Discussion Custom dimension case fans + gpu deshroud

eyso

Efficiency Noob
Original poster
Jun 9, 2023
5
0
The build in question is partpicker list.

I was planning to use t30s in combination with my rtx 3080, however after deshrouding there seems to be a significant gap between the fans and the heatsink that is affecting temps quite negatively. Currently I am using 2 t30s as exhaust at 2000rpm and that yields 85°C under a gaming load, prior to deshrouding iirc I was at 70°C at 1500rpm.

To solve this I was thinking about creating a XXmmx120mm fan while still untilizing the phanteks t30 motor and frame, the only thing custom about the fan would be a fan blade design that is to be extended in the Y axis and a partial shroud where the gap would be between the fans and the heatsink. Is it as straightforward as 3d printing a larger fan blade design and popping it onto the motor assembly or would there be an issue with weight balancing due to a larger/heavier blade design?
 

Phuncz

Lord of the Boards
SFFn Staff
May 9, 2015
5,840
4,906
It's not as simple as spinning plastics will expand, among other characteristics. Look at these videos for more info:


 
  • Like
Reactions: eyso

eyso

Efficiency Noob
Original poster
Jun 9, 2023
5
0
Thanks for the links, will watch later forsure. However this video about the material difference in fdm, resin, and noctua stock fan material might be an interesting watch. The difference between stock and resin was 0.6% in terms of cfm. Not to mention noctua nf a12x25 motor config is a self contained metal shell and fan hub is glue on it, so in hindsight the t30s weren't the best choice but on paper a custom dimension fan is still plausible. Will definitely have to put in the the time and money to actually test it out though.
 

Biowarejak

Maker of Awesome | User 1615
Platinum Supporter
Mar 6, 2017
1,744
2,262
Definitely an interesting engineering problem. The first thing you need to figure out is if you can remove the fan without destroying the hub. This is non-trivial and might be destructive. If you can't recover the bearing and motor electronics, it's not worth it. Industrial glues are pretty difficult to remove in most cases. Maybe the raw parts are available on Alibaba or digikey.

After that you can probably get away with resin printing in Conjure Rigid, it's fairly good and dimensionally stable. However, you might still see warping on long, thin parts. If you're customizing the blade geometry you'd benefit from some CFD analysis, too. Ideally your geometry would be better than stock in regards to air throughput, since your only manufactuability concern is printing it and not injection molding.
 
  • Like
Reactions: Phuncz and eyso