photonblaster

I have seen air cool, TCE cool, H2O cool, Air plus TCE, H2O plus TCE, but not this one: Has anyone seen H2O plus tower air cooling? Change one of the heat pipes in a tower to a water line, put in some heat exchanger grooves in a thicker copper block. You will get some sort of temp drop in the CPU temp at full power. The temp drop will not be as much as one would hope since the relative drop in temp is reduced when what you are cooling is closer in temp to the coolant. I would expect it to be in the 5-10degC range since this is what temp drop I have seen reported for water systems at idle compared to cheap air coolers. Or, alternately, use quieter fans with less air flow to get the same temp as the water cooler by itself with stronger fans. Or use in cases with less total air flow for lower noise that way. Has the advantage that the air cooler portion is backup if your pump fails.

This is cool (sorry I could not help myself). I am researching for my first OC build (I have built a half dozen non-OC systems) Does anyone have some experience to share how to do this intake mod for a tower heat sink? The intake is on the side of the HS, not the top. Also, according the the review of the Sedo case here: Spedo Case Review (Graph about halfway down page) if you put a fan right behind the CPU on the backside of the MB it helps cool the CPU by a similar amount to the intake vent. Has anyone any experience with this case mod to better cool the CPU?

Nice extrapolation. I certainly have to agree with you, otherwise I would be saying my first response was full of it!

Continuing my research for my first OC build. I only found two sites that compared with and without backside CPU cooling. Backside of MB CPU socket cooling is available in two gendre that I know of: a side panel fan in the Thermaltake Spedo, and an auxilary set of cooling fins in the IFX-10 CPU cooler. Both are supposed to be able to reduce the CPU temp by 4-5degC compared to cooling systems without them. This is not a small amount! Combining the two may even add another couple of deg C cooling. Spedo (Graph about halfway down page) IFX-10 I have done some engineering thermal modeling which seems to support backside cooling as a way to lower CPU temp even more than a great tower cooler. In fact beefing up the mounting screws to larger diameter could even help a few deg if backside coolng is used (theoretically at least)(It increased the heat conduction to the backside) Has anyone real life experience with a highly overclocked I7 with and without rear CPU cooling? (could even be frontside H2O and backside air). How much temp delta did you see? Anyone measure the actual temp at the backside? Any other cases or methods to do backside CPU cooling?

thanks for the links. I fogot about the capilary action which can "flow up hill". The first link you gave me ran tests in different orientations with no more than 1- 1/4 degree C difference. The author gave no indication that this small difference could be due entirely to experimental error: however it can easily explained by a 5% difference in thermal impedence of the thermal paste due to thickness/uniformity varuations So probably at most a couple of deg difference regardless of tower geometry as you suggested. I won't worry about it.

Actually, if you just increase the border of the chip with no additional features, I do not think this will improve the situation. Silicon is a good, but not great thermally conductive material. Copper alloys are better. So the distribution of the heat away from the heat source within the chip is better done by the heat spreader than the silicon (assuming the silicon-heat spreader interface has low thermal impedance).

Searched this for an hour did not find any info. I am configuring my first OC system, and being a scientist I am doing my research first. Tall HS with heat pipes will end up with some of the heat pipes having the tip lower than the CPU block. This seems counter intuitive since the liguid inside is supposed to vaporize, rise to the cooling fins where it condenses and falls back down to the block touching the CPU. Has anyone run tests where the case is turned on its side so that the heat sink fins are above the CPU rather than horizontal in classic tower configuration? Does the CPU run cooler?

You do realize that I said that more pure grades of gold will conform to these imperfections, right? It will not defeat the purpose.

Based on what? Engnineering calcs say gold will be equal to or better, and silver definitely better unless it tarnishes and messes up the thermal conductivity.

By conductivity I meant thermal conductivity. Sorry about that. I did run the numbers comparing thermal conductivity. If pure silver does not oxidize too much, it actually would be the way to go, could give a 25% lower thermal conductivity than Arctic Silver 5. AS5 uses very high thermal conductivity fillers, but still not as good as silver, and there is loss in thermal conductivity since the grains of silver do not make contact enough to match a foil. That's why AS5 and gold foild have about the same thermal conductivity. I was hoping someone might have tried this...I do not have the resources to do it myself.

Noobie to the forum, normally just read all the great info. I searched, did not find any threads on this subject. Couple mil thick gold foil, 1" square should cost less than $10. Conductivity should be about 75% of pure silver, but higher than most silver filled pastes. About the same as Arctic Silver 5. Purer grades should be soft enough to conform to surface imperfections. Silver foil would have higher thermal conductivity but may suffer from increase in thermal impedance at silver-heat shink and silver-cpu interfaces because of oxidation. Is this too crazy even to think about trying?