I have a 3500+ Venice running with the stock cooler and it idles around 31C.
Using an 80mm adapter, what's the quietest fan I could get away with using, and still maintain good temps while gaming?
I'm running it at stock and don't plan to overclock if I go for quieter cooling.
What's the least you can get away with?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Re: What's the least you can get away with?
Idle temps are not really indicative. You need to monitor the loadtemps. I think you should just go ahead and experiment. Get a fanmate and some decent 80mm fan and see for yourself how low you can crank it.MrBusiness wrote:I have a 3500+ Venice running with the stock cooler and it idles around 31C.
Using an 80mm adapter, what's the quietest fan I could get away with using, and still maintain good temps while gaming?
The absolute limit on how low you can go depends on the rest of your setup.
As an example, I'm using an Y.S. Tech 80mm fan that normally runs at 2000rpm with its speed reduced to 1600rpm. I could have just gotten a 1600rpm fan, but now I have the option to just improve cooling by turning on a button if the need arises.
Hi, I'm with kesv on this one, get a fan that run 2000-2500rpm and a fan controller or fanmate. Run CPU Burn and turn the fan down till get temp of 60-65C after 30 minutes run time. If you want to push it use Prime95 instead of CPU Burn and you'll get lower temp/slower fan speed.
Panaflo L1A (or whatever they're called now) and Yate Loon D80SL are good 80mm fans that run around 2000rpm, there are many others too.
Seb
Panaflo L1A (or whatever they're called now) and Yate Loon D80SL are good 80mm fans that run around 2000rpm, there are many others too.
Seb
-
MrBusiness
- Posts: 7
- Joined: Thu May 25, 2006 7:46 am
I've found that there are 60mm to 92mm adapters, so now I have two options.
1. An 80mm adapter with a PanaflowL1A 80mm
(does C&Q work with a fan that does not have a fan speed wire?)
2. A 92mm adapter with a NexusRS 92mm
Does the Nexus 92 provide more cooling while producing less noise than the Panaflo?
1. An 80mm adapter with a PanaflowL1A 80mm
(does C&Q work with a fan that does not have a fan speed wire?)
2. A 92mm adapter with a NexusRS 92mm
Does the Nexus 92 provide more cooling while producing less noise than the Panaflo?
Hi, my understanding is that, generally, heat sinks that are designed to be used with small (<80mm fans) expect the fan to spin fast and therefore develop lots of air pressure which they take advantage of with many tightly packed fins.
As you use a bigger and bigger fan you can get the same airflow volume for less noise but you get less and less pressure. I would speculate for a given pressure that larger fans will be noisier.
The result is that I think a 92mm fan will be worse than 80mm through the stock cooler.
Also isn't the stock fan 70mm not 60mm? 80mm -> 70mm adapters do exist. My brief experience with an Athlon64 was the retail cooler at "60%" (~1600rpm, fairly quiet) was capable of keeping the chip under 50°C running prime95 all night, furthermore over clocking from 2GHz to 2.4GHz only lifted temps a few degrees.
I would suggest that if a good 80mm fan adapted onto the stock heatsink isn't quiet enough then you need a bigger/better heatsink.
Hope this helps, Seb
As you use a bigger and bigger fan you can get the same airflow volume for less noise but you get less and less pressure. I would speculate for a given pressure that larger fans will be noisier.
The result is that I think a 92mm fan will be worse than 80mm through the stock cooler.
Also isn't the stock fan 70mm not 60mm? 80mm -> 70mm adapters do exist. My brief experience with an Athlon64 was the retail cooler at "60%" (~1600rpm, fairly quiet) was capable of keeping the chip under 50°C running prime95 all night, furthermore over clocking from 2GHz to 2.4GHz only lifted temps a few degrees.
I would suggest that if a good 80mm fan adapted onto the stock heatsink isn't quiet enough then you need a bigger/better heatsink.
Hope this helps, Seb
-
ultrachrome
- Patron of SPCR
- Posts: 445
- Joined: Mon Apr 03, 2006 10:23 am
- Location: Pacific Northwest
SebRad, I disagree with the pressure comment. Pressure is created by the resistance of squeezing air through the fins on the heatsink. If two different sized fans blow into the same heatsink the same air volume at the same velocity, the pressure would be the same. I don't think pressure is really the issue.
However, I do agree that MrBusiness should first try ramping down the stock fan's speed and see if he can keep it around 50C under load.
You can also try undervolting with RightMark CPU Clock Utility with your fan speed reduction. I just undervolted my 3500+ from 1.5 to 1.2V and my load temp dropped 9C to 35C.
If you go the fan shroud route, you might as well go for the biggest fan. However, watch your costs as it might cost nearly as much to modify your stock cooler as it would be to buy an SPCR-recommended one. I use the Ninja which is currently $40 at NewEgg.
However, I do agree that MrBusiness should first try ramping down the stock fan's speed and see if he can keep it around 50C under load.
You can also try undervolting with RightMark CPU Clock Utility with your fan speed reduction. I just undervolted my 3500+ from 1.5 to 1.2V and my load temp dropped 9C to 35C.
If you go the fan shroud route, you might as well go for the biggest fan. However, watch your costs as it might cost nearly as much to modify your stock cooler as it would be to buy an SPCR-recommended one. I use the Ninja which is currently $40 at NewEgg.
-
tempeteduson
- Posts: 189
- Joined: Sat Jun 05, 2004 12:56 pm
- Location: Houston, Texas
- Contact:
I'm with ultrachrome on getting a heatsink that will accept a larger fan instead of bolting on a fan adapter. I've tried a 60mm-to-80mm adapter on my other computer, and the distance between the fan and the fins is a bit much, causing loss of cooling power. Unless you can get a low-profile adapter, I would recommend a bigger HS for both its compatibility with larger fans and its increased surface area for better heat dissipation.
Hi, I think I've not explained myself very clearly (and I could be wrong 
). To take the case of the 92mm Nexus with 92mm -> 60mm adapter into the stock heatsink because the Nexus is relatively low pressure fan and the heatsink is restrictive the amount of airflow through it will be low. That is the limiting factor in the volume of airflow is the heatsink as the airflow volume is directly related to the pressure applied by the fan divided by the resistance (restriction) posed by the heatsink.
A large slow fan may move more CFM in free air than a smaller high speed one but it's higher pressure potential will allow it to push more CFM through the heatsink.
The airflow delivered by a fan reduces as the pressure (resistance) increases and the pressure required goes up as airflow through the heatsink increases. Think of it as a pressure/volume graph, the fans slope down and the heatsink slopes up, where they cross is the airflow that will result from the combination.
I’ve made a graph using data for Panaflo 25mm thick fans, as only max pressure and max volume where given I’ve interpolated between them, in reality they lines are curves, and I’ve rounded the figures so overall it’s not precise but gives a good picture of how the fans relate to each other. I’ve no idea what shape/angle would be realistic for heatsinks so I’ve pull some numbers out of the air!
Looking at my graph you can see you see the sort of effects I’m trying to describe. The CFM each combo has is the height of the intersection point. True fan curves will also drop more sharply to 0 CFM as the maximum pressure is approached, exaggerating the effect of restrictive heatsinks. The low RPM fans we tend to use also further exaggerate restrictive heatsinks as when you half the RPM of a fan it may only half the CFM in free air but it reduces the fans maximum pressure to only a quarter. (see fan laws here) I believe this is why some of the SPCR tested heatsinks are fine at higher fan speeds yet the performance collapses below a certain fan speed. I also note one of the very best heatsinks, the Scythe Ninja, is clearly designed for very low resistance with widely spaced fins with holes/slots in them and its performance does not collapse at the very lowest fan speeds, hence its popularity.
Not sure if anyone’s managed to follow my arguments but I’ve learnt somethings and got them clearer in my mind. My current theories say:
Max airflow ~ diameter^2 x rpm
Max pressure ~ diameter^2 x rpm^2
The upshot is swapping 80mm for 120mm fan at ½ rpm gives the same airflow but only in free air, the 80mm fan has double the max pressure so if there is significant resistance to airflow the 120mm will drop faster than the 80mm fan. On the other hand the 120mm fan will be quieter. Maybe:
Noise ~ pressure potential
This would fit with the trend for bigger, freer heatsinks and bigger, slower fans to maintain cooling while reducing noise.
I still like the one that says, “When theory and reality conflict, reality always wins!â€
). To take the case of the 92mm Nexus with 92mm -> 60mm adapter into the stock heatsink because the Nexus is relatively low pressure fan and the heatsink is restrictive the amount of airflow through it will be low. That is the limiting factor in the volume of airflow is the heatsink as the airflow volume is directly related to the pressure applied by the fan divided by the resistance (restriction) posed by the heatsink.A large slow fan may move more CFM in free air than a smaller high speed one but it's higher pressure potential will allow it to push more CFM through the heatsink.
The airflow delivered by a fan reduces as the pressure (resistance) increases and the pressure required goes up as airflow through the heatsink increases. Think of it as a pressure/volume graph, the fans slope down and the heatsink slopes up, where they cross is the airflow that will result from the combination.
I’ve made a graph using data for Panaflo 25mm thick fans, as only max pressure and max volume where given I’ve interpolated between them, in reality they lines are curves, and I’ve rounded the figures so overall it’s not precise but gives a good picture of how the fans relate to each other. I’ve no idea what shape/angle would be realistic for heatsinks so I’ve pull some numbers out of the air!
Looking at my graph you can see you see the sort of effects I’m trying to describe. The CFM each combo has is the height of the intersection point. True fan curves will also drop more sharply to 0 CFM as the maximum pressure is approached, exaggerating the effect of restrictive heatsinks. The low RPM fans we tend to use also further exaggerate restrictive heatsinks as when you half the RPM of a fan it may only half the CFM in free air but it reduces the fans maximum pressure to only a quarter. (see fan laws here) I believe this is why some of the SPCR tested heatsinks are fine at higher fan speeds yet the performance collapses below a certain fan speed. I also note one of the very best heatsinks, the Scythe Ninja, is clearly designed for very low resistance with widely spaced fins with holes/slots in them and its performance does not collapse at the very lowest fan speeds, hence its popularity.
Not sure if anyone’s managed to follow my arguments but I’ve learnt somethings and got them clearer in my mind. My current theories say:
Max airflow ~ diameter^2 x rpm
Max pressure ~ diameter^2 x rpm^2
The upshot is swapping 80mm for 120mm fan at ½ rpm gives the same airflow but only in free air, the 80mm fan has double the max pressure so if there is significant resistance to airflow the 120mm will drop faster than the 80mm fan. On the other hand the 120mm fan will be quieter. Maybe:
Noise ~ pressure potential
This would fit with the trend for bigger, freer heatsinks and bigger, slower fans to maintain cooling while reducing noise.
I still like the one that says, “When theory and reality conflict, reality always wins!â€
-
ultrachrome
- Patron of SPCR
- Posts: 445
- Joined: Mon Apr 03, 2006 10:23 am
- Location: Pacific Northwest
Good link. I'll need more time to digest it.
I'm still not sure how you are determining that a smaller fan can create a higher pressure. Clearly there are areas of overlap where smaller fans produce higher CFM than a larger fan. But it seem to me that it can be explained by blade pitch as well as RPM.
Are you coming to this conclusion based on panaflo specs?
I think what we are talking about here is producing the same or comparable level of cooling with a given heatsink at a lower noise level.
I'm still not sure how you are determining that a smaller fan can create a higher pressure. Clearly there are areas of overlap where smaller fans produce higher CFM than a larger fan. But it seem to me that it can be explained by blade pitch as well as RPM.
Are you coming to this conclusion based on panaflo specs?
I think what we are talking about here is producing the same or comparable level of cooling with a given heatsink at a lower noise level.
-
StrongSilentType
- Posts: 22
- Joined: Sat May 20, 2006 10:47 pm
I have a AMD64 3000+ Venice CPU, and I use the stock fan running at 900RPM. I can't hear any difference between this speed and the fan being completely off.
It idles at 40oC and at max load it's at 50oC with 1800RPM. This speed is still fairly quiet, but the graphic card fan is quite loud when playing games.
The stock fan works very well, and I don't think there's any need for a new fan. The CPUs are rated to work up to 70oC, so I think with 65oC at load is OK (what you can get away with).
It idles at 40oC and at max load it's at 50oC with 1800RPM. This speed is still fairly quiet, but the graphic card fan is quite loud when playing games.
The stock fan works very well, and I don't think there's any need for a new fan. The CPUs are rated to work up to 70oC, so I think with 65oC at load is OK (what you can get away with).