Speedfan reports temperatures above 100 C? (Pentium 4 3.0ghz -- Northwood?)

Hi all,

I'm running a Dell Dimension 8300 with a 3.0ghz P4 -- service tag 9YJKX41. I think it is of the "Northwood" variety but I'm having difficulty finding info on this. The machine is from 2004, if that gives an idea of the CPU die size/ voltage/etc.

When the system is idle, Speedfan reports "Core" temperatures between 60C and 70C... which already sounds quite high. But, when I play an intensive game such as Civ 4, if I alt tab out of the game and run Speedfan, I see temperatures above 100C, which sounds insane. I have a "factory overclocked" GeForce 6800 GS by BFG, and I've also wondered if ambient heat put out by this device might be the major culprit. It is just a couple inches away from the CPU.

A few questions:

1) These are awful, destructive temperatures, right?

2) Speedfan is probably giving an accurate report, right? Each sensor appears to be giving a unique, variable temperature report within a believeable range, and I can't think why I would doubt these readings.

3) The particular design of this case has a plastic sleeve which is connected to a fan and which locks in place over the CPU/CPU heatsink. Any thoughts on whether this might interfere with the installation of a better CPU cooler?

4) Is proper installation of a CPU cooler and heatsink compound rocket science? I want to fix this situation, but I'm afraid that if I'm incompetent I will just destroy my CPU or some other expensive component.

 

I would suggest running a different temp monitor just to verify that SpeedFan is accurate; I've had SpeedFan report some erroneous temps myself. Also, check the BIOS and go to the Hardware Monitor (aka PC Health) section and see what temps are reported. Personally, I use HWMonitor to check temps and voltages..... and, yes, the temps you've posted are MUCH too high. Installing a new CPU cooler and thermal paste is not rocket science, but I have seen many people use way too much paste, and not securely mount the cooler to where it was barely hanging on the motherboard and was loose and completely ineffective. Do some quick research online and you'll find PLENTY of how-to guides for cooler/paste setup.
:-D

 

Take the side off the case,run your game until the computer is nice and hot,check your reading in speedfan then touch the CPU heatsink,if its nice and warm/hot you can be pretty sure you don't need fresh thermal paste if your worried about the procedure and the speedfan reading is erroneous.

If its ice cold you may need fresh thermal paste,if you can't touch it and you wouldn't if your CPU temp was 100c you may need cooling installed,but as Dlb said try other temp monitoring programs first.

 

If your CPU really is at 100C, then that is very, very bad! :eek 60-70C sounds like a normal run temperature under load to me... my watercooled C2D sits around 65C when running Prime95 on both cores. However, it is definitely not a good temperature for when the CPU is idle!

SpeedFan is great and I have always found it accurate, after I got it configured right. If you don't get it configured right for your system though, you won't be able to make sense of the numbers. Check that you have the temps labelled right, and that they all get their data from the correct sensors. Definitely get a second opinion from another hardware monitor as well, like the HWMonitor dlb suggested.

Dell did not build their Dimension product line for gaming, so it's entirely possible that the CPU cooling just isn't adequate for that amount of CPU load for long periods of time. If you have some extra cash and you know what you're doing, you might be able to solve your problem by taking dlb's advice and replace the cooling paste and CU cooler. (I was going to suggest that and he beat me to it! )

Arctic Silver is a great cooling paste, and very affordable for the quality you get. A high-end CPU cooler and fan could potentially cost you a lot more, depending on what you get. Generally speaking though, anything you can find is probably going to be an upgrade from a stock Dell cooler...

EDIT:
Also, get a can of compressed air and blow all the dust out of the stock cooler in there, if you haven't already. Dust acts like a fur coat, and if you're lucky, just removing that will get your temperatures down to something much better, while you research a long-term cooling solution.

 

The Dell 8300 is actually quite well engineered. You would have great trouble improving upon their fan and cowling.

However do make sure that they are properly clicked into place and clean.

You can't simple take the side off this case as it is Dell suitcase type. Watch your fingers as it is also very heavy. It can be operated on its side with the clamsheel open however. But you obviously can't use the CD.

Modern pentium processors are claimed to be 'impossible' to overheat, insofar as they have several proprietary thermal shutdown and protection systems. the chips actually run more and more slowly as they heat up, so consuming less power and heating up less. Very cunning. they also have more drastic shutdown sensors.

 

For some reason, I just can't picture that of a Dell... but if you say it is, I'll take your word for it. After all, I don't have an 8300 here to study.

 

dELL says 50-95 degrees,

http://support.dell.com/support/edocs/systems/dim8300/sm/specs.htm

 

http://support.dell.com/support/edocs/systems/dim8300/sm/index.htm

The green plastic in the first picture is a general view with the cowling down, but the main case on its side and open.

I have raised the cowling in the second pic to show the substantial heasink / fan arrangement.

One nice thing about these designs, you don't need a screwdriver to do most of the servicing.

 

Well, I feel like a big idiot. :-o

I was reading the displays wrong. Upon closer inspection it looks like my NVIDIA CARD is the item burning up with temperatures up above 110C (!). This is noticeable from some minor graphical artifacts while playing. The CPU, meanwhile, appears to be chilling along at around 45C while idle, not 70 as I previously thought. Ambient is also at 45C while idle.

I'm not thrilled about the NVIDIA temp (I wonder if this is a result of the "factory overclock" on this card (BFG Geforce 6800 GS "OC")... but then again I seem to be finding other people complaining about their non-overclocked cards running this hot. Well, if this sucker wants to melt itself I guess there's nothing I can do to stop it.

Thanks for the comments, especially the one reminding me to make sure the heatsink assembly is clicked into place. I had removed that once before and didn't think to double check it, but it seems to be snug.

 

Whoa... those temperatures for a graphics card is not good either! Look for dust in the video card fan and heatsink? I have a BFG 7800 OC in my back-up computer that has been running there for a couple of years now. I dust out the heat sink once in a while and it runs just fine.

How long have you had the card? If you registered it on BFG's website within 30 days of buying it you have a limited ten-year warranty on it, if you didn't, your warranty is a year. If you're still in that time period, email them and ask for help and possibly an RMA. Cheaper than buying a new card, for sure.