QuietIce
Disabled
- Joined
- May 7, 2006
- Location
- Anywhere but there
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Theory / Science behind temp per radiator?
- Thread starter gcwebbyuk
- Start date
- Joined
- May 29, 2002
- Location
- Japan/Daytona Beach
not talking about which brand goes where just purely flow pattern this is what i would do.
bottom of case^Filter^Rad^fans.
if you have a Rheo you can actually play with RPMs of both in and out take and observe your systems performance and write the settings down in both voltage and rpm and create sound profiles for performance, then later if you get something like an orbital MX4 you can govern your whole loop from it.
oh for filters a thin low resistance easily removeable and washable filter can make life sooo easy. i actually design cases and cooling systems with this in mind
Last edited:
- Joined
- Mar 26, 2010
- Location
- Raleigh NC
I have that case in my server and its just too small for water cooling is that a mid or a full tower cause the one i have is a mid maby it looks smaller in your pics
OP
- Joined
- Nov 25, 2009
- Thread Starter
- #46
thorilan - the case has filters for the PSU, whole of the top, whole of the front and for the bottom rad bay, so all sorted there 
Currently using SpeedFan to control the fans as I only have two in the case now, but I do have Scythe Kaze Server which I was previously using on my air setup - so may re-mount that to control fan speed for the radiator fan/s.
andrew149 - its the 690-ii case - think its classed as a mid-tower, but it's plenty big enough for water cooling! It has been designed with watercooling mind. Have a read of owenator's watercooled review: http://www.overclockers.com/water-cooling-cooler-master-cm690-ii-advanced/
And hokiealumnus review: http://www.overclockers.com/cooler-master-c690-ii-advanced-review/
Currently using SpeedFan to control the fans as I only have two in the case now, but I do have Scythe Kaze Server which I was previously using on my air setup - so may re-mount that to control fan speed for the radiator fan/s.andrew149 - its the 690-ii case - think its classed as a mid-tower, but it's plenty big enough for water cooling! It has been designed with watercooling mind. Have a read of owenator's watercooled review: http://www.overclockers.com/water-cooling-cooler-master-cm690-ii-advanced/
And hokiealumnus review: http://www.overclockers.com/cooler-master-c690-ii-advanced-review/
OP
- Joined
- Nov 25, 2009
- Thread Starter
- #47
While I wait for the rad to arrive (tomorrow hopefully) I have connected up a couple of S-Flexs - zipped strapped them together and stood them vertically next to the reserator blowing cold air at it - have them running at 5v.
Prime95 running, plus a Acronis Drive Cleanse in the background, current temps low 50s. If I increase the speed, then temps drop, but even at just 5v there is a noticeable difference.
With regard to the Rad, would a shroud help at all? If so what would you recommend?
Prime95 running, plus a Acronis Drive Cleanse in the background, current temps low 50s. If I increase the speed, then temps drop, but even at just 5v there is a noticeable difference.
With regard to the Rad, would a shroud help at all? If so what would you recommend?
QuietIce
Disabled
- Joined
- May 7, 2006
- Location
- Anywhere but there
A shroud helps with almost any rad, for a single-fan on a double rad it should help a lot. I DIY mine out of 1/16-1/8" foam-board (thick paper coating on both sides of sheet foam) and use 3M clear packing tape to hold it together - very tough when assembled. Doesn't look very pretty but it does the job. (There are various colors of foam-board and spray-painting also works.) Shrouds in general need to be ~1" deep to work well ...
- Joined
- Sep 15, 2007
- Location
- Small town Emlenton, PA
A shroud can help. Maybe up to 16%. Thats using high speed 38mm thick fans though, normal fans less than half that. Thats 16% of your DT water temps. If your DT is 10C, then a drop of 1.3C. DT is what it's all about since we are dealing with ambient cooling methods. With lower speed fans it's a lot less of a gain.
http://martin.skinneelabs.com/Radiator-Fan-Orientation-And-Shroud-Testing-Review.html
http://martin.skinneelabs.com/Radiator-Fan-Orientation-And-Shroud-Testing-Review.html
OP
- Joined
- Nov 25, 2009
- Thread Starter
- #50
Thanks QuietIce and Conumdrum, I checked with XSPC at the same time, Paul there, reckons I would notice a small difference, guessing thats going to be like the "6% gains with slow speed fans" from that link.
He also advised against using just one fan, so will be using the two Gentle Typhoons on the rad in a pull config running slow rpm, and then use an S-Flex for my exhaust.
Am guessing I can easily add a shroud later on. I'd rather buy a pre-made one rather than a home-made jobby, as I would want to be sure I get it right.
So far I can only see shrouds designed for specific double rads, or universal for single rads, would two singles next to each other be ok? Am guessing so.
Also been running some tests with the reserator with two sflexs zipstrapped together and then stood on their end blowing cold air at the tower. This lowered temps quite well, although OCCT linpack still came up with errors. I updated the BIOS on the motherboard and the errors went away, so thats that one solved.
Am now running comparison tests (30 mins only) with the sflexs and without. Will post results in a mo in case anyone is interested in how the reserator works with a bit of added airflow.
He also advised against using just one fan, so will be using the two Gentle Typhoons on the rad in a pull config running slow rpm, and then use an S-Flex for my exhaust.
Am guessing I can easily add a shroud later on. I'd rather buy a pre-made one rather than a home-made jobby, as I would want to be sure I get it right.
So far I can only see shrouds designed for specific double rads, or universal for single rads, would two singles next to each other be ok? Am guessing so.
Also been running some tests with the reserator with two sflexs zipstrapped together and then stood on their end blowing cold air at the tower. This lowered temps quite well, although OCCT linpack still came up with errors. I updated the BIOS on the motherboard and the errors went away, so thats that one solved.
Am now running comparison tests (30 mins only) with the sflexs and without. Will post results in a mo in case anyone is interested in how the reserator works with a bit of added airflow.
OP
- Joined
- Nov 25, 2009
- Thread Starter
- #52
OCCT Linpack - External S-Flex fans on 100% -1200rpm
No external fans - max temp - 63c
Silent external fans - max temp - 56.5c
Loud external fans - max temp - 56.5c
Well that shows quite a difference from just having some slow speed fans blowing cold air on the resorator, and that increasing the fans above silent levels made no difference to max load temps (well at least for a 30 min test, didn't have enough time to let the tests run a little longer).
Hoping by adding the rad and having the GTs running silent, those temps will drop a little more. Would like that to be 50c max, but am a little doubtfull.
No external fans - max temp - 63c
Silent external fans - max temp - 56.5c
Loud external fans - max temp - 56.5c
Well that shows quite a difference from just having some slow speed fans blowing cold air on the resorator, and that increasing the fans above silent levels made no difference to max load temps (well at least for a 30 min test, didn't have enough time to let the tests run a little longer).
Hoping by adding the rad and having the GTs running silent, those temps will drop a little more. Would like that to be 50c max, but am a little doubtfull.
- Joined
- Sep 15, 2007
- Location
- Small town Emlenton, PA
You see where the green line drops down?
Overall, the trend didn't look good. It was linear and still climbing.
Thats why we like Prime 95 small ffts, it loads the CPU at 100% all the time.
Still, I can see your worry.
Make a cardboard tube around the fins, make a cardboard shroud and FORCE the air UP over the fins, pulling cool air from the bottom of the fins. Try to get the cardboard tight. I'll think you'll be pleasently surprised.
Overall, the trend didn't look good. It was linear and still climbing.
Thats why we like Prime 95 small ffts, it loads the CPU at 100% all the time.
Still, I can see your worry.
Make a cardboard tube around the fins, make a cardboard shroud and FORCE the air UP over the fins, pulling cool air from the bottom of the fins. Try to get the cardboard tight. I'll think you'll be pleasently surprised.
OP
- Joined
- Nov 25, 2009
- Thread Starter
- #54
Yeah I have seen that mod on another sight, it performs better than the Zalman fan shroud too, looks like they don't sell that anyway.
Was suprised how much of a drop I got from just a little bit of air blowing on the tower.
The radiator and 2 fans should (hoping) perform better than that - and look nicer than having a couple of fans stacked on my desk!
Just running a small FFT run, with HWmonitor on at the same time, but won't be running it for long as off to bed soon!
How long would you suggest to run it for to get a good result?
Was suprised how much of a drop I got from just a little bit of air blowing on the tower.
The radiator and 2 fans should (hoping) perform better than that - and look nicer than having a couple of fans stacked on my desk!
Just running a small FFT run, with HWmonitor on at the same time, but won't be running it for long as off to bed soon!
How long would you suggest to run it for to get a good result?
- Joined
- Jan 29, 2009
+1 to that plan Conum
six inch ducting and some sort of reducer/mount if you can get hold of it.
then pull air upwards.
I messed with shrouding the res1 v2 ugly but beneficial
really have to work that air, the body and base of the res1 are so thick... like a giant toxic thermos flask #Laughs#
buy a cheap 12" domestic fan from B&Q is best
EDIT: and yes the zalman fan is a mythological beast... I never found one for sale
six inch ducting and some sort of reducer/mount if you can get hold of it.
then pull air upwards.
I messed with shrouding the res1 v2 ugly but beneficial
really have to work that air, the body and base of the res1 are so thick... like a giant toxic thermos flask #Laughs#
buy a cheap 12" domestic fan from B&Q is best
EDIT: and yes the zalman fan is a mythological beast... I never found one for sale
- Joined
- Jan 29, 2009
probably out perform it with a 120.2 #Laughs#
- Joined
- Jan 29, 2009
more is better yes.
well youve got about two and half litre of coolant to tackle in the res1 and only need enough to pass through the loop so maybe its a lot of work for the sake as opposed to focusing on max cooling on a smaller volume, im not sure, not in science mode just too busy to focus so its a half assed suggestion
someone here a year(ish) back said they kept some coolant in a pop bottle and froze it every night then dropped it inside the res.
not sure id want coolant near my food but then again I used to keep some horrable frozen fish foods like blood worm and shrimp in my freezer so to each there own.
well youve got about two and half litre of coolant to tackle in the res1 and only need enough to pass through the loop so maybe its a lot of work for the sake as opposed to focusing on max cooling on a smaller volume, im not sure, not in science mode just too busy to focus so its a half assed suggestion
someone here a year(ish) back said they kept some coolant in a pop bottle and froze it every night then dropped it inside the res.
not sure id want coolant near my food but then again I used to keep some horrable frozen fish foods like blood worm and shrimp in my freezer so to each there own.
- Joined
- Jan 29, 2009
I progressed from zalman to swifty rads and phased the zalman out the equation.
after a couple of triples it became pointless to run the res1
its cooling ability lives in another time compared to todays burning chips.
alu just takes so long to lose heat compaired to denser and more conductive copper.
I found more fans on a up to date rad quieter than the airflow needed to drive the res1 as well.
after a couple of triples it became pointless to run the res1
its cooling ability lives in another time compared to todays burning chips.
alu just takes so long to lose heat compaired to denser and more conductive copper.
I found more fans on a up to date rad quieter than the airflow needed to drive the res1 as well.
