Accelero S1 - Side fan enough?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Accelero S1 - Side fan enough?
I have a P182 and 8800GTS-512 with Accelero S1. Currently, I have two 120mm fans near the Accelero:
1. Low-speed Yate Loon ziptied to the Accelero
2. Nexus Real Silent connected to the end of the middle HDD cage via Antec's clips
GPU temps were still within reason after disconnecting fan #1, but should I be concerned about the rest of the card? Will the mosfets or some other part of the card break down quicker because air is coming from the side, rather than directly above it?
1. Low-speed Yate Loon ziptied to the Accelero
2. Nexus Real Silent connected to the end of the middle HDD cage via Antec's clips
GPU temps were still within reason after disconnecting fan #1, but should I be concerned about the rest of the card? Will the mosfets or some other part of the card break down quicker because air is coming from the side, rather than directly above it?
A S1 need very little air flow to cool your VGA. I have an extractor S12 fan @800rpm, a S12@600rpm for CPU, a S12@500-600rpm that you could call an "intake" fan. But the latter is only there to aim air flow to the GPU, 8800GT oced, because of the negative pressure configuration. That's why the extractor is almost at 800rpm at all the time.
500rpm is really moe than enough to cool a 8800GT be it passive or semi passive configration.
I got the following temperature:
idl: 33° for CPU, 50° for GPU, 28° for HDDs and 33° for system (in Celcius);
load: 45° for CPU, 60° for GPU, 33° for HDD and 39° for system.
Note: you have to get rid of several holes to get your aimed air flow.
In a Sonata 3, using the PCI extra bracket from S1 package, I got for the GPU an air flow from intake --> GPU --> out the case. So the main heat from a 880Gt is not used to heatt others components, at least not much.
PS: you shouldnt worry about mosfet heatsink if you have a proper air flow. I have no issue and I'm not using any.
500rpm is really moe than enough to cool a 8800GT be it passive or semi passive configration.
I got the following temperature:
idl: 33° for CPU, 50° for GPU, 28° for HDDs and 33° for system (in Celcius);
load: 45° for CPU, 60° for GPU, 33° for HDD and 39° for system.
Note: you have to get rid of several holes to get your aimed air flow.
In a Sonata 3, using the PCI extra bracket from S1 package, I got for the GPU an air flow from intake --> GPU --> out the case. So the main heat from a 880Gt is not used to heatt others components, at least not much.
PS: you shouldnt worry about mosfet heatsink if you have a proper air flow. I have no issue and I'm not using any.
I actually have heatsinks on the memory and mosfets, but was wondering if they needed air as well. I worried that the bank of capacitors would restrict airflow there.
When I ran ATITool's artifact tester (both Accelero fans on), I noticed that my CPU temp went up 10 degrees to a whopping 52C. That seems awfully high, so I have a feeling the ziptied fan is blowing hot air upward. I'll give it another try without fan 1, and report my findings.
When I ran ATITool's artifact tester (both Accelero fans on), I noticed that my CPU temp went up 10 degrees to a whopping 52C. That seems awfully high, so I have a feeling the ziptied fan is blowing hot air upward. I'll give it another try without fan 1, and report my findings.
I got rid of Fan #1, and the temps are actually worse than I remembered. The GPU now idles at 57C, and hit 67C after 2 minutes of ATITool (at which point I could no longer bear the artifact test.) To add insult to injury, the CPU core temps are exactly what they were before, when the YL was blowing hot air upwards.
Maybe something is wrong with my negative pressure cooling model?
The blue arrows indicate Nexus 120mm fans running at 660rpm idle, 1080rpm load.
Maybe something is wrong with my negative pressure cooling model?
The blue arrows indicate Nexus 120mm fans running at 660rpm idle, 1080rpm load.
Euhm, 52C is pretty good for a graphics card when used (ATItool running). My 8800GT on stock ran at 70C(ish, dont remember exactly) idle and 90C+ when used.nobody wrote:When I ran ATITool's artifact tester (both Accelero fans on), I noticed that my CPU temp went up 10 degrees to a whopping 52C. That seems awfully high
With my S1 it now run 50ish when stressed.
52C was actually the temp from one of my E6600's cores. It was at 42C before ATITool, so it must've been getting a ton of hot air from the card.seraphyn wrote:Euhm, 52C is pretty good for a graphics card when used (ATItool running). My 8800GT on stock ran at 70C(ish, dont remember exactly) idle and 90C+ when used.nobody wrote:When I ran ATITool's artifact tester (both Accelero fans on), I noticed that my CPU temp went up 10 degrees to a whopping 52C. That seems awfully high
With my S1 it now run 50ish when stressed.
Has anyone tried ziptieing a fan to the SI that blows away from the card?
-
- Patron of SPCR
- Posts: 744
- Joined: Tue Mar 04, 2008 4:05 am
- Location: London
- Contact:
My 7900GTO (~80w TDP at stock, slightly overclocked so probably ~85-90w) is passive with my accelero. Just using the intake fan aimed at the heatsink i get temps between 39-48c idle depending on how hot a day it is, and ive never seen it go above 68 under any stress test, which is about what i was getting in idle with the stock cooler.
I use an Antec solo with 1 exhuast, 2 intakes (a 92mm aimed at the Accelero and a kama bay cooler aimed at CPU), and i find that changing the exhuast rpm makes a big difference in CPU and GPU temps. Ive seen 7c drops whilst gaming on my GPU when switching from 5v to 12v. So if you have good airflow (particularly something aimed at your GPU) you can cool very well with a passive S1. The 8800GTS has about 40% more TDP i think, but with a slightly higher intake CFM this shouldnt be a problem.
I use an Antec solo with 1 exhuast, 2 intakes (a 92mm aimed at the Accelero and a kama bay cooler aimed at CPU), and i find that changing the exhuast rpm makes a big difference in CPU and GPU temps. Ive seen 7c drops whilst gaming on my GPU when switching from 5v to 12v. So if you have good airflow (particularly something aimed at your GPU) you can cool very well with a passive S1. The 8800GTS has about 40% more TDP i think, but with a slightly higher intake CFM this shouldnt be a problem.
All your temps are well below the thermal thresholds so I wouldn't be too worried. Could the increase in CPU temps simply be because the CPU is doing more work while ATITool is running? (I am not familiar with ATItool so i don't know)nobody wrote:52C was actually the temp from one of my E6600's cores. It was at 42C before ATITool, so it must've been getting a ton of hot air from the card.seraphyn wrote:Euhm, 52C is pretty good for a graphics card when used (ATItool running). My 8800GT on stock ran at 70C(ish, dont remember exactly) idle and 90C+ when used.nobody wrote:When I ran ATITool's artifact tester (both Accelero fans on), I noticed that my CPU temp went up 10 degrees to a whopping 52C. That seems awfully high
With my S1 it now run 50ish when stressed.
Has anyone tried ziptieing a fan to the SI that blows away from the card?
My bad, but 52C is still a decent temperature, also for a CPU under load.Immortals wrote:52C was actually the temp from one of my E6600's cores. It was at 42C before ATITool, so it must've been getting a ton of hot air from the card.
Has anyone tried ziptieing a fan to the SI that blows away from the card?
Sorry for not posting my findings earlier, but here they are:
Fans used in all of the setups (all @ 6v idle, 9v load):
1. Nexus 120mm exhaust on TRUE120
2. Nexus 120mm exhaust, case top
3. Nexus 120mm exhaust, case rear
4. Nexus 120mm exhaust in between PSU and lower HDD bay
Results:
A. Passive Accelero S1: 70C after 5 minutes of AtiTool
B. D12SL-12 clipped to end of middle HDD bay: 55C...
C. D12SL-12 ziptied to Accelero S1: 55C...
D. Fans from C and D: 45C....
I went with D because the undervolted Yates are inaudible over the humming sound of my 22" Samsung LCD, and don't make a discernible noise difference when the monitor's off at night.
Fans used in all of the setups (all @ 6v idle, 9v load):
1. Nexus 120mm exhaust on TRUE120
2. Nexus 120mm exhaust, case top
3. Nexus 120mm exhaust, case rear
4. Nexus 120mm exhaust in between PSU and lower HDD bay
Results:
A. Passive Accelero S1: 70C after 5 minutes of AtiTool
B. D12SL-12 clipped to end of middle HDD bay: 55C...
C. D12SL-12 ziptied to Accelero S1: 55C...
D. Fans from C and D: 45C....
I went with D because the undervolted Yates are inaudible over the humming sound of my 22" Samsung LCD, and don't make a discernible noise difference when the monitor's off at night.