In 2008, with some advice from this forum, I ordered a lot of bits and my husband built them into a computer for me.
All has been so well in the meantime that I have pretty much forgotten what's under the hood though thankfully I kept notes of it all.
In the last couple of weeks I started to have a problem where the computer would freeze for ten seconds every minute. Gradually it got a bit less frequent but still disturbing. In the event viewer I couldn't see anything significant going on so I trawled around online a bit and looked at some similar problems in various forums. On their advice I installed Speedfan and a trial version of Active SMART.
The speedfan readout is confusing to me but with the program still on it's default settings it seemed the GPU was running too hot (53C) and the CPU was occasionally getting overwarm (highest 62C).
Taking it apart and giving it a good cleanout this morning has reduced the idling temperature to GPU 49C and CPU 46C and the stuttering effect has gone. In my relative ignorance I get the impression all is still not well and all it takes is a quick game and the GPU and CPU fly back to overhot (according to Speedfan).
From what I can read around possible actions would be
Remove the heatsink, clean and then reapply thermal paste?
Get a new fan (this one has always been noisy despite sold as quiet)?
Something else I don't know about?
I'd very much appreciate some guidance here as I feel I'm a bit out of my depth
Actually, the temperatures you have are pretty good. 49C Gpu and 46C Cpu are pretty normal. What temp is your room at?
The freezing probably had to do with the CPU reaching its thermal limit and slowing itself down for a bit to keep temps from running away. I had a laptop that did exactly the same thing for a while and when I took it apart the heatsink was completley clogged with dirt. I would suggest opening the computer every 3-6 months and cleaning out the dust to prevent it from happening again. Just be careful to keep yourself grounded while doing it.
1) Your temperatures are not that bad, particularly for older technology. Graphics cards run hot, and they are built to do so. Do not worry until the temp approaches 100c on the graphics card and 70c on the cpu. You can configure the warning temperatures in speedfan to anything you like.
Note that speedfan will show you the SMART info if you run it as administrator.
2) If a cpu gets too hot, it will downclock itself to preserve itself from damage. That could account for some slowdown, but not a "freeze".
I might suspect some malware/virus or strange started task. Use the task manager to look for tasks you are not expecting. Use cpu-z to monitor the multiplier to see if it is downclocking.
3) Your cooler is ok, but not great. The 92mm fan will be a bit noisy. It is a pushpin mount which can be tricky. If you do try to reseat your cooler, remove the motherboard from the case. You need to be able to look at the back of the motherboard to verify that all 4 pushpins are securely through and locked. Modern cases will have a large opening behind the motherboard to let you do this. Remove all of the old tim, and reapply new. Arctic silver 5 is a good one. If you are so inclined, the CM hyper212 or Xigmatek gaia are better coolers with quieter 120mm fans. They have more secure backplate mounts. Price is about $30.
4) If a fan is noisy, it is usually because it is spinning at high rpm. Some fans will be more efficient and will make less noise at the same cfm.
Do some research at www.silentpcreview.com for reviews of quiet components.
Thanks for the reassurance that the temperatures are not that hot. I hadn't known and my trawl around had not brought me to any specific answers for the CPU and GPU I have, just a forum comment that CPU at 70 for more than a second or two was not good. (This forum actually, now I come to check http://www.tomshardware.co.uk/forum/278950-10-ntel-core...)
Certainly this morning's cleanout did get rid of a lot of dust. Probably around six months since it was last done. I'd hoped for a better reduction in temp from that but at least if it got rid of the stutter it was progress.
My first thought had been malware and I did scans with Avast and Malwarebytes and all clear (I'm pretty paranoid and run NoScript and adblock etc to avoid things getting into the computer)
I'm still considering a problem with my HD and am backing up using TrueImage just to be on the safe side but the fact that the stutter went when I cleaned up probably points to overheating to some extent, even if not severe?
geofelt, thanks for the advice on the cooler itself. As I've never been happy with the noise levels perhaps I will treat myself to a new cooler. The silentpcreview site looks very useful though some of the coolers look even bigger than the one I have which is quite a tight squeeze already! Xigmatek seems promising.
My husband checked all the fans were physically spinning this morning. Speedfan says the CPU fan is running at 1300 rpm or so. A second fan is running at 1562 rpm which I suppose might be the GPU fan. All the other fan options on Speedfan are at 0. There is an intake and exhaust fan. I'll definitely try taking the side off and see if that makes a difference.
Thanks again for giving me your time, I appreciate the guidance
Just to clear up, 70c IS too hot for the cpu. If it's running that hot, there's a problem that needs to be addressed. Just cause it can "handle" 80c before throttling doesn't mean you should run it anywhere near that. And to clear up any further, not all cpu's can handle 80c before throttling. Some AMD cpu's are recommended never to go over 62c.
A good line not to cross with any cpu is 65c and I'm talking about at 100% load using a program like Prime 95 or Intel burn test. If its getting over 65c in normal tasks or playing games, that's a serious issue.
65 under stress tests is a little bit overkill. I assures the life of the CPU but for most people the CPU's heat death at 70C will usually be beyond the usable life of the computer anyways. I might agree with you if it was a server, or was often approaching that kind of load. For most consumers they will never need that kind of temperature control.
If you are setting absolute safe guidelines for all purposes and CPUs I can certainly agree that those are good general guidelines.
You've taken some good steps, but without knowing the loads on your cpu and gpu, and the temps for that load, its hard for us to eliminate heat as a problem. So let's try known loads and a known measurement tool.
1) For the moment, restore your vid card to defaults if you used software to change the fan profile, etc. Easiest way is to uninstall programs that do that kind of thing.
2) Download Prime95 (to drive your cpu at 100%), Furmark (to drive your gpu at "100%"), and CPUID's Hardware Monitor (to measure temps). They are all free.
3) Run Prime95 for about 30 min watching temps. For this purpose, stop the test if temps climb above 75C**. It would not be unusual (though not the best depending on your day to day usage) for a fully driven cpu to hover in the low 70Cs.
4) At 30 minutes, start Furmark in a Window to drive your gpu to 100% and run for at least 10 min. Watch the temps that Furmark shows, as well as Hardware Monitor. The gpu temps may well hit 85C and hover around there. That would be OK. If temps crack 90C**, stop the test.
** these temps are safe for the duration we're talking about here.
At the same time watch your cpu temps. Now that *everything* in the case is getting hot, case temps may rise and force cpu/gpu temps even higher. If this happens, you should investigate getting better air flow thru your case. (You can simulate "best" air flow by removeing the towers side panel and laying it on its side, repeating the tests.)
If your PC passes these tests, it isn't heat that's causing your problem, and we can move on to find what is.