Fractal Design Torrent - Bad GPU temperatures

[Dogzilla07]

15 hours ago, Chunky Moe said:

What I don't understand is how the Phantek P500A Digital is scoring so high on GPU when it has no bottom airflow?
 

Because you don't need bottom intake airflow to have good GPU temps, good static pressure front airflow, with a bit of tinkering is enough (no hard modding). It's just 2 ways to go about this, and you can get each one to do great (And Nvidia 3090 Founders Edition is just the best cooler design overall).

 

Look at @thrasher_565picture above, what properly set-up, what great static pressure fans intake front do, is, just imagine another 3rd arrow going across both sets of red arrows, above, and below the GPU, and to the side of the GPU. As long as that arrow overpowers the red arrows, you're golden (Or take not off the very interesting Enthoo Pro 2 results, if u have tons of empy space below the GPU, that disables the hot pocket from forming as strong as it does).

 

What that does is prevent an air vortex hot pocket from forming, and prevent the GPU from recirculating it's own hot air, by pushing it out the back. But u need to have enough exhaust area out the back for this to work properly. A few, high TDP card designs behave a bit differently (when combined with a specific CPU Air Cooler), so as a general rule of thumb, it's better to pre-emptively de-shroud the card, or spend additional time testing different fan setups to optimize for your exception to the rule GPU.

 

One thing I'm not sure about is that noise-normalized tests use stock configuration in GamersNexus video, are the results noise-normalized when using standardized fans as well ?

[TVwazhere]

9 hours ago, Dogzilla07 said:

One thing I'm not sure about is that noise-normalized tests use stock configuration in GamersNexus video, are the results noise-normalized when using standardized fans as well ?

Someone didnt do the homework  

On 9/14/2021 at 9:02 AM, TVwazhere said:

That's covered EXTENSIVELY in their "we didnt want to do this test but people wouldnt stop asking for it" standardized test video

They do both; one standardized test at a constant speed (I cant find the exact numbers but i think it's 100% fan speed), and one at 36dba

https://www.gamersnexus.net/guides/3477-case-fan-standardization-tests-noise-normalized-thermals

[Dogzilla07]

1 hour ago, TVwazhere said:

Someone didnt do the homework  

 

Hardly xD, as I was one of the people discussing the pitfalls of standardized testing on forums before GN started it, and mentioning it as they were explaining what they were trying to do, before they developed their methodology (in between my first foray into the conundrum of "how many fans and in what way" from the awesome video u guys did in 2015 (with Luke as the presenter), and before GN released their first case review video. Funnily enough that Luke LinusTechTips video was the catalyst for the Alice in the wonderland deep-dive and everything I've learned and wrote in the internet in the past 5 years, and all the walls of texts xD. And, To all the people I unfortunately annoyed due to my often far from tactful delivering of facts and wonders I learned on the way, it's not my fault that it's my fault, it's all Luke's fault :P, and his dastardly enabler Lienus )

 

Paraphrased, I wrote something like this on overclock forum/a domestic forum from my country : "Yes it would be problematic, there's plenty points for and against, but either way it would still be very beneficial for extrapolation and cross-comparing information. It might not be the data we deserve, but it is one we need/want/can use effectively!"

I also, and I'm quote-ing myself almost word for word here said in defense of GN "While it is true they aren't at a level of coolingtechnique, nor VSG, nor jonnyGURU, Aris, nor doyll and others here on the overclock forum, it is my opinion that they will soon reach and perhaps surprass the great reviewers in the respective fields, because I have never seen anyone advance and progress as fast as GN, or as committed to reaching the pinnacle of quality and perfection as Steve is".

 

I'm pretty sure i wrote a youtube comment on a video of theirs, but I might be wrong, it was a while ago. I also said "There's nothing stopping them from co-operating with quality reviewers and specifically",and this was right about when Linus did one of the first LinusTechTips convention videos/and different youtubers from PC tech there, and i said and I'm quote-ing word for word here again in defense of LinusTechTips (because someone was questioning how far from required perfection the data/information and reviews Linus was  doing were:

 

"This, look at this,with this u cannot disagree, all other points aside, no matter how you feel Linus videos are correct enough, or scientifically perfect enough, this organizing and helping of other youtubers is paramount, and will be helpful to the community as a whole. What's stopping Linus from connecting various people in the industry with his influence, and everyone getting better at what they do as result" I also wrote something like this: "This, this is the the exact thing that disproves all the Linus naysayers and negative nancies, and how influence, once built up, can be used for good.

The mini-rant Linus did about the conflict of interest with the Framework laptop, was just full circle for me, and it brought a tear to my left eye. The right one was already reserved for Louis rossmann.

[Dogzilla07]

1 hour ago, TVwazhere said:

They do both; one standardized test at a constant speed (I cant find the exact numbers but i think it's 100% fan speed), and one at 36dba

https://www.gamersnexus.net/guides/3477-case-fan-standardization-tests-noise-normalized-thermals

As for the actual point, heh, I meant for the standardized tests exclusively, specifically if the 3x Noctua fans they use for those tests are noise-normalized to 36dba or what speed they're running, as I couldn't remember on the top of my head, (and quickly searching for the amount of time I head, I also, like you couldn't find the exact numbers) to double check, but pulling from memory when I watched them explain it, was vaguely remembering 100% fan speed for the 3 Noctua fans, like you think as well.

 

Anyway, thank you for a second set of eyes/ears!, hopefully someone else also responds as well, 3 sets of eyes are always better than 1.

 

P.S. I split the original post into two, cause my exposition on the original answer "Hardly xD", got a tiny bit longer than i expected, and most people will miss this, more on-topic post.

[thrasher_565]

6 hours ago, Dogzilla07 said:

As for the actual point, heh, I meant for the standardized tests exclusively, specifically if the 3x Noctua fans they use for those tests are noise-normalized to 36dba or what speed they're running, as I couldn't remember on the top of my head, (and quickly searching for the amount of time I head, I also, like you couldn't find the exact numbers) to double check, but pulling from memory when I watched them explain it, was vaguely remembering 100% fan speed for the 3 Noctua fans, like you think as well.

 

Anyway, thank you for a second set of eyes/ears!, hopefully someone else also responds as well, 3 sets of eyes are always better than 1.

 

P.S. I split the original post into two, cause my exposition on the original answer "Hardly xD", got a tiny bit longer than i expected, and most people will miss this, more on-topic post.

https://www.gamersnexus.net/guides/3477-case-fan-standardization-tests-noise-normalized-thermals

[thrasher_565]

17 hours ago, Dogzilla07 said:

Because you don't need bottom intake airflow to have good GPU temps, good static pressure front airflow, with a bit of tinkering is enough (no hard modding). It's just 2 ways to go about this, and you can get each one to do great (And Nvidia 3090 Founders Edition is just the best cooler design overall).

 

Look at @thrasher_565picture above, what properly set-up, what great static pressure fans intake front do, is, just imagine another 3rd arrow going across both sets of red arrows, above, and below the GPU, and to the side of the GPU. As long as that arrow overpowers the red arrows, you're golden (Or take not off the very interesting Enthoo Pro 2 results, if u have tons of empy space below the GPU, that disables the hot pocket from forming as strong as it does).

 

What that does is prevent an air vortex hot pocket from forming, and prevent the GPU from recirculating it's own hot air, by pushing it out the back. But u need to have enough exhaust area out the back for this to work properly. A few, high TDP card designs behave a bit differently (when combined with a specific CPU Air Cooler), so as a general rule of thumb, it's better to pre-emptively de-shroud the card, or spend additional time testing different fan setups to optimize for your exception to the rule GPU.

 

One thing I'm not sure about is that noise-normalized tests use stock configuration in GamersNexus video, are the results noise-normalized when using standardized fans as well ?

they way i understand it (could be wrong) the hot pocket under the gpu was do to some air being pushed out at the side panel. skinnier case will have less room or wider gpus will also have less room not only that all the crap they put on the side of the gpu blocking the exhaust. now some air gos up towards the back fan(on font air flow case the torrent is different) well some gets re sucked in to the gpu  and repeats. having a bottom mount fan wil not only push air directly in to the gpu but as well push the gpu hot air exhaust up instead of being sucked back in.

 

funny that a dell 3090 has less crap on the vesnt making it a good cool card...

 

most cases have a basement. some have it wide open some have a smaller gap and some are fully closed off well on top could be full honey comb mesh some more restrictive mesh and some closed off. not only that there an hdd rack normally in front of the psu so the air has to go around the hdd and find its way up thow the basement that alot of detours  compared to bottom mount fan case.

 

keep in mind that the standerised case test uses only 3 fans 2 front and 1 back that is. his test bench usese a gtx 1080. not hot runny 3090...or hot running cpus. pluse the gpu fan are at 50%

 

the test bench was saposbly picked for what most people will be running rather then the hi end. imo i think an upgraded test beach is need or run 2.

Edited October 2, 2021 by thrasher_565

[Dogzilla07]

On 10/2/2021 at 4:26 AM, thrasher_565 said:

https://www.gamersnexus.net/guides/3477-case-fan-standardization-tests-noise-normalized-thermals

Yeah, it's as I assumed from a hunch + memory, the standardized tests, do not have the Noctua fans noise-normalized (at least not mentioned), so it's safe to assume they're running at full speed.

 

For the bot fans + hot pocket In theory yes, but for most cases and GPUs (cause they're too high) it doesn't work, check the GamersNexus Lian Li O11 review, and notice how 3 fans intake bot is worse for GPU, then just 3 side intake. You need high feet, good mesh/holed metal bottom plate. How good holes PCI-E has matters. and having the front most intake with short gpus can easily disrupt the closest GPU fan from operating optimally. bottom fan needs to be able to push the air around the GPU as much as into the GPU fans to stop the vortex, otherwise it just happens again.

 

As for the mechanics yes, you're completely correct, it happens because of top-top bottom fin orientation and hot exhaust into the motherboard where the GPU is plugged in and into the side. (The small fan under is a clutch, similar to removing PCI-E brackets above and below the GPU I/O area).

 

And no obstruction high static pressure front intake, does the same as you mention the bot should do, disrupt the hot pocket, air vortex, just prolong the blue lines on your sketch to go all the way to the back, and over the red lines from the bot, and just straighten them and push air out the back, same as above. Same for top it's not all gonna go into the 90mm exhaust.

 

The orange lines are not the full path the air takes, just a part of it. I do agree 3090 cause a problem, and it might be needed to experiment with putting back some PCI-E brackets depending on the card.

Interesting tidbit is some people on OC forums who ran a 80mm/92mm fan below the GPU, like on the picture in your post above, reported that they had better performance with it removed with 3080Ti/3090 cause the GPU wasn't getting enough air.

 

Either way it's case and gpu dependant, and how much of the front of the case has been taped-off so as not to leak air. U can get a negative exhaust pressure setup to work as good as a positive pressure setup with less hassle setting it up, but more dust which is more hassle afterwards.

 

As for a new test bench, it's unfortunately very impractical for GamersNexus, as they would have to scrap all previous results and start from scratch, which is the whole point of my walls of texts here xD (trying to combine all of their tests and findings, extract the essence of that information and their conclusions and personalize it for each case for each person here on the forum).

[thrasher_565]

2 hours ago, Dogzilla07 said:

Yeah, it's as I assumed from a hunch + memory, the standardized tests, do not have the Noctua fans noise-normalized (at least not mentioned), so it's safe to assume they're running at full speed.

 

For the bot fans + hot pocket In theory yes, but for most cases and GPUs (cause they're too high) it doesn't work, check the GamersNexus Lian Li O11 review, and notice how 3 fans intake bot is worse for GPU, then just 3 side intake. You need high feet, good mesh/holed metal bottom plate. How good holes PCI-E has matters. and having the front most intake with short gpus can easily disrupt the closest GPU fan from operating optimally. bottom fan needs to be able to push the air around the GPU as much as into the GPU fans to stop the vortex, otherwise it just happens again.

 

As for the mechanics yes, you're completely correct, it happens because of top-top bottom fin orientation and hot exhaust into the motherboard where the GPU is plugged in and into the side. (The small fan under is a clutch, similar to removing PCI-E brackets above and below the GPU I/O area).

 

And no obstruction high static pressure front intake, does the same as you mention the bot should do, disrupt the hot pocket, air vortex, just prolong the blue lines on your sketch to go all the way to the back, and over the red lines from the bot, and just straighten them and push air out the back, same as above. Same for top it's not all gonna go into the 90mm exhaust.

 

The orange lines are not the full path the air takes, just a part of it. I do agree 3090 cause a problem, and it might be needed to experiment with putting back some PCI-E brackets depending on the card.

Interesting tidbit is some people on OC forums who ran a 80mm/92mm fan below the GPU, like on the picture in your post above, reported that they had better performance with it removed with 3080Ti/3090 cause the GPU wasn't getting enough air.

 

Either way it's case and gpu dependant, and how much of the front of the case has been taped-off so as not to leak air. U can get a negative exhaust pressure setup to work as good as a positive pressure setup with less hassle setting it up, but more dust which is more hassle afterwards.

 

As for a new test bench, it's unfortunately very impractical for GamersNexus, as they would have to scrap all previous results and start from scratch, which is the whole point of my walls of texts here xD (trying to combine all of their tests and findings, extract the essence of that information and their conclusions and personalize it for each case for each person here on the forum).

gn test of the 011 was 3 fans in front and i dont no where hi put the last fan as there's no room for it. so im guessing he just did 3 in the front.

 

the 011 dose not have a back fan that is why front fans and bottom fans are needed. the back fan is needed to exsust the gpu hot air. bottem fans will cool the gpu the best end of destitution. like i siad it pushes the gpu exhaust up. but not having a back fan means it has to go all the way to the top. the tower cooler can they squees the hot air out the back but some wont make it so in that case having a top back fan is needed. its more or less a water cooling case kinda poor for air cooling. the mini is better and the mini air is even better for air cooling.

 

ive owned the 011 razer and i have the 011 air

the razer i think has a mount for 1 80mm fan and the 011 air has an mount for 2 80mm fans. althow the mesh on there is restrictive they did it for looks then pefromace.

 

(the 80mm fan holes are i the same as an ssd mount hose. )

 

if you want to cut and add your own fans to the back of the 011 then 3x 80mm fans will fit or 2x 92mm fans

 

 

 

 

gn standerised testing is 2 120mm fans in front and 1 120mm in the back

if you cant fit it i dose not test standerised stest on it i think

imo 1front 1 bottem 1 back would be idel when posable but he dose not test that

if the case come with fans he test were there stock config is then he might move em around to see it he can make it better normally all in the front

he only needs to cool a gtx 1080 ti and i forget the cpu.

 

he dose not test all test on every case nor dose he fill the case up with fans to see what the best.

 

its posable that the gtx 1080ti runs cool enough with just fans in the front

Edited October 6, 2021 by thrasher_565

[Dogzilla07]

GamersNexus did just 3 in the front yes, no exhaust, and that was the best setup of all the ones tested. Not really, bottom fans are not needed (at most 1 is ok), again for O11. GamersNexus tested this and showed that u absolutely do not need full front/side and full bottom fans. Heck not even partial. Bottom fans being better for GPU is a fallacy, it's true in a  number of situations in specific combinations of cases and gpus, but is in no way, the best way, nor the way that works every time.

 

O11 is not a horrible case for air cooling, it's ok, as long as you a person does what I explained, and GamersNexus proven without a shadow of a doubt, and not otherwise, ...

 

CPU area is gonna be fine in O11, it's the need to remove the PCI-E slots above and below the GPU that's paramount.

 

more than half the cases on market do not have a bottom fan mount, so it is impossible to have a methodology that is consistent. There is nothing worse than filling the case up with fans, it is always gonna be worse than a smaller number of fans. U can read it on the link about their methodology, just because they don't show us, doesn't mean they didn't test a huge variety of fans and fan configuration on their first case. Yes newer cases might be different, so every person needs to try around a bit, but the general things are the same.

 

It's not just 1080Ti, all cards can run fine in proper cases with only front fans, there's thousands of testimonies on reddit, and other forums, over the years. And don't underestimate 1080TI custom AIB in how much heat it produces.

[andrewmp6]

To see more normal temps watch this one

 

[kyasca]

for my understanding, the main problem in this case, is the stock fan configuration, that makes not achieving better gpu temps.

having 5 intake fans, and 0 exhaust, is like blowing hard into a closed bottle. no matter how hard you blow in, you are limited by the exhaust aperture.

 

obviously depends too on the kind of mesh , and filter used at the bottom, and heigh of bottom feets. but with cooler master sl 600m they proved that direct fans to the heatsink of the gpu / cpu with low obstruction, is the best to achieve the best temperatures.

 

for this case, i feel that the best configuration would be:

or 2 intake front, 1 exhaust bottom (rear), 1 exhaust rear

 

or 2 intake bottom, 1 intake rear with tower heatsink blowing its fans to the front of the case, ahd 2 fans exhaust at front.

 

everything intake makes no sense.

[thrasher_565]

1 hour ago, kyasca said:

for my understanding, the main problem in this case, is the stock fan configuration, that makes not achieving better gpu temps.

having 5 intake fans, and 0 exhaust, is like blowing hard into a closed bottle. no matter how hard you blow in, you are limited by the exhaust aperture.

 

obviously depends too on the kind of mesh , and filter used at the bottom, and heigh of bottom feets. but with cooler master sl 600m they proved that direct fans to the heatsink of the gpu / cpu with low obstruction, is the best to achieve the best temperatures.

 

for this case, i feel that the best configuration would be:

or 2 intake front, 1 exhaust bottom (rear), 1 exhaust rear

 

or 2 intake bottom, 1 intake rear with tower heatsink blowing its fans to the front of the case, ahd 2 fans exhaust at front.

 

everything intake makes no sense.

is most case this make sense but in the torren the back is as open as can be with in reason (yes you can remove the pci slot covers, cut out the back fan mesh there still a bit of resistances on the ,mesh) so all the air in wont bonus off the back and it cant build up air presher as it just get push out the back.  so having all intake and 1 rear exhaust is fine. i sapos if there so much air from the front and the rear cant keep up then maybe it might? but it probly wont have much effect if any.

 

althow having intake from the front and bottom for gpu i think there could be a mix there. might be better to have one or the other.

Edited October 19, 2021 by thrasher_565

[DD08]

Hi
Sorry for jumping on this old thread.
I'm using the Torrent with a 420 AIO at the front in Push/Pull config, 2 140mm at the bottom, and 1 at the back. Total volumetric flow for just 3 140mm fans is higher, than the 180mm fans, and the static pressure is higher as well.
MoBo has 4 slot distance between the GPU's(3090's), leaving roughly 2cm / 4/5 inch gap between the cards.
The heavy "overpressure" and "forced" flow towards the lower GPU's, really seems to work quite well, when compared to the old noctua-iPPC-fan-stuffed Phanteks P600S.

Fans are NOT "Air-Lazers", but just create a slight overpressure/flow from them. The air will move in whatever direction there is the least friction/resistance.

Bottom fans were mounted to force the exhausted hot air from the GPU up and out, instead of being circulated back into the GPU. These are 25mm thick, and are very close to the GPU. Considered buying the slim 15mm Arctic/Noctua, but can only get them in 120mm... thereby not covering the full surface of the GPU. They will likely force the GPU fans to turn, but I reckon it will lead to no harm.  

 

The "venturi"-like design of the PSU "shroud", should also help with the flow towards the CPU area and rear/PSU-exhaust.

I hope my gibberish makes a tiny bit of sense. Just wanted to express my happiness about this case... Phanteks or Lian Li is better quality and is easier to build in, though.

I've taken photos of the complete build process, including the few modifications needed... Do you reckon anyone would be interested in a Youtubified oldschool Powerpoint presentation of it, or is it only 4K HDR bling videos with Dolby Atmos surround, people want these days?

Cheers, DD

[thrasher_565]

14 hours ago, DD08 said:

Hi
Sorry for jumping on this old thread.
I'm using the Torrent with a 420 AIO at the front in Push/Pull config, 2 140mm at the bottom, and 1 at the back. Total volumetric flow for just 3 140mm fans is higher, than the 180mm fans, and the static pressure is higher as well.
MoBo has 4 slot distance between the GPU's(3090's), leaving roughly 2cm / 4/5 inch gap between the cards.
The heavy "overpressure" and "forced" flow towards the lower GPU's, really seems to work quite well, when compared to the old noctua-iPPC-fan-stuffed Phanteks P600S.

Fans are NOT "Air-Lazers", but just create a slight overpressure/flow from them. The air will move in whatever direction there is the least friction/resistance.

Bottom fans was mounted to force the exhausted hot air from the GPU up and out, instead of being circulated back into the GPU. These are 25mm thick, and are very close to the GPU. Considered buying the slim 15mm Arctic/Noctua, but can only get them in 120mm... thereby not covering the full surface of the GPU. They will likely force the GPU fans to turn, but I reckon it will lead to no harm.  

 

The "venturi"-like design of the PSU "shroud", should also help with the flow towards the CPU area and rear/PSU-exhaust.

I hope my gibberish makes a tiny bit of sense. Just wanted to express my happiness about this case... Phanteks or Lian Li is better quality and is easier to build in, though.

I've taken photos of the complete build process, including the few modifications needed... Do you reckon anyone would be interested in a Youtubified oldschool Powerpoint presentation of it, or is it only 4K HDR bling videos with Dolby Atmos surround, people want these days?

Cheers, DD

that be funny to do a video like they use to back in the day now haha make it like 800,400... but then upload better vetion later on...

 

gota have the blue back ground an texed and some music thow you need to find not copy right music... maybe start the video with like cs then switch haha

 

basically do what you want...

 

 

Edited January 18, 2022 by thrasher_565

[freeagent]

On 10/19/2021 at 4:59 PM, kyasca said:

everything intake makes no sense.

That how I am running my Meshify C. 4 intake no exhaust and it works great. If you have weak fans it’s probably not a good idea.

[Nasty Nate]

I just got this case and if you use it the way they suggest, the case fans only speed up when your cpu cooler speeds up, meaning when the cpu temp rises.  There are lots of games that will beat up your gpu and the cpu will be almost asleep and cool as a cucumber, so you won't get any gpu cooling.  I made it so I can control my cooler and the case fans seperately in the bios, telling the case fans to go to 100 percent at a very low cpu temp.  I get the best temps ever with this case, stock fan configuration.  Ryzen 5 5600x overclocked to 4700Mhz (pulls up to 115 watts), 3080ftw3 overclocked and pulling up to 450 watts, scythe fuma 2 air cooler, stock configuration.  The case fans don't spin that fast because they are very large, even at 100%, so you don't want to slow them down when you're gaming at maximum levels.  CPU and GPU temps stay below 70c even when they are using 30% more power than stock.