Fan Configurations

MarkVarley

MarkVarley

MarkVarley

Associate
Joined
24 Apr 2013
Posts
7
As part of the planning for my first custom watercooling setup I decided to test three different fan configurations mounted on a radiator, the testing wasn't laboratory standard but the results weren't as expected so I'm curious on peoples thoughts particularly if they've done more accurate tests.

I was using a closed loop thermaltake cpu cooler with a 120mm radiator and I ran my system with prime95 for 10 minutes and noted the max temperature.
The fans I used were Zalman F3's at 12v (I had both fans hooked up to the same (external) supply the whole time so they'd be getting the same power with one or both being tested).

with one fan pushing my max temp was 51.8c
with one fan pulling my max temp was 56.8c
with two fans push&pull max temp was 59.3c

I suspect something wonky with those results, any thoughts folks?

Mark.
 
I don't think 10mins is a long enough test. It needs to long enough for the water to get to max temp.

Your starting water temp may have been higher when starting tests 2 and 3 from the preceding tests.

You do need to be measuring water temps to get any meaning full results from tests like this, which is hard in a closed loop.
 
I watched the temps drop to ~20c with Prime95 off before starting each next test but the incremental increase in temps each time does point to this perhaps being an issue.

I had noticed cpu temps hitting their maximum in just a few minutes so I'd thought 10 minutes indicated a stable temp at that load but again clearly something is amiss and it could be there, if I have more time I'll try and monitor for an hour each time with a long break in between tests.

I hadn't thought to measure water temp, partly as I can't directly (but I could put a probe on a hose) & partly as I didn't feel it was relevant, CPU temp is the difference I'm looking for.

More testing required on this one then..
 
the only way to test watercooling loops is by monitoring the water temp (any metal fitting will do in a pinch, not hose though as it takes much longer for the hose to match the water temp) AND ambient temp

CPU temp only tells part of the story

you will need to run for at least 30 minutes and more like 1 hour to ensure you've reached equilibrium (the bigger the reservoir you have the longer it will take, mine will sometimes take 2 hours before the water temp stops drifing gradually upwards from 2C over ambient to about 5C max)

and run all tests with the system fully warmed up, not from cold

all cooler testing is done by measuring ambient and CPU temp, so that you can see the difference between the 2... if you do one set of tests at 20C and the room raises to 25C in the meantime, all of your readings will go up by 5C, which would invalidate any comparison... my computer room easily fluctuates by 5C within as little as an hour, so just assuming room temp is constant because you've got the heating on the same setting or the weather is broadly the same isn't anywhere near close enough

I would expect the single fan results to be about 1C different from each other and the push pull to be about 2C cooler than either

were they all pointing to push air out of the case in all tests, or all pulling air in to the case?
 
Last edited:
andybird123 said:
the only way to test watercooling loops is by monitoring the water temp (any metal fitting will do in a pinch, not hose though as it takes much longer for the hose to match the water temp) AND ambient temp

CPU temp only tells part of the story

you will need to run for at least 30 minutes and more like 1 hour to ensure you've reached equilibrium (the bigger the reservoir you have the longer it will take, mine will sometimes take 2 hours before the water temp stops drifing gradually upwards from 2C over ambient to about 5C max)

and run all tests with the system fully warmed up, not from cold

all cooler testing is done by measuring ambient and CPU temp, so that you can see the difference between the 2... if you do one set of tests at 20C and the room raises to 25C in the meantime, all of your readings will go up by 5C, which would invalidate any comparison... my computer room easily fluctuates by 5C within as little as an hour, so just assuming room temp is constant because you've got the heating on the same setting or the weather is broadly the same isn't anywhere near close enough

I would expect the single fan results to be about 1C different from each other and the push pull to be about 2C cooler than either

were they all pointing to push air out of the case in all tests, or all pulling air in to the case?
Click to expand...

The test was done on a closed loop cpu cooler so no reservoir.

I have a temp sensor on the metal top of the radiator (opposite end to the hoses) that I currently use for fan control outside of this test, would that be sufficient or is it worth going for the metal hose fittings at the radiator-end?

The fans were getting open-room air and venting away into open room, however room temp could also be a factor, I'll keep that in mind when retesting.

I was just initially looking for a comparative difference as-to performance, if your prediction of 2c lower with a pair of fans as opposed to a single in either configuration is generally accepted then that'll be data-enough at this stage.

I'm also curious about push&pull slim fans compared to a single push or pull as they'll physically occupy the same space. Assuming CFM is a fair comparison of fans air-moving ability when radiator-mounted then it's potentially better.. :confused:
 
sorry, I just meant the more fluid you have in the system the longer it will take to get fully loaded up

a hose fitting might be better

CFM is also not the whole story on fans either - high static pressure is also a good thing, and specs on websites / boxes for fans are also not always comparable as there are no standards on how to test (noise is always the good one - do people test from 1m or 10m... will make a big big difference to a noise test!)
 
Hi, l have tried fan configurations in push, pull, push/pull, being all the same model and about 2'c variation is about right.

If you want to compile fan data, get a cheap thermometer so you know roughly what your room ambient is with out knowing that your data will be unreliable.

I'm using low 900rpm fans in pull configuration on my rad which give good temps, slightly less noise if they where in push but also does not clog the rad up as quick.

Everyone goes on about fan static pressure but in pull configuration this does not apply as much if at all.
 
andybird123 said:
sorry, I just meant the more fluid you have in the system the longer it will take to get fully loaded up

a hose fitting might be better

CFM is also not the whole story on fans either - high static pressure is also a good thing, and specs on websites / boxes for fans are also not always comparable as there are no standards on how to test (noise is always the good one - do people test from 1m or 10m... will make a big big difference to a noise test!)
Click to expand...

Thanks for that :)

One of the problems I'm having in researching without buying (I have about 6 or 7 different kinds of 120mm fans here) is that published cfm is all I have to go on (unless I stumble upon someones side-by-side testing) and as well as being varied across manufacturers, I think it's a different story when pushing air through a radiator, Static Pressure isn't a spec I've seen on fan information.
 
OLDPHART said:
Hi, l have tried fan configurations in push, pull, push/pull, being all the same model and about 2'c variation is about right.

If you want to compile fan data, get a cheap thermometer so you know roughly what your room ambient is with out knowing that your data will be unreliable.

I'm using low 900rpm fans in pull configuration on my rad which give good temps, slightly less noise if they where in push but also does not clog the rad up as quick.

Everyone goes on about fan static pressure but in pull configuration this does not apply as much if at all.
Click to expand...

I have thermometers for redoing this test, I didn't think about varying room temps so hopefully the next test will have better info. :)
 
MarkVarley said:
Thanks for that :)

One of the problems I'm having in researching without buying (I have about 6 or 7 different kinds of 120mm fans here) is that published cfm is all I have to go on (unless I stumble upon someones side-by-side testing) and as well as being varied across manufacturers, I think it's a different story when pushing air through a radiator, Static Pressure isn't a spec I've seen on fan information.
Click to expand...

There's 2 options when it comes to watercooling... overspec the radiator so that you can use any cheap fans and run them slow for total lack of noise... or scythe gentle typhoon fans to make up for a lack of rad space... noise/performance wise they are about the best

Or if you want maximum performance without regard to noise, look out for 38mm thick fans (most 120mm fans are 25mm)
 
You must log in or register to reply here.
Back
Top Bottom