Anyone have any comparisons between the Thermalright XP120 & Ultra120?

[ty1er]

I'm mainly looking for Personal experiences. Was it worth wild upgrade?

Thanks!

 

[BonzaiDuck]

I "jumped on" heatpipe coolers in 2004 when I purchased a ThermalTake PIPE101 cooler for a P4 Northwood machine. I quickly upgraded from there to the XP-120, once I read the reviews on the latter. Some people were saying that heatpipes wouldn't work well because of orientation and gravity, but I found a white-paper published by a company that had developed heatpipes for NASA, and my faith blossomed.

The XP-120 had a thermal resistance (for a fan running around 2,500 rpm) of between 0.14 and 0.16 C/W. It was great for an over-clocked Northwood 3.0C.

I upgraded the processor to a Prescott -- eventually a 3.4E -- and swapped an SI-120 for the XP-120. The SI-120 had a thermal resistance somewhere between 0.12 and 0.14 C/W, and showed a narrower idle-to-load range despite the Prescott's higher TDP.

But I cannot show you a direct comparison for a single processor. The Ultra-120 (original) had a thermal resistance of 0.12 C/W, and over-clocking an E6600 by about 39% brought the load temperature to around 53 or 54C with room ambient around 75+ F degrees. Going from the Ultra 120 to the Ultra-120-Extreme for the same over-clock and processor resulted in a load temperature of about 47C at the same room-ambient, and there was some improvement arising from diamond thermal paste and -- (of course) -- my ducting mod. Also -- the U-120-Extreme was purchased as "custom-lapped." So the improvement includes the benefits of lapping.

My best guess is that the U-120-Extreme has a thermal resistance short of 0.10 C/W -- more precisely estimated at 0.097, but choice of fan and fan-speed would make the precision less meaningful. You could make a computational comparison of idle-to-load spread -- or load temperature reduction -- by applying the ratio of thermal resistances, and it would probably give a decent order of magnitude in what you could expect between the old XP-120 and the U-120-Extreme for, say, a Northwood processor.

Put it this way -- and forget about over-clock settings for a moment. The TDP's of either a Northwood or C2D are known. So with thermal resistance figures for either cooler, you can get an idea of how they would perform with either processor.

 

[ty1er]

Originally posted by: BonzaiDuck
I "jumped on" heatpipe coolers in 2004 when I purchased a ThermalTake PIPE101 cooler for a P4 Northwood machine. I quickly upgraded from there to the XP-120, once I read the reviews on the latter. Some people were saying that heatpipes wouldn't work well because of orientation and gravity, but I found a white-paper published by a company that had developed heatpipes for NASA, and my faith blossomed.

The XP-120 had a thermal resistance (for a fan running around 2,500 rpm) of between 0.14 and 0.16 C/W. It was great for an over-clocked Northwood 3.0C.

I upgraded the processor to a Prescott -- eventually a 3.4E -- and swapped an SI-120 for the XP-120. The SI-120 had a thermal resistance somewhere between 0.12 and 0.14 C/W, and showed a narrower idle-to-load range despite the Prescott's higher TDP.

But I cannot show you a direct comparison for a single processor. The Ultra-120 (original) had a thermal resistance of 0.12 C/W, and over-clocking an E6600 by about 39% brought the load temperature to around 53 or 54C with room ambient around 75+ F degrees. Going from the Ultra 120 to the Ultra-120-Extreme for the same over-clock and processor resulted in a load temperature of about 47C at the same room-ambient, and there was some improvement arising from diamond thermal paste and -- (of course) -- my ducting mod. Also -- the U-120-Extreme was purchased as "custom-lapped." So the improvement includes the benefits of lapping.

My best guess is that the U-120-Extreme has a thermal resistance short of 0.10 C/W -- more precisely estimated at 0.097, but choice of fan and fan-speed would make the precision less meaningful. You could make a computational comparison of idle-to-load spread -- or load temperature reduction -- by applying the ratio of thermal resistances, and it would probably give a decent order of magnitude in what you could expect between the old XP-120 and the U-120-Extreme for, say, a Northwood processor.

Put it this way -- and forget about over-clock settings for a moment. The TDP's of either a Northwood or C2D are known. So with thermal resistance figures for either cooler, you can get an idea of how they would perform with either processor.

Thanks for the reply BonzaiDuck.

I like the proformance of my current XP-120 HS. The only problem is that there is heat buildup in my case. (Well, its not really a problem, but I would just like to get the temps as low as I can.) If I were to purchase the Ultra-120 I would also create a duct that connected the Ultra120 to the rear exhuast fan of my case. Exiting the heat produced from the CPU directly out the back of the case. This should reduce the inturnale case temp by a bit.