Arctic Cooling MX-2 is probably the most popular thermal compound today, and for good reason. It’s inexpensive, readily available, and a really solid performer. In terms of using it as a testing TIM it’s also pretty good because it has a minimal cure (small enough that it shouldn’t induce massive error if used the same way every time) and it responds to contact moderately well. Our tests with utilizing MX-2 are our core tests. MX-2 will be the TIM used to determine which injection plate or nozzle should be used (where applicable); MX-2 will be used for orientation testing (where applicable); and MX-2 is our base of comparison, we feel that how a block performs relative to other blocks with MX-2 will be most indicative of what an end user can expect.
Swiftech Apogee XT rev2 does better than I expected here. While neither the Apogee XT or Supreme HF is the same as the ones originally tested (literally, or even in model number), the Apogee XT rev2 is a lot closer to the EK Supreme HF than it was when I first tested both. Could be sample variation, could be the rev2 is simply better than the rev1 at this pump setting, could be that the Gigabyte X58-Extreme board was causing minor compatibility issues and degrading the performance of the original Apogee XT (it was on the list of incompatible boards that forced an at-launch revision). We will be testing the Apogee XTL in a couple weeks and we’ll test a small sample of Apogee XTs to see if we can clear this up.
It’s important to note that all flowrate/pump setting/hydraulic power spectrum tests are done with Indigo Xtreme (has zero performance change over time, unlike any thermal paste) and then justified to the blocks’ MX-2 performance.
The Apogee XT rev2 performs much differently than I initially expected. The original Apogee XT was basically immune to changes in flowrate (I’d estimate its flow responsiveness to be similar to the Enzotech Stealth’s, maybe even less responsive), especially decreases in flowrate. But here we can see the Apogee XT rev2 have roughly the same scaling as the flow-hungry CPU-370 between Super Low and Medium High pump settings. Between Medium High and Super High pump settings, scaling flattens slightly.
Why is it so different than the original Apogee XT? Little changes in design can mean big changes in performance. Our theory is by reducing the number of micropins the flow is forced through (rev2 provided a bypass), you’re forsaking performance with low flowrate. Let me explain. While there isn’t much performance difference between .5 and 1GPM, there is a big difference in performance between .1GPM and .2GPM, despite being having the same ratio (1:2). So while in the original Apogee XT all the flow went through all the micropins (for argument’s sake), in rev2 you only have a reduced portion of the overall flow going through the micropins with the bypass. So when you reduce flowrate, the flowrate through the micropins with the bypass gets extremely low (since it only has a fraction of the overall flow) and that is likely causing the performance degradation.
Basically, because so much less fluid is moving through the outer micropins as overall flowrate is decreased, you get this more aggressive scaling (and at higher flowrates, you may still get more scaling). It’s not a bad thing, just a very different result from what is otherwise a very similar block. Swiftech must have thought the decrease in restriction was a good trade-off.