View Full Version : water cooling for the first time
well I finally did it went to watercooling! And my temps havent changed a bit?? I have a maze2 block , 2 120mm cooling fans,3.5x5.5x12 cooler,eiheim 315gph pump with a 1 gallon reserve, purple ice,distilled water,arctic silver,I even lapped the maze2,that I must say was pretty warped for being CNC'ed. And after all that the temps are no better than with the swiftech MC462. The temps are 42*c right now on the net and running under a load 44*c,Whats the deal did I miss something?The case is even open right now and nothing changes.
The swiftech 462 is a stout air cooling solution, but I'd think you'd do a little better with water. I just went to water cooling last week and noticed a huge difference in temp on my Duron 650 @ 800. Under air I was around 43 C under normal working load, and now I hover between 30 C and 31 C.
What radiator did you use? Are you sure the water block is mounted correctly and firmly contacting the block? I'm not familiar with your pump, but if it's an inline unit, air bubbles trapped in the lines can cause higher temps so you may want to check that.
Let me know answers to these quesitons and I'll try to help further.
Thanks for the reply, I bled the system of all bubbles, and I took the block off to check if making contact and it is. The cooler is an automotive cooler for oil/transmission up to 450 hp so the cooler should be suffeciant.The pump i may have spelled it wrong eheim 1248.This is really bumming me out!!
Tacoman667
06-09-01, 01:45 PM
Is your radiator mounted in a way that its siphoning air from the hot inside of the case or the cooler air outside the case? That makes a BIG difference. My dangerden cube is inside my case pulling about half outside air and half case air and look at my temps. I need to get some more tubing so that I can put my radiator outside my case and I will see even cooler temps. What water block do you use and do you have a copper block/aluminum radiator going on?
All equiptment is external and the air is drawn in from the outside. I have a dangerden maze2 block and a becooling MBO block, a enheim 315 gph pump inside the reservoire with the water going in the radiator after the pump so its cooling the water after the reservoir,2 x120mm fans. Right now on the net its at 43*c a little high I would say.The The block is making good contact with the chip, that was the first thing I checked but I think I will check it again. I cant believe how warped the block was when I started lapping it.It is alot quieter though!!!The MC462 is sweet but noisy even with the rheostat on it.please keep the suggestion coming.
Tacoman667
06-09-01, 03:41 PM
What is the temperature in the room that the computer is in? I keep my room temps around 70F. I have a 120mm YS-Tech fan sucking air outta my radiator at 131cfm. Fan only cost me about 13 bucks. Best fan I ever used. Try a better fan. Another thing, make sure there are NO kinks in your tubing to inhibit the flow of the water. I have the Eheim 1250 hobby pump 315gph running in my system with about 2 1/2-3 feet of tubing and I can keep very nice temps. these are things you need to think about too. If you have a resavoir, then make sure it's nice and big<1 gallon would be fine and a little smaller is ok too> and have your pump in it to suck the water in. Do not put an adapter on the intake on the pump, that caused me to suck less water through my system.
[Oc]acaridans
06-09-01, 03:52 PM
Im a little confised with the way you have your setup running. Is it
res/pump>>>radiator>>>maze>>>res/pump ?
if it is try arraging it to
res/pump>>>maze>>>radiator>>>res/pump
this way usually works better
I'm running reservoir to radiator to maze2 back to reservoir cuz I read that its better that way cuz it cools the water that the pump heats. I also reversed the direction of the fans,so that they suck air in instead of blowing, but that did not make any difference. 40*c as I write :-( also does anybody think that cuz I'm running a MBO cooling block too that that might be my prob? It's after the maze2. MBO temp is 29*c now, Is that about right for a water cooled MBO chip?
Thelemac
06-10-01, 12:05 AM
Did you case temps change when you installed the watercooling? That would definately lead to that problem. My case temps went up about 10C when I put mine in...course, mine's all internal, but it made it look like I only got about a 4C difference in temps, when I really got like 14C.
Also, do you have your fans set up so that one is on each side of the radiator? If you do, make sure that you have one blowing and one sucking, otherwise the radiator isn't really getting any air going through it (some, but no much).
My case temps are the same about 5* hoter than ambient. I have both 120mm fan running side by side on the radiator.
I totally disassembled the cooling system reassembled it turned the fans around ,rearranged the hoses,no kinks,block is making contact with chip,I cant figure this one out. If all I'm going to gain is quiet I think I'm going to put the MC462 back on and get lower temps.I put temp probed on the hose going in and coming out of the cpu block temp going in is 74.5*f coming out is 77.5*f .Ambient is 72.5*f,case is 75.5*f. what next?
8
RJ (Jun 11, 2001 08:30 a.m.):
I totally disassembled the cooling system reassembled it turned the fans around ,rearranged the hoses,no kinks,block is making contact with chip,I cant figure this one out. If all I'm going to gain is quiet I think I'm going to put the MC462 back on and get lower temps.I put temp probed on the hose going in and coming out of the cpu block temp going in is 74.5*f coming out is 77.5*f .Ambient is 72.5*f,case is 75.5*f. what next?
How many GPH is your pump rates at? If it's up around 300GPH then I'd say slow the water down as it goes through the radiator. I recently tested this and realized a 2 C drop in CPU temps. My Maxijet 1200 is 295GPH. I put a loop (see post Slowing Water down in Watercooling Works!) before the water enters the radiator. Others had suggested this might help temps so I tested it and it did. The loop slows the water down thus giving the radiator more time to cool it.
8
CpuWaterCool.com
06-11-01, 09:44 AM
Radiator cooling is no more! Bong cooling (tower) rocks!
Thanks for the reply my pump is 315gph I cant seem to find that post u were talking about can u give me an idea where it is? I am also going to convert it to an inline instead of submersible see if that may cure a few things
RJ (Jun 11, 2001 09:50 a.m.):
Thanks for the reply my pump is 315gph I cant seem to find that post u were talking about can u give me an idea where it is? I am also going to convert it to an inline instead of submersible see if that may cure a few things
My post on slowing down the water is here.
http://forums.overclockers.ws/forums/UltraBoard.cgi?action=Read&BID=5&TID=15803&SID=107376
If you have a smaller radiator (like I do) then it needs more time to cool the water. I'd stick with a submersible pump. Inlines produce less heat, but air bubbles in the lines can wreak havoc on cooling.
Thanks, I rigged up something kinda like thermostat .that I'm going to install in the return line it will restrict the water volumn coming from the cpu block, hey its worth a shot.it will reduce the hose down to a 1/4 inch shouldnt add to much pressure. It works on the same princible as an auto thermostat.
hey amdguy your setup looks just like mine,however I switched my hose routing to res>chip>radiator>res.I had it the same as u but someone told me that it was the wrong way, I changed but it did not make a difference.Thanks for all your help
Tacoman667
06-11-01, 12:20 PM
Once the water cooling starts getting your temps down to ambient levels, your system temps will drop drastically. My system temps went from 34C down to 23C now. My CPU runs at 28 load and my PSU is the only fluxuating temp in the whole case. Go resuvoir, waterblock, radiator, then resuvoir again. If your intake and outtake temps on your radiator do not have a more significant temp difference, something is wrong with your radiator. Make sure all the fins on it are properly installed, check to make sure you have enough air running through it, and hell, you might even need a different radiator for it to work. I have the DD cube and a 315gph Eheim 1250 hobby pump and I stay at these low temps.
Tacoman667 (Jun 11, 2001 12:20 p.m.):
Once the water cooling starts getting your temps down to ambient levels, your system temps will drop drastically. My system temps went from 34C down to 23C now. My CPU runs at 28 load and my PSU is the only fluxuating temp in the whole case. Go resuvoir, waterblock, radiator, then resuvoir again. If your intake and outtake temps on your radiator do not have a more significant temp difference, something is wrong with your radiator. Make sure all the fins on it are properly installed, check to make sure you have enough air running through it, and hell, you might even need a different radiator for it to work. I have the DD cube and a 315gph Eheim 1250 hobby pump and I stay at these low temps.
23C and 28C?? What's the ambient? What's the room temp?? My room is 25C and ambient is 28-29C. You must have to put a sweater on to sit in the room with your PC LOL.
JoeCrappa
06-11-01, 07:11 PM
most likely it is your block orientation. how tight is your block? also do you have the water inlet above the CPU? any arctic silver? or compound? do you still have your pads on the CPU? do you have a copper shim on there? if your pads are on then TAKE THEM OFF and you'll get better temps guaranteed. i know this can be dangerous but if you get a shim then it could be safer or do what this guy did here Bunker Mentallity (http://www2.apex.net/users/timwhita/protection.html). Please email me and i'll try to help you out. im sure this problem concerns me as much as it concerns you.
JoeCrappa (Jun 11, 2001 07:11 p.m.):
most likely it is your block orientation. how tight is your block? also do you have the water inlet above the CPU? any arctic silver? or compound? do you still have your pads on the CPU? do you have a copper shim on there? if your pads are on then TAKE THEM OFF and you'll get better temps guaranteed. i know this can be dangerous but if you get a shim then it could be safer or do what this guy did here Bunker Mentallity (http://www2.apex.net/users/timwhita/protection.html). Please email me and i'll try to help you out. im sure this problem concerns me as much as it concerns you.
I highly recommend not removing the pads from the CPU. They are there to provide a "cushion" point between the HS install and the die. Removing them greatly increases the chances of a crushed die. My waterblock mounts up fine with the pads in place. I'm leaning toward the radiator and water flow rate.
well looping the return hose did nothing for me either did the restrict,I also put a rheostat on the pump to slow down the volume did nothing.the block is making good contact with the chip I 've had it off 3 times now.I'm idling at 38*c load 41*c ambient right now is 22.2*c case is 23*c. The only other thing that I haven't tried is taking the pump out and running it inline.Does anyone think that would make a big difference?
No I agree I'm not going to remove the pads, I have an 6 pass radiator with 1/2"tubes I thought it would be more than enough If it could handle cooling oil on a 450 hp engine u would think it would cool this thing plus 2 110 cfm fans side by side
something just dawned me see what u guys think I'm using a cooler that was designed to cool oil over 350+ degrees. maybe this thing has too thick of tubing to radiate the lower temps that a cpu puts out ?
JoeCrappa
06-11-01, 08:33 PM
i'm telling you man. its the block orientation. from what i see here thats the only variable left. inline and submersible dont really make that much of a difference. you should try buying a shim, those help you not to crush your CPU. whenever i had problems with my watercooling it was the block orientation. my temp dropped 7c when i took off the pads and put on my shim. watercooling is risky period. if you dont take risks then you dont get good results. its on you RJ i'm just trying to give you my 2 cents which i've learned a lot from.
Well I decided to change the radiator I work at an auto parts store and found a nice copper heater core 7x7x2 hope this will do the trick.
RJ (Jun 12, 2001 11:30 a.m.):
Well I decided to change the radiator I work at an auto parts store and found a nice copper heater core 7x7x2 hope this will do the trick.
Let me know how it turns out. I'm extremely curious now. (Starting to feel like it's my problem) LOL.
You're theory sounds reasonable.
Not a problem Amdguy I soldered up some fittings on it when I went home for lunch I just have to mount it and I'm only using one 107cfm fan will post the results tonight. Thanks for your reply
JoeCrappa
06-12-01, 05:34 PM
cool i wanna know how it turns out also, you guys probably think im a jerk but hey i had to voice my opinion =)
well guys try and figure this one out.New radiator no difference 38*c idle 41*c load without the fan running?? no change if I turn the fan on.If I turn the case fans on high I get a 2-3*drop.any other suggestion?
well this is it I give up here are my stats CPU 40*c idle 43*c load, MBO 28*c, water temp in the reservoir 23.8*c, ambient 22.7*c I've tried everything
Are you sure you tightend down the bolts enough to crush the 4 feet on the cpu? The Maze 2 blocks have springs and you have to compress them somewhat to get enough preasure on the core. It seems like you have tried almost everything else so the problem may very well lie in the thermal transfer from you cpu to the block.
I agree I have the bolts tightened right down as tight as it would get I'm afraid to try tightening any further
JoeCrappa
06-12-01, 07:38 PM
try putting your block back on or reapplying compund (Arctic Silver?). thats the only thing i think you haven't checked=)
oh and i also put a valve to slow down the flow in the radiator, no change sorry if it sounds like im getting an attitude
Have you tried reading the core temp with a Compu-Nurse? There are a few other products like this to basically a digital readout and a very thin thermal probe that will fit in between the waterblock and cpu to read the core temp. The mobo's temp readings may not be anywhere near reliable. I have an A7V and a A7V133 the in socket thermistor is located on the mobo inside the socket. There is no way this can be acurate as it is really reading the air temp in the socket plus the mobo temp. I can very well see why the readings could go up now that you are watercooling. The high CFM fan on the switech could have been helping to cool the mobo which effected the mobo mounted thermistor.
I have an Asus A7V133 board, no idea where the probe is I have a digdoc5 but I think the probe are to thick .But I'm starting to wonder if that is true cuz for one my mother board temps are good and I just over clocked to 1503 and so far everything seems stable. CPU temp right now is 39*c I can turn the cooling fan off and the temp won't even change.
RJ (Jun 12, 2001 08:14 p.m.):
I have an Asus A7V133 board, no idea where the probe is I have a digdoc5 but I think the probe are to thick .But I'm starting to wonder if that is true cuz for one my mother board temps are good and I just over clocked to 1503 and so far everything seems stable. CPU temp right now is 39*c I can turn the cooling fan off and the temp won't even change.
After reading that you have a Asus A7V I'm beginning to wonder if the probe isn't off as well. I just can't see it being that warm with water. I'd try Hoot's suggestion on using that digital meter to check temps. It's a cheap device (~$14) and would give you peace of mind. I'm sure your cooler than that is reading. Asus is known for poor temp readings from their thermistors. I've heard of them being off as much as 10 C!
The article hoot did on the device he found works well is here: http://www.overclockers.com/tips491/
Hope this helps you out!
Cool thanks AMDguy I have a digitaldoc5 I will try it .
hey Amdguy I hope this is right I put the probe next to the cpu core like the article said u sent me. trimmed the edge off and I have sisoftsandra running for the last 4hr and the temp is 31.7*c I hope that was the problem. The MBO5 shows 41*c. So do the temps seem more realistic 28.5*c idle 31.7*c load for water cooling?
RJ (Jun 13, 2001 03:02 p.m.):
hey Amdguy I hope this is right I put the probe next to the cpu core like the article said u sent me. trimmed the edge off and I have sisoftsandra running for the last 4hr and the temp is 31.7*c I hope that was the problem. The MBO5 shows 41*c. So do the temps seem more realistic 28.5*c idle 31.7*c load for water cooling?
That sounds more like it. I'd run Prime95 and see what full load temps you get. Run Prime95 for an hour or so under Torture Mode and Priority 10. Then list your full load temp as well as room temp and ambient temp (ambient is usually considered the air temp 1" from the processor). I'm sure that was you're problem. Just think, all that testing and stressing out because Asus can't get a good thermistor reading.
AMDguy I tried prime95 but it says harware problem and wont go any further, that is why I use sisoftsandra.
RJ (Jun 13, 2001 03:41 p.m.):
AMDguy I tried prime95 but it says harware problem and wont go any further, that is why I use sisoftsandra.
If it says that then you need to either step the Mhz down a notch or increase the voltage. That means you're not getting a clean overclock.
Well, I clocked it at 1500mhz stabile so far, idle temps are 31*c load 34*c Ambient temp is 21.9*c case 25.4*c, vcore 1.79v set on auto.Still can't get Prime95 to stop throwing harware failure error, I tries lowering the bus and still get the error. how do these temps look now? I also hooked probes up to the water into the cpu and out cpu, those temps are:In 23.6* out 24* MBO temp is 30*c
Just a note to all the "slow down the flow through your radiator" people, while I don't doubt that when you slowed the flow your temps dropped, I can state that it had nothing to do with the radiator itself. The reason is fairly simple. If the water is moving more slowly through the radiator, it must move more slowly through the ENTIRE system(just like electricity), which means the water will pick up more heat as it crosses the cpu, decreasing thermal efficiency, and increasing the amount of work the radiator has to do. The radiator doesn't mind though, cuz it has more time to do it. However, radiators are more efficient with hotter water(as all things are, as the thermal difference increases). This MAY be the case with very small and very inefficient radiators(the kind that fit one 120mm fan, or the kind that are very very thin, etc). I had thought about trying to slow down the flow through my radiator as well, but I realized very quickly that it couldn't be done without slowing the flow through the waterblock.
RJ (Jun 13, 2001 09:58 p.m.):
Well, I clocked it at 1500mhz stabile so far, idle temps are 31*c load 34*c Ambient temp is 21.9*c case 25.4*c, vcore 1.79v set on auto.Still can't get Prime95 to stop throwing harware failure error, I tries lowering the bus and still get the error. how do these temps look now? I also hooked probes up to the water into the cpu and out cpu, those temps are:In 23.6* out 24* MBO temp is 30*c
Those temps look good, but I'd turn the voltage off of "Auto" and on to "User Define". Set it to 1.825 or even 1.85V. You'll likely get more out of the CPU plus Prime95 will run. Prime95 is giving you a hardware error likely because it is finding a problem in the calculation. This occurs when the CPU can't process the information correctly. My system would run at 966 just fine in Windows and Sandra, but Prime95 gave me a hardware error. I dropped it back to 952 (because my voltage was maxed), and then Prime95 worked fine.
The important thing is, if Prime95 will not run, the the CPU is not stable at that speed and voltage.
I set the voltage to1.85v and got p95 to work it did a series of tests until test 36 then errored hardware failure.Should I step down the mhz, or is there something else I should try? the load temp was 34.5*c during the test. I'm at 11.5x130 now
RJ (Jun 14, 2001 11:20 a.m.):
I set the voltage to1.85v and got p95 to work it did a series of tests until test 36 then errored hardware failure.Should I step down the mhz, or is there something else I should try? the load temp was 34.5*c during the test. I'm at 11.5x130 now
I'd set it to 11x133. This will give you 1463. This will probably run stable in Prime95 and will likely be faster in the long run as the FSB will be 133 instead of 130. You could also try upping the FSB 1mhz at a time until Prime95 is no longer stable. Your temps are good.
Thanks I'll keep u posted
I ran P95 for about 41/2 hrs the temps were doing good ambient was 22.7*c load 35.2*c I went out for a little bit came back and my wife shut the door and its hotter than hell here right now and the ambient went up to 28.7 in the the room and guess what ?so did my cpu temp 39.1*c So after I chewed the wife out and opened the door the temp came down a degree or two. so I guess still not to bad. Thanks for all the help guy's
New temps I think I finally got it right cpu idle 32*c load 34.5*c running P95 for 24hrs no errors :-) ambient 22.7*c case 25.2*c went back to dual fans and a 8x5.5x2 copper radiator.low noise and cool temps I luv it.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.