• Please review our updated Terms and Rules here

Airflow question

hunterjwizzard

hunterjwizzard

Veteran Member
Joined
Mar 20, 2020
Messages
2,293
When replacing fans, how important is it to match the CFM? I am trying to replace some cooler fans with a "quiet" model, but the existing fan is 33cfm and the quiet ones I'm finding are like 8-16cfm. I know in a system with a functioning fan controller the max speed will cycle down depending on the cooling requirements but I'm not sure if I can go under the max.
 
I'm not a huge fan of that solution since it means buying multiple fans until I get one right. Also this is going into an Opteron system where mistakes mean mysterious blue smoke.
 
Color me confused. :rolleyes: How will you know if a lower airflow fan will work if you don't observe the temperature? Personally, if I were unsure, I'd install a high-airflow fan and throttle the speed down and watch the temperature. So without and detailed data, how do you expect to get an answer if we don't even know what it is that you're trying to cool?
 
Ok, I am trying to cool a dual socket F system with AMD Opteron processors. I installed a pair of Coolermaster E2U-K65CS-01 socket F coolers. Per this spec sheet I found on teh interwebs: https://download.datasheets.com/pdfs/2005/10/21/c/cs/colr/ds/e2u-k65cs-01.pdf them things are pumping out 45 DB, which explains why I can barely stand to be in the same room with the system when its running.

The fans are 60mm which is easy enough to find in "quiet" variants, but I'm concerned the airflow won't be high enough.
 
I wouldn't risk it personally.

The reason they are probably quieter is because they shift less air...

Dave
 
daver2 said:
The reason they are probably quieter is because they shift less air...
Click to expand...
They definitely shift less air, unless I am confused about what the CFM measurement is.

My problem is I am otherwise stuck for a way to quiet this system down without spending $$$$$ on cooling. And for some reason my wife won't let me set up the tower in her office and then attach it to my system by cutting a hole in the wall to run the cables through.
 
CFM = cubic feet per minute. Basically, the less air the fans shift, the quieter they are.

>>> And for some reason my wife won't let me set up the tower in her office and then attach it to my system by cutting a hole in the wall to run the cables through.

Hmmmm, I don't need to think much about that one!

Dave
 
daver2 said:
CFM = cubic feet per minute. Basically, the less air the fans shift, the quieter they are.
Click to expand...

I feel like there's more to the "quiet fan" technology than "just make it shift less air", but the quieter ones definitely have lower CFMs. I've seen a few get as high as 18 CFM while only clocking 27 db. I have no concept of how loud that is, but my grade school math tells me 27 is less than 45. But Alexa claims 45 db is not very loud.

I'm slowly going insane in case that wasn't clear. The sad part is working on this project is supposed to be my break from more stressful exercises.
 
hunterjwizzard said:
how important is it to match the CFM?
Click to expand...
Very, as you well know.

hunterjwizzard said:
I feel like there's more to the "quiet fan" technology than "just make it shift less air", but the quieter ones definitely have lower CFMs
Click to expand...
You sound like you are getting stuck on what you want to believe toward the solution that you want.

Recommend you concentrate on the fundamentals. You have a CPU that runs at a certain temperature - can you not monitor that?

You have a CPU with a knowable operational temperature range (overclockers go nuts about this stuff). You need those numbers AND you need to know how close or far you are now, with the working, albeit noisy, system.

Nobody can tell you the precise relationship between CFM and temperature unless all (or most) of the other factors are kept constant - it is reasonable to think that you can keep all those factors constant since you have one system in a room. Given that case, how can you vary CFM?

Chuck(G) said:
I'd install a high-airflow fan and throttle the speed down and watch the temperature.
Click to expand...

So, if you can figure out how to increase or decrease your CFM via messing with PWM or something even simpler, then you can see what effect it has on temperature. A SPL meter can be used in concert with the manipulations. I agree with the suggestion.

Too much? Then buy the quieter fan and wait until the CPU overheats. More seriously, if you monitor CPU temp and know the max recommended temp, you could watch it and have a good idea, before it shut down.

You may also be able to add to other areas of the case fans, but that seems somewhat indirect to me.

I can't believe you can't measure CPU temp!
 
Last edited:
syzygy said:
I can't believe you can't measure CPU temp!
Click to expand...
Oh I can see the CPU temp easy. From the BIOS or a util in windows.

Problem is the motherboard is a salvage from an old HP workstation tat appears to have been designed for a water cooling system. There are no headers on the motherboard for CPU fans. As such, I have to run the fans directly off the PSU, which means no automated throttling. It just runs at full power all the time. The CPU and case fans are quieter.

Is there maybe a fan controller I can get that interfaces with the board?
 
You may be able to quieten the fan down by blade design - and still push the same airflow. You can also make it worse! There are other techniques - but they just add cost to the solution...

Dave
 
Well at the moment the machine has stopped posting. I did hook one HSF fan up to an external power supply and ho daddy is it loud. I do have a manual fan control somewhere, if I manage to find it I'll hook it up and see what throttling it down does to the noise. Its possible at 80% it'll be worlds quieter while still cooling well.

The HSFs I have are designed for 2U rack servers, an environment notoriously unconcerned with acoustics, but also one known to generally be warmer than say a comfortable office.
 
We had a (relatively small) server at work with a set of fans that would give the space shuttle a run for its money!

When we closed the office I thought about taking it home. Then thought about the wife's comments... Decided against it!

Dave
 
hunterjwizzard said:
Oh I can see the CPU temp easy. From the BIOS or a util in windows.

Problem is the motherboard is a salvage from an old HP workstation tat appears to have been designed for a water cooling system. There are no headers on the motherboard for CPU fans. As such, I have to run the fans directly off the PSU, which means no automated throttling. It just runs at full power all the time. The CPU and case fans are quieter.

Is there maybe a fan controller I can get that interfaces with the board?
Click to expand...
ugh so no rpm feedback and no pwm.

How close are you now to a max temp and can you look at it when you know the CPU is at a max load?

You can control the fan speed using a pot, a mosfet and a resistor, and I would NOT suggest that you do it that way, which is by varying voltage and a simple search on such a setup will be accompanied by EEs explaining why it is not a good idea. Experimentally, I would futz with it but yours is a different situation.

You don't want to hear it but a water cooled system is consistent with the design. Remember, happy wife, happy life.
 
syzygy said:
How close are you now to a max temp and can you look at it when you know the CPU is at a max load?
Click to expand...
Oh I'm not far enough into this build to have any idea about that. When it posts(the machine is persnickety like that) it boots into XP x64. I ordered a pair of graphics cards for it late the night before last. I have not started installing software because its noisy and I was waiting on a proper GPU setup before I continued. I started this thread thinking if I can fix the noise problem for ~$15 we'd be in fine shape.

syzygy said:
You don't want to hear it but a water cooled system is consistent with the design. Remember, happy wife, happy life.
Click to expand...
It would be an interesting project to try and find a water cooling system for socket F. In my experience obsolete water cooling parts have 0 after-market. I've always thought it'd be a hell of a lot of fun to build custom water blocks so I could water-cool every system I own(and water-cool every warm internal component). But that's a long ways out. For the moment I think the little woman would be more upset with me if I spent $200 on AIO liquid coolers for the motherboard I got for free :ROFLMAO:

This does however have me thinking about an alternative solution - what if instead of looking at "quiet fans" I start looking for comparable fans and see if maybe I can find one that happens to be quieter while also moving a similar amount of air?

I feel like it can't be tough to beat the fans that are in there.
 
If one *really* wanted to move the right amount of air through the case more quietly I'd be considering:

(1) reducing the ambient (intake) air temperature where the computer is located.
(2) in effect pressurizing the case interior by adding forced intake fan(s) -- probably not feasible given typical case design.
(3) building an outflow duct to host two or more quiet(er) fans in parallel to achieve the original rated CFM.

If noise is more important than beauty, then #2 seems quite feasible. One might be able to use a set of smaller fans to achieve required CFM.
 
hunterjwizzard said:
Ok, I am trying to cool a dual socket F system with AMD Opteron processors. I installed a pair of Coolermaster E2U-K65CS-01 socket F coolers. Per this spec sheet I found on teh interwebs: https://download.datasheets.com/pdfs/2005/10/21/c/cs/colr/ds/e2u-k65cs-01.pdf them things are pumping out 45 DB, which explains why I can barely stand to be in the same room with the system when its running.

The fans are 60mm which is easy enough to find in "quiet" variants, but I'm concerned the airflow won't be high enough.
Click to expand...
You could party like it's early 2000s a dump everything in a tank of mineral oil.
 
My rule for noisy fans is to replace a smaller fan with a bigger fan with a similar or greater air flow.
Another alternative I used once was to use two fans one front/one back, one sucking air into the case, the other exhausting air from it (although I'm not 100% certain how effective that was).
 
Plasma said:
You could party like it's early 2000s a dump everything in a tank of mineral oil.
Click to expand...

Expensive but also somewhere on my bucket list. I knew a guy back in the day who had his dual single core Xeon overclocked from 2.0 to 4.0GHz this way. Chilled mineral oil delivered by a sump pump directly to the Jakes CPUs. Quite a site.
 
You must log in or register to reply here.
Back
Top