- Jul 15, 2018
- 2
- 0
- 1,510
I'm finally in a position where I can build a desktop, and was lucky enough to snag a 3070. I want to liquid cool my build primarily for aesthetic/fun project reasons with a secondary focus on performance and noise reduction. Therefore I'm going to build a primarily Barrow and alphacool loop with a budget of about 200$ like in this post but also with this waterblock connecting the GPU into the loop.
After some research, I've discovered in 2014 it was somewhat popular to use a universal waterblock on your GPU, and then attach generic heatsinks to the RAM and VRMs with thermal pads. The obvious benefits of this approach being reusability and price. Full cover blocks are in vogue right now, and my first assumption would be for economic reasons. Cooling companies have figured out they can sell 200$ blocks every 2 years instead of 40$ blocks once. I also understand that modern graphic cards draw nearly double the power.
I've read posts claiming shrouds choke VRMs and memory as secondary citizens, and that with a water cooled GPU not leaching heat through the PCB and no hardware constricting airflow over the VRMs/mem they actually have lower temps. I've also read posts claiming you need to have active airflow (read: internally mounted fan) next to your card if you want to use standard heatsinks.
Lastly, and most interesting, alphacool has a solution wherein they stem off smaller tubes from the main loop and individually cool each of the surrounding chips. One, now defunct, German webiste reviewed it and claimed that it's a good option for cards that don't have a full waterblock available.
So, my question would be am I right that the drive towards custom full cover waterblocks is primarily economic and passive cooling + universal waterblock is still very viable? Or do modern VRMs and memory produce so much more heat that passive cooling is unrealistic? Anyone have thoughts on the MCX one? They look sick, but likely introduce a lot of potential leaks.
Thanks so much for any input!
After some research, I've discovered in 2014 it was somewhat popular to use a universal waterblock on your GPU, and then attach generic heatsinks to the RAM and VRMs with thermal pads. The obvious benefits of this approach being reusability and price. Full cover blocks are in vogue right now, and my first assumption would be for economic reasons. Cooling companies have figured out they can sell 200$ blocks every 2 years instead of 40$ blocks once. I also understand that modern graphic cards draw nearly double the power.
I've read posts claiming shrouds choke VRMs and memory as secondary citizens, and that with a water cooled GPU not leaching heat through the PCB and no hardware constricting airflow over the VRMs/mem they actually have lower temps. I've also read posts claiming you need to have active airflow (read: internally mounted fan) next to your card if you want to use standard heatsinks.
Lastly, and most interesting, alphacool has a solution wherein they stem off smaller tubes from the main loop and individually cool each of the surrounding chips. One, now defunct, German webiste reviewed it and claimed that it's a good option for cards that don't have a full waterblock available.
So, my question would be am I right that the drive towards custom full cover waterblocks is primarily economic and passive cooling + universal waterblock is still very viable? Or do modern VRMs and memory produce so much more heat that passive cooling is unrealistic? Anyone have thoughts on the MCX one? They look sick, but likely introduce a lot of potential leaks.
Thanks so much for any input!
Facebook
Twitter
Instagram