first time watercooling.. temps too high???

marcuri · 08-18-2005, 11:27 PM

i just got done installing my setup in my computer after running leak tests and such. what interesting is when i booted up i noticed a lower temp the program that comes with my DFI NF4 lanparty mobo however its not significantly lower. approximately 3-4 degrees cooler than the stock HSF that i had on the amd 64bit processor. the temp is now 28C , or ~83F... i do know motherboard temp sensors tend to be off and i didn't adjust them... however the stock HSF had a temp of approx 32-33C If I Recall Correctly.

My current setup includes:
Black Ice Pro II, 2x120mm panaflo m1a fans with shrouds installed, externally mounted
DD Mag pump
DD TDX a64 block
1 5.25" bay res.

As far as i can tell all air is out of the system with maybe the exception of a few bubbles in the radiator. Flow seems good as well. Theres a few things i think might be happening:

1. Put too much arctic silver 5 on the tdx block thus causing ineffecient cooling. (i had put a size alittle more than twice that recommended by the instructions because i noticed how poaste arctic silver 5 was to the other stuff and didn't think it'd spread out as well... the instructions recommened a size around a small grain of uncooked rice.. i put alittle more than twice that i'd say).
2. Some Alunimun fileings from cutting alittle bit on my case may have found their way into my tubing which was near by and is causing issues with the copper. I did make sure it was all gone (the few peices that i found in there) but there could've been some very small, pin head sized, specs of AL still left over in there.

If theres one place i know that can help me fix this problem its here. thanks for any advice ahead of time.

marcuri · 08-19-2005, 12:00 AM

well i just re did the artic silver 5 application and remounted the waterblock. temps are still as high as they were before and i definately put significantly less arctic silver and i followed the directions carefully.

maybe i'm expecting temps to be better than they can be? i thought with water cooling i'd have the temps down close to ambient (75F, 23.8 C). right now idle i'm still getting 28C-29C.... which considering the ambient temp i thought might be high.

redleader · 08-19-2005, 12:45 AM

Quote:

Originally Posted by marcuri

well i just re did the artic silver 5 application and remounted the waterblock. temps are still as high as they were before and i definately put significantly less arctic silver and i followed the directions carefully.

maybe i'm expecting temps to be better than they can be? i thought with water cooling i'd have the temps down close to ambient (75F, 23.8 C). right now idle i'm still getting 28C-29C.... which considering the ambient temp i thought might be high.

Thats not bad at all, though generally load temps are much more important.

marcuri · 08-19-2005, 12:58 AM

well while running prime 95 (in place large FFts) i get a boost of approx 10 degrees F. temps jump from 80-82F to 89-91F.

Seems alittle high or is about a 10 degree jump normal? figured with water cooling it'd jump around +5 F(+3.5C) or so... so +10 F (+6C) is about twice what i expected.

so is that good or no for my setup?

Brians256 · 08-19-2005, 01:24 AM

30C under load isn't bad at all, unless your ambient temps are -40C.

Assuming your ambient temps are 75F, your 80-82F idle temps are very respectable. It should also be a lot quieter. What is even better is that if you OC your computer, you should be able to keep your temps down much better than if you had air.

Long Haired Git · 08-19-2005, 01:29 AM

I don't have the CW or PQ graphs for those components, so can't judge.
However, keep in mind that you used to have a fair bit of air flow around the CPU socket, and often the thermal probe for the CPU is impacted by this. So, grab a fan and point it at the CPU socket, and see if the temps drop.

BTW, did you do a flow rate test and hence know the approx flow rate? I am gobsmacked most people don't do this...

Next, I've approximated your rig using:
CSP Mag, 1.2m of 1/2" tubing, Thermochill 120.2, LRR G4
and I get 5.6 LPM and an air-to-CPU deltaT of 13.374 degrees for a 70W CPU.
That's approximately 0.19 degrees per Watt.

Yes, the TDX is more LRR White Water than LRR G4, but my simulator (called the Approximator) is not yet handing the multiple outputs of the LRR.

Anyhoo, according to you ambient = "23.8 C, right now idle i'm still getting 28C-29C".
Lets call it 28.8 and 23.8, which means 5 degrees off ambient.

So, according to my model, your CPU consumes approximately 25W at idle.

billbartuska · 08-19-2005, 01:36 AM

That's about what I would expect for that setup. Are you using the 1/2" barbs? You could try changing nozzles (I'd start out trying less restriction) and/or lapping the cpu and block. Other than that its more air (Panaflo 120x38mm Ultra High Speed Fan FBA12G12U), but you'll pay a heavy noise penalty. Also the AS5 will take a few days of heat cycles to start working at it's best.
All in all a 10C temp rise under load is pretty good. Are you monitoring water temps to see if they have stabalized when you see the 10C temp rise?
Is your system MUCH quieter with the watercooling? Case and MOBO temps much lower too?

Long Haired Git · 08-19-2005, 02:03 AM

I bet it was 32-33 on stock air under heavy load.
My FOP-"oh my god ears bleeding eyes watering"-38 was like ambient + 22°C under load, same as my fanless-watercooled rig (which was quiet with a ten foot high Q).

marcuri · 08-19-2005, 02:56 AM

haven't oc'ed this baby yet but i will very soon, its a 3000+ a64 (forgot the name, but its 90nm and 512 L2) and i've got mushkin 3500 redline memory and screwed around with the case alittle bit to have a 80mm panaflo blowing air onto the chips and a 120mm m1a panaflo exhausting air out the case right next to it. either way its good to hear that the temps won't go as high while ocing compared to air. its not that much quieter than air mostly due to the fact that i have the rad mounted externally using the thermaltake radbox and those 2 panaflo fans which are quiet but still audible, plus an additional 80mm panaflo 120mm panaflo and vantec 120mm stealth on intake. also i still have the crappy stock chipset fan on my DFI nf4 lanparty sli-dr mobo (which i don't think i'll be able to change due to the fact that it would cause problems with mounting a videocard into the pci-e slots if i put anything thats thicker in there.. and definately no chipset water cooling unless i can find one with barbs on the side) and the stock fan on my video card is kinda loud too.

i haven't yet tested the flow rate... i actually have no idea how to. i was thinking of buying this thing (Forgot the name of it... i'll blame it on being tired) that you put in-line that shows you your flow rate but i'm going to wait on it and maybe pick it up when i have more money. what ways are there to go about measureing the flow rate? also: whats the best way to measure the water temp?

also after running the pc for alittle while i'm getting 80-84 degrees F (up to 30+ degrees C) idle and up to 95 degrees F or 35C under load with prime 95. not sure why they went up alittle bit but my guess is it has to do with the arctic silver needing to be broken in.

yea i'm using 1/2" barbs with tygon 1/2" id 3/8" od tubing and those metal clamps with screws on all tubing/barbs tightened down really hard. i don't know what you mean by less restriction.. were you assuming i was using 3/8" id tubing?

quick question: am i decreasing my flow rates/effectivness of the cooling system by having the mag pump pumping the water "up" (the tubing goes up above the hieght of the res in the 5.25" bay and then loops down into the cpu block)?

p.s. cooling system is going from res -> pump -> block -> rad -> res... and i'll post pics of it soon cause its alittle unusual the way this whole thing is setup.

Dave · 08-19-2005, 06:25 AM

^ Your flow path is correct, and as stated you can not judge from idle temps.

I would suggest you check your block for any obstructions.
I found hair in a test set up, which caused a major decrese in flow.

**That had me confused for about 1/2 day, as the hair was under the nozzle plate and in between the vanes

marcuri · 08-19-2005, 10:40 AM

i might give that a try dave... i did manage to find a hair after i took the system apart from the leak test i did... although i don't know if its still nessicary as i'm seeing, based on what everyone has told me, normal load temps. after doing about 30 mins of prime 95 (large ffts to get max heat dissipation from the cpu) the cpu gets between 91F and 95F (between 32-35C)... which.. i think... seems about right.

to save myself the headache of taking the block apart what should i expect in terms of flow rate from the mag pumps? and whats the easiest way to measure the flow rate in my system?

in mother board monitor i'm getting a rpm reading from where the mag pump is connected to the the mobo but i don't know if its anything meaningful. can that be used for the flow rate?

Long Haired Git · 08-19-2005, 04:00 PM

How to test flow rates: http://www.overclockers.com.au/wiki/...te_measurement

08-18-2005, 11:27 PM	#1
marcuri Cooling Neophyte Join Date: Jul 2005 Location: Florida Posts: 15	first time watercooling.. temps too high??? i just got done installing my setup in my computer after running leak tests and such. what interesting is when i booted up i noticed a lower temp the program that comes with my DFI NF4 lanparty mobo however its not significantly lower. approximately 3-4 degrees cooler than the stock HSF that i had on the amd 64bit processor. the temp is now 28C , or ~83F... i do know motherboard temp sensors tend to be off and i didn't adjust them... however the stock HSF had a temp of approx 32-33C If I Recall Correctly. My current setup includes: Black Ice Pro II, 2x120mm panaflo m1a fans with shrouds installed, externally mounted DD Mag pump DD TDX a64 block 1 5.25" bay res. As far as i can tell all air is out of the system with maybe the exception of a few bubbles in the radiator. Flow seems good as well. Theres a few things i think might be happening: 1. Put too much arctic silver 5 on the tdx block thus causing ineffecient cooling. (i had put a size alittle more than twice that recommended by the instructions because i noticed how poaste arctic silver 5 was to the other stuff and didn't think it'd spread out as well... the instructions recommened a size around a small grain of uncooked rice.. i put alittle more than twice that i'd say). 2. Some Alunimun fileings from cutting alittle bit on my case may have found their way into my tubing which was near by and is causing issues with the copper. I did make sure it was all gone (the few peices that i found in there) but there could've been some very small, pin head sized, specs of AL still left over in there. If theres one place i know that can help me fix this problem its here. thanks for any advice ahead of time.

08-19-2005, 12:00 AM	#2
marcuri Cooling Neophyte Join Date: Jul 2005 Location: Florida Posts: 15	well i just re did the artic silver 5 application and remounted the waterblock. temps are still as high as they were before and i definately put significantly less arctic silver and i followed the directions carefully. maybe i'm expecting temps to be better than they can be? i thought with water cooling i'd have the temps down close to ambient (75F, 23.8 C). right now idle i'm still getting 28C-29C.... which considering the ambient temp i thought might be high. Last edited by marcuri; 08-19-2005 at 12:09 AM.

08-19-2005, 12:58 AM	#4
marcuri Cooling Neophyte Join Date: Jul 2005 Location: Florida Posts: 15	well while running prime 95 (in place large FFts) i get a boost of approx 10 degrees F. temps jump from 80-82F to 89-91F. Seems alittle high or is about a 10 degree jump normal? figured with water cooling it'd jump around +5 F(+3.5C) or so... so +10 F (+6C) is about twice what i expected. so is that good or no for my setup? Last edited by marcuri; 08-19-2005 at 01:05 AM.

08-19-2005, 01:29 AM	#6
Long Haired Git Cooling Savant Join Date: Jan 2003 Location: Sydney, Oz Posts: 336	I don't have the CW or PQ graphs for those components, so can't judge. However, keep in mind that you used to have a fair bit of air flow around the CPU socket, and often the thermal probe for the CPU is impacted by this. So, grab a fan and point it at the CPU socket, and see if the temps drop. BTW, did you do a flow rate test and hence know the approx flow rate? I am gobsmacked most people don't do this... Next, I've approximated your rig using: CSP Mag, 1.2m of 1/2" tubing, Thermochill 120.2, LRR G4 and I get 5.6 LPM and an air-to-CPU deltaT of 13.374 degrees for a 70W CPU. That's approximately 0.19 degrees per Watt. Yes, the TDX is more LRR White Water than LRR G4, but my simulator (called the Approximator) is not yet handing the multiple outputs of the LRR. Anyhoo, according to you ambient = "23.8 C, right now idle i'm still getting 28C-29C". Lets call it 28.8 and 23.8, which means 5 degrees off ambient. So, according to my model, your CPU consumes approximately 25W at idle. __________________ Long Haired Git "Securing an environment of Windows platforms from abuse - external or internal - is akin to trying to install sprinklers in a fireworks factory where smoking on the job is permitted." (Prof. Gene Spafford) My Rig, in all its glory, can be seen best here AMD XP1600 @ 1530 Mhz \| Soyo Dragon + \| 256 Mb PC2700 DDRAM \| 2 x 40 Gb 7200rpm in Raid-0 \| Maze 2, eheim 1250, dual heater cores! \| Full specifications (PCDB)

08-19-2005, 02:03 AM	#8
Long Haired Git Cooling Savant Join Date: Jan 2003 Location: Sydney, Oz Posts: 336	I bet it was 32-33 on stock air under heavy load. My FOP-"oh my god ears bleeding eyes watering"-38 was like ambient + 22°C under load, same as my fanless-watercooled rig (which was quiet with a ten foot high Q). __________________ Long Haired Git "Securing an environment of Windows platforms from abuse - external or internal - is akin to trying to install sprinklers in a fireworks factory where smoking on the job is permitted." (Prof. Gene Spafford) My Rig, in all its glory, can be seen best here AMD XP1600 @ 1530 Mhz \| Soyo Dragon + \| 256 Mb PC2700 DDRAM \| 2 x 40 Gb 7200rpm in Raid-0 \| Maze 2, eheim 1250, dual heater cores! \| Full specifications (PCDB)

08-19-2005, 01:24 AM	#5
Brians256 Pro/Staff Join Date: Oct 2001 Location: Klamath Falls, OR Posts: 1,439	30C under load isn't bad at all, unless your ambient temps are -40C. Assuming your ambient temps are 75F, your 80-82F idle temps are very respectable. It should also be a lot quieter. What is even better is that if you OC your computer, you should be able to keep your temps down much better than if you had air.

08-19-2005, 01:36 AM	#7
billbartuska Cooling Savant Join Date: Jun 2005 Location: Skokie, Illinois Posts: 322	That's about what I would expect for that setup. Are you using the 1/2" barbs? You could try changing nozzles (I'd start out trying less restriction) and/or lapping the cpu and block. Other than that its more air (Panaflo 120x38mm Ultra High Speed Fan FBA12G12U), but you'll pay a heavy noise penalty. Also the AS5 will take a few days of heat cycles to start working at it's best. All in all a 10C temp rise under load is pretty good. Are you monitoring water temps to see if they have stabalized when you see the 10C temp rise? Is your system MUCH quieter with the watercooling? Case and MOBO temps much lower too?

08-19-2005, 02:56 AM	#9
marcuri Cooling Neophyte Join Date: Jul 2005 Location: Florida Posts: 15	haven't oc'ed this baby yet but i will very soon, its a 3000+ a64 (forgot the name, but its 90nm and 512 L2) and i've got mushkin 3500 redline memory and screwed around with the case alittle bit to have a 80mm panaflo blowing air onto the chips and a 120mm m1a panaflo exhausting air out the case right next to it. either way its good to hear that the temps won't go as high while ocing compared to air. its not that much quieter than air mostly due to the fact that i have the rad mounted externally using the thermaltake radbox and those 2 panaflo fans which are quiet but still audible, plus an additional 80mm panaflo 120mm panaflo and vantec 120mm stealth on intake. also i still have the crappy stock chipset fan on my DFI nf4 lanparty sli-dr mobo (which i don't think i'll be able to change due to the fact that it would cause problems with mounting a videocard into the pci-e slots if i put anything thats thicker in there.. and definately no chipset water cooling unless i can find one with barbs on the side) and the stock fan on my video card is kinda loud too. i haven't yet tested the flow rate... i actually have no idea how to. i was thinking of buying this thing (Forgot the name of it... i'll blame it on being tired) that you put in-line that shows you your flow rate but i'm going to wait on it and maybe pick it up when i have more money. what ways are there to go about measureing the flow rate? also: whats the best way to measure the water temp? also after running the pc for alittle while i'm getting 80-84 degrees F (up to 30+ degrees C) idle and up to 95 degrees F or 35C under load with prime 95. not sure why they went up alittle bit but my guess is it has to do with the arctic silver needing to be broken in. yea i'm using 1/2" barbs with tygon 1/2" id 3/8" od tubing and those metal clamps with screws on all tubing/barbs tightened down really hard. i don't know what you mean by less restriction.. were you assuming i was using 3/8" id tubing? quick question: am i decreasing my flow rates/effectivness of the cooling system by having the mag pump pumping the water "up" (the tubing goes up above the hieght of the res in the 5.25" bay and then loops down into the cpu block)? p.s. cooling system is going from res -> pump -> block -> rad -> res... and i'll post pics of it soon cause its alittle unusual the way this whole thing is setup. Last edited by marcuri; 08-19-2005 at 03:05 AM.

08-19-2005, 06:25 AM	#10
Dave Cooling Savant Join Date: Oct 2004 Location: Canada Posts: 260	^ Your flow path is correct, and as stated you can not judge from idle temps. I would suggest you check your block for any obstructions. I found hair in a test set up, which caused a major decrese in flow. **That had me confused for about 1/2 day, as the hair was under the nozzle plate and in between the vanes

08-19-2005, 10:40 AM	#11
marcuri Cooling Neophyte Join Date: Jul 2005 Location: Florida Posts: 15	i might give that a try dave... i did manage to find a hair after i took the system apart from the leak test i did... although i don't know if its still nessicary as i'm seeing, based on what everyone has told me, normal load temps. after doing about 30 mins of prime 95 (large ffts to get max heat dissipation from the cpu) the cpu gets between 91F and 95F (between 32-35C)... which.. i think... seems about right. to save myself the headache of taking the block apart what should i expect in terms of flow rate from the mag pumps? and whats the easiest way to measure the flow rate in my system? in mother board monitor i'm getting a rpm reading from where the mag pump is connected to the the mobo but i don't know if its anything meaningful. can that be used for the flow rate? Last edited by marcuri; 08-19-2005 at 11:05 AM.

08-19-2005, 04:00 PM	#12
Long Haired Git Cooling Savant Join Date: Jan 2003 Location: Sydney, Oz Posts: 336	How to test flow rates: http://www.overclockers.com.au/wiki/...te_measurement __________________ Long Haired Git "Securing an environment of Windows platforms from abuse - external or internal - is akin to trying to install sprinklers in a fireworks factory where smoking on the job is permitted." (Prof. Gene Spafford) My Rig, in all its glory, can be seen best here AMD XP1600 @ 1530 Mhz \| Soyo Dragon + \| 256 Mb PC2700 DDRAM \| 2 x 40 Gb 7200rpm in Raid-0 \| Maze 2, eheim 1250, dual heater cores! \| Full specifications (PCDB)

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)