Fractal Design Celsius+ S28 Prisma Review

👤by Matthew Hodgson Comments 📅19-07-20
Test Setup & Methodology
Due to popular demand, we have retired the long-standing cooling rig that we once used and replaced it with a shiny new system with an Intel Core i7 9700K at its heart. This should benefit you, the reader, in a couple of ways. The first being it can provide you with results on a CPU that is currently available to buy and secondly, the mounting instructions on our installation page will be far more relevant than the outgoing LGA 2066 was, with its pre-installed backplate.

We perform two individual testing routines with each CPU cooler that we receive - temperatures and acoustics. The system used is as follows and all tests are performed at stock frequencies.




To keep our tests consistent, we always review CPU coolers as they come out of the box - with their factory cooling fans, as the manufacturer intended and as the consumer expects. It should, however, be noted that we do not use supplied thermal paste but instead, Noctua NT-H1 paste to, again, ensure consistency. The paste is applied and the CPU allowed to run for a couple of hours, at idle, to ensure any and all updates or processes are completed to give a true idle reading.

In the first test we record the idle and load temperatures across a 60 second period, with the results logged to file. This way we can provide you with an accurate average temperature over a longer period, instead of, for example, a maximum or minimum temperature. We believe this is the fairest possible method and most representative of what is going on as temperatures will fluctuate even with a consistent load on the CPU. We have three case fans on the front of the PC running with the door closed for this test.

We aim to measure acoustics with our second test, this time with the three case fans on the front disabled and all other ambient noise reduced to an absolute minimum. Our testing room measures around 25dB(A) without the system running. We measure the system acoustics for 60 seconds with the microphone in a set and unchanging position with the chassis door removed completely. An average is then taken across the 60 seconds.

9 pages « < 5 6 7 8 > »

Comments