How much heat will you tolerate for lower volume?

crispyfish · Post by **crispyfish** » Wed Feb 28, 2007 11:56 am

The title pretty much says it all... how hot are you willing to let your CPU, GPU, hard drives, etc. get in order to make your system more quiet?

Right now my temps are about as follows:
- CPU: 33 C
- GPU: 53 C (passive)
- Hard disks: 30 C

This is all before undervolting my Nexus fans... it'll be interesting to see how much the temps go up when I do, and whether I'll be willing to live with it.

-CF

Erssa · Post by **Erssa** » Wed Feb 28, 2007 12:48 pm

crispyfish wrote:The title pretty much says it all... how hot are you willing to let your CPU, GPU, hard drives, etc. get in order to make your system more quiet?

Right now my temps are about as follows:
- CPU: 33 C
- GPU: 53 C (passive)
- Hard disks: 30 C

This is all before undervolting my Nexus fans... it'll be interesting to see how much the temps go up when I do, and whether I'll be willing to live with it.

-CF

I'm very careful about temps and try to do everything possible to stay at safe temps. I would never want to overheat my system! So I'm happy, if my CPU is under 70c at load, gpu under 110c at load, and hard drive under 50c. That said, my cpu idles at 43c, gpu 57c and enclosed raptor at 44c.

pyogenes · Post by **pyogenes** » Wed Feb 28, 2007 12:51 pm

CPU: Mine idles in the 50's and goes up to 62C under load. I can't reduce fans any lower without risking them stopping. But that's with a hefty overclock - I'm sacrificing heat for performance, not silence.

GPU: Stuck a HR-03 on my X1950XTX using a 120mm Yate Loon suspended below the heatsink with some string. It's a ton cooler and much, much quieter than the stock solution. No compromise here, it's a win-win situation.

HD: Mine idles around 40C. Drive enclosure in the lower compartment of a P180B with only the PSU moving air so it's a tad high due to minimal airflow. Well below the maximum operating temps (60ish I think) so I'm not concerned at all.

jaganath · Post by **jaganath** » Wed Feb 28, 2007 1:22 pm

something went wrong with my CPU fan recently and my Sempy 3000 started idling at 46C. That to me is unacceptable, considering how cool a chip it is; so I put a Sharkoon Silent Eagle 80mm on it and now it's almost inaudible and much cooler.

CPU: 31C
GPU: integrated; very cool
HDD: 36C

Good fans go a long way toward combining performance and quiet.

sjoukew · Post by **sjoukew** » Wed Feb 28, 2007 3:21 pm

I consider the highest room temperature in summer when it can be hot. In that environment at full load I want the temperature at least 10 degrees below the maximum operating temperature of each component.

This makes that I never ever have to worry myself about my computer overheating.

The Asus Q-Fan (motherboard) and AtiTool(software) are doing the job slowing down the fans.

andyb · Post by **andyb** » Wed Feb 28, 2007 4:02 pm

Ignore my signature.

Some VIA based ECS mobo.
3200+ 939
x800XT PE
400GB Samsung SATA
500GB Samsung PATA

My GPU is idleing currently at 49C passively using an AC VGA Silencer 5.

My 400GB HDD is idling at 33C, and my 500GB is not spinning and is at 23C.

According to speedfan, the core is at 39C, TEMP1 is at 35C, another TEMP1 is at 40C, TEMP2 is at 33C.

the CPU fan on the ninja is a Nexus at is annoyingly loud at 12V, the top hole on the P180 is blocked, the case fan at the back is a Nexus at 5v, and the HDD/PSU cooling fan is a Nexus at 5v.

No complaints from me for my temps. As far as temps are concerned so long as nothing gets within 5C of its operating temps I am happy except HDD's which I dont want to let get beyond 45C ideally and 50C at the absolute most.

Andy

crispyfish · Post by **crispyfish** » Wed Feb 28, 2007 4:47 pm

After undervolting my Nexus case fans (2x92mm intake, 2x80mm exhaust), the CPU cores have gone up to 40C. That's acceptable IMO. I left the CPU fan (also a Nexus) at 12V because there doesn't appear to be a noticeable acoustic difference. Slowing down the 80mm and 92mm versions makes a big acoustic difference though.

dhanson865 · Post by **dhanson865** » Wed Feb 28, 2007 4:49 pm

I've seen too many drive failures due to heat so I take HD temps very seriously.

I don't want my HD temp over 40c under any circumstances (load, summer, case open or closed, etc). I consider the ideal idle temp for the HD to be equal to the lowest temp you can monitor by software assuming that temp is below 30C. For example my current temps in speedfan are:

33C (CPU)
31C (Drive 0)
40C IT8712F (sensor)
38C IT8712F (sensor)
39C (Fintek F75383 (sensor))
28C (Fintek F75383 (sensor))
40C ACPI

The HD is 31C and is higher than the 28C seen elsewhere in the case. The difference is because I have the fan in that drive cage running at 7v. If it were running at 12v the HD temp would be below 30C and would actually be the lowest temp in the case once the system is on long enough to reach stable temps.

If I was the type of user that never put sustained loads on my hard drive and I had noise issues with the fan speed required to keep it below 30C I would be willing to slow a fan some but only if the idle temp could remain below 35C.

CPUs are more forgiving than hard drives. I watch temperature and stability though there is no single temp range that is safe. Each CPU has its own safe range. You have to know detailed info to be sure you know the safe temp for a specific processor.

I'm currently running a Athlon 3000+ slightly overclocked with speedfan set such that no matter how long I run an app the CPU won't go above 42C. My CPU could tolerate much higher temps but short of running prime95 or rthdribl.exe or some such for extended periods I can't hear the fan on my stock CPU heatsink. I suppose you could say I'm not willing to go above 50C for the CPU but my current build does not come anywhere near testing that issue.

Video Cards are more forgiving and I go by system stability and graphics quality. I usually don't care how hot my video card gets. Really, I've never had a video card get too hot even when overclocking. Of course, it does matter that you buy a card with a decent thermal profile to begin with but if you get a good card, heat won't be an issue.

The coldest day this winter I've seen stable idle temps for my CPU around 26C and my HD around 27C. I guess I'll find out this summer if I'll have to adjust anything to address heat or noise...

Erssa · Post by **Erssa** » Wed Feb 28, 2007 5:45 pm

dhanson865: Your hard drives are dangerously cool! I suggest you turn that fan down to increase your hard drive temps for better reliability.

viewtopic.php?t=38343

http://labs.google.com/papers/disk_failures.pdf

Sarcasm doesn't translate so well over internet, so I will just to clarify, that I think you are cooling your hard drives for no gains. If it makes you feel safer, fine. But cooler temps won't make them any more reliable as shown by the google study.

PopCorn · Post by **PopCorn** » Wed Feb 28, 2007 5:48 pm

well i dont really have "limits" cuz i try to get every thing as cold as i can with out having noise issues....i myself perfer the cold... its winter and i sleep with my window open and a fan blowin ( not on me, just to move air around)

CPU 19.4C (in 18C room)

( i said i liked the cold)
Hard drive 0 24C
Hard drive 1 23C
GPU N/A (My video card is like 10 years old not kidding the thing doesnt even need a heatsink on it, im waiting for my 8800gts to arrive because i didnt have a PCIe video card to put in untill the 8800 arrived)

this is while its cold enough to snow out... wer gonna get snow on friday

jaganath · Post by **jaganath** » Wed Feb 28, 2007 5:49 pm

CPU 19.4C (in 18C room)

what are you using, ice cubes??

PopCorn · Post by **PopCorn** » Wed Feb 28, 2007 6:08 pm

no where i live its below 0C and my windows always open, i <3 the cold

... my cpu is a 3800+ X2 with a sweet ThermalTake TMG A1 heatsink running passive..... this is my coldest running pc ever and ive done water cooling before... (wasnt that great) of course that will change once my new kinda shiny 8800GTS gets here... will be orderd this weekend

Erssa · Post by **Erssa** » Wed Feb 28, 2007 6:36 pm

PopCorn wrote:no where i live its below 0C and my windows always open

Lol, talk about wasted heat. How can you stay healthy with the draft in your room?

PopCorn · Post by **PopCorn** » Wed Feb 28, 2007 6:44 pm

well the only room that has the window open is my bedroom and its also the room my computer is in so its a win-win ... and i try to save the other house heat by keeping my door closed... health-wise im perfectly fine... i have been sleeping with my window open since i was like 8... im 17 now so.... and i havent been sick yet this year unless u call hang overs sick

niels007 · Post by **niels007** » Wed Feb 28, 2007 6:54 pm

The danger in the topic starters question is that this can be purely subjective. Temperatures that 'feel safe' to you might not have anything to do with the thermal limits of your pc parts.

CPU:
The only real limit you can set that is objective, is, *find* the limit. Run that prime95 torture test (two if dual core) and see how slow you can spin your fans and at which temperature does it begin to fail. The chances of burning the chip this way are pretty much zero. I had an Athlon xp at a reported 98c this way before it gave an error. Another one I couldn't get beyond 82c (ran it passive on a SI97) and it kept stable.. Only if you do this, you find the actual limit. It would be silly if you subjectively want your cpu to stay sub 50c when you can totally safely get away with 70. Thats probably half the fan speed required. If you would turn up your fan because of subjective fears, you're just exposing yourself to excessive noise! It has been said on this site before that it is not heard off that a chip dies of 'age' sooner at a higher temperature. They all outlast their effective lifespan even at high temperatures. Perhaps, only perhaps, extreme overvolting combined with high overclocks could 'reduce life' but probably even then, providing you run a bit below 'max stable temp', its probably going to last much longer than you will use it!

GPU:
Same approach, reduce cooling slowly and watch for artifacts or lockups. Once they appear / happen, providing they didn't happen before you started slowing the fan down, you know its too hot. Passive sollutions are not uncommonly seen at around 100c! shows you that a gfx card chip can safely get pretty damn hot. Whether this high temperature, comapred to most CPUs, could influence the lifespan of the circuit board / soldering remains to be seen. Again, as with the CPU, chances of the card dying before you replace it anyway are very small. Now gfx cards can and DO die sometimes, its just highly unlikely that you can be *sure* that it died of excessive temperature.

HD:
Tricky one this.. You don't want to stress this to where it corrupts data.. A recent report by Google showed that there was no relation between their disks failing and the operating temperature. Though it was probably not a perfect statistically valid report, it does seem to indicate that HDs will and do die, but again running them at a higher temperature seems to have very little to do with them dying during the period you use the thing. This probably means you can safely run a HD at the specced max temp (often 50c, which sometimes even seems to mean 50c 'abmient air' temp i.e. a much hotter drive?).. Anyway it should not be a problem keeping a <10W consuming HD sub 50 degrees unless you really isolate it..

Overall:
Don't get me wrong, when people say temperature has an effect on component lifespan, they are right. However there is almost no relationship between the failing of parts and their operating temperature, providing they where running stable until they died.

If I drive a car into a concrete wall at 40mph, it'll be totalled. If I do this in Finland at -20c it will be totalled. If I do that in the saharah desert, it will be totalled. Was there a difference in temperature? Yep. Did that cause the car being destroyed? Nope..

Hope that made some sense, I'm all for a 'less subjective' approach to pc operating temperatures!

psiu · Post by **psiu** » Wed Feb 28, 2007 8:08 pm

Well the above post is somewhat how I base my limits...I do try to find the temp limits by the manufacturers for CPU...so my Athlon XP's had upper limits of 85 and 90 C....staying under 75 was my goal with a bit of "insurance" if you will for warmer days. Currently my wife's idles at ~30C, load of 45. My Athlon 64 runs at ~30 idle and 45-50C.

Only my X800 has thermal reporting and I aim for 100% fan usage at 75C...it rarely gets to 60C however (AC Silencer on it with ATITool controlling fan usage). My wife's 9700 has a Z80D on it and doesn't artifact or have any issues with Age of Empires 3 or Company of Heroes.

I aim for under 45C for hard drives.

ronrem · Post by **ronrem** » Thu Mar 01, 2007 12:20 am

You could scrap out a refrigerator,stick some 5000 rpm Deltas on a Ninja and move to the Yukon...and you'd get some real low temps....that and $2 will get you a cappuchino (in the Yukon,San Francisco prices higher)

Here's something to note. In WW II the Germans had some equipment failures due to heat. However,at the Battle of Stalingrad,soldiers sometimes pissed on their rifles because they were too frozen to work. Tanks,trucks and aircraft often were unusable because they were too cold. The moral is don't fight in nasty weather but don't sweat it being somewhat warm. If a German can stand 115 F in the summer-so can a Seagate.
Pretty much anything in the puter can be 50C. If you have temps in the 30's or below and a bunch of fans making a bunch of noise....we have failed you,or you don't get it. The exception would be if you've adapted/mutated and freezing temps don't bother you. In that case a Sempron does not need a Ninja,you could cool it with a bottle cap and some artic silver

Bluefront · Post by **Bluefront** » Thu Mar 01, 2007 3:25 am

I simply don't buy much of what has been claimed in this thread......particularly the matter of HD temps. The vast majority of personal computers I have repaired, suffered hard drive failure. Second biggest cause of failure....power supplies.

The common problem found in every dead computer I ever worked on....dust. Dust clogged the intakes, clogged the PSUs, clogged the heatsinks, covered the fan blades......all this caused a heat problem. And what dies first is the HD, followed by the PSU.

Now most people posting on SPCR, probably keep their computers clean enough to avoid dust/heat problems. But if the temperatures posted in this thread are any indication, they are facing temperature related failures caused by low airflow/heat problems. Sure you can run your GPU 100C.....most you'll lose is your expensive card. Try that high-temperature thing with your HD, and you stand to lose much more, much more than money.

You can make a cool-running, quiet computer.....I do it every time.

niels007 · Post by **niels007** » Thu Mar 01, 2007 5:51 am

Agreed, with a couple of 500rpm Nexus fans you can often get a dozen or two degrees below 'stability threshold'. It is also pretty hard to get a 7200rpm HD to run above 40c with any sort of minimal airflow.

SockToy · Post by **SockToy** » Thu Mar 01, 2007 2:53 pm

How do people clean their cases? I got told off for saying I hoovered my GPU cooler once... apparently the moving air over plastic causes static.

dhanson865 · Post by **dhanson865** » Thu Mar 01, 2007 3:03 pm

Erssa wrote:dhanson865: Your hard drives are dangerously cool! I suggest you turn that fan down to increase your hard drive temps for better reliability.

viewtopic.php?t=38343

http://labs.google.com/papers/disk_failures.pdf

Sarcasm doesn't translate so well over internet, so I will just to clarify, that I think you are cooling your hard drives for no gains. If it makes you feel safer, fine. But cooler temps won't make them any more reliable as shown by the google study.

I've read both of those and had them in mind when I replied in this thread. It is widely considered that the below 30C hard drive failures were not temp related. None of the studies discussed are single variable.

The failure rate for drives that are several years old climbs quickly in the 35 to 40C range so below the start of that curve is where I want to be.

I'm using hard wired voltages for my case fans so I'm not going to change the voltage of the HD cage fan more than a few times a year.

If my summer conditions allow me to slow my fans and stay below 35C drive temps then I will do so. But I'm not rich enough to willingly shorten the life of my hard drives just for the minor noise difference between a single 80mm case fan at 7v vs a single 80mm nexus at 7v. The nexus claims 1500 rpm at 12v and the other fan I'm using claims 2000 RPM at 12v. There is a noticeable difference in sound between the two but I'm willing to live with the difference to secure my data and save money.

I consider the failures below 30C to be noise in the data. If you think they make sense then feel free to add resistant heating or some sort of heating/air conditioning logic to your drive enclosure to make sure it won't get below 30C.

crispyfish · Post by **crispyfish** » Thu Mar 01, 2007 3:13 pm

dhanson865 wrote:I consider the failures below 30C to be noise in the data.

That is HIGHLY unlikely due to the number of drives Google is monitoring.

dhanson865 · Post by **dhanson865** » Thu Mar 01, 2007 3:32 pm

crispyfish wrote:
dhanson865 wrote:I consider the failures below 30C to be noise in the data.
That is HIGHLY unlikely due to the number of drives Google is monitoring.

I'm not going to waste bandwidth here discussing what others have said well. Read most of the threads all over the web related to this study and someone will say similar things. Look at what "whiic" said here on SPCR if you don't feel like reading the rest.

The study wasn't designed around temp vs failure. Until you understand the variables and how to interpret the data you are just not going to be able to make a proper statement about that study.

crispyfish · Post by **crispyfish** » Thu Mar 01, 2007 3:39 pm

dhanson865 wrote:
crispyfish wrote:
dhanson865 wrote:I consider the failures below 30C to be noise in the data.
That is HIGHLY unlikely due to the number of drives Google is monitoring.
I'm not going to waste bandwidth here discussing what others have said well. Read most of the threads all over the web related to this study and someone will say similar things. Look at what "whiic" said here on SPCR if you don't feel like reading the rest.

The study wasn't designed around temp vs failure. Until you understand the variables and how to interpret the data you are just not going to be able to make a proper statement about that study.

I read whiic's comments and they make sense. I hadn't considered the wide variety and ages of drive models being used...

sjoukew · Post by **sjoukew** » Fri Mar 02, 2007 1:34 pm

Disk drive components are generally rated for an operating temperature range of 5 to 55Â°C. Within these limits, component life is specified to be at least 5 years. If a drive gets too hot, that is beyond its rated temperature range, the components will start to degrade and the danger of the electronics malfunctioning increases.

Source: interview with maxtor at http://www.nextgenelectronics.com/maxto ... view.shtml

How much heat will you tolerate for lower volume?

How much heat will you tolerate for lower volume?

Re: How much heat will you tolerate for lower volume?