Ever wondered what would happen if your CPU fan died? I felt suicidal enough to test!

RainMotorsports

Leadership
Leadership
Joined
Mar 6, 2011
Messages
8,649
I was worried about not having ASUS AI Suite when I install windows 8.1 later but finally figured out how to configure speedfan so this test was now easily possible.

CPU was at 4.3Ghz @ 1.38v. Cooler is the compact but well suited Zalman CNPS9900 Max. All intake fans in the case were dead, top 200mm exhaust unplugged. Cooling was provided by the 120 exhaust which is near the cpu cooler and throttled up to 2200 rpm (I normally have it at 1500).

Image:
Untitled.jpg

Under heavy (not extreme) load the temps when to around 85 at peak (did not sustain for more than a few seconds). Survivable but not ideal. Sandy bridge has a rated tcase of 72 degrees Celsius now this is at the heat spreader. The per core sensors will read higher than that but one would would assume package would be closer to the heatspreader but its not. Seeing as how it never throttled it obviously never reached the tjmax which is assumed at 98 but I believe not officially published as such since Intel likes T Case now.

So basically I get temps like cplmac does on his stock cooler while playing BF4 which is less intensive than this burn test :p
 
over here when the cpu fan dies, we just replace the thermal paste...
 
Intel burn test, pfft.. Prime is where stability testing is at.

Also, most modern CPUs will usually downclock and shut off if there's not enough cooling I think. Not that I've tried. But I'm guessing if you run it without a cooler altogether, it might still explode, old style. (reference: Ventrilo Harassment - Duke Nukem Style - YouTube )
 
Intel burn test, pfft.. Prime is where stability testing is at.

Also, most modern CPUs will usually downclock and shut off if there's not enough cooling I think. Not that I've tried. But I'm guessing if you run it without a cooler altogether, it might still explode, old style. (reference: Ventrilo Harassment - Duke Nukem Style - YouTube )

To generate high temperatures in a short period of time for a short test Prime is not the answer. This was not a stability test. The entire point was to test what temperatures a game would reach.

Shutdown is the ONLY answer, downclocking = death or damage. Nvidia employed that whole shutdown at tjmax thing and I have had the experience of trying it out. In the absence of cooling the downclock does not provide adequate protection in an AFK scenario.
 
Back
Top