View Single Post
  #9 (permalink)  
Old January 10, 2012, 09:06 AM
Notagiant Notagiant is offline
Rookie
 
Join Date: Jan 2012
Location: Ottawa
Posts: 10
Default

There is no risk in "freezing". Almost every extreme overclocking contest uses liquid nitrogen at well below 0C to cool their CPU, GPU and chipsets to reach their 5+ GHz clock speeds. The risk lies in condensation. Even though the air is considerably dryer in the winter than in our lovely Canadian summers, it still contains a fair amount of moisture in the form of water vapour. If any of the surfaces of your components hits the dew point, water droplets would form on them and you could run the risk of shorting something out.

Now, the good news is that you would probably have to go below 0C on the surface of one of your components for this to happen, so it is highly unlikely. For example:

Today, in Ottawa, the relative humidity is about 70%. From your info, your air temperature is about 10C. The dew point at that temperature and humidity is somewhere between -5C and -10C. It is not 100% impossible for this to happen, but is is unprobable. Where you run into real risk is near your hotter components (heat sinks, RAM, GPU). The higher the ambient, the higher the dew point. A quick blast of cold air over a component that is running with air around 20C would form dew at only 12C! YIKES!!

The ideal way to do this and eliminate risk would be to run the air intake over a mole sieve or dessicant system in order to remove moisture from the air, dropping its relative humidity. Knocking 10% off of the humidity could buy you an additional 7-10C of buffer on the dew point (depending on your temperatures and average humidity). EDIT: Or using dielectric grease, as mentioned above. Some people have even gone as far as to melt candle wax around problematic condensation spots. :)

Long story short, I like when others take advantage of what we are stuck with. Just be careful! No one wants to see you short out your whole system or even start a fire.
Reply With Quote