Pump Top Performance Testing

With the pump separately tested, it is now time to see how the integrated pump top of the CORSAIR Hydro XD7 performs. Note that this has been done before, but is still apt for testing how the top performs assuming this Xylem D5 is no different from other recent offerings—something the data on the previous page clearly supports. For this test, I used the same pump on various Xylem D5 tops from various companies, some standalone, some integrated into reservoirs. The testing methodology to generate a P-Q curve was the same as before, though only with 100% PWM duty cycle this time around.


There is a fair amount of variation from top to top, with some focusing on providing higher maximum head at the expense of flow rate and others going the other way. A good balance of the two is the best overall in realistic scenarios, and with excellent overall performance, the integrated pump top of the Hydro XD7 does not disappoint. If anything, the result was within error margins of the Hydro XD5 to where I did not bother showing it here separately. It is highly likely that CORSAIR has the same volute and pump top design in both cases.

A Realistic Loop Test

So far, we have seen the results of extremely controlled tests, often to minimize the effects of variables. However, this does result in an incomplete picture by focusing on the possible maximum as opposed to realistically achievable values. Take a pump's P-Q curve, for instance. The Xylem D5 is rated at 1500 LPH (liters per hour, 6.6 GPM) of maximum flow, which is great to look at but corresponds to an open loop with no liquid-flow restriction. Close the loop and it is reduces to ~4.1-4.2 GPM at most. CORSAIR is being realistic by stating 800 LPH flow rate since the Hydro XD7 has its own corners and twists. Add your everyday closed-loop components and you are looking at a maximum flow rate of ~1–1.2 GPM for most setups. This is where the budget pumps come in and why you should consider all options.

I used an XSPC Raystorm Pro CPU block, Swiftech Komodo R9-LE GPU block, Black Ice Nemesis GTS 360 radiator, two sets of Koolance QD3 quick disconnects with six Bitspower 1/2" x 3/4" fittings, and Primochill Advanced LRT tubing in the same size. The blocks are one of the least restrictive in each category, and the addition of the QDCs and relatively more restrictive radiator brings the average restriction close to what most loops would have. At the same time, I measured pump noise alone by placing the setup inside an anechoic chamber at 19 dBA, with a sound probe 6" from the pump. No other sound sources were inside the chamber when the measurements were taken.


Things are a lot closer here, and the CORSAIR Hydro XD7 does a very good job at keeping things flowing quietly. It is bested by the newer Alphacool VPP755 at the higher end and in noise output, but the Alphacool VPP755 does not yet have the proven long-term reliability the Xylem D5 in the CORSAIR Hydro XD7 enjoys, which Alphacool argues against by stating that the VPP755 has been updated over time, too. There are a few more budget pumps that do a decent job in terms of performance to noise, but those wanting the best-possible performance at reasonable noise levels will look at the Xylem D5, which makes the CORSAIR offering a strong option. Once again, as expected, the Hydro XD5 tested near-identical to the Hydro XD7, so I am not showing it separately here.

Given the nature of so many moving parts inside a pump, reliability is hard to talk about within a review's time frame. It is thus helpful that the Xylem D5 and DDC (reflected above in the CORSAIR Hydro XD3) have both been in use for years in some of my systems and those of others. The D5 especially has lasted through the test of time, which is another plus for the CORSAIR Hydro XD7. Since there are so many hard-to-quantify variables when it comes to a pump, I will refrain from making an attempt at a relative performance chart and, as such, a price-to-performance chart that would put this pump, reservoir, and distro plate combo unit at a disadvantage owing to all the non-quantifiable factors involved.