The Intel core i5-14600KF processor we received for the Tweak League, although not overclockable, runs at 5.3 Ghz. But if you take a closer look, this Max Turbo frequency can’t be maintained for very long if you are thermal throttling, even with a great cooler like the DeepCool AK620.
A few months ago, YouTuber Optimum made a video on cooling using 3D printing. His focus was on lowering temperatures. His experiment worked well, with -10 degrees on the 13900K in play.
And I said to myself: for the Tweak League, if I could lower the temperature, I’d gain in performance..
The best way to test? Cinebench!
So we run a series of 5 Cinebench tests with 10 minutes’ rest between each. You’ll notice that in 2 seconds, you’re already at the thermal throttle of the processor, which is constantly at 100 degrees. The results? Cinebench R23 Multi Core averages 22,290 pts.
Now lets be honest, I hasn’t delt the best silicone, since it’s common for 14600KF to run around 25000 pts, but I wanted to find solutions to remedy that.
3D PRINTING
Now all we have to do is open Autodesk Fusion and some measurements to create our air ducts. And given the size of the parts, it’ll take about ten hours to print the biggest one… assuming I didn’t screw up my measurements..
I’ve got 2 FDM printers at my disposal, the BambuLab A1 Mini which we’ve just received and which will soon be put to the test, but also the Elegoo Neptune 4 Plus which we’ll be working on the largest part.
And in the end I was good, but unfortunately it was the printing that went wrong. Too bad, let’s improve a bit and print again.
I’ve designed the ducts so as to be able to embed the fan’s external fan inside them. Because that makes 4 fans in series… clearly too much redundancy, so I wanted to test with and wit
RESULTS
After assembly, we launch Cinebench. The room is 3 degrees warmer than for the previous test, but that shouldn’t make any difference. And indeed, we’re now running at an average of 23’500 pts in Multi Core on Cinebench R23. That’s a gain of 5.4%.
Even though we removed one of the fans from the CPU cooler. Because yes, with this extra fan, performances were slightly worse.
We’ll stay with this completely scientific equation:
“AIR DUCTS – 1 FAN = 5.4% PERFORMANCE”
Voila, I’m just blown away, even aesthetically it’s clean. The experience doesn’t generate any additional noise nuisance – quite the contrary. And I’ve already got lots of ideas for optimization, but the clock is ticking and we’ve got to tease you about the next 3D printing content we’ll be doing on overclocking.com 😉
