Pro/Forums - View Single Post

Cathar · 10-09-2002, 09:18 AM

Just completed a round up of 7 waterblocks of various designs (maze, open pool, channelled, multi-channeled, maze 3).

In order to guage the best indication of the performance difference than a 1250 and a 1048 make, I had both pumps in the reservoir. The system would be bled and primed with the 1048. While keeping the pump outlet hose under the water at all times, I'd pull it off and move it over to the 1250 and plug it in.

This way there is no issues with rebleeding the system, remounting the CPU, moving any components about, changing air-flow, etc. The only thing that changes is the pump.

The net effect of moving from a 1048 to a 1250 is around an estimated 0.2C drop in temperatures for an over-volted over-clocked CPU. Most of the blocks had indetectable differences between the two pumps (performed exactly the same across 3 different CPU power levels), while a very few saw a 0.5C drop. Given that the CPU diode resolution is 0.5C, this doesn't say much. It just means that the block that saw the drop was on the edge of the 0.5C switchover with the 1048 when the 1250's small extra performance caused the drop.

If we average out all the readings, I stand by the 0.2C difference in performance between the two pumps as stated above.

If people are seeing larger differences, then I believe that they altered something else without realising it.

I measured the flow rate differences. In all cases through a complete system (waterblock + radiator), the Eheim 1048 saw 69-73% of the flow rate through the system that the Eheim 1250 was getting.

The Eheim 1048 in my mind is a perfectly acceptable pump for any in-case in-line solution that doesn't have excessive restrictions.

10-09-2002, 09:18 AM	#3
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	Just completed a round up of 7 waterblocks of various designs (maze, open pool, channelled, multi-channeled, maze 3). In order to guage the best indication of the performance difference than a 1250 and a 1048 make, I had both pumps in the reservoir. The system would be bled and primed with the 1048. While keeping the pump outlet hose under the water at all times, I'd pull it off and move it over to the 1250 and plug it in. This way there is no issues with rebleeding the system, remounting the CPU, moving any components about, changing air-flow, etc. The only thing that changes is the pump. The net effect of moving from a 1048 to a 1250 is around an estimated 0.2C drop in temperatures for an over-volted over-clocked CPU. Most of the blocks had indetectable differences between the two pumps (performed exactly the same across 3 different CPU power levels), while a very few saw a 0.5C drop. Given that the CPU diode resolution is 0.5C, this doesn't say much. It just means that the block that saw the drop was on the edge of the 0.5C switchover with the 1048 when the 1250's small extra performance caused the drop. If we average out all the readings, I stand by the 0.2C difference in performance between the two pumps as stated above. If people are seeing larger differences, then I believe that they altered something else without realising it. I measured the flow rate differences. In all cases through a complete system (waterblock + radiator), the Eheim 1048 saw 69-73% of the flow rate through the system that the Eheim 1250 was getting. The Eheim 1048 in my mind is a perfectly acceptable pump for any in-case in-line solution that doesn't have excessive restrictions. Last edited by Cathar; 10-09-2002 at 09:23 AM.