Watercooling: how bad is it if the pump fails?

[Pandamonium]

Say you're watercooling a CPU and the pump dies on you. How devastating is that? Does your system just run hotter than normal or do you actually run the risk of damaging the CPU?

I'm thinking you could get away with it, assuming your CPU didn't need watercooling to begin with. The way I see it, the same water is in your radiator and waterblocks - your CPU should heat that medium to a point where the temperature difference within the water cannot be increased. Does this make sense or am I babbling?

 

[Varun]

Well, you are babbling, as the water in your water block would heat up a lot more than the water in your radiator if your pump quit. I don't think you would damage the CPU though - most have safeguards against that now.

I believe you can get a piece that has a turbine in it with a speed sensor. Basically you can hook that up to your motherboard and set a minimum speed (like a CPU fan) and if it drops below that it will shut down.

It won't be a catastrophic failure in any event.

 

[Matthias99]

Actually, there were some posts on the forums at www.overclockers.com (in the Watercooling forum) a while back about a guy who had inadvertently had his pump fail, but his system kept going for months before he realized it! What he found is that if you have a relatively low-resistance setup, and the reservoir/radiators (which contain cooler water) are above the heat sources, convection will keep a low level of circulation going, and if your system doesn't need much cooling in the first place, it'll work fine. They may have even done a front-page article about it.

 

[Mavrick007]

It can't be a very good thing if your machine is highly overclocked though.
That would be one of my main reasons to use a waterblock (not to mention
to tone it down but my machine is pretty quiet now).