If your pump is only a few weeks old and you're having issues, it's defective. That has nothing to do with galvanic corrosion. But on the subject of galvanic corrosion, that's why the AIO's use propylene glycol.
If you're going to keep bringing up your delidding/fan shroud results in nearly every AIO thread, could you at least do a fair comparison with an AIO on a delidded processor and better fans because as is that's a completely misleading statement.
Well, the OP asked about a choice between a D15 and his Eisbauer 240. this was the same choice I faced for comparing my TR LGM to an EKWB Predator 240 -- the troublesome model prior to the Phoenix kit I thought you may have mentioned in another thread. Or maybe it was another veteran member -- Aikouka.
It's true that I should give references to the reviews I turned up in my own comparisons. They usually overlap reviews shown at the Frosty-Tech site. Often, you find reviews that only compare air coolers to air coolers and AiOs to AiOs. But when there is a mix, and when you have at least two reviews, you can extrapolate relative rankings by simple transitivity inference.
To further clarify, I don't have a "fair comparison with an AiO on a delidded processor." Instead, I have the comparison under a 140W overclocked testbed between a Predator 240, an LGM and the NH-D15 without delidding modification. I know my CLU relid gave me a 12C advantage over the stock processor at a binned clock. That result -- my actual measurement under the same 4.7 Ghz clock -- with the CLU -- compared to the same but stock-original processor at 4.7 and a review test result with an external 240mm EXOS cooler. I'd like to remember which review that was, or the review site. I think it was Hardware Secrets.
So the CLU relid i7-6700K with the LGM @ 4.7 compares to the same unmodified stock processor (@4.7) using the EXOS water cooler -- the type that sits on top of a midtower or full-tower case -- an external assembly. I was running LinX with AVX affinitized, maximum problem size and memory -- a test so extreme that you couldn't imagine what other test would provide more stress. The EXOS stock-processor system @4.7 showed 82C, and my system was showing between 80 and 82C.
In the 140W testbed configurations, the Predator 240 outperformed the LGM by 5C, and the LGM outperformed the NH-D15 by 1C. And again -- this was a testbed with no delid/relid modification.
But I didn't collect all this information to make an inferential argument of a review here. It was -- according to my mind and foresight -- the research I used in making final parts purchases. And I backed away from the Predator 240 at the last minute when I could see what would likely happen with an LGM and a CLU-relidded processor.
I'm sorry if this is insufficient for your own confirmation of it. Also, the more I think of it, the OP would have no interest in CLU_relid modification of his processors for whatever number he's arranging to ship. I think he made that more clear even if it were possible that it would pique his interest. So in that case, he'd likely get results with a D15 that were 5 or 6C short on an AiO of a 240mm size in performance.
I'm totally OK that I have no prospect of failures, leaks, or more than dust-removal maintenance with my rig. But as far as I can see, I come out at least even with an AiO of the sizes discussed here, or even a little bit better just for the relidding.
Facebook
Twitter