELETE KEY RIGHT AFTER YOU PUSH THE POWER BUTTON) then go into "PC Health Status" use the arrow keys to navigate and click enter to go into setting. It will tell you the temp of your cpu. Click esc (escape) and click y to leave without messing around with anything. That will tell you the real temps. First do some CPU intensive programs, such as a game.

I know the BIOS shows its around 36C just like the first core showing in Sisoftware Sandra. if you look at the second image I posted in my first post you will see Gigabyte's monitoring program Easy Tune says its about 32C - but if you look at the left side, you can see it ONLY lists VCoreA - What happened to VcoreB?? if the second core is not faulty it should show its voltages too but it doesnt.
When I ran many programs to see if the temps jump, 34C or so jumped to about 41-43C and 70C jumped to 71C (71C is the max according to AMD).

I have the same board and CPU. The temps given by ET5 are 100% accurate and aren't bad. Play a game or some intensive CPU usage then check the temp in say, 30 Min's and see what it is saying. Along as it doesn't go over 60 C you should be fine. I will do some more research and post back. But I think it is nothing to worry about and the software is reporting back faulty temps.

The other day I ran "SuperPI" and my temps jumped like I said to 42C and 71C it cant go over 71C



Since you have the same CPU, does Easy Tune lists VcoreB for you?? and if you have Sisandra what does it tell you under "motherboard information" catagory? I'm just worried that the 2nd core is defective because if it is and I'm not getting the most of my CPU or it is risky then I would prefer to try and replace it with a good one.

Well I think the problems aren't with the CPU, it the psu. It is not giving out sufficient amount of power on the 5v rail. You are going to need a newone, man how did this go unnotice. Everyone in this thread is getting sited (including me) .( lol that means nothing)

I thought the 12V rails are most important.

This is my PSU by Thermaltake, it says 28A on the +5V rail in max load.
http://www.thermaltake.com/nl/produc.../w006162-1.asp

In all your pics it says about 3.5v on the 5v rail. That is just to low, even fot that. Oh, and in Easy Tune 5 it says nothing about the second core. Only states VCoreA.

But maybe nothing much is using the +5V rail, the components use mostly the 12V rail. what should be the average voltage on the +5V rail? if I were to go and replace it I need to give them a real reason.

Eveything that has a red wire from the psu connected to it is using +5v, that includes all hard drives and all optical drives as well as your mobo.

Perhaps this may help [url]http://www.extreme.outervision.com/index.jsp[/url]

Thanks, but I already used the PSU calculator before I bought the PSU and it gives me 361W and my PSU is 420W with 2 +12V rails of 25A together and +5V rail of 28A. shouldnt 28A be enough?? I have just one SATA II HD and 2 DVD burners.

My PSU cables are black, you can see here in the pictures:
http://www.thermaltake.com/nl/produc.../w006162-1.asp

Had a look at your psu link, it states Load Reg. ± 5% and Ripple V(p-p) 50mV. The readings you are getting are way outside their tolerances so it could well be a faulty psu.

Well I don't really understand what is Ripple V(p-p) 50mV. what should the average +5V voltage be? I mean how higher than what I get now.

P/S- If you look at the picture above of Everest, you can see it also has "+5V Standby" as well as in the SpeedFan screenshot I have TWO +5V readings. maybe they should be added up.

The 5v standby shouldn't be added, its a separate supply that allows the psu to constantly be in "standby" mode so that it can power on when you press the power button or start your pc from keyboard, mouse, or modem when they are set up to do so.

The Ripple V(p-p) 50mV. P-p means peak to peak. In an ideal world your psu would give a constant 5v with no variation, this is close to impossible with out spending a large amount of money. Your 5v is a dc (direct current) supply, p-p 50mv means that its regulated so that no power spike should exceed 50mv under normal operation.

The thing is they also stated 5v +/- 5% which means your 5v should only vary from 4.750v to 5.250v, if its dropping to 3.5v that is way outside the 5%.

The best thing you could do is connect a digital volt meter up between a red wire and any black on a power molex that's unused and monitor it closely while your pc is running and doing things that will stress it like playing games or running benchmarking software like aquamark3d. If the 5v varies outside the voltages i mentioned earlier then it'd be best to take it back and get another one under warranty.

So you're saying it always needs to be from 4.750v to 5.250v. I do see it sometimes at 4.5V or even 5V+ but I guess usually around 3.5V. could I have low voltages because this PSU has PFC function (improved energy efficiency and reduced current load on AC delivery systems)? its suppose to reduce so maybe its doing just that.

I dont really have a digital volt tester so the only way is probably to take it to the store. I would hate to think something is wrong with the PSU or the CPU, that would mean waiting quite a while for them to first check and verify it and I'm left with no PC to use during this time.

The 5v should always stay close to 5v, if it doesn't then there is a problem. If it was meant to be at 3.5v then it would be called 3.5v. If the psu is faulty (and i believe it is) then it could damage your system if gets any worse. If it was me, i wouldn't take that risk.

I will try to call them tomorrow and see what they say. maybe next week I will take it to them, I just hope it wouldnt take them long to give me a new one