Rule number one in any testing.
Establish a sense of repeatability in your methodology (and therefore results).
Mount each block at least 10 times. Record the results of each mount under controlled conditions. Be extremely careful with the mounting. Be conscious of anything that could have changed other than the waterblock. This could be something as trivial as having the door to the room being open, closed, or half-open. As trivial as routing the hosing differently, or non-trivial such as the radiator being mounted in a different location, or a fan sitting differently or not turned on.
Measure the water-temperature at some fixed point for all tests. Measure the air temperature entering the radiator at some fixed point for all tests.
Good water-block testing is extremely time consuming.
Not saying that even doing all of the above will satisfy all, but if you establish repeatability and control your variances to a degree, you'll be somewhat on the path of getting a clearer picture.
I say to mount each block 10 times because variances of 3-5C are actually fairly easy to achieve between different mountings of a block, and that's more than the difference between the blocks that you've reported.
Once you've done all of the above, then consider lapping the bases of each block. The Maze 3's I know can be somewhat uneven on their bases, which affects both temperatures, but moreso, overclockability to a large degree.
Just some suggestions.
|