I've heard of overheating, but is there such a thing as over-COOLING?

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.

skykiii

kiwifarms.net
Joined
Jun 17, 2018
I'm just curious.

In computer building there's always a lot of emphasis placed on heat management, but I wonder if its possible to overdo it?

Like for example, if I had a desktop PC placed in front of a window unit air conditioner so cold air was always blowing on it, would that be bad? What if I somehow found a way to run a gaming PC that was inside a refridgerator?

(And no, I'm not actually going to do either of those things. I'm crazy, not stupid. I do want to hear thoughts on what the likely result would be, though--would it break the computer?)
 
IIRC when you overcool the CPU with liquid nitrogen during extreme overclocking you can crash your system, but not necessarily cause any permanent hardware damage. But don't quote me on that, I don't remember if the XOC crashes were due to overcooling or just due to unstable overclocks.

Though this is if we're speaking strictly about the operating temperature of the hardware. The methods through which you cool your system can definitely cause more damage. Like stated by Vecr, water condensation. Or for example, a faulty custom closed loop water cooling system that leaks onto the components. Or using some industrial blower that's so strong that it can rip SMD components off of your hardware.

Basically the sort of mechanical damage that you're not going to encounter if you're a civilized white man and you use a good case, good case fans in a good configuration, and a good air cooler, all with the right fan curves and the right expectations as for what your temperatures should be and your personal preference as for how loud your system should be under a given load with given temperatures. But that's fairly straightforward once you've built your PC, just run some games/benchmarks, use whatever you want to monitor the temperatures, use Fan Control to edit the fan curves and hand tune it as you go along. Also undervolting your GPU helps a ton, no risk of hardware damage, at worst you'll get instabilities you fix by disabling the UV, and with the right voltages and the right silicon lottery ticket, you can even overclock your GPU while lowering it's power draw, and by extent heat output. Undervolting at stock clocks is also good, and can also give a performance boost by the merit of your GPU not doing a constant juggle of boost clocks, but instead sticking to one boost clock. Do some GPU clock monitoring under benchmarks with and without undervolting and you should understand.
 
I've seen plenty of videos of computers still operating while in a literal block of ice. Till it starts melting and shorting shit out, anyways. My phone becomes noticably slower if its on me in icy weather and gets cold, so I imagine excessive cooling could do the same to a computer. I don't think excessive cooling will damage it in anyway so long as no moisture is involved, at worst you'll get some slowdown. As far as my experience dictates, you need to go deep into the negatives for a computer to give a shit about cold. It tends to take extreme and prolonged cold for a material to become warped or damaged from the exposure, heat, not so much.
 
Water condensation. Cooling with a chiller or dry ice or liquid nitrogen can also lead to weird effects.
LinusTroonTips had a video 10 years ago about phase change cooling (refrigerator for your CPU). In that video they used kneaded eraser to seal the socket from condensate. The kit didn't look practical at all, you're better off just using the money to buy a better CPU (unless you're Richie Rich trying to squeeze more juice out of the most powerful consumer CPU on the market).
 
Are you being for real?
I am 100% serious. If it tries to boot up or restart without some time in the ice box it freezes and then boots into automatic repair which tells me to restart it where it freezes again. It also freezes if it gets unplugged

As such I have replaced it with a new laptop and relegated it to second monitor duty. Once its up and running it works fine but I can't let it sleep because it also tends not to wake up from sleep so instead of sleeping I open BlackScreen.png and set it to full.

That said the other day I think it actually successfully restarted itself for an update so maybe its cured now but I'm not about to test it and see because the whole freezer thing is a pain in the ass and every time I grow concerned it might be the last time it works.
 
There’s a point where the material properties of the doped silicon change too much to be used for computation, but it’s not achievable by normal means. This would happen around like 40-60 K, while liquid nitrogen is like 75 K. Getting much colder quickly gets more difficult and expensive.

I think you’d see problems in the solder sooner, though. The chip slots sit on a bed of solder connecting them to the mobo. Solder gets brittle at like -30 degrees C and a bit colder leads to extreme susceptibility to fracture (idk exactly, a few 10s of degrees), but would be especially weakened before instant failure by thermal cycling. The repeated swings in extreme temps makes the solder bed flex (this is partly intended, but only as much as is expected at normal operating temps). Once part of it burns out you get a cascading failure across the rest of cpu/gpu socket. Liquid nitrogen is easily cold enough to do this, but even a stack of peltiers can achieve the embrittlement temp.

This failure due to solder flex under the chip is actually the cause of the red ring of death in the hex box 360 and a similar (less publicized) defect in the fat PS3. In both consoles the GPU was fixed in place with a design and solder mix that could not withstand the thermal cycling in the newer technology.
 
There’s a point where the material properties of the doped silicon change too much to be used for computation, but it’s not achievable by normal means. This would happen around like 40-60 K, while liquid nitrogen is like 75 K. Getting much colder quickly gets more difficult and expensive.

I think you’d see problems in the solder sooner, though. The chip slots sit on a bed of solder connecting them to the mobo. Solder gets brittle at like -30 degrees C and a bit colder leads to extreme susceptibility to fracture (idk exactly, a few 10s of degrees), but would be especially weakened before instant failure by thermal cycling. The repeated swings in extreme temps makes the solder bed flex (this is partly intended, but only as much as is expected at normal operating temps). Once part of it burns out you get a cascading failure across the rest of cpu/gpu socket. Liquid nitrogen is easily cold enough to do this, but even a stack of peltiers can achieve the embrittlement temp.

This failure due to solder flex under the chip is actually the cause of the red ring of death in the hex box 360 and a similar (less publicized) defect in the fat PS3. In both consoles the GPU was fixed in place with a design and solder mix that could not withstand the thermal cycling in the newer technology.
WHOA YOU'RE AWESOME
 
LinusTroonTips had a video 10 years ago about phase change cooling (refrigerator for your CPU). In that video they used kneaded eraser to seal the socket from condensate. The kit didn't look practical at all, you're better off just using the money to buy a better CPU (unless you're Richie Rich trying to squeeze more juice out of the most powerful consumer CPU on the market).
Vaseline is cheaper and easier to handle.
 
Back